General Relativity: Matthias Bartelmann Institut F Ur Theoretische Astrophysik Universit at Heidelberg

General Relativity
Matthias Bartelmann
Institut f ur Theoretische Astrophysik
Universit at Heidelberg
Contents
1 Introduction 1
1.1 The Idea behind General Relativity . . . . . . . . . . . 1
1.2 Fundamental Properties of Gravity . . . . . . . . . . . 3
1.2.1 Scales . . . . . . . . . . . . . . . . . . . . . . 3
1.2.2 The Equivalence Principle . . . . . . . . . . . 4
1.3 Immediate Consequences of the Equivalence Principle 6
1.3.1 Gravitational Redshift . . . . . . . . . . . . . 6
1.3.2 Gravitational Light Deection . . . . . . . . . 7
1.4 Impossibility of a Theory of Gravity with Flat Spacetime 8
1.4.1 Gravitational Redshift . . . . . . . . . . . . . 8
1.4.2 A Scalar Theory of Gravity and the Perihelion
Shift . . . . . . . . . . . . . . . . . . . . . . . 8
2 Dierential Geometry I 13
2.1 Dierentiable Manifolds . . . . . . . . . . . . . . . . 13
2.2 Vectors and Tensors . . . . . . . . . . . . . . . . . . . 16
2.2.1 The tangent space . . . . . . . . . . . . . . . . 16
2.2.2 Coordinate basis and transformation under co-
ordinate changes . . . . . . . . . . . . . . . . 17
2.2.3 Tangent vectors, curves, and innitesimal trans-
formations . . . . . . . . . . . . . . . . . . . 19
2.2.4 Dual vectors and tensors . . . . . . . . . . . . 20
2.2.5 The metric . . . . . . . . . . . . . . . . . . . 23
3 Dierential Geometry II 25
i
CONTENTS ii
3.1 Connections, Covariant Derivatives and Geodesics . . 25
3.1.1 Linear Connections . . . . . . . . . . . . . . . 25
3.1.2 Parallel Transport and Geodesics . . . . . . . . 27
3.1.3 Normal Coordinates . . . . . . . . . . . . . . 28
3.1.4 Covariant derivative of tensor elds . . . . . . 29
3.2 Curvature . . . . . . . . . . . . . . . . . . . . . . . . 31
3.2.1 The Torsion and Curvature Tensors . . . . . . 31
3.2.2 The Bianchi Identities . . . . . . . . . . . . . 33
3.3 Riemannian Connections . . . . . . . . . . . . . . . . 34
3.3.1 Denition and Uniqueness . . . . . . . . . . . 34
3.3.2 Symmetries. The Einstein Tensor . . . . . . . 35
4 Physical Laws in External Gravitational Fields 38
4.1 Motion of Particles and Light . . . . . . . . . . . . . . 38
4.1.1 Freely Falling Particles . . . . . . . . . . . . . 38
4.1.2 Maxwells Equations . . . . . . . . . . . . . . 39
4.1.3 Geometrical Optics . . . . . . . . . . . . . . . 41
4.1.4 Redshift . . . . . . . . . . . . . . . . . . . . . 44
4.2 Energy-Momentum Conservation . . . . . . . . . . 44
4.2.1 Contracted Christoel Symbols . . . . . . . . 44
4.2.2 Covariant Divergences . . . . . . . . . . . . . 46
4.2.3 Charge Conservation . . . . . . . . . . . . . . 46
4.2.4 Energy-Momentum Conservation . . . . . . 47
4.3 The Newtonian Limit . . . . . . . . . . . . . . . . . . 48
4.3.1 Metric and Gravitational Potential . . . . . . . 48
4.3.2 Gravitational Light Deection . . . . . . . . . 49
5 Dierential Geometry III 52
5.1 The Lie Derivative . . . . . . . . . . . . . . . . . . . 52
5.1.1 The Pull-Back . . . . . . . . . . . . . . . . . 52
5.1.2 The Lie Derivative . . . . . . . . . . . . . . . 54
CONTENTS iii
5.1.3 Killing Vector Fields . . . . . . . . . . . . . . 57
5.2 Dierential Forms . . . . . . . . . . . . . . . . . . . . 58
5.2.1 Denition . . . . . . . . . . . . . . . . . . . . 58
5.2.2 The Exterior Derivative . . . . . . . . . . . . . 60
5.3 Integration . . . . . . . . . . . . . . . . . . . . . . . . 61
5.3.1 The Volume Form and the Codierential . . . 61
5.3.2 Example: Maxwells Equations . . . . . . . . 63
5.3.3 Integrals and Integral Theorems . . . . . . . . 64
6 Einsteins Field Equations 66
6.1 The physical meaning of the curvature tensor . . . . . 66
6.1.1 Congruences of time-like geodesics . . . . . . 66
6.1.2 The curvature tensor and the tidal eld . . . . . 68
6.2 Einsteins eld equations . . . . . . . . . . . . . . . . 70
6.2.1 Heuristic derivation . . . . . . . . . . . . . 70
6.2.2 Uniqueness . . . . . . . . . . . . . . . . . . . 71
6.2.3 Lagrangian formulation . . . . . . . . . . . . 72
6.2.4 The energy-momentum tensor . . . . . . . . . 74
6.2.5 Equations of motion . . . . . . . . . . . . . . 77
7 Weak Gravitational Fields 79
7.1 Linearised Theory of Gravity . . . . . . . . . . . . . . 79
7.1.1 The linearised eld equations . . . . . . . . . 79
7.1.2 Gauge transformations . . . . . . . . . . . . . 81
7.2 Nearly Newtonian gravity . . . . . . . . . . . . . . . . 82
7.2.1 Newtonian approximation of the metric . . . . 82
7.2.2 Gravitational lensing and the Shapiro delay . . 83
7.2.3 The gravitomagnetic eld . . . . . . . . . . . 84
7.3 Gravitational Waves . . . . . . . . . . . . . . . . . . . 87
7.3.1 Polarisation states . . . . . . . . . . . . . . . 87
7.3.2 Generation of gravitational waves . . . . . . . 89
CONTENTS iv
7.3.3 Energy transport by gravitational waves . . . . 91
8 The Schwarzschild Solution 93
8.1 Cartans structure equations . . . . . . . . . . . . . . . 93
8.1.1 The curvature and torsion forms . . . . . . . . 93
8.2 Stationary and static spacetimes . . . . . . . . . . . . 96
8.3 The Schwarzschild solution . . . . . . . . . . . . . . . 98
8.3.1 Form of the metric . . . . . . . . . . . . . . . 98
8.3.2 Connection and curvature forms . . . . . . . . 99
8.3.3 Components of the Ricci and Einstein tensors . 102
8.3.4 Solution of the vacuum eld equations . . . . . 102
8.3.5 Birkhos theorem . . . . . . . . . . . . . . . 104
9 Physics in the Schwarzschild Space-Time 106
9.1 Motion in the Schwarzschild space-time . . . . . . . . 106
9.1.1 Equation of motion . . . . . . . . . . . . . . . 106
9.1.2 Comparison to the Kepler problem . . . . . . . 109
9.1.3 The perihelion shift . . . . . . . . . . . . . . . 111
9.1.4 Light deection . . . . . . . . . . . . . . . . . 113
9.1.5 Spin precession . . . . . . . . . . . . . . . . . 114
10 Schwarzschild Black Holes 119
10.1 The singularity at r = 2m . . . . . . . . . . . . . . . . 119
10.1.1 Free fall towards the centre . . . . . . . . . . . 119
10.1.2 Problems with the Schwarzschild coordinates . 120
10.1.3 Curvature at r = 2m . . . . . . . . . . . . . . 122
10.2 The Kruskal Continuation of the Schwarzschild Space-
Time . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
10.2.1 Transformation to Kruskal coordinates . . . . . 122
10.2.2 Physical meaning of the Kruskal continuation . 126
10.2.3 Eddington-Finkelstein coordinates . . . . . . . 127
10.2.4 Redshift approaching the Schwarzschild radius 129
CONTENTS v
11 Charged, Rotating Black Holes 132
11.1 Metric and electromagnetic eld . . . . . . . . . . . . 132
11.1.1 The Reissner-Nordstrm solution . . . . . . . 132
11.1.2 The Kerr-Newman solution . . . . . . . . . . . 134
11.1.3 Schwarzschild horizon, ergosphere and Killing
horizon . . . . . . . . . . . . . . . . . . . . . 137
11.2 Motion near a Kerr black hole . . . . . . . . . . . . . 141
11.2.1 Keplers third law . . . . . . . . . . . . . . . . 141
11.2.2 Accretion ow onto a Kerr black hole . . . . . 142
11.3 Entropy and temperature of a black hole . . . . . . . . 144
12 Homogeneous, Isotropic Cosmology 148
12.1 The generalised Birkho theorem . . . . . . . . . . . 148
12.1.1 Spherically-symmetric space-times . . . . . . 148
12.1.2 Birkhos theorem . . . . . . . . . . . . . . . 150
12.2 Cosmological solutions . . . . . . . . . . . . . . . . . 152
12.2.1 Homogeneity and isotropy . . . . . . . . . . . 152
12.2.2 Spaces of constant curvature . . . . . . . . . . 153
12.2.3 Form of the metric and Friedmanns equations 155
12.2.4 Density evolution . . . . . . . . . . . . . . . . 158
12.2.5 Cosmological redshift . . . . . . . . . . . . . 159
12.2.6 Alternative forms of the metric . . . . . . . . . 160
13 Two Examples of Relativistic Astrophysics 162
13.1 Light Propagation . . . . . . . . . . . . . . . . . . . . 162
13.1.1 Arbitrary spacetimes . . . . . . . . . . . . . . 162
13.1.2 Specialisation to homogeneous and isotropic
spacetimes . . . . . . . . . . . . . . . . . . . 166
13.2 Relativistic Stellar Structure Equations . . . . . . . . . 169
13.2.1 The Tolman-Oppenheimer-Volko equation . . 169
13.2.2 The mass of non-rotating neutron stars . . . . . 171
CONTENTS vi
A Electrodynamics 173
A.1 Electromagnetic eld tensor . . . . . . . . . . . . . . 173
A.2 Maxwells equations . . . . . . . . . . . . . . . . . . 173
A.3 Lagrange density and energy-momentum tensor . . . . 174
B Summary of Dierential Geometry 176
B.1 Manifold . . . . . . . . . . . . . . . . . . . . . . . . 176
B.2 Tangent and Dual Spaces . . . . . . . . . . . . . . . . 176
B.3 Tensors . . . . . . . . . . . . . . . . . . . . . . . . . 177
B.4 Covariant Derivative . . . . . . . . . . . . . . . . . . 178
B.5 Parallel Transport and Geodesics . . . . . . . . . . . . 179
B.6 Torsion and Curvature . . . . . . . . . . . . . . . . . . 179
B.7 Pull-Back, Lie Derivative and Killing Vector Fields . . 181
B.8 Dierential Forms . . . . . . . . . . . . . . . . . . . . 182
B.9 Cartans Structure Equations . . . . . . . . . . . . . . 183
B.10 Dierential Operators and Integration . . . . . . . . . 184
Chapter 1
Introduction
1.1 The Idea behind General Relativity
There was no need for general relativity when Einstein started
working on it. There was no experimental data signalling any
failure of the Newtonian theory of gravity, except perhaps for the
minute advance of the perihelion of Mercurys orbit by 47
per
century, which researchers at the time tried to explain by pertur-
bations not included yet into the calculations of celestial mechan-
ics in the Solar System.
Einstein found general relativity essentially because he was
deeply dissatised with some of the concepts of the Newtonian
theory, in particular the concept of an inertial system, for which
no experimental demonstration could be given.
After special relativity, he was convinced quite quickly that try-
ing to build a relativistic theory of gravitation led to conclusions
which were in conict with experiments. Action at a distance is
impossible in special relativity because the absolute meaning of
space and time had to be given up. The most straightforward way
to combine special relativity with Newtonian gravity seemed to
start from Poissons equation for the gravitational potential and
to add time derivatives to it so as to make it relativistically invari-
ant.
However, it was then unclear how the law of motion should be
modied because, according to special relativity, energy and mass
are equivalent and thus the mass of a body should depend on its
position in a gravitational eld.
This led Einstein to a result which raised his suspicion. In Newto-
nian theory, the vertical acceleration of a body in a vertical gravi-
tational eld is independent of its horizontal motion. In a special-
1
CHAPTER 1. INTRODUCTION 2
relativistic extension of Newtons theory, this would no longer be
the case: the vertical gravitational acceleration would depend on
the kinetic energy of a body, and thus not be independent of its
horizontal motion.
This was in striking conict with experiment, which says that all
bodies experience the same gravitational acceleration. At this
point, the equivalence of inertial and gravitational mass struck
Einstein as a law of deep signicance. It became the heuristic
guiding principle in the construction of general relativity.
This line of thought leads to the fundamental concept of gen-
eral relativity. It says that it must be possible to introduce lo-
cal, freely-falling frames of reference in which gravity is locally
transformed away. The directions of motion of dierent freely-
falling reference frames will generally not be parallel: Einstein
elevators released at the same height above the Earths surface
but over dierent locations will fall towards the Earths centre
and thus approach each other.
This leads to the idea that space-time is a four-dimensional mani-
fold instead of the rigid, four-dimensional Euclidean space. As
will be explained in the following two chapters, manifolds can
locally be mapped onto Euclidean space. In a freely-falling ref-
erence frame, special relativity must hold, which implies that the
Minkowskian metric of special relativity must locally be valid.
The same operation must be possible in all freely-falling refer-
ence frames individually, but not globally, as is illustrated by the
example of the Einstein elevators falling towards the Earth.
Thus, general relativity considers the metric of the space-time
manifold as a dynamical eld. The necessity to match it with
the Minkowski metric in freely-falling reference frames means
that the signature of the metric must be (, +, +, +) or (+, , , ).
A manifold with a metric which is not positive denite is called
pseudo-Riemannian, or Lorentzian if the metric has the signature
of the Minkowski metric.
The lecture starts with an introductory chapter describing the fun-
damental characteristics of gravity, their immediate consequences
and the failure of a specially-relativistic theory of gravity. It then
introduces in two chapters the mathematical apparatus necessary
for general relativity, which are the basics of dierential geome-
try, i.e. the geometry on curved manifolds. After this necessary
mathematical digression, we shall return to physics when we in-
troduce Einsteins eld equations in chapter 4.
1.2 Fundamental Properties of Gravity
1.2.1 Scales
The rst remarkable property of gravity is its weakness . It is by
far the weakest of the four known fundamental interactions. To
see this, compare the gravitational and electrostatic forces acting
between two protons at a distance r. We have
gravity
electrostatic force
=
_
_
Gm
2
p
r
2
_
_
_
e
2
r
2
_
1
=
Gm
2
p
e
2
= 8.1 10
37
!
(1.1)
This leads to an interesting comparison of scales. In quantum
physics, a particle of mass m can be assigned the Compton wave-
length
=
mc
. (1.2)
We ask what the mass of the particle must be such that its gravi-
tational potential energy equals its rest mass mc
2
, and set
Gm
2
!
= mc
2
. (1.3)
The result is the Planck mass ,
m = M
Pl
=
_
c
G
= 2.2 10
5
g = 1.2 10
19
GeV
c
2
, (1.4)
which, inserted into (1.2), yields the Planck length
Pl
=
_
G
c
3
= 1.6 10
33
cm (1.5)
and the Planck time
t
Pl
=
Pl
c
=
_
G
c
5
= 5.3 10
44
s . (1.6)
As Max Planck noted already in 1900
1
, these are the only scales
for mass, length and time that can be assigned an objective mean-
ing.
The Planck mass is huge in comparison to the mass scales of el-
ementary particle physics. The Planck length and time are com-
monly interpreted as the scales where our classical description
of space-time is expected to break down and must be replaced by
an unknown theory combining relativity and quantum physics.
1

Uber irreversible Strahlungsvorg ange, Annalen der Physik 306 (1900) 69
Using the Planck mass, the ratio from (1.1) can be written as
Gm
2
p
e
2
=
1
m
2
p
M
2
Pl
, (1.7)
where = e
2
/c 1/137 is the ne-structure constant.
This suggests that gravity will dominate all other interactions
once the mass of an object is suciently large. A mass scale
important for the astrophysics of stars is set by the ratio
M
Pl
M
2
Pl
m
2
p
= 1.7 10
38
M
Pl
= 3.7 10
33
g , (1.8)
which is almost two solar masses. We shall see at the end of this
lecture that stellar cores of this mass cannot be stabilised against
gravitational collapse.
1.2.2 The Equivalence Principle
The observation that inertial and gravitational mass cannot be ex-
perimentally distinguished is a highly remarkable nding. It is by
no means obvious that the ratio between any force exerted on a
body and its consequential acceleration should have anything to
do with the ratio between the gravitational force and the bodys
acceleration in a gravitational eld.
The experimentally well-established fact that inertial and gravi-
tational mass are the same at least within our measurement accu-
racy was raised to a guiding principle by Einstein, the principle of
equivalence, which can be formulated in several dierent ways.
The weaker and less precise statement is that the motion of a test
body in a gravitational eld is independent of its mass and com-
position,
which can be cast into the more precise form that in an arbi-
trary gravitational eld, no local non-gravitational experiment
can distinguish a freely falling, non-rotating system from a uni-
formly moving system in absence of the gravitational eld.
The latter is Einsteins Equivalence Principle, which is the heuris-
tic guiding principle for the construction of general relativity.
It is important to note the following remarkable conceptual
advance: Newtonian mechanics starts from Newtons axioms,
which introduce the concept of an inertial reference frame, saying
that force-free bodies in inertial systems remain at rest or move
at constant velocity, and that bodies in inertial systems experi-
ence an acceleration which is given by the force acting on them,
divided by their mass.
Firstly, inertial systems are a deeply unsatisfactory concept be-
cause they cannot be realised in any strict sense. Approximations
to inertial systems are possible, but the degree to which a refer-
ence frame will approximate an inertial system will depend on the
precise circumstances of the experiment or the observation made.
Secondly, Newtons second axiom is, strictly speaking, circular in
the sense that it denes forces if one is willing to accept inertial
systems, while it denes inertial systems if one is willing to ac-
cept the relation between force and acceleration. A satisfactory,
non-circular denition of force is not given in Newtons theory.
The existence of inertial frames is postulated.
Special relativity replaces the rigid Newtonian concept of abso-
lute space and time by a space-time which carries the peculiar
light-cone structure demanded by the universal value of the light
speed. Newtonian space-time can be considered as the Cartesian
product R R
3
. An instant t R in time uniquely identies the
three-dimensional Euclidean space of all simultaneous events.
Of course, it remains possible in special relativity to dene si-
multaneous events, but the three-dimensional hypersurface from
four-dimensional Euclidean space R
4
identied in this way de-
pends on the motion of the observer relative to another observer.
Independent of their relative motion, however, is the light-cone
structure of Minkowskian space-time. The future light cone en-
closes events in the future of a point p in space-time which can
be reached by material particles, and its boundary is dened by
events which can be reached from p by light signals. The past
light cone encloses events in the past of p from which material
particles can reach p, and its boundary is dened by events from
which light signals can reach p.
Yet, special relativity makes use of the concept of inertial ref-
erence frames. Physical laws are required to be invariant under
transformations from the Poincar e group, which translate from
one inertial system to another.
General relativity keeps the light-cone structure of special relativ-
ity, even though its rigidity is given up: the orientation of the light
cones can vary across the space-time. Thus, the relativity of dis-
tances in space and time remains within the theory. However, it is
one of the great achievements of general relativity that it nally
replaces the concept of inertial systems by something else which
can be experimentally demonstrated: the principle of equivalence
replaces inertial systems by non-rotating, freely-falling frames of
reference.
1.3 Immediate Consequences of the Equiva-
lence Principle
Without any specic form of the theory, the equivalence principle
immediately allows us to draw conclusions on some of the conse-
quences any theory must have which is built upon it. We discuss
two here to illustrate its general power, namely the gravitational
redshift and gravitational light deection.
1.3.1 Gravitational Redshift
We enter an Einstein elevator which is at rest in a gravitational
eld at t = 0. The elevator is assumed to be small enough for the
gravitational eld to be considered as homogeneous within it, and
the (local) gravitational acceleration be g.
According to the equivalence principle, the downward gravita-
tional acceleration felt in the elevator cannot locally be distin-
guished from a constant upward acceleration of the elevator with
the same acceleration g. Adopting the equivalence principle, we
thus assume that the gravitational eld is absent and that the ele-
vator is constantly accelerated upward instead.
At t = 0, a photon is emitted by a light source at the bottom of
the elevator, and received some time t later by a detector at the
ceiling. The time interval t is determined by
h +
g
2
t
2
= ct , (1.9)
where h is the height of elevator. This equation has the solution
t
=
1
g
_
c
_
c
2
2gh
_
=
c
g
_
_
1
_
1
2gh
c
2
_
_

h
c
; (1.10)
the other branch makes no physical sense.
When the photon is received at the ceiling, the ceiling moves with
the velocity
v = gt =
gh
c
(1.11)
compared to the oor when the photon was emitted. The photon is
thus Doppler shifted with respect to its emission, and is received
with the longer wavelength

_
1 +
v
c
_
=
_
1 +
gh
c
2
_
. (1.12)
The gravitational acceleration is given by the gravitational poten-
tial through
g =
gh , (1.13)
where
h is the change in from the oor to the ceil-

ing of the elevator. Thus, the equivalence principle demands a
gravitational redshift of
z

c
2
. (1.14)
1.3.2 Gravitational Light Deection
Similarly, it can be concluded from the equivalence principle that
light rays should be curved in gravitational elds. To see this,
consider again the Einstein elevator from above which is at rest
in a gravitational eld g =
at t = 0.
As before, the equivalence principle asserts that we can consider
the elevator as being accelerated upwards with the acceleration g.
Suppose now that a horizontal light ray enters the elevator at t = 0
from the left and leaves it at a time t = w/c to the right, if w is
the horizontal width of the elevator.
As the light ray leaves the elevator, the elevators velocity has
increased to
v = gt =
w
c
(1.15)
such that, in the rest frame of the elevator, it leaves at an angle
=
v
c
=
w
c
2
(1.16)
downward from the horizontal because of the aberration due to
the nite light speed. Since the upward accelerated elevator cor-
responds to an elevator at rest in a downward gravitational eld,
this leads to the expectation that light will be deected towards
gravitational elds.
Although it is possible to construct theories of gravity which obey
the equivalence principle and do not lead to gravitational light
deection, the bending of light in gravitational elds is by now a
well-established experimental fact.
1.4 Impossibility of a Theory of Gravity with
Flat Spacetime
1.4.1 Gravitational Redshift
We had seen before that the equivalence principle implies a grav-
itational redshift, which has been demonstrated experimentally.
We must thus require from a theory of gravity that it does lead to
gravitational redshift.
Suppose we wish to construct a theory of gravity which retains
the Minkowski metric
. In such a theory, how ever it may look

in detail, the proper time measured by observers moving along a
world line x
() from
1
to
2
is
=
_

2
1
d
_
, (1.17)
where the minus sign under the square root appears because we
choose the signature of
to be (1, 1, 1, 1).
Now, let a light ray propagate from the oor to the ceiling of the
elevator in which we have measured gravitational redshift before.
Specically, let the light source shine between coordinates times
t
1
and t
2
. The emitted photons will propagate to the receiver at
the ceiling along world lines which may be curved, but must be
parallel because the metric is constant. The time interval within
which the photons arrive at the receiver must thus equal the time
interval t
2
t
1
within which they left the emitter. Thus there cannot
be gravitational redshift in a theory of gravity in at spacetime.
1.4.2 A Scalar Theory of Gravity and the Perihelion
Shift
Let us now try and construct a scalar theory of gravity starting
from the eld equation
= 4GT , (1.18)
where is the gravitational potential and T = T
is the trace of
the energy-momentum tensor. Note that is made dimensionless
here by dividing it by c
2
.
In the limit of weak elds and non-relativistic matter, this reduces
to Poissons equation
2
= 4G , (1.19)
since then the time derivatives in dAlemberts operator and the
pressure contributions to T can be neglected.
Let us further adopt the Lagrangian
[(x
, x
) = mc
_
(1 + ) , (1.20)
which is the ordinary Lagrangian of a free particle in special rel-
ativity, multiplied by the factor (1 + ). This is the only possible
Lagrangian that yields the right weak-eld (Newtonian) limit.
We can write the square root in (1.20) as
_
c
2
v
2
= c
_
1
2
, (1.21)
where

= v/c is the velocity in units of c. The weak-eld limit
of (1.20) for non-relativistic particles is thus
[(x
, x
) mc
2
_
1
v
2
2c
2
_
(1 + ) mc
2
+
m
2
v
2
mc
2
,
(1.22)
which is the right expression in Newtonian gravity.
The equations of motion can now be calculated inserting (1.20)
into Eulers equations,
d
dt
[
x
=
[
x
. (1.23)
On the right-hand side, we nd
[
x
= mc
2
_
1
2

x
. (1.24)
On the left-hand side, we rst have
[
x
=
1
c
[
=
mc
_
1
2
(1 + ) , (1.25)
and thus
d
dt
[
x
= mc(1 + )
_
_
1
2
+
(1 )
3/2
_
_
+
mc
_
1
2
. (1.26)
We shall now simplify these equations assuming that the potential
is static,

= 0, and that the motion is non-relativistic, < 1.
Then, (1.26) becomes
d
dt
[
v
mc(1 + )
_
1 +
2
2
_
m(1 + )
x , (1.27)
and (1.24) turns into
[
x
mc
2
_
1 +
2
2
_

mc
2
. (1.28)
The equation of motion thus reads, in this approximation,
(1 + )
x = c
2
(1.29)
or
x = c
2
(1 ) = c
2
_

2
2
_
. (1.30)
Compared to the equation of motion in Newtonian gravity, there-
fore, the potential is augmented by a quadratic perturbation.
For a static potential and non-relativistic matter, the potential is
given by Poissons equation.
We now proceed to work out the perihelion shift expected for
planetary orbits around the Sun in such a theory of gravity. As we
know from the discussion of Keplers problem in classical me-
chanics, the radius r and the polar angle of such orbits are char-
acterised by
dr
d
=
mr
2
L
_
2
m
(E V
L
) , (1.31)
where V
L
is the eective potential energy
V
L
= V +
L
2
2mr
2
, (1.32)
and L is the (orbital) angular momentum. Thus,
dr
d
=
r
2
L
_
2m(E V)
L
2
r
2
. (1.33)
The perihelion shift is the change in upon integrating once
around the orbit, or integrating twice from the perihelion radius
r
0
to the aphelion radius r
1
,
= 2
_
r
1
r
0
dr
d
dr
. (1.34)
Inverting (1.33), it is easily seen that (1.34) can be written as
= 2

L
_
r
1
r
0
dr
_
2m(E V)
L
2
r
2
. (1.35)
Now, we split the potential energy V into the Newtonian contri-
bution V
0
and a perturbation V < V and expand the integrand to
lowest order in V, which yields
2

L
_
r
1
r
0
dr
_
A
0
_
1
mV
A
0
_
(1.36)
where the abbreviation
A
0
2m(E V
0
)
L
2
r
2
(1.37)
was inserted for convenience.
We know that orbits in Newtonian gravity are closed, so that the
rst term in the integrand of (1.36) must vanish. Thus, we can
write
2

L
_
r
1
r
0
dr
mV
A
0
. (1.38)
Next, we transform the integration variable from r to , using that
dr
d
=
r
2
L
_
A
0
(1.39)
to leading order in V, according to (1.31). Thus, (1.38) can be
written as
=
L
2m
L
_

0
d r
2
V . (1.40)
Finally, we specialise the potential energy. Since Poissons equa-
tion for the gravitational potential remains valid, we have
V
0
= mc
2
=
GM
,
m
r
, (1.41)
where M
,
is the Suns mass, and following (1.30), the potential
perturbation is
V = mc
2
2
2
=
mc
2
2
V
2
0
m
2
c
4
=
G
2
M
2
,
m
2c
2
r
2
. (1.42)
Inserting this into (1.40) yields the perihelion shift
=
L
m
L
G
2
M
2
,
m
c
2
=
G
2
M
2
,
m
2
c
2
L
2
. (1.43)
The angular momentum L can be expressed by the semi-major
axis a and the eccentricity of the orbit,
L
2
= GM
,
m
2
a(1
2
) , (1.44)
which allows us to write (1.43) in the form
=
GM
,
ac
2
(1
2
)
. (1.45)
For the Sun, M
,
= 2 10
33
g, thus
GM
,
c
2
= 1.5 10
5
cm . (1.46)
For Mercury, a = 5.8 10
12
cm and the eccentricity = 0.2 can
be neglected because it appears quadratic in (1.45). Thus, we nd
= 8.1 10
8
radian = 0.017
(1.47)
per orbit. Mercurys orbital time is 88 d, i.e. it completes about
415 orbits per century, so that the perihelion shift predicted by the
scalar theory of gravity is
100
= 7
(1.48)
per century.
This turns out to be wrong: Mercurys perihelion shift is six times
as large, and not even the sign is right. Therefore, our scalar
theory of gravity fails in its rst comparison with observations.
Chapter 2
Dierential Geometry I
2.1 Dierentiable Manifolds
By the preceding discussion of how a theory of gravity may be
constructed which is compatible with special relativity, we are led
to the concept of a space-time which looks like Minkowskian
space-time locally, but may globally be curved. This concept is
cast into a mathematically precise form by the introduction of a
manifold.
An n-dimensional manifold M is a topological Hausdor space
with a countable base, which is locally homeomorphic to R
n
. This
means that for every point p M, an open neighbourhood U of p
exists together with a homeomorphism h which maps U onto an
open subset U
of R
n
,
h : U U
. (2.1)
A trivial example for an n-dimensional manifold is the R
n
itself,
on which h may be the identity map id. Thus, h is a specialisation
of a map from one manifold M to another manifold N, : M
N.
The homeomorphism h is called a chart or a coordinate system
in the language of physics. U is the domain or the coordinate
neighbourhood of the chart. The image h(p) of a point p M
under the chart h is expressed by the n real numbers (x
1
, . . . x
n
),
the coordinates of p in the chart h.
A set of charts h
is called an atlas of M if the domains of the

charts cover all of M.
An example for a manifold is the n-sphere S
n
, for which the two-
sphere S
2
is a particular specialisation. It cannot be continuously
mapped to R
2
, but pieces of it can.
13
CHAPTER 2. DIFFERENTIAL GEOMETRY I 14
We can embed the two-sphere into R
3
and describe it as the point
set
S
2
= (x
1
, x
2
, x
3
) R(x
1
)
2
+ (x
2
)
2
+ (x
3
)
2
= 1 ; (2.2)
then, the six half-spheres U
i
dened by
U
i
= (x
1
, x
2
, x
3
) S
2
x
i
> 0 (2.3)
can be considered as domains of maps which cover S
2
com-
pletely, and the charts can be the projections of the half-spheres
onto open disks
D
i j
= (x
i
, x
j
) R
2
(x
i
)
2
+ (x
j
)
2
< 1 , (2.4)
such as
f
+
1
: U
+
1
D
23
, f
+
1
(x
1
, x
2
, x
3
) = (x
2
, x
3
) . (2.5)
Thus, the charts f
i
, i 1, 2, 3, together form an atlas of the
two-sphere.
Let now h
and h
be two charts, and U
be
the intersection of their domains. Then, the composition of charts
h
h
1
exists and denes a map between two open sets in R

n
which describes the change of coordinates or a coordinate trans-
formon the intersection of domains U
and U
. An atlas of a man-
ifold is called dierentiable if the coordinate changes between all
its charts are dierentiable. A manifold, combined with a dier-
entiable atlas, is called a dierentiable manifold.
Using charts, it is possible to dene dierentiable maps between
manifolds. Let M and N be dierentiable manifolds of dimension
m and n, respectively, and : M N be a map from one mani-
fold to the other. Introduce further two charts h : M M
R
m
and k : N N
R
n
whose domains cover a point p M and its
image (p) N. Then, the combination k h
1
is a map from
the domain M
to the domain N
, for which it is clear from ad-

vanced calculus what dierentiability means. Unless stated oth-
erwise, we shall generally assume that coordinate changes and
maps between manifolds are C
, i.e. their derivatives of all or-

ders exist and are continuous.
We return to the two-sphere S
2
and the atlas of the six pro-
jection charts 7 = f
1
, f
2
, f
3
described above and investigate
whether it is dierentiable. For doing so, we arbitrarily pick the
charts f
+
3
and f
+
1
, whose domains are the northern and eastern
half-spheres, respectively, which overlap on the north-eastern
quarter-sphere. Let therefore p = (p
1
, p
2
, p
3
) be a point in the
domain overlap, then
f
+
3
(p) = (p
1
, p
2
) , f
+
1
(p) = (p
2
, p
3
) ,
f
+1
3
(p
1
, p
2
) =
_
p
1
, p
2
,
_
1 (p
1
)
2
(p
2
)
2
_
,
f
+
1
f
+1
3
(p
1
, p
2
) =
_
p
2
,
_
1 (p
1
)
2
(p
2
)
2
_
, (2.6)
which is obviously dierentiable. The same applies to all other
coordinate changes between charts of 7, and thus S
2
is a dier-
entiable manifold.
As an example for a dierentiable map, let : S
2
S
2
be a
map which rotates the sphere by 45
around its z axis. Let us

further choose a point p on the positive quadrant of S
2
in which
all coordinates are positive. We can combine with the charts f
+
3
and f
+
1
to dene the map
_
f
+
1
f
+1
3
_
(p
1
, p
2
) =
_
p
1
+ p
2
2
,
_
1 (p
1
)
2
(p
2
)
2
_
, (2.7)
which is also evidently dierentiable.
Finally, we introduce product manifolds in a straightforward way.
Given two dierentiable manifolds M and N of dimension m and
n, respectively, we can turn the product space MN consisting of
all pairs (p, q) with p M and q N into an (m+n)-dimensional
manifold as follows: if h : M M
and k : N N
are charts of
M and N, a chart h k can be dened on M N such that
h k : M N M
, (h k)(p, q) =
_
h(p), k(q)
_
. (2.8)
In other words, pairs of points from the product manifold are
mapped to pairs of points from the two open subsets M
R
m
and N
R
n
.
Many manifolds which are relevant in General Relativity can be
expressed as product manifolds of the Euclidean space R
m
with
spheres S
n
. For example, we can construct the product manifold
RS
2
composed of the real line and the two-sphere. Points on this
product manifold can be mapped onto R R
2
for instance using
the identical chart id on R and the chart f
+
3
on the northern half-
sphere of S
2
,
(id f
+
3
) : R S
2
R D
12
, (p, q) (p, q
2
, q
3
) ; (2.9)
2.2 Vectors and Tensors
2.2.1 The tangent space
Now we have essentially introduced ways how to construct local
coordinate systems, or charts, on a manifold, how to change be-
tween them, and how to use charts to dene what dierentiable
functions on the manifold are. We now proceed to see how vec-
tors can be introduced on a manifold.
Recall the denition of a vector space: a set V, combined with a
eld (K orper in German) F, an addition,
+ : V V V , (v, w) v + w , (2.10)
and a multiplication,
: F V V , (, v) v , (2.11)
is an F-vector space if V is an Abelian group under the addition
+ and the multiplication is distributive and associative. In other
words, a vector space is a set of elements which can be added and
multiplied with scalars (i.e. numbers from the eld F).
On a curved manifold, this vector space structure is lost because it
is not clear how vectors at dierent points on the manifold should
be added. However, it still makes sense to dene vectors locally
in terms of innitesimal displacements within suciently small
neighbourhood of a point p, which are tangential to the mani-
fold at p.
This leads to the concept of the tangential space of a manifold,
whose elements are tangential vectors, or directional derivatives
of functions. We denote by / the set of C
functions f from the

manifold into R.
An example for a function from the manifold S
2
R could be
the average temperature on Earth.
Generally, the tangent space V
p
of a dierentiable manifold M at
a point p is the set of derivations of /(p). A derivation v is a map
from /(p) into R,
v : /(p) R , (2.12)
which is linear,
v(f + g) = v( f ) + v(g) (2.13)
for f , g /(p) and , R, and satises the product rule (or
Leibniz rule)
v( f g) = v( f )g + f v(g) . (2.14)
Note that this denition immediately implies that the derivation
of a constant vanishes: let h / be a constant function, h(p) = c
for all p M, then v(h
2
) = 2cv(h) from (2.14) and v(h
2
) = v(ch) =
cv(h) from (2.13), which is possible only if v(h) = 0.
Together with the real numbers R and their addition and multipli-
cation laws, V
p
does indeed have the structure of a vector space,
with
(v + w)( f ) = v( f ) + w( f ) and (v)( f ) = v( f ) (2.15)
for v, w V
p
, f / and R.
2.2.2 Coordinate basis and transformation under co-
ordinate changes
We now construct a basis for the vector space V
p
, i.e. we provide a
complete set e
i
of linearly independent basis vectors. For doing
so, let h : U U
R
n
be a chart with p U and f /(p)
a function. Then, f h
1
: U
R is C
by denition, and we
introduce n vectors e
i
V
p
, 1 i n, by
e
i
( f ) =
x
i
( f h
1
)
h(p)
, (2.16)
where x
i
are the usual cartesian coordinates of R
n
.
The function ( f h
1
) is applied to the image h(p) R
n
of p
under the chart h, i.e. ( f h
1
) carries the function f from the
manifold M to the locally isomorphic manifold R
n
.
To show that these vectors span V
p
, we rst state that for any C
function F : U
R dened on an open neighbourhood U
of
the origin of R
n
, there exist n C
functions H
i
: U
R such that
F(x) = F(0) +
n
i=1
x
i
H
i
(x) . (2.17)
Note the equality! This is not a Taylor expansion. This is easily
seen using the identity
F(x) F(0) =
_
1
0
d
dt
F(tx
1
, . . . , tx
n
)dt
=
n
i=1
x
i
_
1
0
D
i
F(tx
1
, . . . , tx
n
)dt , (2.18)
where D
i
is the partial derivative with respect to the i-th argument
of F. Thus, it suces to set
H
i
(x) =
_
1
0
D
i
F(tx
1
, . . . , tx
n
)dt (2.19)
to prove (2.17). For x = 0 in particular, we nd
H
i
(0) =
_
1
0
F
x
i
0
dt =
F
x
i
0
. (2.20)
Now we substitute F = f h
1
and choose a chart h : U U
such that h(q) = x and h(p) = 0, i.e. x = h

1
(q). Then, we rst
obtain from (2.17)
f (q) = f (p) +
n
i=1
(x
i
h)(q) (H
i
h)(q) , (2.21)
and from (2.20)
H
i
(0) = (H
i
h)(p) =
x
i
( f h
1
)
h(p)
= e
i
( f ) . (2.22)
Next, we apply a tangent vector v V
p
to (2.20),
v( f ) = v[ f (p)] +
n
i=1
_
v(x
i
h) (H
i
h)
p
+ (x
i
h)
p
(H
i
h)
_
=
n
i=1
v(x
i
h) e
i
( f ) , (2.23)
where we have used that v applied to the constant f (p) vanishes,
that (x
i
h)(p) = 0 and that H
i
(0) = e
i
(p) according to (2.22).
Thus, setting v
i
= v(x
i
h), we nd that any v V
p
can be written
as a linear combination of the basis vectors e
i
. This also demon-
strates that the dimension of the tangent space V
p
equals that of
the manifold itself.
The basis e
i
, which is often simply denoted as /x
i
or
i
, is
called a coordinate basis of V
p
. If we choose a dierent chart h
instead of h, we obtain of course a dierent coordinate basis e
i
.
Denoting the i-th coordinate of the map h
h
1
with x
, the chain
rule applied to f h
1
= ( f h
1
) (h
h
1
) yields
e
i
=
n
j=1
x
j
x
i
e
j
=: J
j
i
e
j
. (2.24)
which shows that the two dierent coordinate bases are related by
the Jacobian matrix of the coordinate change, which has the ele-
ments J
j
i
= x
j
/x
i
. Its inverse has the elements J
i
k
= x
i
/x
j
.
This relates the present denition of a tangent vector to the tra-
ditional denition of a vector as a quantity whose components
transform as
v
i
= v(x
i
h
) =
n
j=1
v
j
e
j
(x
i
h
) =
n
j=1
x
i
x
j
v
j
= J
i
j
v
j
. (2.25)
Repeating the construction of a tangent space at another point
q M, we obtain a tangent space V
q
which cannot be identied
in any way with the tangent space V
p
given only the structure of
a dierentiable manifold that we have so far.
Consequently, a vector eld is dened as a map v : p v
p
which
assigns a tangent vector v
p
V
p
to every point p M. If we apply
a vector eld v to a C
function f , its result (v( f ))(p) is a number

for each point p. The vector eld is called smooth if the function
(v( f ))(p) is also smooth.
Since we can write v = v
i
i
with components v
i
in a local coordi-
nate neighbourhood, the function v( f ) is
(v( f ))(p) = v
i
(p)
i
f (p) , (2.26)
and thus it is called the derivative of f with respect to the vector
eld v.
2.2.3 Tangent vectors, curves, and innitesimal trans-
formations
We can give a geometrical meaning to tangent vectors as in-
nitesimal displacements on the manifold. First, we dene a
curve on M through p M as a map from an open interval I R
with 0 I into M,
: I M , (2.27)
such that (0) = p.
Next, we introduce a one-parameter group of dieomorphisms
t
as a C
map,
t
: R M M , (2.28)
such that for a xed t R,
t
: M M is a dieomorphism and,
for all t, s R,
t

s
=
t+s
. Note the latter requirement implies
that
0
is the identity map.
For a xed t,
t
maps points p M to other points q M in a
dierentiable way. As an example on the two-sphere S
2
,
t
could
be the map which rotates the sphere about an (arbitrary) z axis by
an angle parameterised by t, such that
t
is the rotation by zero
degrees.
We can now associate a vector eld v to
t
as follows: For a xed
point p M, the map
t
: R M is a curve as dened above
which passes through p at t = 0. This curve is called an orbit of
t
.
Then, we assign to p the tangent vector v
p
to this curve at t = 0.
Repeating this operation for all points p M denes a vector
eld v on M which is associated with
t
and can be considered as
the innitesimal generator of the transformations
t
.
In our example on S
2
, we x a point p on the sphere whose orbit
under the map
t
is a part of the latitude circle through p. The
tangent vector to this curve in p denes the local direction of
motion under the rotation expressed by
t
. Applying this to all
points p S
2
denes a vector eld v on S
2
.
Conversely, given a vector eld v on M, we can construct curves
through all points p M whose tangent vectors are v
p
. This
is most easily seen in a local coordinate neighbourhood, h(p) =
(x
1
, . . . x
n
), in which the curves are the unique solutions of the
system
dx
i
dt
= v
i
(x
1
, . . . x
n
) (2.29)
of ordinary, rst-order dierential equations. Thus, tangent vec-
tors can be identied with innitesimal transformations of the
manifold.
Given two vector elds v, w on M, we can dene the commutator
of the two elds as
[v, w]( f ) = vw( f ) wv( f ) , (2.30)
which can easily be shown to have the following properties
(where v, w, x are vector elds and f , g are functions on M):
[v + w, x] = [v, x] + [w, x]
[v, w] = [w, v]
[ f v, gw] = f g[v, w] + f v(g)w gw( f )v
[v, [w, x]] + [x, [v, w]] + [w, [x, v]] = 0 , (2.31)
where the latter equation is called the Jacobi identity. In coordi-
nates, we can write v = v
i
i
and w = w
j
j
, and the commutator
can be written
[v, w] =
_
v
i
i
w
j
w
i
i
v
j
_
j
. (2.32)
2.2.4 Dual vectors and tensors
We had introduced the tangent space V as the set of derivations of
functions / on M, which were certain linear maps from / into
R. We now introduce the dual vector space V
to V as the set of
linear maps
f : V R (2.33)
from V into R. Dening addition of elements of V
and their mul-

tiplication with scalars in the obvious way, V
obtains the struc-

ture of a vector space; the elements of V
are called dual vectors.

Let now f be a C
function on M and v V an arbitrary tangent

vector. Then, we dene the dierential of f by
d f : V R , d f (v) = v( f ) . (2.34)
It is obvious that, by denition of the dual space V
, d f is an
element of V
and thus a dual vector. Choosing a coordinate rep-

resentation, we see that
d f (v) = v
i
i
f . (2.35)
Specically letting f = x
i
be the ith coordinate function, we see
that
dx
i
(
j
) =
j
x
i
=
i
j
, (2.36)
which shows that the n-tuple e
i
= dx
i
forms a basis of V
,
which is called the dual basis to the basis e
i
=
i
of the tangent
space V.
Starting the same operation leading from V to the dual space V
with V
instead, we arrive at the double-dual vector space V
as
the vector space of all linear maps from V
R. It can be shown
that V
is isomorphic to V and can thus be identied with V.

Tensors T of rank (r, s) can now be dened as multilinear maps
T : V
. . . V
.,,.
r
V . . . V
.,,.
s
R , (2.37)
in other words, given r dual vectors and s tangent vectors, T re-
turns a real number, and if all but one vector or dual vector are
xed, the map is linear in the remaining argument. If a tensor of
rank (r, s) is assigned to every point p M, we have a tensor eld
of rank (r, s) on M.
According to this denition, tensors of rank (0, 1) are simply dual
vectors, and tensors of rank (1, 0) are elements of V
and can thus

be identied with tangent vectors.
For one specic example, a tensor of rank (1, 1) is a bilinear map
from V
V R. If we x a vector v V, T(, v) is a linear

map V
R and thus an element of V
, which can be identied

with a vector. In this way, given a vector v V, a tensor of
rank (1, 1) produces another vector V, and vice versa for dual
vectors. Thus, tensors of rank (1, 1) can be seen as linear maps
from V V, or from V
.
With the obvious rules for adding linear maps and multiplying
them with scalars, the set of tensors 7
r
s
of rank (r, s) attains the
structure of a vector space of dimension n
r+s
.
Given a tensor t of rank (r, s) and another tensor t
of rank (r
, s
),
we can construct a tensor of rank (r + r
, s + s
) called the outer

product t t
of t and t
by simply multiplying their results on the

r + r
dual vectors v
i
and the s + s
vectors v
j
, thus
(t t
)(v
1
, . . . , v
r+r
, v
1
, . . . , v
s+s
) = (2.38)
t(v
1
, . . . , v
r
, v
1
, . . . , v
s
)t
(v
r+1
, . . . , v
r+r
, v
s+1
, . . . , v
s+s
) .
In particular, it is thus possible to construct a basis for tensors of
type (r, s) out of the bases e
i
of the tangent space and e
j
of the
dual space by taking the tensor products. Thus, a tensor of rank
(r, s) can be written in the form
t = t
i
1
...i
r
j
1
... j
s
_
i
1
. . .
i
r
_
_
dx
j
1
. . . dx
i
s
_
, (2.39)
where the numbers t
i
1
...i
r
j
1
... j
s
are its components with respect to the
coordinate system h.
The transformation law (2.25) for the basis vectors under coordi-
nate changes implies that the tensor components transform as
t
i
1
...i
r
j
1
... j
s
= J
i
1
k
1
. . . J
i
r
k
r
J
l
1
j
1
. . . J
l
r
j
s
t
k
1
...k
r
l
1
...l
s
, (2.40)
a property which is often used to dene tensors in the rst place.
The contraction Ct of a tensor of rank (r, s) is a map which re-
duces both r and s by unity,
Ct : 7
r
s
7
r1
s1
, Ct = t(. . . , e
k
, . . . , e
k
, . . .) , (2.41)
where e
k
and e
k
are bases of the tangent and dual spaces, as
before, and the summation over all 1 k n is implied. The
basis vectors e
k
and e
k
are inserted as the ith and jth arguments
of the tensor t.
Expressing the tensor in a coordinate basis, we can write the ten-
sor in the form (2.39), and thus its contraction with respect to the
i
a
th and j
b
th arguments reads
Ct = t
i
1
...i
r
j
1
... j
s
dx
i
k
(
j
k
)
_
i
1
. . .
i
a
1

i
a
+1
. . .
i
r
_
_
dx
j
1
. . . dx
j
b
1
dx
j
b
+1
. . . dx
j
s
_
= t
i
1
...i
a1
i
k
i
a+1
...i
r
j
1
... j
b1
i
k
j
b+1
... j
s
_
i
1
. . .
i
a
1

i
a
+1
. . .
i
r
_
_
dx
j
1
. . . dx
j
b
1
dx
j
b
+1
. . . dx
j
s
_
. (2.42)
For a simple example, let v V be a tangent vector and w V
a
dual vector, and t = v w a tensor of rank (1, 1). Its contraction
results in a tensor of rank (0, 0), i.e. a real number, which is
Ct = C(v w) = dx
k
(v) w(
k
) = v
k
w
k
. (2.43)
At the same time, this can be written as
Ct = C(v w) = w(v) (2.44)
= (w
j
dx
j
)(v
i
i
) = w
j
v
i
dx
j
(
i
) = w
j
v
i
i
x
j
= w
j
v
i
j
i
= w
i
v
i
.
In that sense, the contraction amounts to applying the tensor (par-
tially) on itself.
2.2.5 The metric
We need some way to dene and measure the distance between
two points on a manifold. A metric is introduced as the innites-
imal squared distance between two neighbouring points on the
manifold.
We had seen above that tangent vectors v V
p
are closely related
to innitesimal displacements around a point p on the manifold.
Moreover, the innitesimal squared distance between two neigh-
bouring points p and q should be quadratic in the displacement
leading from one point to the other. Thus, we construct the metric
g as a bi-linear map
g : V V R , (2.45)
which means that the g is a tensor of rank (0, 2). The metric thus
assigns a number to two elements of a vector eld V on M. The
metric g thus assigns to two vectors their scalar product, which is
not necessarily positive. We abbreviate the scalar product of two
vectors v, w V by
g(v, w) (v, w) . (2.46)
In addition, we require that the metric be symmetric and non-
degenerate, which means
g(v, w) = g(w, v) v, w V
p
,
g(v, w) = 0 v V
p
w = 0 . (2.47)
In a coordinate basis, the metric can be written in components as
g = g
i j
dx
i
dx
j
(2.48)
or, omitting the tensor-product sign and introducing the line ele-
ment ds
2
,
ds
2
= g
i j
dx
i
dx
j
. (2.49)
Given a coordinate basis e
i
, the metric g can always be chosen
such that
g(e
i
, e
j
) = (e
i
, e
j
) =
i j
, (2.50)
where the number of positive and negative signs is independent
of the coordinate choice and is called the signature of the metric.
Positive-(semi)denite metrics, which have only positive signs,
are called Riemannian, and pseudo-Riemannian metrics have pos-
itive and negative signs.
Perhaps the most common pseudo-Riemannian metric is the
Minkowski metric known from special relativity, which can be
chosen to have the signature (, +, +, +) and has the line element
ds
2
= c
2
dt
2
+ (dx
1
)
2
+ (dx
2
)
2
+ (dx
3
)
2
. (2.51)
A metric with the same signature as for the space-time is called
Lorentzian.
Given a tangent vector v, the metric can also be seen as a linear
map from V into V
,
g : V V
, v g(, v) . (2.52)
This is an element of V
because it linearly maps vectors into

R. Since the metric is non-degenerate, the inverse map g
1
also
exists, and the metric can be used to establish a one-to-one cor-
respondence between vectors and dual vectors, and thus between
the tangent space V and its dual space V
.
Chapter 3
Dierential Geometry II
3.1 Connections, Covariant Derivatives and
Geodesics
3.1.1 Linear Connections
The curvature of the two-dimensional sphere S
2
can be described
by embedding the sphere into Euclidean space of the next-higher
dimension, R
3
. However, (as far as we know) there is no natural
embedding of our four-dimensional curved space-time into R
5
,
and thus we need a description of curvature which is intrinsic to
the manifold.
There is a close correspondence between the curvature of a man-
ifold and the transport of vectors along curves.
As we have seen before, the structure of a manifold does not
trivially allow to compare vectors which are elements of tangent
spaces at two dierent points. We will thus have to introduce ad-
ditional structure which allows us to meaningfully shift vectors
from one point to another on the manifold.
Even before we do so, it is intuitively clear how vectors can be
transported along closed paths in Euclidean space, R
3
say. There,
the vector arriving at the starting point after the transport will be
identical to the vector before the transport.
However, this will no longer be so on the two-sphere: starting
on the equator with a vector pointing north, we can shift it along
a meridian to the north pole, then back to the equator along a
dierent meridian, and nally back to its starting point on the
equator. There, it will point into a dierent direction than the
original vector.
25
CHAPTER 3. DIFFERENTIAL GEOMETRY II 26
Curvature can thus be dened from this misalignment of vectors
after transport along closed curves. In order to work this out, we
thus rst need some way for transporting vectors along curves.
We start by generalising the concept of a directional derivative
from R
n
by dening a linear connection or covariant dierenti-
ation on a manifold as a mapping which assigns to every pair
v, y of C
vector elds another vector eld

v
y which is bilinear
in v and y and satises
f v
y = f
v
y
v
( f y) = f
v
y + v( f )y , (3.1)
where f / is a C
function on M.
In a local coordinate basis e
i
, we can describe the linear connec-
tion by its action on the basis vectors,
i
(
j
)
k
i j

k
, (3.2)
where the n
3
numbers
k
i j
are called the Christoel symbols or
connection coecients of the connection in the given chart.
Elwin Christoel (18291900)
The Christoel symbols are not the components of a ten-
sor, which is seen from their transformation under coordinate
changes. Let x
i
and x
i
be two dierent coordinate systems, then
we have on the one hand, by denition,
a
(
b
) =
c
ab

c
=
c
ab
x
k
x
k
=
c
ab
J
k
c

k
, (3.3)
where J
k
c
is the Jacobian matrix of the coordinate transform as
dened in (2.24). On the other hand, the axioms (3.1) imply, with
f represented by the elements J
k
i
of the Jacobian matrix,
a
(
b
) =
J
i
a
i
(J
j
b
j
) = J
i
a
i
(J
j
b
j
)
= J
i
a
_
J
j
b
j
+
i
J
j
b
j
_
= J
i
a
J
j
b
k
i j

k
+ J
i
a
i
J
k
b
k
. (3.4)
Comparison of the two results (3.3) and (3.4) shows that
c
ab
J
k
c
= J
i
a
J
j
b
k
i j
+ J
i
a
i
J
k
b
, (3.5)
or, after multiplying with the inverse Jacobian matrix J
c
k
,
c
ab
= J
i
a
J
j
b
J
c
k

k
i j
+ J
c
k
J
i
a
i
J
k
b
. (3.6)
While the rst term on the right-hand side reects the tensor
transformation law (2.40), the second term diers from it.
Let now y and v be vector elds on M and w a dual vector eld,
then the covariant derivative y is a tensor eld of rank (1, 1)
which is dened by
y(v, w) = w[
v
(y)] . (3.7)
In a coordinate basis
i
, we write
y = y
i
i
and y = y
i
; j
dx
j

i
, (3.8)
and have
y
i
; j
= y(
j
, dx
i
) = dx
i
_
j
(y
k
k
)
_
= dx
i
_
y
k
, j
k
+ y
k
l
jk

l
_
= y
k
, j
i
k
+ y
k
l
jk

i
l
= y
i
, j
+ y
k
i
jk
. (3.9)
An ane connection is symmetric if
v
w
w
v = [v, w] , (3.10)
which a short calculation shows to be equivalent to the symmetry
property
k
i j
=
k
ji
(3.11)
of the Christoel symbols in a coordinate basis.
3.1.2 Parallel Transport and Geodesics
Given a linear connection, it is now straightforward to introduce
parallel transport . To begin, let : I M with I R a curve in
M with tangent vector (t). A vector eld v is called autoparallel
along if

v = 0 . (3.12)
The vector

v is the covariant derivative of v along , and it is
often denoted by

v =
Dv
dt
=
v
dt
. (3.13)
In the coordinate basis
i
, the covariant derivative along reads

v =
x
i
i
(v
j
j
) = x
i
i
(v
j
j
)
= x
i
_
v
j
i
(
j
) +
i
v
j
j
_
=
_
v
k
+
k
i j
x
i
v
j
_
k
, (3.14)
and if this is to vanish identically, (3.12) implies
v
k
+
k
i j
x
i
v
j
= 0 . (3.15)
The existence and uniqueness theorems for ordinary dierential
equations imply that (3.15) has a unique solution once v is given at
one point along the curve (t). The parallel transport of a vector
along a curve is then uniquely dened.
If the tangent vector of a curve is autoparallel along ,

= 0 , (3.16)
the curve is called a geodesic. In a local coordinate system, this
condition reads
x
k
+
k
i j
x
i
x
j
= 0 . (3.17)
3.1.3 Normal Coordinates
Geodesics allowthe introduction of a special coordinate systemin
the neighbourhood of a point p M. First, given a point p = (0)
and a vector (0) V
p
from the tangent space in p, the exis-
tence and uniqueness theorems for ordinary dierential equations
ensures that (3.17) has a unique solution, which implies that a
unique geodesic exists through p into the direction (0).
Obviously, if
v
(t) is a geodesic with initial velocity v = (0),
then
v
(at) is also a geodesic with initial velocity av = a (0), or
av
(t) =
v
(at) . (3.18)
Thus, given some neighbourhood U V
p
of p = 0, unique
geodesics (t) with t [0, 1] can be constructed through p into
any direction v U, i.e. such that (0) = p and (0) = v U.
Using this, we dene the exponential map at p,
exp
p
: v U
v
(1) , v exp
p
(v) =
v
(1) , (3.19)
which maps any vector v from U V
p
into a point along the
geodesic through p into direction v at distance t = 1.
Now, we choose a coordinate basis e
i
of V
p
and use the n ba-
sis vectors in the exponential mapping (3.19). Then, the neigh-
bourhood of p can uniquely be represented by the exponential
mapping along the basis vectors, exp
p
(x
i
e
i
), and the x
i
are called
normal coordinates.
Since exp
p
(tv) =
tv
(1) =
v
(t), the curve
v
(t) has the normal
coordinates x
i
= tv
i
, with v = v
i
e
i
. In these coordinates, x
i
is
linear in t, thus x
i
= 0, and (3.17) implies
k
i j
v
i
v
j
= 0 , (3.20)
and thus
k
i j
+
k
ji
= 0 . (3.21)
If the connection is symmetric as dened in (3.11), the connection
coecients must vanish,
k
i j
= 0 . (3.22)
Thus, at every point p M, local coordinates can uniquely be
introduced by means of the exponential map, the normal coordi-
nates, in which the coecients of a symmetric connection vanish.
This will turn out to be important.
3.1.4 Covariant derivative of tensor elds
Extending the concept of the covariant derivative to tensor elds,
we start with a simple tensor of rank (1, 1) which is the tensor
product of a vector eld v and a dual vector eld w,
t = v w , (3.23)
and we require that
x
satisfy the Leibniz rule,
x
(v w) =
x
v w + v
x
w , (3.24)
and commute with the contraction,
C [
x
(v w)] =
x
[w(v)] . (3.25)
We now contract (3.24) and use (3.25) to nd
C [
x
(v w)] = C (
x
v w) + C (v
x
w)
= w(
x
v) + (
x
w)(v)
=
x
[w(v)] = xw(v) , (3.26)
where (3.1) was used in the nal step (note that w(v) is a real-
valued function). Thus, we nd an expression for the covariant
derivative of a dual vector,
(
x
w)(v) = xw(v) w(
x
v) . (3.27)
Introducing the coordinate basis
i
, it is straightforward to show
(and a useful exercise!) that this result can be expressed as
(
x
w)(v) =
_
w
j,i

k
i j
w
k
_
x
i
v
j
. (3.28)
As before, we now dene the covariant derivative t of a tensor
eld as a map from the tensor elds of rank (r, s) to the tensor
elds of rank (r, s + 1),
: 7
r
s
7
r
s+1
(3.29)
by setting
(t)(w
1
, . . . , w
r
, v
1
, . . . , v
s
, v
s+1
)
(
v
s+1
t)(w
1
, . . . , w
r
, v
1
, . . . , v
s
) , (3.30)
where the v
i
are vector elds and the w
j
dual vector elds.
We nd a general expression for t, with t 7
r
s
, by taking the
tensor product of t with r vector elds v
i
and s dual vector elds
w
j
and applying
x
to the result, using the Leibniz rule,
x
(w
1
. . . w
r
v
1
. . . v
s
t)
=
x
w
1
. . . t + . . . w
1
. . .
x
v
1
. . . t
+ w
1
. . .
x
t , (3.31)
and then taking the total contraction, using that it commutes with
the covariant derivative, which yields
x
[t(w
1
, . . . , w
r
, v
1
, . . . , v
s
)]
= t(
x
w
1
, . . . , w
r
, v
1
, . . . , v
s
) + . . . + t(w
1
, . . . , w
r
, v
1
, . . . ,
x
v
s
)
+ (
x
t)(w
1
, . . . , w
r
, v
1
, . . . , v
s
) . (3.32)
Therefore, the covariant derivative
x
t of t is
(
x
t)(w
1
, . . . , w
r
, v
1
, . . . , v
s
)
= xt(w
1
, . . . , w
r
, v
1
, . . . , v
s
)
t(
x
w
1
, . . . , v
s
) . . . t(w
1
, . . . ,
x
v
s
) . (3.33)
We now work out the last expression for the covariant derivative
of a tensor eld in a local coordinate basis
i
and its dual basis
dx
j
for the special case of a tensor eld t of rank (1, 1). The
result for tensor elds of higher rank are then easily found by
induction.
We can write the tensor eld t as
t = t
i
j
(
i
dx
j
) , (3.34)
and the result of its application to v
1
=
a
and w
1
= dx
b
is
t(
a
, dx
b
) = t
i
j
dx
b
(
i
)dx
j
(
a
) = t
b
a
. (3.35)
Therefore, we can write (3.33) as
(
x
t)(
a
, dx
b
) (3.36)
= x
c
c
t
a
b
t
i
j
dx
j
(
x
a
)dx
b
(
i
) t
i
j
dx
j
(
a
)
x
dx
b
(
i
) .
The second term on the right-hand side is
t
i
j

b
i
x
c
(
a
)(x
j
) = x
c
t
b
j
k
ca

k
x
j
= x
c
t
b
j
j
ca
, (3.37)
and the third term is
t
i
j

j
a
_
(x
c
c
)(
j
x
b
) dx
b
(x
c
i
)
_
= x
c
t
i
a
k
c j

k
x
b
= x
c
t
i
a
b
ci
.
(3.38)
Summarising, the components of
x
t are
t
b
a;c
= t
b
a,c
+
b
ci
t
i
a

j
ca
t
b
j
, (3.39)
showing that the covariant indices are transformed with the neg-
ative, the contravariant indices with the positive Christoel sym-
bols.
In particular, the covariant derivatives of tensors of rank (0, 1)
(dual vectors w) and of tensors of rank (1, 0) (vectors v) have com-
ponents
w
i;k
= w
i,k

j
ki
w
j
,
v
i
;k
= v
i
,k
+
i
k j
v
j
. (3.40)
3.2 Curvature
3.2.1 The Torsion and Curvature Tensors
The torsion T maps two vector elds x and y into another vector
eld,
T : V V V , (3.41)
such that
T(x, y) =
x
y
y
x [x, y] ; (3.42)
obviously, the torsion vanishes if and only if the connection is
symmetric, cf. (3.10).
The torsion is antisymmetric,
T(x, y) = T(y, x) , (3.43)
and satises
T( f x, gy) = f g T(x, y) (3.44)
with arbitrary C
functions f and g.
The map
V
V V R , (w, x, y) w[T(x, y)] (3.45)

with w V
and x, y V is a tensor of rank (1, 2) called the

torsion tensor .
According to (3.45), the components of the torsion tensor in the
coordinate basis
i
and its dual basis dx
i
are
T
k
i j
= dx
k
_
T(
i
,
j
)
_
=
k
i j

k
ji
. (3.46)
The curvature R maps three vector elds x, y and v into a vector
eld,
R : V V V V , (3.47)
such that
R(x, y)v =
x
(
y
v)
y
(
x
v)
[x,y]
v . (3.48)
Since the covariant derivatives
x
and
y
represent the innites-
imal parallel transports along the integral curves of the vector
elds x and y, the curvature R quanties directly the change of
the vector v when it is parallel-transported around an innitesi-
mal, closed loop.
Exchanging x and y and using the antisymmetry of the commuta-
tor [x, y], we see that R is antisymmetric in x and y,
R(x, y) = R(y, x) . (3.49)
Also, if f , g and h are C
functions on M,
R( f x, gy)hv = f gh R(x, y)v , (3.50)
which follows immediately from the dening properties (3.1) of
the connection.
Obviously, the map
V
V V V R , (w, x, y, v) = w[R(x, y)v] (3.51)

with w V
and x, y, v V denes a tensor of rank (1, 3) called

the curvature tensor.
To work out the components of R in a local coordinate basis
i
,
we rst note that
i
(
k
) =
i
_
j
(
k
)
_
=
i
_
l
jk

l
_
=
l
jk,i

l
+
l
jk
m
il

m
. (3.52)
Interchanging i and j yields the coordinate expression for
y
(
x
v). Since the commutator of the basis vectors vanishes,
[
i
,
j
] = 0, the components of the curvature tensor are
R
i
jkl
= dx
i
[R(
k
,
l
)
j
]
=
i
l j,k

i
k j,l
+
m
l j
i
km

m
k j
i
lm
. (3.53)
The Ricci tensor is a contraction of the curvature tensor whose
components are
R
jl
= R
i
jil
=
i
l j,i

i
i j,l
+
m
l j
i
im

m
i j
i
lm
. (3.54)
3.2.2 The Bianchi Identities
The curvature and the torsion together satisfy the two Bianchi
identities. The rst Bianchi identity is
cyclic
_
R(x, y)z
_
=
cyclic
T[T(x, y), z] + (
x
T)(y, z) , (3.55)
where the sums extend over all cyclic permutations of the vectors
x, y and z. The second Bianchi identity is
cyclic
(
x
R)(y, z) + R[T(x, y), z] = 0 . (3.56)
They are important because they dene symmetry relations of the
curvature and the curvature tensor. In particular, for a symmetric
connection, T = 0 and the Bianchi identities reduce to
cyclic
_
R(x, y)z
_
= 0 ,
cyclic
(
x
R)(y, z) = 0 . (3.57)
Before we go on, we have to clarify the meaning of the covariant
derivatives of the torsion and the curvature. We have seen that T
denes a tensor eld

T of rank (1, 2). Given a dual vector eld
w V
, we dene the covariant derivative of the torsion T such

that
w
_
v
T(x, y)
_
= (
v

T)(w, x, y) . (3.58)
Using (3.33), we can write the right-hand side as
(
v

T)(w, x, y) = v

T(w, x, y)

T(
v
w, x, y)

T(w,
v
x, y)

T(w, x,
v
y) . (3.59)
The rst two terms on the right-hand side can be combined using
(3.27),
T(
v
w, x, y) =
v
w[T(x, y)] = vw[T(x, y)] w[
v
T(x, y)]
= v

T(w, x, y) w[
v
T(x, y)] , (3.60)
which yields
(
v

T)(w, x, y) = w[
v
T(x, y)]

T(w,
v
x, y)

T(w, x,
v
y) (3.61)
or, dropping the common argument w from all terms,
(
v
T)(x, y) =
v
[T(x, y)] T(
v
x, y) T(x,
v
y) . (3.62)
Similarly, we nd that
(
v
R)(x, y) =
v
[R(x, y)] R(
v
x, y) R(x,
v
y) R(x, y)
v
.
(3.63)
For symmetric connections, T = 0, the rst Bianchi identity is
easily proven. Its left-hand side reads
y
z
y
x
z +
y
z
x
z
y
x +
z
x
y
x
z
y

[x,y]
z
[y,z]
x
[z,x]
y
=
x
(
y
z
z
y) +
y
(
z
x
x
z) +
z
(
x
y
y
x)

[x,y]
z
[y,z]
x
[z,x]
y
=
x
[y, z]
[y,z]
x +
y
[z, x]
[z,x]
y +
z
[x, y]
[x,y]
z
= [x, [y, z]] + [y, [z, x]] + [z, [x, y]] = 0 , (3.64)
where we have used the relation (3.10) and the Jacobi identity
(2.31).
3.3 Riemannian Connections
3.3.1 Denition and Uniqueness
Up to now, the ane connection is not uniquely dened. We
shall now see that a unique connection can be introduced on each
pseudo-Riemannian manifold (M, g).
A connection is called metric if the parallel transport along any
smooth curve in M leaves the inner product of two autoparallel
vector elds x and y unchanged. This is the case if and only if the
covariant derivative of g vanishes,
g = 0 . (3.65)
Because of (3.33), this condition is equivalent to the Ricci identity
xg(y, z) = g(
x
y, z) + g(y,
x
z) , (3.66)
where x, y, z are vector elds.
It can now be shown that a unique connection can be introduced
on each pseudo-Riemannian manifold such that is symmetric or
torsion-free, and metric, i.e. g = 0. Such a connection is called
the Riemannian or Levi-Civita connection.
Suppose rst that such a connection exists, then (3.66) and the
symmetry of allow us to write
xg(y, z) = g(
y
x, z) + g([x, y], z) + g(y,
x
z) . (3.67)
Taking the cyclic permutations of this equation, summing the sec-
ond and the third and subtracting the rst (3.67), we obtain the
Koszul formula
2g(
z
y, x) = xg(y, z) + yg(z, x) + zg(x, y) (3.68)
+ g([x, y], z) g([y, z], x) g([z, x], y) .
Since the right-hand side is independent of , and g is non-
degenerate, this result implies the uniqueness of . The existence
of an ane, symmetric and metric connection can be proven by
explicit construction.
The Christoel symbols for a Riemannian connection can now
be determined specialising the Koszul formula (3.68) to the basis
vectors
i
of a local coordinate system. We choose x =
k
, y =
j
and z =
i
and use that their commutator vanishes, [
i
,
j
] = 0,
and that g(
i
,
j
) = g
i j
.
Then, (3.68) implies
2g(
j
,
k
) =
k
g
i j
+
j
g
ik
+
i
g
jk
, (3.69)
thus
g
mk
m
i j
=
1
2
_
g
ik, j
+ g
jk,i
g
i j,k
_
. (3.70)
If (g
i j
) denotes the matrix inverse to (g
i j
), we can write
l
i j
=
1
2
g
lk
_
g
ik, j
+ g
jk,i
g
i j,k
_
. (3.71)
3.3.2 Symmetries. The Einstein Tensor
In addition to (3.49), the curvature tensor of a Riemannian con-
nection has the following symmetry properties:
(R(x, y)v, w) = (R(x, y)w, v) , (R(x, y)v, w) = (R(v, w)x, y) .
(3.72)
The rst of these relations is easily seen noting that the antisym-
metry is equivalent to
(R(x, y)v, v) = 0 . (3.73)
From the denition of R and the antisymmetry (3.49), we rst
have
(v, R(x, y)v) = (v,
x
y
v
y
x
v
[x,y]
v) . (3.74)
Replacing y by
y
v and z by v, the Ricci identity (3.66) allows us
to write
(v,
x
y
v) = x(
y
v, v) (
y
v,
x
v) (3.75)
and, replacing x by y and both y and z by v,
(
y
v, v) =
1
2
y(v, v) . (3.76)
Hence, the rst two terms on the right-hand side of (3.74) yield
(v,
x
y
v
y
x
v) = (v, x(
y
v, v) y(
x
v, v))
=
1
2
(v, xy(v, v) yx(v, v))
=
1
2
(v, [x, y](v, v)) . (3.77)
By (3.76), this is the negative of the third term on the right-hand
side of (3.74), which proves (3.73).
The symmetries (3.49) and (3.72) imply
R
i jkl
= R
jikl
= R
i jlk
, R
i jkl
= R
kli j
, (3.78)
where R
i jkl
g
im
R
m
jkl
. Of the 4
4
= 256 components of the Rie-
mann tensor in four dimensions, the rst symmetry relation (3.78)
leaves 66 = 36 independent components, while the second sym-
metry relation (3.78) reduces their number to 6+5+4+3+2+1 =
21.
Bernhard Riemann (18261866)
In a coordinate basis, the Bianchi identities (3.57) for the curva-
ture tensor of a Riemannian connection read
( jkl)
R
i
jkl
= 0 ,
(klm)
R
i
jkl;m
= 0 , (3.79)
where ( jkl) denotes the cyclic permutations of the indices en-
closed in parentheses. In four dimensions, the rst Bianchi iden-
tity establishes one further relation between the components of
the Riemann tensor which is not covered yet by the symmetry
relations (3.78), namely
R
0123
+ R
0231
+ R
0312
= 0 , (3.80)
and thus leaves 20 independent components of the Riemann ten-
sor. These are
_
_
R
0101
R
0102
R
0103
R
0112
R
0113
R
0202
R
0203
R
0212
R
0213
R
0223
R
0303
R
0312
R
0313
R
0323
R
1212
R
1213
R
1223
R
1313
R
1323
R
2323
_
_
, (3.81)
where R
0123
is determined by (3.80).
Using the symmetries (3.78) and the second Bianchi identity from
(3.79), we can obtain an important result. We rst contract
R
i
jkl;m
+ R
i
jlm;k
+ R
i
jmk;l
= 0 (3.82)
by multiplying with
k
i
and use the symmetry relations (3.78) to
nd
R
jl;m
+ R
i
j ml;i
R
jm;l
= 0 . (3.83)
Next, we contract again by multiplying with g
jm
, which yields
R
m
l;m
+ R
i
l;i
R
;l
= 0 , (3.84)
where R
i j
are the components of the Ricci tensor and R = R
i
i
is the Ricci scalar or the scalar curvature. Renaming dummy
indices, the last equation can be brought into the form
_
R
i
j

R
2
i
j
_
;i
= 0 , (3.85)
which is the contracted Bianchi identity. Moreover, the Ricci ten-
sor can easily be shown to be symmetric,
R
i j
= R
ji
. (3.86)
We nally introduce the symmetric Einstein tensor by
G
i j
R
i j

R
2
g
i j
, (3.87)
which has vanishing divergence because of the contracted Bianchi
identity,
G
i
j;i
= 0 . (3.88)
Chapter 4
Physical Laws in External
Gravitational Fields
4.1 Motion of Particles and Light
4.1.1 Freely Falling Particles
In special relativity, the action of a free particle was
S = mc
_
b
a
ds = mc
_
b
a
_
dx
dx
, (4.1)
which can be rewritten as follows: rst, we parameterise the tra-
jectory of the particle as a curve () and write the four vec-
tor dx = d. Second, we introduce the Minkowski metric
= diag(1, 1, 1, 1) and use the notation (2.46)

( , ) = ( , ) (4.2)
to cast the action into the form
S = mc
_
b
a
_
( , ) d . (4.3)
In general relativity, the metric is replaced by the dynamic met-
ric g. We thus expect that the motion of a free particle will be
described by the action
S = mc
_
b
a
_
g( , ) d = mc
_
b
a
_
( , ) d . (4.4)
To see what this equation implies, we now carry out the variation
of S and set it to zero,
S = mc
_
b
a
_
g( , ) d = 0 . (4.5)
38
CHAPTER4. PHYSICAL LAWS INEXTERNAL GRAVITATIONAL FIELDS39
If we parameterise the curve by the proper time, we see from
=
_
b
a
d
_
( , ) (4.6)
that the integrand must equal unity, and thus the four-velocity
must satisfy
( , ) = 1 . (4.7)
This allows us to write the variation (4.5) as
S =
mc
2

_
b
a
d
_
g
,
x
+ 2g
_
= 0 . (4.8)
We can integrate the second term by parts to nd
2
_
b
a
d g
= 2
_
b
a
d
d
d
_
g
_
x
= 2
_
b
a
d
_
g
,
x
+ g
_
x
. (4.9)
Interchanging the summation indices and and inserting the
result into (4.8) yields
_
g
,
2g
,
_
x
2g
= 0 (4.10)
or, after multiplication with g
,
x
+
1
2
g
_
2g
,
g
,
_
x
= 0 . (4.11)
Comparing the result (4.11) to (3.17) and using the symmetry of
the Christoel symbols (3.69), we see that trajectories extremis-
ing the action (4.4) are geodesic curves. We thus arrive at the
conclusion that freely falling particles follow geodesic curves.
4.1.2 Maxwells Equations
As an example for how physical laws can be carried from special
to general relativity, we now formulate the equations of classical
electrodynamics in a gravitational eld.
In terms of the eld tensor F, Maxwells equations read
= 0 ,
=
4
c
j
, (4.12)
where j
is the current four-vector. The homogeneous equations

are identically satised introducing the potentials A
, in terms of
which the eld tensor is
F
. (4.13)
We can impose a gauge condition, such as the Lorentz gauge
= 0 , (4.14)
which allows to write the inhomogeneous Maxwell equation in
the form
=
4
c
j
. (4.15)
Indices are raised with the (inverse) Minkowski metric,
F
. (4.16)
Finally, the equation for the Lorentz force can be written as
m
du
d
=
e
c
F
, (4.17)
where u
= dx
/d is the four-velocity.
Moving to general relativity, we rst replace the partial by covari-
ant derivatives in Maxwells equations and nd
= 0 ,
=
4
c
j
. (4.18)
However, it is easy to see that the identity
+ cyclic
+ cyclic (4.19)
holds because of the antisymmetry of the eld tensor F and the
symmetry of the connection .
Indices have to be raised with the inverse metric g
1
now,
F
= g
. (4.20)
Equation (4.17) for the Lorentz force has to be replaced by
m
_
du
d
+
_
=
e
c
F
. (4.21)
We thus arrive at the following general rule: In presence of
a gravitational eld, the physical laws of special relativity are
changed simply by substituting the covariant derivative for the
partial derivative, , by raising indices with g
instead of
and by lowering them with g
instead of
, and by replac-
ing the motion of free particles along straight lines by the motion
along geodesics. Note that this is a rule, not a law, because ambi-
guities may occur in presence of second derivatives, as we shall
see shortly.
We can impose a gauge condition such as the generalised Lorentz
gauge
= 0 , (4.22)
but now the inhomogeneous wave equation (4.15) becomes more
complicated. We rst note that
F
(4.23)
identically. Inserting (4.23) into the inhomogeneous Maxwell
equation rst yields
) =
4
c
j
, (4.24)
but now the term
does not vanish despite the Lorentz

gauge condition because the covariant derivatives do not com-
mute. Instead, we have to use
_
_
A
= R
(4.25)
by denition of the curvature tensor, and thus
+ R
= R
(4.26)
inserting the Lorentz gauge condition. Thus, the inhomogeneous
wave equation in general relativity reads
=
4
c
j
. (4.27)
Had we started directly from the wave equation (4.15) from spe-
cial relativity, we would have missed the curvature term! This il-
lustrates the ambiguities that may occur applying the rule
when second derivatives are involved.
4.1.3 Geometrical Optics
We now study how light rays propagate in a gravitational eld.
As in geometrical optics, we assume that the wavelength of the
electromagnetic eld is very much smaller compared to the scale
L of the space within which we study light propagation. In a grav-
itational eld, which causes space-time to curve on another scale
R, we have to further assume that is also very small compared
to R, thus
< L and < R . (4.28)
Consequently, we introduce an expansion of the four-potential in
terms of a small parameter /min(L, R) < 1 and write the
four-potential as a product of a slowly varying amplitude and a
quickly varying phase,
A
= T
_
(a
+ b
)e
i/
_
, (4.29)
where the amplitude is understood as the two leading-order terms
in the expansion, and the phase carries the factor
1
because
it is inversely proportional to the wave length. The real part is
introduced because the amplitude is complex.
As in ordinary geometrical optics, the wave vector is the gradient
of the phase, thus k
. We further introduce the scalar am-

plitude a (a
)
1/2
, where the asterisk denotes complex conju-
gation, and the polarisation vector e
/a.
We rst impose the Lorentz gauge and nd the condition
T
__
(a
+ b
) + (a
+ b
)
i
_
e
i/
_
= 0 . (4.30)
To leading order (
1
), this implies
k
= 0 , (4.31)
which shows that the wave vector is perpendicular to the polari-
sation vector. The next-higher order yields
+ ik
= 0 . (4.32)
Next, we insert the ansatz (4.29) into Maxwells equation (4.27)
in vacuum, i.e. setting the right-hand side to zero. This yields
T
__
(a
+ b
) +
2i
(a
+ b
)
+
i
(a
+ b
2
k
(a
+ b
)
R
(a
+ b
)
_
e
i/
_
= 0 . (4.33)
To leading order (
2
), this implies
k
= 0 , (4.34)
which yields the general-relativistic eikonal equation
g
= 0 . (4.35)
Trivially, (4.34) implies
0 =
(k
) = 2k
. (4.36)
Recall that the wave vector is the gradient of the scalar phase .
The second covariant derivatives of commute,
(4.37)
as is easily seen by direct calculation, using the symmetry of the
connection. Thus,
, (4.38)
which, inserted into (4.36), leads to
k
= 0 or
k
k = 0 . (4.39)
In other words, in the limit of geometrical optics, Maxwells
equations imply that light rays follow null geodesics.
The next-higher order (
1
) gives
2i
_
k
+
1
2
a
_
k
= 0 (4.40)
and, with (4.34), this becomes
k
+
1
2
a
= 0 . (4.41)
We use this to derive a propagation law for the amplitude a. Ob-
viously, we can write
2ak
a = 2ak
a = k
(a
2
) = k
(a
+ a
) .
(4.42)
By (4.41), this can be transformed to
k
(a
+ a
) =
1
2
(a
+ a
) = a
2
.
(4.43)
Combining (4.43) with (4.42) yields
k
a =
a
2
, (4.44)
which shows how the amplitude is transported along light rays:
the change of the amplitude in the direction of the wave vector is
proportional to the negative divergence of the wave vector, which
is a very intuitive result.
Finally, we obtain a law for the propagation of the polarisation.
Using a
= ae
in (4.41) gives
0 = k
(ae
) +
1
2
ae
= ak
+ e
_
k
a +
a
2
_
= ak
, (4.45)
where (4.44) was used in the last step. This shows that
k
= 0 or
k
e = 0 , (4.46)
in other words, the polarisation is parallel-transported along light
rays.
4.1.4 Redshift
Suppose now that a light source moving with four-velocity u
s
is
sending a light ray to an observer moving with four-velocity u
o
,
and another light ray after a proper-time interval
s
. The phases
of the rst and second light rays be
1
and
2
=
1
+ , respec-
tively.
Clearly, the phase dierence measured at the source and at the
observed must equal, thus
u
s
(
)
s
s
= = u
o
(
)
o
o
. (4.47)
Using k
, and assigning frequencies

s
and
o
to the light
rays which are indirecly proportional to the time intervals
s
and
o
, we nd
s
=
o
=
(k, u
o
)
(k, u
s
)
, (4.48)
which gives the combined gravitational redshift and the Doppler
shift of the light rays. Any distinction between Doppler shift and
gravitational redshift has no invariant meaning in general relativ-
ity.
4.2 Energy-Momentum Conservation
4.2.1 Contracted Christoel Symbols
From (3.69), we see that the contracted Christoel symbol can be
written as
=
1
2
g
_
g
,
+ g
,
g
,
_
. (4.49)
Exchanging the arbitrary dummy indices and and using the
symmetry of the metric, we can simplify this to
=
1
2
g
g
,
. (4.50)
We continue by using Cramers rule from linear algebra, which
states that the inverse of a matrix A has the components
(A
1
)
i j
=
C
ji
det A
, (4.51)
where the C
ji
are the cofactors (signed minors) of the matrix A.
Thus, the cofactors are
C
ji
= det A(A
1
)
i j
. (4.52)
The determinant of A can be expressed using the cofactors as
det A =
n
j=1
C
ji
A
ji
, (4.53)
where n is the dimension of the (square) matrix. This follows
after multiplying (4.51) with the matrix A
jk
.
The metric g is expressed by the matrix g
, its inverse by g
.
Cramers rule then implies that the cofactors of g
are g g
, and
we can specialise (4.53) to write
g =
3
=0
g
gg
= g
3
=0
g
, (4.54)
which allows us to write
g
g
= gg
(4.55)
and thus
g
,
=
g
g
g
,
= gg
g
,
. (4.56)
Comparing this with the expression (4.50) for the contracted
Christoel symbol, we see that
gg
g
,
= 2g
=
1
2
g
g
,
=
1
2g
g
,
=
1
g , (4.57)
which is a very convenient expression for the contracted Christof-
fel symbol, as we shall see.
4.2.2 Covariant Divergences
The covariant derivative of a vector with components v
has the
components
. (4.58)
Using (4.57), the covariant divergence of this vector can thus be
written
+
1
g
v
g =
1
gv
) . (4.59)
Similarly, for a tensor A of rank (2, 0) with components A
, we
have
. (4.60)
Again, by means of (4.57), we can combine the rst and third
terms on the right-hand side to write
=
1
gA
) +
. (4.61)
If the tensor A
is antisymmetric, the second term on the right-

hand side vanishes because then the symmetric Christoel sym-
bol
is contracted with the antisymmetric tensor A
. If A
is symmetric, however, this nal term remains, with important

consequences.
4.2.3 Charge Conservation
Since the electromagnetic eld tensor F
is antisymmetric,
(4.61) implies
=
1
gF
) . (4.62)
On the other hand, replacing the vector v
by
in (4.59), we
see that
=
1
) =
1
gF
) ,
(4.63)
where we have used (4.62) in the nal step. But the partial deriva-
tives commute, so that once more the antisymmetric tensor F
is
contracted with the symmetric symbol
. Thus, the result must

vanish, allowing us to conclude
= 0 . (4.64)
However, by Maxwells equation (4.18),
=
4
c

, (4.65)
which implies, by (4.59)
gj
) = 0 . (4.66)
This is the continuity equation the electric four-current, implying
charge conservation. We thus see that the antisymmetry of the
electromagnetic eld tensor is necessary for charge conservation.
4.2.4 Energy-Momentum Conservation
In special relativity, energy-momentum conservation can be
expressed by the vanishing four-divergence of the energy-
momentum tensor T,
= 0 . (4.67)
For example, the energy-momentum tensor of the electromag-
netic eld is, in special relativity
T
=
1
4
_
F
+
1
4
_
, (4.68)
and for = 0, the vanishing divergence (4.67) yields the energy
conservation equation
t
_
E
2
+

B
2
8
_
_
+

_
c
4
(
E

B)
_
= 0 , (4.69)
in which the Poynting vector

S = (c/4)
E

B represents the
energy current density.
According to our general rule for moving results from special rel-
ativity to general relativity, we can replace the partial derivative
in (4.67) by the covariant derivative,
= 0 , (4.70)
and obtain an equation which is covariant and thus valid in
all reference frames. Moreover, we would have to replace the
Minkowski metric in (4.68) by the metric g if we wanted to con-
sider the energy-momentum tensor of the electromagnetic eld.
From our general result (4.61), we know that we can rephrase
(4.70) as
1
gT
) +
= 0 . (4.71)
If the second term on the left-hand side was absent, this equa-
tion would imply a conservation law. It remains there, however,
because the energy-momentum tensor is symmetric. In presence
of this term, we cannot convert (4.71) to a conservation law any
more. This result expresses the fact that energy is not conserved
in general relativity. This is not surprising because energy can
now be exchanged with the gravitational eld.
4.3 The Newtonian Limit
4.3.1 Metric and Gravitational Potential
Finally, we want to see under which conditions on the metric the
Newtonian limit for the equation of motion in a gravitational eld
is reproduced, which is
x =
(4.72)
to very high precision in the Solar System.
We rst restrict the gravitational eld to be weak and to vary
slowly with time. This implies that the Minkowski metric of at
space is perturbed by a small amount,
g
+ h
, (4.73)
with h
< 1.
Moreover, we restrict the consideration to bodies moving much
slower than the speed of light, such that
dx
i
d
<
dx
0
d
1 . (4.74)
Under these conditions, the geodesic equation reduces to
d
2
x
i
c
2
dt
2

d
2
x
i
d
2
=
i
dx
d
dx
d

i
00
. (4.75)
By denition (3.69), the remaining Christoel symbols read
i
00
= h
0i,0

1
2
h
00,i

1
2
h
00,i
(4.76)
due to the assumption that the metric changes slowly in time so
that its time derivative can be ignored compared to its spatial
derivatives. Equation (4.75) can thus be reduced to
d
2
x
dt
2

c
2
2
h
00
, (4.77)
which agrees with the Newtonian equation of motion (4.72) if we
identify
h
00

2
c
2
+ const. . (4.78)
The constant can be set to zero because both the deviation from
the Minkowski metric and the gravitational potential vanish at
large distance from the source of gravity. Therefore, the metric in
the Newtonian limit has the 00 element
g
00
1
2
c
2
. (4.79)
4.3.2 Gravitational Light Deection
Based on this result, we might speculate that metric in Newtonian
approximation could be written as
g = diag
_
_
1 +
2
c
2
_
, 1, 1, 1
_
. (4.80)
We shall now work out the gravitational light deection by the
Sun in this metric, which was one of the rst observational tests
of general relativity.
Since light rays propagate along null geodesics, we have
k
k = 0 or k
, (4.81)
where k = (/c,
k) is the wave four-vector which satises

(k, k) = 0 thus = c
k , (4.82)
which is the ordinary dispersion relation for electromagnetic
waves in vacuum. We introduce the unit vector e in the direction
of
k by
k =
ke = e/c.
Assuming that the gravitational potential does not vary with
time,
0
= 0, the only non-vanishing Christoel symbols of the
metric (4.80) are
0
0i
=
1
c
2
i
=
i
00
. (4.83)
For = 0, (4.81) yields
_
1
c
t
+e

_
+
e

c
2
= 0 , (4.84)
which shows that the frequency changes with time only because
the light path can run through a spatially varying gravitational
potential. Thus, if the potential is constant in time, the frequencies
of the incoming and the outgoing light must equal.
Using this result, the spatial components of (4.81) read
_
1
c
t
+e

_
e =
de
cdt
=
1
c
2
_
e(e

)
_
=
1
c
2
;
(4.85)
in other words, the total time derivative of the unit vector in the di-
rection of the light ray equals the negative perpendicular gradient
of the gravitational potential.
For calculating the light deection, we need to know the total
change in e as the light ray passes the Sun. This is obtained by
integrating (4.85) along the actual (curved) light path, which is
quite complicated. However, due to the weakness of the gravita-
tional eld, the deection will be very small, and we can evaluate
the integral along the unperturbed (straight) light path.
We choose a coordinate system centred on the Sun and rotated
such that the light ray propagates parallel to the z axis from
to at an impact parameter b. Outside the Sun, its gravitational
potential is
c
2
=
GM
,
c
2
r
=
GM
,
c
2
b
2
+ z
2
. (4.86)
The perpendicular gradient of is
b
e
b
=
GM b
c
2
(b
2
+ z
2
)
3/2
e
b
, (4.87)
where e
b
is the radial unit vector in the x-y plane from the Sun to
the light ray.
Thus, the deection angle is
e = e
b
_

dz
GM
c
2
(b
2
+ z
2
)
3/2
=
2GM
c
2
b
e
b
. (4.88)
Evaluating (4.88) at the rim of the Sun, we insert M
,
= 2 10
33
g
and R
,
= 7 10
10
cm to nd
e = 0.87
. (4.89)
For several reasons, this is a remarkable result. First, it had al-
ready been derived by the German astronomer Soldner in the 19th
century who had assumed that light was a stream of material par-
ticles to which celestial mechanics could be applied just as well
as to planets. Before general relativity, a strict physical mean-
ing could not be given to the trajectory of light in presence of
a gravitational eld because the interaction between electromag-
netic elds and gravity was entirely unclear. The statement of
general relativity that light propagates along null geodesics for
the rst time provided a physical law for the propagation of light
rays in gravitational elds.
Second, the result (4.89) is experimentally found to be wrong. In
fact, the measured value is twice as large. This is a consequence
of our assumption that the metric in the Newtonian limit is given
by (4.80), while the true Newtonian limit is
g = diag[(1+2/c
2
), 12/c
2
, 12/c
2
, 12/c
2
] . (4.90)
Chapter 5
Dierential Geometry III
5.1 The Lie Derivative
5.1.1 The Pull-Back
Following (2.27), we considered one-parameter groups of dieo-
morphisms
t
: R M M (5.1)
such that points p M can be considered as transported along
curves
: R M (5.2)
with (0) = p. Similarly, the dieomorphism
t
can be taken at
xed t R, dening a dieomorphism
t
: M M (5.3)
which maps the manifold onto itself and satises
t

s
=
s+t
.
We have seen the relationship between vector elds and one-
parameter groups of dieomorphisms before. Let now v be a vec-
tor eld on M and from (5.2) be chosen such that the tangent
vector (t) dened by
( (t))( f ) =
d
dt
( f )(t) (5.4)
is identical with v, = v. Then is called an integral curve of v.
If this is true for all curves obtained from
t
by specifying initial
points (0), the result is called the ow of v.
The domain of denition 1 of
t
can be a subset of R M. If
1 = R M, the vector eld is said to be complete and
t
is called
the global ow of v.
52
CHAPTER 5. DIFFERENTIAL GEOMETRY III 53
If 1 is restricted to open intervals I R and open neighbour-
hoods U M, thus 1 = I U R M, the ow is called
local.
Let now M and N be two manifolds and : M N a map from
M onto N. A function f dened at a point q N can be dened
at a point p M with q = (p) by
f : M R , (
f )(p) = ( f )(p) = f [(p)] . (5.5)

The map
pulls functions f on N back to M and is thus

called the pull-back.
Similarly, the pull-back allows to map vectors v from the tangent
space V
p
of M in p onto vectors from the tangent space V
q
of N
in q. We can rst pull-back the function f dened in q N to
p M and then apply v on it, and identify the result as a vector
v applied to f ,
: V
p
V
q
, v
v = v
, (5.6)
such that (
v)( f ) = v(
f ) = v( f ). This denes a vector from

the tangent space of N in q = (p). The map
pushes vectors
from the tangent space of M in p to the tangent space of N in q
and is thus called the push-forward.
In a natural generalisation to dual vectors, we dene their pull-
back
by
: V
q
V
p
, w
w = w
, (5.7)
such that (
w)(v) = w(
v) = w(v
), where w V
q
is an
element of the dual space of N in q. This operation pulls back
the dual vector w from the dual space in q = (p) N to p M.
The pull-back
and the push-forward
can now be extended

to tensors. Let T be a tensor eld of rank (0, r) on N, then its
pull-back is dened by
: 7
0
r
(N) 7
0
r
(M) , T
T = T
, (5.8)
such that
T(v
1
. . . , v
r
) = T(
v
1
, . . . ,
v
r
). Similarly, we can
dene the pull-back of a tensor eld of rank (r, 0) on N by
: 7
r
0
(N) 7
r
0
(M) , T
T (5.9)
such that
T(
w
1
, . . . ,
w
r
) = T(w
1
, . . . , w
r
).
If the pull-back
is a dieomeorphism, which implies in partic-

ular that the dimensions of M and N are equal, the pull-back and
the push-forward are each others inverses,
= (
)
1
. (5.10)
Irrespective of the rank of a tensor, we now denote by
the pull-
back of the tensor and by
its inverse, i.e.
: 7
r
s
(N) 7
r
s
(M) ,
: 7
r
s
(M) 7
r
s
(N) . (5.11)
The important point is that if
: M M is a dieomorphism
and T is a tensor eld on M, then
T can be compared to T. If
T = T,
is a symmetry transformation of T because T stays

the same even though it was moved by
. If the tensor eld

is the metric g, such a symmetry transformation of g is called an
isometry.
5.1.2 The Lie Derivative
Let now v be a vector eld on M and
t
be the ow of v. Then, for
an arbitrary tensor T 7
r
s
, the expression
[
v
T = lim
t0
t
T T
t
(5.12)
is called the Lie derivative of the tensor T with respect to v.
Note that this denition naturally generalises the ordinary deriva-
tive with respect to time t. The manifold M is innitesimally
transformed by one element
t
of a one-parameter group of dif-
feomorphisms. This could, for instance, represent an innitesimal
rotation of the two-sphere S
2
. The tensor T on the manifold af-
ter the transformation is pulled back to the manifold before the
transformation, where it can be compared to the original tensor T
before the transformation.
Obviously, the Lie derivative of rank-(r, s) tensor is itself a rank-
(r, s) tensor. It is linear,
[
v
(t
1
+ t
2
) = [
v
(t
1
) + [
v
(t
2
) , (5.13)
satises the Leibniz rule
[
v
(t
1
t
2
) = [
v
(t
1
) t
2
+ t
1
[
v
(t
2
) , (5.14)
and it commutes with contractions. So far, these properties are
easy to verify in particular after choosing local coordinates.
The application of the Lie derivative to a function f follows di-
rectly from the denition (5.4) of the tangent vector ,
[
v
f = lim
t0
t
f f
t
= lim
t0
( f
t
) ( f
0
)
t
=
d
dt
( f ) = f = v f = df (v) . (5.15)
The additional convenient property
[
x
y = [x, y] (5.16)
for vector elds y is non-trivial to prove.
Given two vector elds x and y, the Lie derivative further satises
the linearity relations
[
x+y
= [
x
+ [
y
, [
x
= [
x
, (5.17)
with R, and the commutation relation
[
[x,y]
= [[
x
, [
y
] = [
x
[
y
[
y
[
x
. (5.18)
If and only if two vector elds x and y commute, so do the re-
spective Lie derivatives,
[x, y] = 0 [
x
[
y
= [
y
[
x
. (5.19)
If and are the ows of x and y, the following commutation
relation is equivalent to (5.19),
s

t
=
t

s
. (5.20)
Let t 7
0
r
be a rank-(0, r) tensor eld and v
1
, . . . , v
r
be vector
elds, then
([
x
t)(v
1
, . . . , v
r
) = x(t(v
1
, . . . , v
r
))
i=1
t(v
1
, . . . , [x, v
i
], . . . , v
r
) . (5.21)
To demonstrate this, we apply the Lie derivative to the tensor
product of t and all v
i
and use the Leibniz rule (5.13),
[
x
(t v
1
. . . v
r
) = [
x
t v
1
. . . v
r
+ t [
x
v
1
. . . v
r
+ . . .
+ t v
1
. . . [
x
v
r
. (5.22)
Then, we take the complete contraction and use the fact that the
Lie derivative commutes with contractions, which yields
[
x
(t(v
1
, . . . , v
r
)) = ([
x
t)(v
1
, . . . , v
r
) (5.23)
+ t([
x
v
1
, . . . , v
r
) + . . . + t(v
1
, . . . , [
x
v
r
) .
Inserting (5.16), we now obtain (5.21).
As an example, we apply (5.21) to a tensor of rank (0, 1), i.e. a
dual vector w:
([
x
w)(y) = xw(y) w([x, y]) . (5.24)
One particular dual vector is the dierential of a function f , de-
ned in (2.34). Inserting df for w in (5.24) yields the useful rela-
tion
([
x
d f )(y) = xd f (y) d f ([x, y])
= xy( f ) [x, y]( f ) = yx( f )
= y[
x
f = d[
x
f (y) , (5.25)
and since this holds vor any vector eld v, we nd
[
x
d f = d[
x
f . (5.26)
Using the latter expression, we can derive coordinate expressions
for the Lie derivative. We introduce the coordinate basis
i
and
its dual basis dx
i
and apply (5.26) to dx
i
,
[
v
dx
i
= d[
v
x
i
= dv(x
i
) = dv
j
j
x
i
= dv
i
=
j
v
i
dx
j
. (5.27)
The Lie derivative of the basis vectors
i
is
[
v
i
= [v,
i
] =
i
v
j
j
, (5.28)
where (2.32) was used in the second step.
To illustrate the components of the Lie derivative of a tensor, we
take a tensor t of rank (1, 1) and apply the Lie derivative to the
tensor product t dx
i
x
j
,
[
v
(t dx
i

j
) = ([
v
t) dx
i

j
+ t [
v
dx
i

j
+ t dx
i
[
v
j
,(5.29)
and now contract completely. This yields
[
v
t
i
j
= ([
v
t)
i
j
+ t(
k
v
i
dx
k
,
j
) t(dx
i
,
j
v
k
k
)
= ([
v
t)
i
j
+ t
k
j
k
v
i
t
i
k
j
v
k
. (5.30)
Solving for the components of the Lie derivative of t, we thus
obtain
([
v
t)
i
j
= v
k
k
t
i
j
t
k
j
k
v
i
+ t
i
k
j
v
k
, (5.31)
and similarly for tensors of higher ranks.
In particular, for a tensor of rank (0, 1), i.e. a dual vector w,
([
v
w)
i
= v
k
k
w
i
+ w
k
i
v
k
. (5.32)
5.1.3 Killing Vector Fields
A Killing vector eld K is a vector eld along which the Lie
derivative of the metric vanishes,
[
K
g = 0 . (5.33)
This implies that the ow of a Killing vector eld denes a sym-
metry transformation of the metric, i.e. an isometry.
To nd a coordinate expression, we use (5.31) to write
([
K
g)
i j
= K
k
k
g
i j
+ g
k j
i
K
k
+ g
ik
j
K
k
= K
k
(
k
g
i j

i
g
k j

j
g
ik
) +
i
(g
k j
K
k
) +
j
(g
ik
K
k
)
=
i
K
j
+
j
K
i
= 0 , (5.34)
where we have identied the Christoel symbols (3.69) in the
last step. This is the Killing equation.
Let be a geodesic, i.e. a curve satisfying

= 0 , (5.35)
then the projection of a Killing vector K on the tangent to the
geodesic is constant along the geodesic,

( , K) = 0 . (5.36)
This is easily seen as follows. First,

( , K) = (

, K) + ( ,

K) = ( ,

K) (5.37)
because of the geodesic equation (5.35).
Writing the last expression explicitly in components yields
( ,

K) = g
ik

i

j
j
K
k
=
i

j
j
K
i
, (5.38)
changing indices and using the symmetry of the metric, we can
also write it as
( ,

K) = g
jk

j

i
i
K
k
=
j

i
i
K
j
. (5.39)
Adding the latter two equations and using the Killing equation
(5.34) shows
2( ,

K) =
i

j
_
i
K
j
+
j
K
i
_
= 0 , (5.40)
which proves (5.36). More elegantly, we have contracted the sym-
metric tensor
i

j
with the tensor
i
K
j
which is antisymmetric
because of the Killing equation, thus the result must vanish.
Equation (5.36) has a profound meaning: freely-falling particles
and light rays both follow geodesics. The constancy of ( , K)
along geodesics means that each Killing vector eld gives rise
to a conserved quantity for freely-falling particles and light rays.
Since a Killing vector eld generates an isometry, this shows that
symmetry transformations of the metric give rise to conservation
laws.
5.2 Dierential Forms
5.2.1 Denition
Dierential p-forms are totally antisymmetric tensors of rank
(0, p). The most simple example are dual vectors w V
since
they are tensors of rank (0, 1). A general tensor t of rank (0, 2) is
not antisymmetric, but can be antisymmetrised dening the two-
form
(v
1
, v
2
)
1
2
[t(v
1
, v
2
) t(v
2
, v
1
)] , (5.41)
with two vectors v
1
, v
2
V.
To generalise this for tensors arbitrary ranks (0, r), we rst dene
the alternation operator by
(7t)(v
1
, . . . , v
r
) =
1
r!
sgn()t(v
(1)
, . . . , v
(r)
) , (5.42)
where the sum extends over all permutations of the integer num-
bers from 1 to r. The sign of a permutation, sgn(), is negative if
the permutation is odd and positive otherwise.
In components, we briey write
(7t)
i
1
...i
r
= t
[i
1
...i
r
]
(5.43)
so that p-forms are dened by the relation
i
1
...i
p
=
[i
1
...i
p
]
(5.44)
between their components. For example, for a 2-form we have
i j
=
[i j]
=
1
2
_
i j

ji
_
. (5.45)
Taking the product of two tensors 7
0
p
and 7
0
q
, yields a
tensor of rank (0, p + q) which is not antisymmetric, but can be
antisymmetrised by means of the alternation operator. The result

(p + q)!
p!q!
7( ) (5.46)
is called the exterior product . Evidently, it turns the tensor
7
0
p+q
into a p + q-form.
The denition of the exterior product implies that it is bilinear,
associative, and satises
= (1)
pq
. (5.47)
The vector space of p-forms is denoted by
_
p
. A basis for it can
be constructed from the basis dx
i
, 1 i n, of the dual space
V
by taking
dx
i
1
. . . dx
i
p
with 1 i
1
< . . . < i
p
n , (5.48)
which shows that the dimension of
_
p
is
_
n
p
_

n!
p!(n p)!
(5.49)
for p n and zero otherwise. The skewed commutation relation
(5.47) implies
dx
i
dx
j
= dx
j
dx
i
. (5.50)
Given two vector spaces V and W above the same eld F, the
Cartesian product V W of the two spaces can be turned into a
vector space by dening the vector-space operations component-
wise. Let v, v
1
, v
2
V and w, w
1
, w
2
W, then the operations
(v
1
, w
1
)+(v
2
, w
2
) = (v
1
+v
2
, w
1
+w
2
) , (v, w) = (v, w) (5.51)
with F give V W the structure of a vector space V W
which is called the direct sum of V and W. Similarly, we dene
the vector space of dierential forms
_

n
p=0
_
p
(5.52)
as the direct sum of the vector spaces of p-forms with arbitrary p.
Recalling that a vector space V attains the structure of an algebra
by dening a vector-valued product between two vectors,
: V V V , (v, w) v w , (5.53)
we see that the exterior product gives the vector space
_
of
dierential forms the structure of a Grassmann algebra,
:
_
_

_
, (, ) . (5.54)
The interior product of a p-form with a vector v V is a map-
ping
V
_
p
_
p1
, (v, ) i
v
(5.55)
dened by
(i
v
)(v
1
, . . . , v
p1
) (v, v
1
, . . . , v
p1
) (5.56)
and i
v
= 0 if is 0-form (a number or a function).
5.2.2 The Exterior Derivative
If a connection is dened on the manifold, we can use it to
dene a map
_
p
_
p+1
, d (5.57)
with
d (p + 1)7() (5.58)
such that
d(v
1
, . . . , v
p+1
) = (p + 1)7
_
v
1
(v
2
, . . . , v
p+1
)
_
. (5.59)
This is called the exterior derivative of .
The result (3.33) for the covariant derivative if a tensor eld and
the antisymmetrisation imply the convenient formula for the ex-
terior derivative of a p-form
d(v
1
, . . . , v
p+1
) =
p+1
i=1
(1)
i+1
v
i
(v
1
, . . . , v
i
, . . . , v
p+1
) (5.60)
+
i<j
(1)
i+j
([v
i
, v
j
], v
1
, . . . , v
i
, . . . , v
j
, . . . , v
p+1
) ,
where the hat over a symbol means that this object is to be left
out.
If the connection is symmetric, in particular if we use the Rie-
mannian connection, it is straightforward to conclude that all
terms in (5.59) containing Christoel symbols cancel because
of the antisymmetrisation and the symmetry of the connection.
Thus, we can replace the covariant by ordinary partial dieren-
tiation, . The components of the exterior derivative of a
p-form can thus be written
(d)
i
1
...i
p+1
= (p + 1)
[i
1
i
2
...i
p+1
]
. (5.61)
Since
i
2
...i
p+1
is itself antisymmetric, this last expression can be
brought into the simpler form
(d)
i
1
...i
p+1
=
p+1
k=1
(1)
k+1
i
k
i
1
,...,
i
k
,...i
p+1
, (5.62)
with 1 i
0
< i
1
< . . . < i
p
n.
For an example, let us apply these relations to a 1-form . For it,
the denition (5.58) implies
(d)(v
1
, v
2
) = (
v
1
)(v
2
) (
v
2
)(v
1
) , (5.63)
while (5.60) specialises to
d(v
1
, v
2
) = v
1
(v
2
) v
2
(v
1
) ([v
1
, v
2
]) . (5.64)
With (5.61) or (5.62), we nd the components
d
i j
=
i
j

j
i
. (5.65)
An alternative way of calculating the exterior derivative proceeds
as follows. A p-form can be expanded in the basis (5.48) as
=
i
1
...i
p
dx
i
1
. . . dx
i
p
. (5.66)
Its exterior derivative is thus
d = d
i
1
...i
p
dx
i
1
. . . dx
i
p
. (5.67)
For example, for the 1-form =
i
dx
i
, we nd
d = d
i
dx
i
=
j
i
dx
j
dx
i
. (5.68)
In R
3
, this turns into
d = (
1
2

2
1
) dx
1
dx
2
+ (
1
3

3
1
) dx
1
dx
3
+ (
2
3

3
2
) dx
2
dx
3
. (5.69)
As (5.61) shows, we can calculate the exterior derivative using the
ordinary partial derivative instead of the covariant derivative. This
implies immediately that dd must vanish identically because the
partial derivatives commute and thus cancel by the antisymmetri-
sation.
A dierential p-form is called exact if a p1-form exists such
that = d. If d = 0, the p-form is called closed. obviously,
an exact form is closed because of d d = 0.
5.3 Integration
5.3.1 The Volume Form and the Codierential
An atlas of a dierentiable manifold is called oriented if for every
pair of charts h
1
on U
1
M and h
2
on U
2
M with U
1
U
2

0, the Jacobi determinant of the coordinate change h
2
h
1
1
is
positive.
An n-dimensional, paracompact manifold M is orientable if and
only if a C
, n-form exists on M which vanishes nowhere. This

is called a volume form.
The canonical volume form on a pseudo-Riemannian manifold
(M, g) is dened by

_
g dx
1
. . . dx
n
. (5.70)
This denition is independent of the coordinate system because
it transforms proportional to the Jacobian determinant upon coor-
dinate changes.
Equation (5.70) implies that the components of the canonical vol-
ume form in n dimensions is proportional to the n-dimensional
Levi-Civita symbol,
i
1
...i
n
=
_
g
i
1
...i
n
, (5.71)
which is dened such that it is +1 for even permutations of the
i
1
, . . . , i
n
, 1 for odd permutations, and vanishes if any two of its
indices are equal. A very useful relation is
j
1
... j
q
k
1
...k
p
j
1
... j
q
i
1
...i
p
= p!q!
k
1
[i
1
k
2
i
2
. . .
k
p
i
p
]
, (5.72)
where the square brackets again denote the complete antisym-
metrisation. In three dimensions, one specic example for (5.72)
is the familiar formula
i jk
klm
=
ki j
klm
=
i
l
j
m

i
m
j
l
. (5.73)
Note that p = 1 and q = 2 here, but the factor 2! = 2 is cancelled
by the antisymmetrisation.
The Hodge star operator (-operation) turns a p form into an
(n p)-form (),
:
_
p
_
np
, , (5.74)
which is uniquely dened by the its application to the dual basis.
For the basis dx
i
of the dual space V
,
(dx
i
1
. . . dx
i
p
) =
_
g
(n p)!
i
1
...i
p
i
p+1
...i
n
dx
i
p+1
. . . dx
i
n
.
(5.75)
If the dual basis e
i
is orthonormal, this simplies to
(e
i
1
. . . e
i
p
) = e
i
p+1
. . . e
i
n
. (5.76)
In components, we can write
()
i
p+1
...i
n
=
1
p!

i
1
...i
n
i
1
...i
p
, (5.77)
i.e. () is the volume form contracted with the p-form .
Straightforward calculation shows that
() = sgn(g)(1)
p(np)
. (5.78)
For a 1-form =
i
dx
i
in R
3
, we can use
dx
1
= dx
2
dx
3
, dx
2
= dx
3
dx
1
, dx
3
= dx
1
dx
2
(5.79)
to nd the Hodge-dual 2-form
=
1
dx
2
dx
3

2
dx
1
dx
3
+
3
dx
1
dx
2
, (5.80)
while the 2-form d (5.69) has the Hodge dual 1-form
d = (
2
3

3
2
)dx
1
(
1
3

3
1
)dx
2
+ (
1
2

2
1
)dx
3
=
jk
i

j
k
dx
i
. (5.81)
The codierential is a map
:
_
p
_
p1
, (5.82)
dened by
sgn(g)(1)
n(p+1)
(d) . (5.83)
d d = 0 immediately implies = 0.
By successive application of (5.72) and (5.62), it can be shown
that the coordinate expression for the codierential is
()
i
1
...i
p1
=
1
_
g
k
_ _
g
ki
1
...i
p1
_
. (5.84)
Comparing this with (4.62), we see that this is the divergence of
. To see this more explicitly, let us work out the codierential
of a 1-form in R
3
by rst taking the exterior derivative of from
(5.80),
d = (
1
1
+
2
2
+
3
3
) dx
1
dx
2
dx
3
, (5.85)
whose Hodge dual is
=
1
1
+
2
2
+
3
3
. (5.86)
5.3.2 Example: Maxwells Equations
The Faraday 2-form is dened by
F
1
2
F
dx
dx
. (5.87)
Application of (5.62) shows that
(dF)
= 0 , (5.88)
i.e. the homogeneous Maxwell equations can simply be expressed
by
dF = 0 . (5.89)
Similarly, the components of the codierential of the Faraday
form are, according to (5.84) and (4.62)
(F)
=
1
gF
_
=
=
4
c
j
. (5.90)
Introducing further the current 1-form by j = j
dx
, we can thus
write the inhomogeneous Maxwell equations as
F =
4
c
j . (5.91)
5.3.3 Integrals and Integral Theorems
The integral over an n-form ,
_
M
, (5.92)
is dened in the following way: Suppose rst that the support
U M of is contained in a single chart which denes positive
coordinates (x
1
, . . . , x
n
) on U. Then, if = f dx
1
. . . dx
n
with
a function f /(U),
_
M
=
_
U
f (x
1
, . . . , x
n
)dx
1
. . . dx
n
. (5.93)
Note that this denition is independent of the coordinate system
because upon changes of the coordinate system, both f and the
volume element dx
1
. . . dx
n
change in proportion to the Jacobian
determinant of the coordinate change.
If the domain of the n-form is contained in multiple maps, the
integral (5.93) needs to be dened piece-wise, but the principle
remains the same.
The integration of functions f /(M) is achieved using the
canonical volume form ,
_
M
f
_
M
f . (5.94)
Stokes theorem can now be formulated as follows: let M be an
n-dimensional manifold and the region D M have a smooth
boundary D such that

D D D is compact. Then, for every
n 1-form , we have
_
D
d =
_
D
. (5.95)
Likewise, Gauss theorem can be brought into the form
_
D
x
=
_
D
x
, (5.96)
where x V is a vector eld on M and x
is the 1-form belonging

to this vector eld. It is dened by
x
(y) = g(x, y) (5.97)

and has the components
(x
)
i
= g
i j
x
j
. (5.98)
Chapter 6
Einsteins Field Equations
6.1 The physical meaning of the curvature
tensor
6.1.1 Congruences of time-like geodesics
Having walked through the introductory chapters, we are now
ready to introduce Einsteins eld equations, i.e. the equations de-
scribing the dynamics of the gravitational eld; Einstein searched
for these equations essentially between early August 1912, when
he moved back from Prague to Zurich, and November 25, 1915,
when he published them in their nal form. We shall give a
heuristic argument for the form of the eld equations, which
should not be mistaken for a derivation, and later show that these
equations follow from a suitable Lagrangian.
First, however, we shall investigate into the physical role of the
curvature tensor. As we have seen, gravitational elds can locally
be transformed away by choosing normal coordinates, in which
the Christoel symbols (the connection coecients) all vanish.
by its nature, this does not hold for the curvature tensor which, as
we shall see, is related to the gravitational tidal eld. thus, in this
sense, the gravitational tidal eld has a more profound physical
signicance as the gravitational eld itself.
Let us begin with a congruence of geodesics. This is a bundle
of time-like geodesics imagined to run through every point of a
small environment U M of a point p M.
Let the geodesics be parameterised by the proper time along
them, and introduce a curve transversal to the congruence, pa-
rameterised by a curve parameter . transversal means that the
curve is nowhere parallel to the congruence.
66
CHAPTER 6. EINSTEINS FIELD EQUATIONS 67
When normalised, the tangent vector to one of the time-like
geodesics can be written as
u =
with (u, u) = 1 . (6.1)

Since it is tangent to a geodesic, it is parallel-transported along
the geodesic,
u
u = 0 . (6.2)
Similarly, we introduce a unit tangent vector v along the curve ,
v = =
. (6.3)
Since the partial derivatives with respect to the curve parameters
and commute, so do the vectors u and v, and thus v is Lie-
transported (or Lie-invariant) along u,
0 = [u, v] = [
u
v . (6.4)
Now, we project v on u and dene a vector n which is perpendic-
ular to u,
n = v + (v, u)u , (6.5)
which does indeed satisfy (n, u) = 0 because of (u, u) = 1. This
vector is also Lie-transported along u, as we shall verify now.
First, we have
[
u
n = [u, n] = [u, v] + [u, (v, u)u]
= u((v, u))u = (
(v, u))u , (6.6)

where (6.4) was used in the rst step. Since (u, u) = 1, we have
0 =
(u, u) = v(u, u) = 2(
v
u, u) , (6.7)
if we use the Ricci identity (3.64).
But the vanishing commutator between u and v and the symmetry
of the connection imply
v
u =
u
v, and thus
(u, v) = u(u, v) = (
u
u, v) + (u,
u
v) = (u,
v
u) = 0 , (6.8)
where the Ricci identity was used in the second step, the geodesic
property (6.2) in the third, and (6.7) in the last. Returning to (6.6),
this proves that n is Lie-transported,
[
u
n = 0 . (6.9)
The perpendicular separation vector between neighbouring
geodesics of the congruence is thus Lie-invariant along the con-
gruence.
6.1.2 The curvature tensor and the tidal eld
Now, we take the second derivative of v along u,
2
u
v =
u
u
v =
u
v
u = (
u
v

v
u
)u , (6.10)
where we have used again that u and v commute and that u is a
geodesic. With [u, v] = 0, the curvature tensor (3.47) applied to u
and v reads
R(u, v)u = (
u
v

v
u
)u , (6.11)
and thus we see that the second derivative of v along u is deter-
mined by the curvature tensor through the Jacobi equation
2
u
v = R(u, v)u . (6.12)
Let us now use this result to nd a similar equation for n. First,
we observe that
u
n =
u
v +
u
((v, u)u) =
u
v + (
(v, u))u =
u
v (6.13)
because of (6.8). Thus
2
u
n =
2
u
v and
2
u
n = R(u, v)u . (6.14)
Next, we use
R(u, n) = R(u, v+(u, v)u) = R(u, v)+(u, v)R(u, u) = R(u, v) (6.15)
to nd
2
u
n = R(u, n)u . (6.16)
This is called the equation of geodesic deviation because it
describes directly how the separation between neighbouring
geodesics evolves along the geodesics according to the curvature.
Finally, let us introduce a coordinate basis e
i
in the subspace
perpendicular to u which is parallel-transported along u. Since n
is conned to that subspace, we can write
n = n
i
e
i
(6.17)
and thus
u
n = (un
i
)e
i
+ (n
i
u
)e
i
=
dn
i
d
e
i
. (6.18)
Since u is normalised and perpendicular to the space spanned by
the triad e
i
, we can form a tetrad from the e
i
and e
0
= u. The
equation of geodesic deviation (6.14) then implies
d
2
n
i
d
2
e
i
= R(e
0
, n
j
e
j
)e
0
= n
j
R(e
0
, e
j
)e
0
= n
j
R
i
00j
e
i
. (6.19)
Thus, dening a matrix K by
d
2
n
i
d
2
= R
i
00 j
n
j
K
i
j
n
j
, (6.20)
we can write (6.19) in matrix form
d
2
n
d
2
= Kn . (6.21)
Note that K is symmetric because of the symmetries (3.77) of the
curvature tensor.
Moreover, the trace of K is
trK = R
i
00i
= R
00
= R
00
= R
, (6.22)
where we have inserted R
0
000
= 0 and the denition of the Ricci
tensor (3.53).
Let us now compare this result to the motion of test bodies in
Newtonian theory. At two neighbouring points x and x + n, we
have the equations of motion
x
i
= (
i
)
x
(6.23)
and, to rst order in a Taylor expansion,
x
i
+ n
i
= (
i
)
x+n
(
i
)
x
(
i
j
)
x
n
j
. (6.24)
Subtracting (6.23) from (6.24) yields the evolution equation for
the separation vector
n
i
= (
i
j
)n
j
, (6.25)
which we can compare to the result (6.21).
Taking into account that
d
2
n
i
d
2
=
n
i
c
2
=
_
c
2
_
n
j
, (6.26)
we see that the matrix K in Newtons theory is
K
(N)
i j
=
c
2
, (6.27)
and its trace is
trK
(N)
=
c
2
=
c
2
, (6.28)
i.e. the negative Laplacian of the Newtonian potential, scaled by
the squared light speed.
Thus, we observe the correspondences
R
i
0 j0

i
c
2
(6.29)
and
R
c
2
. (6.30)
this conrms the assertion that the curvature represents the gravi-
tational tidal eld, describing the relative accelerations of freely-
falling test bodies; (6.29) and (6.30) will provide useful guidance
in guessing the eld equations.
6.2 Einsteins eld equations
6.2.1 Heuristic derivation
We start from the eld equation from Newtonian gravity, i.e. the
Poisson equation
4G =

2
= c
2
trK
(N)
. (6.31)
The density can be expressed by the energy-momentum tensor.
For an ideal uid, we have
T
= (c
2
+ p)u
+ pg
, (6.32)
from which we nd because of (u, u) = 1
T
= c
2
. (6.33)
Moreover, its trace is
trT T = T
= c
2
+ 3p c
2
(6.34)
because p < c
2
under Newtonian conditions (the pressure is
much less than the energy density).
Hence, let us take a constant , put
c
2
= T
+ (1 )g
T (6.35)
and insert this into the eld equation (6.31), using (6.22) for the
trace of K. We thus obtain
R
=
4G
c
4
_
T
+ (1 )g
T
_
u
. (6.36)
Since this equation should hold for any observer and thus for ar-
bitrary four-velocities u, we nd
R
=
4G
c
4
_
T
+ (1 )g
T
_
, (6.37)
where R remains to be determined.
We take the trace of (6.37), obtain
R =
4G
c
4
(T + 4(1 )T) (6.38)
and combine this with (6.37) to assemble the Einstein tensor
(3.80),
G
= R

1
2
g
R
=
4G
c
4
_
T

2
2
g
T
_
. (6.39)
We have seen in (3.81) that the Einstein tensor G satises the
contracted Bianchi identity
= 0 . (6.40)
Likewise, the divergence of the energy-momentum tensor must
vanish in order to guarantee local energy-momentum conserva-
tion,
= 0 . (6.41)
These two conditions are generally compatible with (6.39) only if
we choose = 2, from which we nd the eld equations either in
the form
G
=
8G
c
4
T
(6.42)
or, from (6.37),
R
=
8G
c
4
_
T

1
2
g
T
_
. (6.43)
These are Einsteins eld equations as he published them on
November 25, 1915.
6.2.2 Uniqueness
In the appropriate limit, Einsteins equations satisfy Newtons
theory by construction and are thus one possible set of gravita-
tional eld equations. A remarkable theorem due to Lovelock
states that they are the only possible eld equations under certain
very general conditions.
It is reasonable to assume that the gravitational eld equations
can be written in the form
1[g] = T , (6.44)
where the tensor 1[g] is a functional of the metric tensor g and
T is the energy-momentum tensor. This equation says that the
source of the gravitational eld is assumed to be expressed by the
energy-momentum tensor of all matter and energy contained in
space-time. Now, Lovelocks theorem says,
if 1[g] depends on g and its derivatives only up to second order,
then it must be a linear combination of the Einstein and metric
tensors,
1[g] = G + g , (6.45)
with , R. This absolutely remarkable theorem says that G
must be of the form
G = T + g , (6.46)
with and are constants. The correct Newtonian limit then
requires that = 8Gc
4
, and is the cosmological constant
introduced by Einstein for reasons which will become clear later.
6.2.3 Lagrangian formulation
The remarkable uniqueness of the tensor 1shown by Lovelocks
theorem lets us suspect that a Lagrangian formulation of general
relativity should be possible starting from a scalar constructed
from 1, most naturally its contraction 1
, which is simply pro-

portional to the Ricci scalar R = R
if we ignore the cosmological

term proportional to for now.
Writing down the action, we have to take into account that we
require an invariant volume element, which we obtain from the
canonical volume form introduced in (5.70). Then, according
to (5.93) and (5.94), we can represent volume integrals as
_
M
=
_
U
g d
4
x , (6.47)
where

g is the square root of the determinant of g, and U M
admits a single chart. Recall that, if we need integrate over a
domain covered by multiple charts, a sum over the domains of
the individual charts is understood.
Thus, we suppose that the action of general relativity in a compact
region D M with smooth boundary D is
S
GR
[g] =
_
D
R[g] =
_
D
R[g]
g d
4
x . (6.48)
Working out the variation of this action, we write explicitly
R = g
(6.49)
and thus
S
GR
=
_
D
_
g
g
_
d
4
x (6.50)
=
_
D
R
g d
4
x +
_
D
R
_
g
g
_
d
4
x .
We evaluate the variation of the Ricci tensor rst, using its expres-
sion (3.54) in terms of the Christoel symbols. Matters simplify
considerably if we introduce normal coordinates, which allow us
to ignore the terms in (3.54) which are quadratic in the Christoel
symbols. Then, the Ricci tensor specialises to
R
, (6.51)
and its variation is
R
) . (6.52)
Although the Christoel symbols do not transform as tensors,
their variation does, as the transformation law (3.6) shows. Thus,
we can locally replace the partial by the covariant derivatives and
write
R
) , (6.53)
which is a tensor identity, called the Palatini identity, and thus
holds in all coordinate systems everywhere. It implies
g
_
g
_
, (6.54)
where the indices and were swapped in the last term. Thus,
the variation of the Ricci tensor, contracted with the metric, can
be expressed by the divergence of a vector W,
g
, (6.55)
whose components W
are dened by term in parentheses on the

right-hand side of (6.54).
From Cramers rule in the form (4.55), we see that
g =
g
g
= gg
. (6.56)
moreover, since g
= const. = 4, we conclude
g
= g
. (6.57)
Using these expressions, we obtain for the variation of

g
g =
g
2
g
=
gg
g
=
g
2
g
, (6.58)
or, in terms of the canonical volume form ,
=
1
2
g
=
1
2
g
. (6.59)
Now, we put (6.58) and (6.55) back into (6.50) and obtain
S
GR
=
_
D
+
_
D
_
R

1
2
Rg
_
g
=
_
D
G
+
_
D
!
= 0 . (6.60)
Varying g
only in the interior of D, the divergence term vanishes

by Gau theorem, and admitting arbitrary variations g
implies
G
= 0 . (6.61)
Including the cosmological constant and using (6.58) once more,
we see that Einsteins vacuum equations, Gg = 0, follow from
the variational principle
_
D
(R + 2) = 0 . (6.62)
6.2.4 The energy-momentum tensor
In order to include matter (where matter summarises all kinds
of matter and non-gravitational energy) into the eld equations,
we assume that the matter elds are described by a Lagrangian
[ depending on , its gradient and the metric g,
[(, , g) , (6.63)
where may be a scalar or tensor eld.
The eld equations are determined by the variational principle
_
D
[ = 0 , (6.64)
where the Lagrangian is varied with respect to the elds and
their derivatives . Thus,
_
D
[ =
_
D
_
[
+
[
_
= 0 . (6.65)
As usual, we can express the second term by the dierence
[
=
_
[
_

[
_
, (6.66)
of which the rst termis a divergence which vanishes according to
Gau theorem upon volume integration. Combining (6.66) with
(6.65), and allowing arbitrary variations of the matter elds,
then yields the Euler-Lagrange equations for the matter elds,
[
= 0 . (6.67)
To give an example, suppose we describe a neutral scalar eld
with the Lagrangian
[ =
1
2
(, )
m
2
2

2
=
1
2

m
2
2

2
, (6.68)
where m
2
/2 is a self-interaction potential with a constant pa-
rameter m. The Euler-Lagrange equations then imply the eld
equations
_
+ m
2
_
=
_
+ m
2
_
= 0 , (6.69)
which can be interpreted as the Klein-Gordon equation for a par-
ticle with mass m.
Similarly, we can vary the action with respect to the metric, which
requires care because the Lagrangian may depend on the metric
explicitly and implicitly through the covariant derivatives of
the elds, and the canonical volume form depends on the metric
as well because of (6.59). Thus,
_
D
[ =
_
D
_
([) + [
_
=
_
D
_
[
1
2
g
[g
_
.
(6.70)
If the Lagrangian does not implicitly depend on the metric, we
can write
[ =
[
g
. (6.71)
If there is an implicit dependence on the metric, we can use intro-
duce normal coordinates to evaluate the variation of the Christof-
fel symbols,
=
1
2
g
_
, (6.72)
which is a tensor, as remarked above, whence (6.72) holds in
all coordinate frames everywhere. The derivatives can then be
moved away from the variations of the metric by partial integra-
tion, and expressions proportional to g
remain.
Thus, it is possible to write the variation of the action with respect
to the metric in the form
_
D
[ =
1
2
_
D
T
, (6.73)
in which the tensor T is the energy-momentum tensor. If there
are no implicit dependences on the metric, its components are
T
= 2
[
g
+ [g
. (6.74)
Let us show by an example that this identication does indeed
make sense. We start from the Lagrangian of the free electromag-
netic eld,
[ =
1
16
F
=
1
16
F
. (6.75)
We know from (4.23) that the covariant derivatives in the eld
tensor F can be replaced by partial derivatives, thus there is no
implicit dependence on the metric. Then, the variation [ is
[ =
1
8
F
(6.76)
With (6.70), this implies
_
D
[ =
1
8
_
D
_
F
+
1
4
F
_
g
(6.77)
and, from (6.73), the familiar energy-momentum tensor
T
=
1
4
_
F

1
4
g
_
(6.78)
of the electromagnetic eld.
Therefore, Einsteins eld equations and the matter equations fol-
low from the variational principle
_
D
_
Rc
4
16G
+
c
4
8G
+ [
_
= 0 (6.79)
Since, as we have seen before, the variation of the rst two terms
yields G g, and the variation of the third term yields minus
one-half of the energy-momentum tensor, T/2. In components,
the variation yields
G
=
8G
c
4
T
+ g
. (6.80)
This shows that the cosmological constant can be considered as
part of the energy-momentum tensor,
T
+ T
, T

c
4
8G
g
. (6.81)
6.2.5 Equations of motion
Suppose space-time is lled with an ideal uid whose pressure p
can be neglected compared to the energy density c
2
. then, the
energy-momentum tensor (6.32) can be written as
T
= u
. (6.82)
Conservation of the uid can be expressed in the following way:
the amount of matter contained in a domain D of space-time must
remain the same, even if the domain is mapped into another do-
main
t
(D) by the ow
t
of the vector eld u. Thus
_
D
=
_
t
(D)
. (6.83)
This expression just says that, if the domain D is mapped along
the ow lines of the uid ow, it will encompass a constant
amount of material independent of time t.
Now, we can use the pull-back to write
_
t
(D)
=
_
D
t
() , (6.84)
and take the limit t 0 to see the equivalence of (6.83) and (6.84)
with the vanishing Lie derivative of along u,
[
u
() = 0 . (6.85)
The Leibniz rule (5.14) yields
([
u
) + [
u
= 0 . (6.86)
Due to (5.15), the rst term yields
([
u
) = (u) = (u
i
i
) = (u
i
i
) = (
u
) . (6.87)
For the second term, we can apply equation (5.30) for the com-
ponents of the Lie derivative of a rank-(0, 4) tensor, and use the
antisymmetry of to see that
[
u
= (
) = ( u) . (6.88)
Accordingly, (6.86) can be written as
0 = (
u
+ u) = (u) , (6.89)
or
(u
) = 0 . (6.90)
At the same time, the divergence of T must vanish, hence
0 =
(u
) =
(u
) u
+ u
. (6.91)
The rst term vanishes because of (6.90), and the second implies
u
= 0
u
u = 0 . (6.92)
In other words, the ow lines have to be geodesics. For an ideal
uid, the equation of motion thus follows directly from the van-
ishing divergence of the energy-momentum tensor, which is re-
quired in general relativity by the contracted Bianchi identity
(3.88).
Chapter 7
Weak Gravitational Fields
7.1 Linearised Theory of Gravity
7.1.1 The linearised eld equations
We begin our study of solutions for the eld equations with situ-
ations in which the metric is almost Minkowskian, writing
g
+ h
, (7.1)
where h
is considered as a perturbation of the Minkowski metric
such that
h
< 1 . (7.2)
This condition is excellently satised e.g. in the Solar System,
where
h

c
2
10
6
. (7.3)
Note that small perturbations of the metric do not necessarily im-
ply small perturbations of the matter density, as the Solar System
illustrates. Also, the metric perturbations may change rapidly in
time.
First, we write down the Christoel symbols for this kind of met-
ric. Starting from (4.69) and ignoring quadratic terms in h
, we
can write
=
1
2
_
=
1
2
_
_
. (7.4)
79
CHAPTER 7. WEAK GRAVITATIONAL FIELDS 80
Next, we can ignore the terms quadratic in the Christoel symbols
in the components of the Ricci tensor (3.53) and nd
R
. (7.5)
Inserting (7.4) yields
R
=
1
2
_
_
=
1
2
_
h
_
, (7.6)
where we have introduced the dAlembert operator and abbrevi-
ated the trace of the metric perturbation,
=
, h h
. (7.7)
The Ricci scalar is the contraction of R
,
R =
h , (7.8)
and the Einstein tensor is
G
=
1
2
_
h h
h
_
. (7.9)
Neglecting terms of order h
2
, the contracted Bianchi identity
reduces to
= 0 , (7.10)
which, together with the eld equations, implies
= 0 . (7.11)
One could now insert the Minkowski metric in T
, search for
a rst solution h
(0)
of the linearised eld equations and iterate

replacing
by
+ h
(0)
in T
to nd a corrected solution h
(1)
,
and so forth. This procedure is useful as long as the back-reaction
of the metric on the energy-momentum tensor is small.
If we specialise (7.11) for pressure-less dust and insert (6.78), we
nd the equation of motion
u
= 0 , (7.12)
which means that the uid elements follow straight lines.
The eld equations simplify considerably when we substitute

1
2
h (7.13)
for h
. Since
= h, we can solve (7.13) for h
and insert
h

1
2
(7.14)
into (7.9) to obtain the linearised eld equations
=
16G
c
4
T
. (7.15)
7.1.2 Gauge transformations
Let be a dieomorphism of M, then the metric is equivalent to
the pulled-back metric
g. In particular, this holds for a one-

parameter group
t
of dieomorphisms which is the (local) ows
of some vector eld v. By the denition of the Lie derivative, we
have, to rst order in t,
g = g + t[
v
g . (7.16)
Now, set g = + h and dene the innitesimal vector tv.
Then, the transformation (7.16) implies
h
h = h + t[
v
+ t[
v
h = h + [
+ [
h . (7.17)
For weak elds, the second term can be neglected. Using (5.31),
we see that
([
. (7.18)
Thus, the weak metric perturbation h
admits the gauge trans-

formation
h
. (7.19)
which transforms the tensor
as
. (7.20)
We can now arrange matters to enforce the Hilbert gauge
= 0 . (7.21)
The gauge transformation (7.20) shows that the divergence of
is transformed as
, (7.22)
such that, if (7.21) is not satised yet, it can be achieved by choos-
ing for
a solution of the inhomogeneous wave equation
, (7.23)
which, as we know from electrodynamics, can be obtained by
means of the retarded Greens function of the dAlembert operator.
Enforcing the Hilbert gauge in this way simplies the linearised
eld equation dramatically,
=
16G
c
4
T
. (7.24)
These equations are formally identical to Maxwells equations in
Lorentz gauge, and therefore admit the same solutions. Dening
the Greens function of the dAlembert operator by
G(x, x
) = G(t, t
, x, x
) = 4
D
(t t
, x x
) , (7.25)
we nd the retarded Greens function
G(x, x
) =
1
x x
D
_
x
0
x
0
x x
_
. (7.26)
Using it, we arrive at the particular solution
(x) =
4G
c
4
_
T
_
x
0
x x
, x
_
x x
d
3
x
(7.27)
for the linearised eld equation. Of course, arbitrary solutions of
the homogeneous (vacuum) wave equation can be added.
Thus, similar to electrodynamics, the metric perturbation consists
of the eld generated by the source plus wave-like vacuum solu-
tions propagating at the speed of light.
7.2 Nearly Newtonian gravity
7.2.1 Newtonian approximation of the metric
A nearly Newtonian source of gravity can be described by the
approximations T
00
> T
0j
and T
00
> T
i j
, which express that
mean velocities are small, and the rest-mass energy dominates the
kinetic energy. Then, we can also neglect retardation eects and
write
00
(x) =
4G
c
2
_
(t, x
) d
3
x
x x
= 4
(x)
c
2
, (7.28)
where (x) is the ordinary Newtonian gravitational potential. All
other components of the metric perturbation
vanish,
0j
= 0 =
i j
. (7.29)
Then, the full metric
g
+ h
+
_

1
2
_
(7.30)
has the components
g
00
=
_
1 +
2
c
2
_
, g
0 j
= 0 , g
i j
=
_
1
2
c
2
_
i j
, (7.31)
creating the line element
ds
2
=
_
1 +
2
c
2
_
dt
2
+
_
1
2
c
2
_
(dx
2
+ dy
2
+ dz
2
) . (7.32)
Far away fromthe source, the monopole termin (7.28) dominates,
which yields
ds
2
=
_
1
2GM
rc
2
_
dt
2
+
_
1 +
2GM
rc
2
_
(dx
2
+ dy
2
+ dz
2
) . (7.33)
7.2.2 Gravitational lensing and the Shapiro delay
Two interesting conclusions can be drawn directly from (7.32).
Since light follows null geodesics, light propagation is charac-
terised by ds
2
= 0 or
_
1 +
2
c
2
_
dt
2
=
_
1
2
c
2
_
dx
2
, (7.34)
which implies that the light speed in a (weak) gravitational eld
is
c
= c
dx
dt
=
_
1 +
2
c
2
_
c (7.35)
to rst order in .
Since 0 if normalised such that 0 at innity, c

c, which we can express by the index of refraction for a weak
gravitational eld,
n =
c
c
= 1
2
c
2
. (7.36)
This can be used to calculate light deection using Fermats prin-
ciple, which asserts that light follows a path along which the
light-travel time between a xed source and a xed observer is
extremal, thus
_
dt =
_
dx
c

_
n(x )dx = 0 . (7.37)
Introducing a curve parameter , we can write x = x(), thus
dx = (
x
2
)
1/2
d and
_
n(x)(
x
2
)
1/2
d = 0 , (7.38)
where the overdot denotes derivation with respect to .
The variation leads to the Euler-Lagrange equation
L
x

d
d
L
x
= 0 with L n(x)(
x
2
)
1/2
. (7.39)
Thus, we nd
(
x
2
)
1/2

n
d
d
n
x
_
x
= 0 . (7.40)
We can simplify this expression by choosing the curve parameter
such that

x is a unit vector e, hence
n e

n n
e = 0 . (7.41)
The rst two terms are the component of

n perpendicular to e,
and

e is the change of direction of the tangent vector along the
light ray. Thus,
e =

ln n =
2
c
2
(7.42)
to rst order in . The total deection angle is obtained by inte-
grating

e along the light path.
As a second consequence, we see that the light travel time along
an innitesimal path length dl is
dt =
dl
c
= n
dl
c
=
_
1
2
c
2
_
dl
c
. (7.43)
Compared to light propagation in vacuum, there is thus a time
delay
(dt) = dt
dl
c
=
2
c
3
dl , (7.44)
which is called the Shapiro delay.
7.2.3 The gravitomagnetic eld
At next order in powers of c
1
, the current terms in the energy-
momentum tensor appear, but no stresses yet. That is, we now
approximate T
i j
= 0 and use the eld equations
i j
= 0 ,
0
=
16G
c
4
T
0
. (7.45)
Now, we set A

0
/4 and obtain the Maxwell-type equations
A
=
4
c
2
j
, (7.46)
where the current density j
GT
0
/c
2
was introduced. Ac-
cording to our earlier result (7.28), A
0
= /c
2
. This similarity
to electromagnetic theory naturally leads to the introduction of
electric and magnetic components of the gravitational eld.
Suppose now that the eld is quasi-stationary, so that time deriva-
tives of the metric
can be neglected. Then,

i j
= 0 ev-
erywhere because T
i j
= 0 was assumed, thus
i j
= 0, and the
potentials A
determine the eld completely. They are

A
0
=
c
2
, A
i
=
G
c
2
_
T
0i
(x
) d
3
x
x x
, (7.47)
and the components of the metric g are, according to (7.30),
g
00
= 1 +2A
0
, g
0i
=
0i
= 4A
i
, g
i j
= (1 +2A
0
)
i j
. (7.48)
The most direct approach to the equations of motion starts from
the variational principle (4.4), or
_
_
g
_
1/2
dt = 0 . (7.49)
The radicand is
c
2
2c
2
A
0
8c
A v v
2
, (7.50)
where we have neglected terms of order v
2
since the velocities
are assumed small compared to the speed of light.
Using (7.50), we can reduce the least-action principle (7.49) to
the Euler-Lagrange equations with the eective Lagrangian
[ =
v
2
2
+ A
0
c
2
+ 4c
A v . (7.51)
We rst nd
[
v
=v + 4c
A
d
dt
[
v
=
dv
dt
+ 4c(v

)
A . (7.52)
Using the vector identity
(a
b) = (a

)
b + (
b

)a +a

b +
b

a , (7.53)
we further obtain
[
x
= c
2
A
0
+ 4c
_
(v

)
A +v (

A)
_
, (7.54)
from which we obtain the equations of motion
dv
dt

f = c
2
A
0
+ 4cv (

A) , (7.55)
in which the (specic) force term on the right-hand side corre-
sponds to the Lorentz force in electrodynamics.
Let us consider now a small body characterised by its density
suspended in a gravitomagnetic eld; small means that the eld
can be considered constant across it. It experiences the torque
about its centre-of-mass
M =
_
d
3
x x
f (7.56)
= c
2
A
0

_
d
3
x x + 4c
_
d
3
x x (
j

B) ,
where

j = v is the matter current density and

B =

A is the
gravitomagnetic eld. Since the coordinates are centred on the
centre-of-mass, the rst term vanishes, and the second gives
M = 4c
__
d
3
x x
j
_

B = 4cs

B , (7.57)
where s is the intrinsic angular momentum of the body, i.e. its
spin.
Thus, the bodys spin changes according to
s =

M = 4cs

B . (7.58)
Let us now orient the coordinate frame such that

B = Be
3
,
i.e. B
1
= 0 = B
2
. Then,
s
1
= 4cBs
2
, s
2
= 4cs
1
B . (7.59)
Introducing = s
1
+ is
2
turns this into the single equation
= 4cBi , (7.60)
which is solved by the ansatz =
0
exp(it) if = 4cB.
This shows that a spinning body in a gravitomagnetic eld will
experience spin precession with the angular frequency
= 4c
B = 4c

A , (7.61)
which is called the Lense-Thirring eect.
7.3 Gravitational Waves
7.3.1 Polarisation states
As shown in (7.24), the linearised eld equations in vacuum are
= 0 , (7.62)
if the Hilbert gauge condition (7.21) is enforced,
= 0 . (7.63)
Within the Hilbert gauge class, we can further require that the
trace of
vanish,
=
= 0 . (7.64)
To see this, we return to the gauge transformation (7.20), which
implies
+ 2
= 2
, (7.65)
i.e. if 0, we can choose the vector
such that
2
= . (7.66)
Moreover, (7.22) shows that the Hilbert gauge condition remains
preserved if
satises the dAlembert equation
= 0 (7.67)
at the same time. It can be generally shown that vector elds
can be constructed which indeed satisfy (7.66) and (7.67) at the

same time. If we arrange things in this way, (7.14) shows that
then h
.
All functions propagating with the light speed satisfy the
dAlembert equation (7.62). In particular, we can describe them
as superpositions of plane waves
= h
= T
_
e
i(k,x)
_
(7.68)
with amplitudes given by the so-called polarisation tensor
.
They satisfy the dAlembert equation if
k
2
= (k, k) = k
= 0 . (7.69)
The Hilbert gauge condition then requires
0 =
= 0 , (7.70)
and (7.64) is satised if the trace of
vanishes,
= 0 . (7.71)
The ve conditions (7.70) and (7.71) imposed on the originally
ten independent components of
leave ve independent compo-

nents. Suppose the wave propagates into the positive z direction,
then
k
= (k, 0, 0, k) , (7.72)
and (7.70) implies
0
=
3
; (7.73)
specically,
00
=
30
=
03
=
33
and
01
=
31
,
02
=
32
, (7.74)
while (7.71) means

00
+
11
+
22
+
33
= 0 . (7.75)
Since
33
=
00
, this last equation means
11
+
22
= 0 . (7.76)
Therefore, all components of
can be expressed by ve of them,

as follows:
=
_
00
01
02
00
01
11
12
01
02
12
11
02
00
01
02
00
_
_
. (7.77)
Now, a gauge transformation belonging to a vector eld
= T
_
i
e
i(k,x)
_
(7.78)
which keeps the metric perturbation h
trace-less,
= 0 , (7.79)
changes the polarisation tensor according to
+ k
+ k
. (7.80)
for the k vector specied in (7.72), we thus have
00

00
+ 2k
0
,
01

01
+ k
1
,
02

02
+ k
2
,
11

11
,
12

12
. (7.81)
The condition (7.79) implies that k
= 0, hence
0
=
3
. We
can then use (7.81) to make
00
,
01
and
02
vanish, and only the
gauge-invariant components
11
and
12
are left. Then
=
1
2k
_
00
, 2
01
2
02
,
00
_
(7.82)
xes the gauge transformation, and the polarisation tensor is re-
duced to
=
_
_
0 0 0 0
0
11
12
0
0
12
11
0
0 0 0 0
_
_
. (7.83)
Thus, there are only two linearly independent polarisation states,
as for light.
Under rotations about the z axis by an arbitrary angle , the po-
larisation tensor changes according to
= R
, (7.84)
where R is the rotation matrix with the components
R() =
_
_
1 0 0 0
0 cos sin 0
0 sin cos 0
0 0 0 1
_
_
. (7.85)
Carrying out the matrix multiplication yields
11
=
11
cos 2 +
12
sin 2
12
=
11
sin 2 +
12
cos 2 . (7.86)
Dening

11
i
12
, this can be written as
= e
2i
, (7.87)
which shows that the two polarisation states
have helicity 2,
and thus that they correspond to left and right-handed circular
polarisation.
7.3.2 Generation of gravitational waves
We return to (7.27) to see how gravitational waves can be emit-
ted. From the start, we introduce the two simplications that the
source is far away and changing with a velocity small compared
to the speed-of-light. Then, we can replace the distance x x
by
x x
x = r (7.88)
because far away means that the source is small compared to
its distance. Moreover, we can approximate the retarded time as
follows:
x
0
x x
= x
0
_
(x x
)
2
= x
0
x
2
+ x
2
2x x
x
0
r + x
e
r
, (7.89)
where e
r
is the unit vector in radial direction. Then, we obtain
(t, x) =
4G
rc
4
_
T
_
t
r x
e
r
c
, x
_
d
3
x
. (7.90)
Under the assumption of slow motion, we can further ignore the
directional dependence of the retarted time, thus approximate x
e
r
= 0, and write
(t, x) =
4G
rc
4
_
T
_
t
r
c
, x
_
d
3
x
. (7.91)
By means of the local conservation law
= 0, we can now
simplify the integral on the right-hand side of (7.91):
0 =
_
x
k
d
3
x =
1
c
t
_
x
k
T
0
d
3
x +
_
x
k
l
T
l
d
3
x
=
1
c
t
_
x
k
T
0
d
3
x
_
T
l
k
l
d
3
x , (7.92)
where the second term on the right-hand side was partially inte-
grated, assuming that boundary terms vanish (i.e. enclosing the
source completely in the integration boundary). Thus, we see that
_
T
k
d
3
x =
1
c
t
_
x
k
T
0
d
3
x . (7.93)
According to Gau theorem, we know that the volume integral
over a divergence within the same surface must vanish,
_

j
(T
j0
x
l
x
k
)d
3
x = 0 . (7.94)
This enables us to write
1
c
t
_
T
00
x
l
x
k
d
3
x =
_

(T
0
x
l
x
k
)d
3
x =
_
T
0
(x
l
x
k
)d
3
x
=
_
T
0
(
k
x
l
+ x
k
)d
3
x
=
_
(T
k0
x
l
+ T
l0
x
k
)d
3
x . (7.95)
Taking the partial time derivative of (7.95) yields
1
2c
2
2
t
_
(T
00
x
k
x
l
)d
3
x =
1
2c
t
_
(T
k0
x
l
+ T
l0
x
k
)d
3
x (7.96)
=
1
2
_
(T
kl
+ T
lk
)d
3
x =
_
T
kl
d
3
x ,
where (7.93) was used. In other words, the spatial components of
the metric perturbation
are
jk
(t, x) =
2G
rc
6
2
t
_
T
00
_
t
r
c
, x
_
x
j
x
k
d
3
x
=
2G
rc
4
2
t
_

_
t
r
c
, x
_
x
j
x
k
d
3
x
, (7.97)
where we have used that the T
00
component of the energy-
momentum tensor is well approximated by the matter density if
the sources material is moving slowly.
The sources quadrupole tensor is
Q
jk
=
_
_
3x
j
x
k
r
2
jk
_
(x)d
3
x , (7.98)
which allows us to write the metric perturbation as
jk
=
2G
3rc
4
_
2
t
Q
jk
_
t
r
c
, x
_
+
jk
2
t
_
r
2
_
t
r
c
, x
_
d
3
x
_
.
(7.99)
7.3.3 Energy transport by gravitational waves
It can be shown that the energy-momentum tensor in gravitational
waves propagating into the x
1
-direction is
T
01
GW
=
1
32G
_
2
2
23
+
1
2
(
22

33
)
2
_
, (7.100)
which can be written with the help of (7.99) as
T
01
GW
=
G
72r
2
_
2
...
Q
2
23
+
1
2
_
...
Q
22

...
Q
33
_
2
_
, (7.101)
showing one of the rare cases of a third time derivative in physics.
The transversal quadrupole tensor is
Q
T
=
_
Q
22
Q
23
Q
32
Q
33
_
(7.102)
because the direction of propagation was chosen as the x
1
axis.
dening the transversal trace-free quadrupole tensor by
Q
TT
= Q
T
J
2
trQ
T
=
1
2
_
Q
22
Q
33
2Q
23
2Q
23
(Q
22
Q
33
)
_
, (7.103)
we see that an invariant expression for the right-hand side of
(7.101) is given by
tr(Q
TT
Q
TT
) =
1
2
(Q
22
Q
33
)
2
+ 2Q
2
23
, (7.104)
and thus the energy current in gravitational waves is
T
0i
GW
=
G
72r
2
_
tr(Q
TT
Q
TT
)
_
. (7.105)
a nal integration over all directions yields Einsteins famous
quadrupole formula for the gravitational-wave luminosity,
L
GW
=
G
45
_
...
Q
kl
...
Q
kl
_
. (7.106)
Chapter 8
The Schwarzschild Solution
8.1 Cartans structure equations
8.1.1 The curvature and torsion forms
Karl Schwarzschild (18731916)
Elie Cartan (18691951)

This section deals with a generalisation of the connection coef-
cients, and the torsion and curvature tensor components, to ar-
bitrary bases. This will prove enormously ecient in our further
discussion of the Schwarzschild solution.
Let M be a dierentiable manifold, e
i
an arbitrary basis for vec-
tor elds and
i
an arbitrary basis for dual vector elds, or 1-
forms. In analogy to the Christoel symbols, we introduce the
connection forms by
v
e
i
=
j
i
(v)e
j
. (8.1)
Since
v
e
i
is a vector,
j
i
(v) R is a real number, and thus
j
i

_
1
is a dual vector, or a one-form.
Since, by denition (3.2) of the Christoel symbols
j
=
i
k j

i
=
i
j
(
k
)
i
(8.2)
in the coordinate basis
i
, we have in that particular basis,
i
j
=
i
k j
dx
k
. (8.3)
Since (
j
, e
i
) is a constant (which is either zero or unity), we must
have
0 =
v
(
i
, e
j
) = (
v
i
, e
j
) + (
i
,
v
e
j
)
= (
v
i
, e
j
) + (
i
,
k
j
(v)e
k
)
= (
v
i
, e
j
) +
i
j
(v) . (8.4)
93
CHAPTER 8. THE SCHWARZSCHILD SOLUTION 94
Thus, we conclude
i
=
i
j
(v)
j
(8.5)
or, without specifying the vector v,
i
=
j

i
j
. (8.6)
Let now
_
1
be a one-form such that =
i
i
with arbitrary
functions
i
. Then, the equations we have derived so far imply
v
= v(
i
)
i
+
i
i
= (d
i

k
k
i
, v)
i
, (8.7)
where we have used the dierential of the function
i
, dened
in (2.34) by d
i
(v) = v( f ), together with the notation (w, v) =
w(v) for a vector v and a dual vector w. This expression can be
generally written as
=
i
(d
i

k
k
i
) . (8.8)
Similarly, for a vector eld x = x
i
e
i
, we nd
v
(x) = (dx
i
+ x
k
i
k
, v)e
i
(8.9)
or
x = e
i
(dx
i
+
i
k
x
k
) . (8.10)
We are now in a position to introduce the torsion and curvature
forms. By denition, the torsion T(x, y) is a vector, which can be
written in terms of the torsion forms
i
as
T(x, y) =
i
(x, y)e
i
. (8.11)
Obviously,
i
_
2
is a two-form, such that
i
(x, y) R is a real
number.
In the same manner, we express the curvature by the curvature
forms
i
j

_
2
,
R(x, y)e
j
=
i
j
(x, y)e
i
. (8.12)
The torsion and curvature 2-forms satisfy Cartans structure
equations,
i
= d
i
+
i
j

j
i
j
= d
i
j
+
i
k

k
j
. (8.13)
Their proof is straightforward.
To prove the rst structure equation, we insert the denition
(3.42) of the torsion to obtain as a rst step
i
(x, y) =
x
y
y
x [x, y]
=
x
(
i
(y)e
i
)
y
(
i
(x)e
i
)
i
([x, y])e
i
, (8.14)
where we have expanded the vectors x, y and [x, y] in the basis
e
i
according to x = (x,
i
)e
i
=
i
(x)e
i
. Then, we continue by
using the connection forms,
i
(x, y) =
x
(
i
(y)e
i
)
y
(
i
(x)e
i
)
i
([x, y])e
i
= x
i
(y)e
i
+
i
(y)
j
i
(x)e
j
y
i
(x)e
i

i
(x)
j
i
(y)e
j

i
([x, y])e
i
=
_
x
i
(y) y
i
(x)
i
([x, y])
_
e
i
+
_
i
(y)
j
i
(x)
i
(x)
j
i
(y)
_
e
j
. (8.15)
According to (5.64), the rst termcan be expressed by the exterior
derivatives of the
i
, and since the second term is antisymmetric
in x and y, we can write this as
i
(x, y) = d
i
(x, y)e
i
+ (
i
j

j
)(x, y)e
i
, (8.16)
from which the rst structure equation follows immediately.
The proof of the second structure equation proceeds similarly, us-
ing the denition (3.48) of the curvature. Thus,
i
j
(x, y)e
i
=
x
y
e
j

y
x
e
j

[x,y]
e
j
=
x
(
i
j
(y)e
i
)
y
(
i
j
(x)e
i
)
i
j
([x, y])e
i
= x
i
j
(y)e
i
+
i
j
(y)
x
e
i
y
i
j
(x)e
i

i
j
(x)
y
e
i

i
j
([x, y])e
i
=
_
x
i
j
(y) y
i
j
(x)
i
j
([x, y])
_
e
i
+
_
i
j
(y)
k
i
(x)
i
j
(x)
k
i
(y)
_
e
k
= d
i
j
(x, y)e
i
+ (
k
i

i
j
)(x, y)e
k
, (8.17)
which proves the second structure equation.
Now, let us use the curvature forms
i
j
to dene tensor compo-
nents R
i
jkl
by
i
j

1
2
R
i
jkl

k

l
, (8.18)
whose antisymmetry in the last two indices is obvious by deni-
tion,
R
i
jkl
= R
i
jlk
. (8.19)
In an arbitrary basis e
i
, we then have
(
i
, R(e
k
, e
l
)e
j
) = (
i
,
s
j
(e
k
, e
l
)e
s
) =
i
j
(e
k
, e
l
) = R
i
jkl
. (8.20)
Comparing this to the components of the curvature tensor in the
coordinate basis
i
given by (3.52) shows that the functions R
i
jkl
are indeed the components of the curvature tensor in the arbitrary
basis e
i
.
A similar operation shows that the functions T
i
jk
dened by
1
2
T
i
jk

j

k
(8.21)
are the elements of the torsion tensor in the basis e
i
, since
(
i
, T(e
j
, e
k
)) = (
i
,
s
(e
j
, e
k
)e
s
) =
i
j
(e
j
, e
k
) = T
i
jk
. (8.22)
Thus, Cartans structure equations allow us to considerably sim-
plify the computation of curvature and torsion for an arbitrary
metric, provided we nd a base in which the metric appears sim-
ple (e.g. diagonal and constant).
We mention without proof that the connection is metric if and
only if
i j
+
ji
= dg
i j
, (8.23)
where the denitions
i j
g
ik
k
j
and g
i j
g(e
i
, e
j
) (8.24)
were used, i.e. the g
i j
are the components of the metric in the
arbitrary basis e
i
.
8.2 Stationary and static spacetimes
Stationary spacetimes (M, g) are dened to be spacetimes which
have a time-like Killing vector eld K. This means that observers
moving along the integral curves of K do not notice any change.
This denition implies that we can introduce coordinates in which
the components g
of the metric do not depend on time. To see

that, suppose we choose a space-like hypersurface M and
construct the integral curves of K through .
We further introduce arbitrary coordinates on and carry them
into M as follows: let
t
be the ow of K, p
0
and p =
t
(p
0
),
then the coordinates of p are chosen as (t, x
1
(p
0
), x
2
(p
0
), x
3
(p
0
)).
These are the so-called Lagrange coordinates of p.
In these coordinates, K = (
0
, 0, 0, 0). From (5.34), we further
have that the components of the Lie derivative of the metric are
([
K
g)
= K
+ g
+ g
=
0
g
= 0 , (8.25)
which proves that the g
do not depend on time in these so-called

adapted coordinates.
We can straightforwardly assign a one-form to the vector K by
dening
= g
. This one-form obviously satises

(K) = (K, K) 0 . (8.26)
Suppose that we have a stationary spacetime in which we have
introduced adapted coordinates and in which also g
0i
= 0. Then,
the Killing vector eld is orthogonal to the spatial sections, for
which t = const. Then, the one-form is obviously
= g
00
dt = (K, K) dt , (8.27)
because K =
t
. This then trivially implies the Frobenius condi-
tion
d = 0 (8.28)
because the exterior derivative d satises d d 0.
Conversely, it can be shown that if the Frobenius condition holds,
the one-form can be written in the form (8.27). For a vector
eld v tangential to a spacelike section dened by t = const., we
have
(K, v) = (v) = (K, K)dt(v) = (K, K)v(t) = 0 (8.29)
because t = const., and thus K is then perpendicular to the spatial
section. Thus, K =
t
and
g
0i
= (
t
,
i
) = (K,
i
) = 0 . (8.30)
Thus, in a stationary spacetime with time-like Killing vector eld
K, the Frobenius condition (8.28) for the one-form correspond-
ing to K is equivalent to the condition g
0i
= 0 in adapted coordi-
nates. Such spacetimes are called static. In other words, station-
ary spacetimes are static if and only if the Frobenius condition
holds.
In static spacetimes, the metric can thus be written in the form
g = g
00
(x )dt
2
+ g
i j
(x )dx
i
dx
j
. (8.31)
8.3 The Schwarzschild solution
8.3.1 Form of the metric
Technically speaking, the Schwarzschild solution is a static,
spherically symmetric solution of Einsteins eld equations for
vacuum spacetime.
From our earlier considerations, we know that a static spacetime
is a stationary spacetime whose (time-like) Killing vector eld
satises the Frobenius condition (8.28).
From the spacetimes being (globally) stationary, we know that
we can introduce spatial hypersurfaces perpendicular to the
Killing vector eld which, in adapted coordinates, is K =
t
. The
manifold (M, g) can thus be foliated as M = R .
From (8.31), we then know that, also in adapted coordinates, the
metric assumes the form
g =
2
dt
2
+ h , (8.32)
where is a smoothly varying function on and h is the metric
of the spatial sections . Under the assumption that K is the only
time-like Killing vector eld which the space-time admits, t is a
uniquely distinguished time coordinate, and we can write

2
= (K, K) . (8.33)
The stationarity of the space-time, expressed by the existence of
the single Killing vector eld K, thus allows a convenient folia-
tion of the space-time.
Furthermore, the spatial hypersurfaces are expected to be spher-
ically symmetric. This means that the group SO(3) (i.e. the group
of rotations in three dimensions) must be an isometry group of
the metric h. The orbits of SO(3) are two-dimensional, space-
like surfaces on . Thus, SO(3) foliates the spacetime (, h) into
invariant two-spheres.
Let the surface of these two-spheres be A, then we dene a radial
coordinate for the Schwarzschild metric requiring
4r
2
= A (8.34)
as in Euclidean geometry. Moreover, the spherical symmetry im-
plies that we can introduce spherical polar coordinates (, ) on
one particular orbit of SO(3) which can then be transported along
geodesic lines perpendicular to the orbits. Then, the spatial metric
h can be written in the form
h = e
2b(r)
dr
2
+ r
2
_
d
2
+ sin
2
d
2
_
, (8.35)
where the exponential factor was introduced to allow a scaling of
the radial coordinate.
Due to the stationarity of the metric and the spherical symmetry
of the spatial sections, (K, K) can only depend on r. We set
2
= (K, K) = e
2a(r)
. (8.36)
The full metric g is thus characterised by two radial functions a(r)
and b(r) which we need to determine.
The spatial sections are now foliated according to
= I S
2
, I R
+
, (8.37)
the metric is
g = e
2a(r)
dt
2
+ e
2b(r)
dr
2
+ r
2
(d
2
+ sin
2
d
2
) (8.38)
in the Schwarzschild coordinates (t, r, , ). The functions a(r)
and b(r) are constrained by the requirement that the metric should
asymptotically turn at, which means
a(r) 0 , b(r) 0 for r . (8.39)
They must be determined by inserting the metric (8.38) into the
vacuum eld equations, G = 0.
8.3.2 Connection and curvature forms
In order to evaluate Einsteins eld equations for the
Schwarzschild metric, we need to compute the Riemann, Ricci,
and Einstein tensors. Traditionally, one would begin this step with
computing all Christoel symbols of the metric (8.38). this very
lengthy and error-prone procedure can be considerably shortened
using Cartans structure equations (8.13) for the torsion and cur-
vature forms
i
and
i
j
.
To do so, we need to introduce a suitable basis, or tetrad e
i
or dual basis
i
. Guided by the form of the metric (8.38), we
choose
0
= e
a
dt ,
1
= e
b
dr ,
2
= rd ,
3
= r sin d . (8.40)
In terms of these, the metric attains the simple diagonal,
Minkowskian form
g = g
, g
= diag(1, 1, 1, 1) . (8.41)
Obviously, dg = 0, and thus (8.23) implies that the connection
forms
are antisymmetric,
. (8.42)
Given the dual tetrad
, we must take their exterior derivatives.

For this purpose, we apply the expression (5.57) with p = 1 and
nd, for d
0
,
d
0
=
r
0
0
dt dr = a
e
a
dt dr (8.43)
because
0
= (e
a
, 0, 0, 0) according to (8.40), so that i
0
= 0 and
i
1
= 1 are the only possible indices to be summed over.
Similarly, we nd
d
1
= 0 , (8.44)
further
d
2
=
r
2
2
dr d = dr d (8.45)
and
d
3
=
r
3
3
dr d +
3
3
d d
= sin dr d + r cos d d . (8.46)
Using (8.40), we can also express the coordinate dierentials by
the dual tetrad,
dt = e
a
0
, dr = e
b
1
, d =
2
r
, d =
3
r sin
, (8.47)
so that we can write the exterior derivatives of the dual tetrad as
d
0
= a
e
b
1

0
, d
1
= 0 , d
2
=
1
re
b

1

2
,
d
3
=
1
re
b

1

3
+
cot
r

2

3
. (8.48)
Since the torsion must vanish,
i
= 0, Cartans rst structure
equation from (8.13) implies
d
. (8.49)
With (8.48), this suggests that the connection forms of the
Schwarzschild metric are
0
1
=
1
0
=
a
0
e
b
,
0
2
=
2
0
= 0 ,
0
3
=
3
0
= 0 ,
2
1
=
1
2
=
2
re
b
,
3
1
=
1
3
=
3
re
b
,
3
2
=
2
3
=
cot
3
r
. (8.50)
They satisfy the antisymmetry condition (8.42) and Cartans rst
structure equation (8.49) for a torsion-free connection.
For evaluating the curvature forms
, we rst recall that the ex-

terior derivative of a one-form multiplied by a function f is
d( f ) = df + f d
= (
i
f )dx
i
+ f d (8.51)
according to the Leibniz rule.
Thus, we have for d
0
1
d
0
1
= (a
e
b
)
dr
0
+ a
e
b
d
0
= (a
e
b
a
e
b
)e
b
1

0
+ (a
e
b
)
2
1

0
= e
2b
(a
+ a
2
)
0

1
, (8.52)
where we have used (8.47) and (8.48).
In much the same way, we nd
d
2
1
=
b
re
2b

1

2
,
d
3
1
=
b
re
2b

1

3
+
cot
r
2
e
b

2

3
,
d
3
2
=
1
r
2

2

3
. (8.53)
This yields the curvature two-forms according to (8.13),
0
1
= d
0
1
=
a
+ a
2
e
2b

0

1
=
1
0
0
2
=
0
1

1
2
=
a
re
2b

0

2
=
2
0
0
3
=
0
1

1
3
=
a
re
2b

0

3
=
3
0
1
2
= d
1
2
+
1
3

3
2
=
b
re
2b

1

2
=
2
1
1
3
= d
1
3
+
1
2

2
3
=
b
re
2b

1

3
=
3
1
2
3
= d
2
3
+
2
1

1
3
=
1 e
2b
r
2

2

3
=
3
2
. (8.54)
The other curvature two-forms follow from antisymmetry since
= g
, (8.55)
because of the antisymmetry properties of the curvature.
8.3.3 Components of the Ricci and Einstein tensors
The components of the curvature tensor are given by (8.20), and
thus the components of the Ricci tensor in the tetrad e
are
R
= R
(e
, e
) . (8.56)
Thus, the components of the Ricci tensor in the Schwarzschild
tetrad are
R
00
=
1
0
(e
1
, e
0
) +
2
0
(e
2
, e
0
) +
3
0
(e
3
, e
0
)
=
a
+ a
2
+ 2a
/r
e
2b
,
R
11
=
0
1
(e
0
, e
1
) +
2
1
(e
2
, e
1
) +
3
1
(e
3
, e
1
)
=
a
+ a
2
2b
/r
e
2b
(8.57)
and
R
22
=
0
2
(e
0
, e
2
) +
1
2
(e
1
, e
2
) +
3
2
(e
3
, e
2
)
=
a
re
2b
+
e
2b
1
r
2
e
2b
R
33
=
0
3
(e
0
, e
3
) +
1
3
(e
1
, e
3
) +
2
3
(e
2
, e
3
) = R
22
(8.58)
The Ricci scalar becomes
R = 2
a
+ a
2
e
2b
4
a
re
2b
+ 2
e
2b
1
r
2
e
2b
, (8.59)
such that the components of the Einstein tensor in the tetrad e
read
G
00
= R
00

R
2
g
00
=
1
r
2
e
2b
_
1
r
2

2b
r
_
G
11
=
1
r
2
+ e
2b
_
1
r
2
+
2a
r
_
G
22
= e
2b
_
a
+ a
2
+
a
r
_
= G
33
, (8.60)
and all other components of G
vanish identically.
8.3.4 Solution of the vacuum eld equations
The vacuum eld equations now require that all components of
the Einstein tensor vanish. In particular, then,
0 = G
00
+ G
11
=
2e
2b
r
(a
+ b
) (8.61)
shows that a
+b
= 0. Since a+b 0 asymptotically for r ,

integrating a+b from r indicates that a+b = 0 everywhere,
or b = a.
The equation G
00
= 0 itself implies that
e
2b
_
2b
r

1
r
2
_
+
1
r
2
= 0 . (8.62)
Multiplication by r
2
gives
e
2b
(1 2rb
) = 1
_
re
2b
_
= 1 . (8.63)
Therefore, (8.62) is equivalent to
re
2b
= r 2m e
2b
= 1
2m
r
, (8.64)
where 2m has been chosen as an integration constant.
Since a = b, this also means that
e
2a
= e
2b
= 1
2m
r
. (8.65)
We thus obtain the Schwarzschild solution for the metric,
ds
2
=
_
1
2m
r
_
dt
2
+
dr
2
1
2m
r
+ r
2
_
d
2
+ sin
2
d
2
_
. (8.66)
We have seen before in (4.79) that the 00 element of the metric
must be related to the Newtonian gravitational potential as g
00
=
(1+2/c
2
) in order to meet the Newtonian limit. The Newtonian
potential of a point mass M at a distance r is
=
GM
r
. (8.67)
Together with (8.66), this shows that the Newtonian limit is
reached by the Schwarzschild solution if the integration constant
m has the value
2m
r
=
2GM
rc
2
m =
GM
c
2
1.5 km
_
M
M
,
_
. (8.68)
The Schwarzschild metric (8.66) has an (apparent) singularity at
r = 2m or
r = R
s

2GM
c
2
, (8.69)
the so-called Schwarzschild radius. We shall clarify later the
meaning of this singularity.
In order to illustrate the geometrical meaning of the spatial part
of the Schwarzschild metric, we need to nd a geometrical inter-
pretation for its radial dependence. Specialising to the equatorial
plane of the Schwarzschild solution, = /2 and t = 0, we nd
the induced spatial line element
dl
2
=
dr
2
1 2m/r
+ r
2
d
2
(8.70)
on that plane.
On the other hand, consider a surface of rotation in the three-
dimensional Euclidean space E
3
. If we introduce the adequate
cylindrical coordinates (r, , z) on E
3
and rotate a curve z(r) about
the z axis, we nd the induced line element
dl
2
= dz
2
+ dr
2
+ r
2
d
2
=
_
dz
dr
_
2
dr
2
+ dr
2
+ r
2
d
2
= (1 + z
2
)dr
2
+ r
2
d
2
. (8.71)
We can now try and identify the two induced line elements from
(8.70) and (8.71) and nd that this is possible if
z
=
_
1
1 2m/r
1
_
1/2
=
_
2m
r 2m
, (8.72)
which is readily integrated to yield
z =
_
8m(r 2m) + const. or z
2
= 8m(r 2m) , (8.73)
if we set the integration constant to zero.
This shows that the geometry on the equatorial plane of the spa-
tial section of the Schwarzschild solution can be identied with
a rotational paraboloid in E
3
. In other words, the dependence of
radial distances on the radius r is equivalent to that on a rotational
paraboloid.
8.3.5 Birkhos theorem
Suppose now we had started from a spherically symmetric vac-
uum space-time, but with explicit time dependence of the func-
tions a and b, such that the space-time could either expand or
contract. Then, a repetition of the derivation of the connection
and curvature forms, and the components of the Einstein tensor
following from them, had resulted in the new components

G
G
00
= G
00
,

G
11
= G
11
G
22
= G
22
e
2a
_
b
2
a
b

b
_
=

G
33
G
10
=
2
b
r
e
ab
(8.74)
and

G
= 0 for all other components.

The vacuum eld equations imply

G
10
= 0 and thus

b = 0, hence
b must be independent of time. From

G
00
= 0, we can again
conclude (8.64), i.e. b retains the same form as before. Similarly,
since

G
00
+

G
11
= G
00
+ G
11
, the requirement a
+ b
= 0 must
continue to hold, but now the time dependence of a allows us to
conclude only that
a = b + f (t) , (8.75)
where f (t) is an otherwise unconstrained function of time only.
Thus, the line element then reads
ds
2
= e
2 f
_
1
2m
r
_
dt
2
+ dl
2
, (8.76)
where dl
2
is the unchanged line element of the spatial sections.
Introducing the new time coordinate t
by
t
=
_
e
2f
dt (8.77)
converts (8.76) back to the original form (8.66) of the
Schwarzschild metric.
This is Birkhos theorem, which says that a spherically symmet-
ric solution of Einsteins vacuum equations is necessarily static
for r > 2m.
Chapter 9
Physics in the Schwarzschild
Space-Time
9.1 Motion in the Schwarzschild space-time
9.1.1 Equation of motion
According to (4.3), the motion of a particle in the Schwarzschild
space-time is determined by the Lagrangian
[ =
_
( , ) d , (9.1)
where is the proper time dened by (4.6), satisfying
d =
_
( , ) d (9.2)
and thus requiring that the four-velocity be normalised,
( , ) = 1 . (9.3)
Note that we have to dierentiate and integrate with respect to the
proper time rather than the coordinate time t because the latter
has no invariant physical meaning. In the Newtonian limit, = ct.
This implies that, instead of varying the action
S = mc
_
b
a
_
( , ) d , (9.4)
we can just as well require that the variation of
S =
1
2
_
b
a
( , ) d (9.5)
106
CHAPTER9. PHYSICS INTHE SCHWARZSCHILDSPACE-TIME107
vanish. In fact, from S = 0, we have
0 =
_
b
a
_
( , ) d =
1
2
_
b
a
( , )
_
( , )
d
=
_
1
2
_
b
a
( , ) d
_
. (9.6)
because of the normalisation condition (9.3).
Thus, we can obtain the equation of motion just as well from the
Lagrangian
[ =
1
2
( , ) =
1
2
g
(9.7)
=
1
2
_
(1 2m/r)
t
2
+
r
2
1 2m/r
+ r
2
_
2
+ sin
2
2
_
_
,
where it is important to recall that the overdot denotes dierentia-
tion with respect to proper time . In addition, (9.3) immediately
implies that 2[ = 1.
The Euler-Lagrange equation for is
d
d
[
= 0 =
d
d
(r
2
) r
2
2
sin cos . (9.8)
Suppose the motion starts in the equatorial plane, = /2 and
= 0. Should this not be the case, we can always rotate the

coordinate frame so that this is satised. Then, (9.8) shows that
r
2
= const. = 0 . (9.9)
Without loss of generality, we can thus restrict the discussion to
motion in the equatorial plane, which simplies the Lagrangian
to
[ =
1
2
_
(1 2m/r)
t
2
+
r
2
1 2m/r
+ r
2
2
_
. (9.10)
Obviously, t and are cyclic, thus
[
= r
2
L = const. (9.11)
and
[
t
= (1 2m/r)
t E = const. . (9.12)
Now, we insert
=
L
r
2
,

t =
E
1 2m/r
(9.13)
into the Lagrangian (9.10), use 2[ = 1 and nd
1 = (12m/r)
t
2
+
r
2
1 2m/r
+r
2
2
=
r
2
E
2
1 2m/r
+
L
2
r
2
, (9.14)
which can be cast into the form
r
2
+ V(r) = E
2
, (9.15)
where V(r) is the eective potential
V(r)
_
1
2m
r
_ _
1 +
L
2
r
2
_
. (9.16)
Since it is our primary goal to nd the orbit r(), we use r
=
dr/d = r/
to transform (9.15) to
r
2
+ V(r) =

2
r
2
+ V(r) =
L
2
r
4
r
2
+ V(r) = E
2
. (9.17)
Now, we substitute u 1/r and u
= r
/r
2
= u
2
r
and nd
L
2
u
4
u
2
u
4
+ V(1/u) = L
2
u
2
+ (1 2mu)(1 + L
2
u
2
) = E
2
(9.18)
or, after dividing by L
2
and rearranging terms,
u
2
+ u
2
=
E
2
1
L
2
+
2m
L
2
u + 2mu
3
. (9.19)
Dierentiation with respect to cancels the constant rst term on
the right-hand side and yields
2u
+ 2uu
=
2m
L
2
u
+ 6mu
2
u
. (9.20)
the trivial solution is u
= 0, which implies a circular orbit. If

u
0, this equation can be simplied to read

u
+ u =
m
L
2
+ 3mu
2
. (9.21)
The fact that t and are cyclic coordinates in the Schwarzschild
space-time can be studied from a more general point of view. Let
() be a geodesic curve with tangent vector (), and let further
be a Killing vector eld of the metric. Then, we know from (5.36)
that the projection of the Killing vector eld on the geodesic is
constant along the geodesic,

( , ) = 0 ( , ) = constant along (9.22)
Due to its stationarity and the spherical symmetry, the
Schwarzschild space-time has the Killing vector elds
t
and
.
Thus,
( ,
t
) = (
t
t
,
t
) =
t
(
t
,
t
) = g
00

t
=
_
1
2m
r
_

t = const.
(9.23)
and
( ,
) =
) = g
= r
2
sin
2
= r
2
= const. ,
(9.24)
where we have used = /2 without loss of generality. this
reproduces (9.11) and (9.12).
9.1.2 Comparison to the Kepler problem
It is instructive to compare this to the Newtonian case. There, the
Lagrangian is
[ =
1
2
_
r
2
+ r
2
2
_
(r) , (9.25)
where (r) is some centrally-symmetric potential and the dots
denote the derivative with respect to the coordinate time t now
instead of the proper time . In the Newtonian limit, = ct. For
later comparison of results obtained in this and the previous sec-
tions, we thus replace coordinate-time derivatives by derivatives
with respect to proper time and introduce the appropriate factors
of c
2
, thus
[
[
c
2
=
1
2
_
r
2
+ r
2
2
_
(r)
c
2
, (9.26)
where the dots now mean derivatives with respect to , as in the
previous section.
Since is cyclic,
[
= r
2
L = const . (9.27)
The Euler-Lagrange equation for r is
d
dt
[
r

[
r
= 0 = r r
2
+
1
c
2
d
dr
. (9.28)
Since
r =
dr
dt
= r
= r
L
r
2
= Lu
, (9.29)
we can write the second time derivative of r as
r = L
du
dt
= L
du
= Lu
L
r
2
= L
2
u
2
u
. (9.30)
Thus, the equation of motion (9.28) can be written as
L
2
u
2
u
r
L
2
r
4
+
1
c
2
d
dr
= 0 (9.31)
or, after dividing by u
2
L
2
,
u
+ u =
1
L
2
u
2
c
2
d
dr
. (9.32)
In the Newtonian limit of the Schwarzschild solution,
=
GM
r
,
1
c
2
d
dr
=
GM
c
2
r
2
=
GM
c
2
u
2
= mu
2
, (9.33)
so that the equation of motion becomes
u
+ u =
m
L
2
. (9.34)
Compared to this, the equation of motion in the Schwarzschild
case (9.21) has the additional term 3mu
2
. We have seen in (8.68)
that m 1.5 km in the Solar system. There, the ratio of the two
terms on the right-hand side of (9.21) is
3mu
2
m/L
2
= 3u
2
L
2
=
3
r
2
r
4
2
= 3(r
)
2
=
3v
2
c
2
7.7 10
8
(9.35)
for the innermost planet Mercury. Here, v
is the tangential ve-

locity along the orbit, v
= r
.
The equation of motion (9.21) in the Schwarzschild space-time
can thus be reduced to a Kepler problem with a potential which,
according to (9.32), is given by
1
L
2
u
2
c
2
d(r)
dr
=
m
L
2
+ 3mu
2
(9.36)
or
1
c
2
d(r)
dr
= mu
2
+ 3mL
2
u
4
=
m
r
2
+
3mL
2
r
4
, (9.37)
which leads to
(r)
c
2
=
m
r

mL
2
r
3
. (9.38)
As a function of x r/R
s
= r/2m, the eective potential V(r)
from (9.16) depends in an interesting way on L/R
s
= L/2m .
The function
V(x) =
_
1
1
x
_ _
1 +
2
x
2
_
(9.39)
corresponding to the eective potential (9.16) asymptotically be-
haves as V(x) 1 for x and V(x) for x 0.
For the potential to have a minimum, V(x) must have a vanishing
derivative, V
(x) = 0. This is the case where

0 = V
(x) =
1
x
2
_
1 +
2
x
2
_
_
1
1
x
_
2
2
x
3
(9.40)
or, after multiplication with x
4
,
x
2
2
2
x + 3
2
= 0 x
=
2

2
3 . (9.41)
Real solutions require

3. If <
3, particles with E
2
< 1
will crash directly towards r = R
s
.
The last stable orbit must thus have =
3 and thus x
= 3,
i.e. at r = 3R
s
, or three Schwarzschild radii. There, the eective
potential is
V(x = 3) =
2
3
_
1 +
3
9
_
=
8
9
. (9.42)
For >
3, the eective potential has a minimum at x

+
and a
maximum at x
which reaches the height V = 1 for = 2 at

x
= 2 and is higher for larger . This means that particles with

E 1 and L < 2R
s
will fall unimpededly towards r = R
s
.
9.1.3 The perihelion shift
The treatment of the Kepler problemin classical mechanics shows
that closed orbits in the Newtonian limit are described by
u
0
() =
1
p
(1 + e cos ) , (9.43)
where the parameter p is
p = a(1 e
2
) =
L
2
m
(9.44)
in terms of the semi-major axis a and the eccentricity e of the
orbit.
Assuming that the perturbation 3mu
2
in the equation of motion
(9.21) is small, we can approximate it by 3mu
2
0
, thus
u
+ u =
m
L
2
+
3m
p
2
(1 + e cos )
2
. (9.45)
The solution of this equation turns out to be simple because dif-
ferential equations of the sort
u
+ u =
_
_
A
Bcos
C cos
2
, (9.46)
which are driven harmonic-oscillator equations, have the particu-
lar solutions
u
1
= A , u
2
=
B
2
sin , u
3
=
C
2
_
1
1
3
cos 2
_
. (9.47)
Since the unperturbed equation u
+ u = m/L
2
has the Keplerian
solution u = u
0
, the complete solution is thus the sum
u = u
0
+ u
1
+ u
2
+ u
3
(9.48)
=
1
p
(1 + e cos ) +
3m
p
2
_
1 + e sin +
e
2
2
_
1
1
3
cos 2
__
.
This solution of (9.45) has its perihelion at = 0 because the
unperturbed solution u
0
was chosen to have it there. This can be
seen by taking the derivative with respect to ,
u
=
e
p
sin +
3me
p
2
_
sin + cos +
e
3
sin 2
_
(9.49)
and verifying that u
= 0 at = 0, i.e. the orbital radius r = 1/u

still has an extremum at = 0.
We now use equation (9.49) in the following way. Starting at the
perihelion at = 0, we wait for approximately one revolution at
= 2 + and see what has to be for u
to vanish again.
Thus, the condition for the next perihelion is
0 = sin +
3m
p
_
sin + (2 + ) cos +
e
3
sin 2
_
(9.50)
or, to rst order in the small angle ,

3m
p
_
2 + 2 +
2e
3

_
. (9.51)
Sorting terms, we nd
_
1
6m
p
_
1 +
e
3
_
_

6m
p
=
6m
a(1 e
2
)
. (9.52)
Substituting the Schwarzschild radius from (8.69), we can write
this result as

3R
s
a(1 e
2
)
. (9.53)
This turns out to be 6 times the result (1.45) from the scalar
theory of gravity discussed in 1.4.2, or
42
(9.54)
per century for Mercurys orbit, which reproduces the measure-
ment extremely well.
9.1.4 Light deection
For light rays, the condition 2[ = 1 that we had before for
material particles is replaced by 2[ = 0. Then, (9.14) changes to
r
2
E
2
1 2m/r
+
L
2
r
2
= 0 (9.55)
or
r
2
+
L
2
r
2
_
1
2m
r
_
= E
2
. (9.56)
Changing again the independent variable to and substituting
u = 1/r leads to the equation of motion for light rays in the
Schwarzschild space-time
u
2
+ u
2
=
E
2
L
2
+ 2mu
3
, (9.57)
which should be compared to the equation of motion for material
particles, (9.19). Dierentiation nally yields
u
+ u = 3mu
2
. (9.58)
Compared to u on the left-hand side, the term 3mu
2
is very small.
In the Solar system,
3mu
2
u
= 3mu =
3R
s
2r

R
s
R
,
10
6
. (9.59)
Thus, the light ray is almost given by the homogeneous solution
of the harmonic-oscillator equation u
+ u = 0, which is u
0
=
Asin + Bcos . We require that the closest impact at u
0
= 1/b
be reached when = /2, which implies B = 0 and A = 1/b, or
u
0
=
sin
b
r
0
=
b
sin
. (9.60)
Note that this is a straight line in plane polar coordinates, as it
should be!
Inserting this lowest-order solution as a perturbation into the
right-hand side of (9.58) gives
u
+ u =
3m
b
2
sin
2
=
3m
b
2
(1 cos
2
) , (9.61)
for which particular solutions can be found using (9.46) and
(9.47). Combining this with the unperturbed solution (9.60) gives
u =
sin
b
+
3m
b
2

3m
2b
2
_
1
1
3
cos 2
_
. (9.62)
Given the orientation of our coordinate system, i.e. with the clos-
est approach at = /2, we have 0 for a ray incoming from
the left at large distances. Then, sin and cos 2 1, and
(9.62) yields
u

b
+
2m
b
2
. (9.63)
In the asymptotic limit r , or u 0, this gives the angle

2m
b
. (9.64)
The total deection angle is twice this result,
= 2
4m
b
= 2
R
s
b
1.74
. (9.65)
This is twice the result from our simple consideration leading to
(4.83) which did not take the eld equations into account yet.
9.1.5 Spin precession
Let us now study how a gyroscope with spin s is moving along
a geodesic in the Schwarzschild space-time. Without loss of
generality, we assume that the orbit falls into the equatorial plane
= 0, and we restrict the motion to circular orbits.
Then, the four-velocity of the gyroscope is characterised by u
1
=
0 = u
2
because both r = x
1
and = x
2
are constant.
The equations that the spin s and the tangent vector u = of the
orbit have to satisfy are
(s, u) = 0 ,
u
s = 0 ,
u
u = 0 . (9.66)
The rst is because s falls into a spatial hypersurface perpendicu-
lar to the time-like four-velocity u, the second because the spin is
parallel transported, and the third because the gyroscope is mov-
ing along a geodesic curve.
We work in the same tetrad
(8.40) that we used to derive the

Schwarzschild solution. From (8.9), we know that
(
u
s)
= (ds
+ s
, u) = u(s
) +
(u)s
= s
(u)s
= 0 . (9.67)
With the connection forms in the Schwarzschild tetrad given in
(8.50), and taking into account that a = b and cot = 0, we nd
for the components of s
s
0
=
0
1
(u)s
1
= b
e
b
u
0
s
1
,
s
1
=
1
0
(u)s
0

1
2
(u)s
2

1
3
(u)s
3
= b
e
b
u
0
s
0
+
e
b
r
u
3
s
3
,
s
2
=
2
1
(u)s
1

2
3
(u)s
3
= 0 ,
s
3
=
3
1
(u)s
1

3
2
(u)s
2
=
e
b
r
u
3
s
1
, (9.68)
where we have repeatedly used that
1
(u) = u
1
= 0 = u
2
=
2
(u) (9.69)
and
2
3
= 0 because cot = 0.
Similarly, the geodesic equation
u
u = 0, specialised to u
1
= 0 =
u
2
, leads to
u
0
= b
e
b
u
0
u
1
= 0 ,
u
1
= b
e
b
(u
0
)
2
+
e
b
r
(u
3
)
2
= 0 ,
u
2
= 0 ,
u
3
=
e
b
r
u
1
u
3
= 0 . (9.70)
The second of these equations implies
_
u
0
u
3
_
2
=
1
b
r
. (9.71)
We now introduce a set of basis vectors orthogonal to u, namely
e
1
= e
1
, e
2
= e
2
, e
3
= u
3
e
0
+ u
0
e
3
. (9.72)
The orthogonality of e
1
and e
2
to u is obvious because of u
1
= 0 =
u
2
, and
(u, e
3
) = u
3
u
0
+ u
0
u
3
= 0 (9.73)
shows the orthogonality of u and e
3
. Recall that u
0
= u
0
, but
u
3
= u
3
because the metric is g = diag(1, 1, 1, 1) in this basis.
Since the basis e
i
spans the three-space orthogonal to u, the spin
s of the gyroscope can be expanded into this basis as s = s
i
e
i
. We
nd
s
0
= ( s
i
e
i
, e
0
) = u
3
s
3
, s
1
= s
1
, s
2
= s
2
, s
3
= u
0
s
3
,
(9.74)
which we can insert into (9.68) to nd
u
3
s
3
+ u
3
s
3
= u
3
s
3
= b
e
b
u
0
s
1
,
s
1
=
_
b
e
b
+
e
b
r
_
u
0
u
3
s
3
,
s
2
= 0 ,
u
0
s
3
+ u
0
s
3
= u
0
s
3
=
e
b
r
u
3
s
1
. (9.75)
Note that u
1
= 0 = u
3
according to (9.70). Using (9.71) and the
normalisation relation (u
0
)
2
(u
3
)
2
= 1, we obtain
s
1
= b
e
b
_
1
(u
0
)
2
(u
3
)
2
_
u
0
u
3
s
3
= b
e
b
u
0
u
3
s
3
,
s
2
= 0 ,
s
3
= b
e
b
u
0
u
3
s
1
. (9.76)
From now on, we shall drop the overbar, understanding that the s
i
denote the components of the spin with respect to the basis e
i
.
Next, we transform the time derivative from the proper time to
the coordinate time t. Since
u
0
= (
0
, u) = (e
a
dt, u) =
dt
d
e
a
, (9.77)
we have
t =
dt
d
= u
0
e
a
= u
0
e
b
, (9.78)
or
ds
i
dt
=
s
i
t
=
s
i
u
0
e
b
. (9.79)
Inserting this into (9.76) yields
ds
1
dt
=
b
u
3
e
2b
s
3
,
ds
2
dt
= 0 ,
ds
3
dt
=
b
u
3
e
2b
s
1
. (9.80)
Finally, using (8.40), we have
u
3
= (
3
, u) = (r sin d, u) = r(d, u
) = r
, (9.81)
which yields the angular frequency

d
dt
=
t
=
e
b
r
u
3
u
0
, (9.82)
which can be rewritten by means of (9.71),
2
=
_
u
3
u
0
_
2
1
r
2
e
2b
=
b
r
e
2b
=
1
2r
_
e
2b
_
. (9.83)
Now, the exponential factor was
e
2b
=
_
1
2m
r
_
(e
2b
)
=
2m
r
2
, (9.84)
whence
2
=
m
r
3
, (9.85)
which is Keplers third law.
Taking another time derivative of (9.80), we can use r = 0 for
circular orbits and u
3
= 0 because of (9.70). Thus,
d
2
s
1
dt
2
=
b
u
3
e
2b
ds
3
dt
=
b
2
e
4b
(u
3
)
2
s
1
(9.86)
and likewise for s
3
. This is an oscillator equation for s
1
with the
squared angular frequency
2
=
b
2
e
4b
(u
3
)
2
= b
2
e
4b
(u
0
)
2
(u
3
)
2
(u
3
)
2
= b
2
e
4b
_
1
1
b
r
_
=
b
e
4b
r
(1 + b
r) . (9.87)
Now, we use (9.82) to substitute the factor out front and nd
2
=
2
e
2b
(b
r + 1) . (9.88)
From (8.63), we further know that
rb
=
1
2
_
1 e
2b
_
=
1
2
_
1
1
1 2m/r
_
=
m
r
1
1 2m/r
, (9.89)
thus
rb
+ 1 =
r 3m
r 2m
=
1 3m/r
1 2m/r
(9.90)
and
2
=
2
e
2b
1 3m/r
1 2m/r
=
2
_
1
3m
r
_
. (9.91)
In vector notation, we can write (9.80) as
ds
dt
=

s ,

=
_
_
0
0
_
_
. (9.92)
Recall that we have projected the spin into the three-dimensional
space perpendicular to the direction of motion. Thus, the result
(9.92) shows that s precesses retrograde in that space about an
axis perpendicular to the plane of the orbit, since u
2
= 0.
After a complete orbit, the projection of s onto the plane of the
orbit has advanced by an angle
2
_
1
3m
r
_

2

s
, (9.93)
and thus the precession frequency is
s
=
_
m
r
3
_
_
1
_
1
3m
r
_
1/2
_
_

_
GM
r
3
_
1/2
GM
2r
=
3
2
(GM)
3/2
r
5/2
.
(9.94)
If we insert the Earths mass and radius here, we nd
s
8.4
_
R
earth
r
_
5/2
. (9.95)
Chapter 10
Schwarzschild Black Holes
10.1 The singularity at r = 2m
10.1.1 Free fall towards the centre
Before we can continue discussing the physical meaning of the
Schwarzschild metric, we need to clarify the nature of the singu-
larity at the Schwarzschild radius, r = 2m. Upon closer inspec-
tion, it seems to lead to contradictory conclusions.
Let us begin with an observer falling freely towards the centre
of the Schwarzschild space-time along a radial orbit. Since

=
0, the angular momentum vanishes, L = 0, and the equation of
motion (9.15) reads
r
2
+
_
1
2m
r
_
= E
2
. (10.1)
Suppose the observer was at rest at r = R, then E
2
= 1 2m/R
and E
2
< 1, and we have
r
2
=
_
1
2m
R
_
_
1
2m
r
_
= 2m
_
1
r

1
R
_
, (10.2)
which yields
dr
_
2m
_
1
r

1
R
_
= d , (10.3)
where is the proper time.
This equation admits a parametric solution. Starting from
r =
R
2
(1 + cos ) , dr =
R
2
sin d , (10.4)
119
CHAPTER 10. SCHWARZSCHILD BLACK HOLES 120
we rst see that
1
r

1
R
=
2
R(1 + cos )

1
R
=
1
R
_
1 cos
1 + cos
_
=
1
R
(1 cos )
2
sin
2
, (10.5)
where we have used in the last step that 1 cos
2
= sin
2
. This
result allows us to translate (10.3) into
Rsin dr
2m(1 cos )
=
R
R
2
2m
sin
2
d
1 cos
=
_
R
3
8m
(1 + cos )d . (10.6)
Integrating, we nd that this solves (10.3) if
=
_
R
3
8m
( + sin ) , d =
_
R
3
8m
(1 + cos ) . (10.7)
At = 0, the proper time is = 0 and r = R, i.e. the proper
time starts running when the free fall begins. The centre r = 0 is
reached when = , i.e. after the proper time
0
=
_
R
3
8m
. (10.8)
This indicates that the observer falls freely within nite time
through the singularity at r = 2m without encountering any
(kinematic) problem.
10.1.2 Problems with the Schwarzschild coordinates
However, let us now describe the radial coordinate r as a function
of the coordinate time t. Using (9.13), we rst nd
r =
dr
dt

t =
dr
dt
E
1 2m/r
. (10.9)
Next, we introduce a new, convenient radial coordinate r such that
d r =
dr
1 2m/r
, (10.10)
which can be integrated as follows,
dr
1 2m/r
=
r/2m 1 + 1
r/2m 1
dr = dr +
dr
r/2m 1
= dr + 2md ln
_
r
2m
1
_
, (10.11)
giving
r = r + 2mln
_
r
2m
1
_
. (10.12)
With this, we nd
r =
E
1 2m/r
dr
dt
= E
d r
dt
, (10.13)
and thus, from the equation of motion (10.1),
E
2
_
d r
dt
_
2
= E
2
+
2m
r
1 . (10.14)
Approaching the Schwarzschild radius from outside, i.e. in the
limit r 2m+, we have
lim
r2m+
r = lim
r2m+
2m
_
1 + ln
_
r
2m
1
__
= , (10.15)
However, in the same limit, the equation of motion says
E
2
_
d r
dt
_
2
E
2
, (10.16)
and thus
d r
dt
1 . (10.17)
Of the two signs, we have to select the negative because of r
, as (10.15) shows. Therefore, an approximate solution of the
equation of motion near the singularity is
r t + const. (10.18)
or, substituting r for r,
t + const. = r + 2m ln
_
r
2m
1
_
2m
_
1 + ln
_
r
2m
1
__
. (10.19)
Solving for r, we nd
ln
_
r
2m
1
_

t + const.
2m
1 (10.20)
or
r 2m + const. e
t/2m
, (10.21)
showing that the Schwarzschild radius is only reached after in-
nitely long coordinate time!
Finally, radial light rays are described by radial null geodesics,
thus satisfying
0 = ds
2
=
_
1
2m
r
_
dt
2
+
dr
2
1
2m
r
(10.22)
or
dr
dt
=
_
1
2m
r
_
, (10.23)
suggesting that the light cones become innitely narrow as r
2m+.
These results are obviously dissatisfactory: while a freely falling
observer reaches the Schwarzschild radius and even the centre of
the Schwarzschild space-time after nite proper time, the coordi-
nate time becomes innite even for reaching the Schwarzschild
radius, and the attening of the light cones as one approaches the
Schwarzschild radius is totally unwanted.
10.1.3 Curvature at r = 2m
Moreover, consider the components of the Ricci tensor given in
(8.57) and (8.58) near the Schwarzschild radius. Since a = b
and
b =
1
2
ln
_
1
2m
r
_
= a (10.24)
from (8.65), the needed derivatives are
a
=
m
r(r 2m)
= b
, a
=
2m(r m)
r
2
(r 2m)
2
= b
. (10.25)
Thus,
R
00
=
_
a
+ 2a
2
+
2a
r
_ _
1
2m
r
_
= 0 = R
11
(10.26)
and
R
22
=
2a
r
_
1
2m
r
_
+
1
r
2
_
1 e
2b
_
=
2m(r 2m)
r
3
(r 2m)
+
2m
r
3
= 0 = R
33
, (10.27)
i.e. the components of the Ricci tensor in the Schwarzschild tetrad
remain perfectly regular at the Schwarzschild radius!
10.2 The Kruskal Continuation of the
Schwarzschild Space-Time
10.2.1 Transformation to Kruskal coordinates
Martin Kruskal (19252006)
We shall now try to remove the obvious problems with the
Schwarzschild coordinates by transforming (t, r) to new coordi-
nates (u, v), leaving and , requiring that the metric can be writ-
ten as
g = f
2
(u, v)(dv
2
du
2
) + r
2
(d
2
+ sin
2
d
2
) (10.28)
with a function f (u, v) to be determined.
Provided f (u, v) 0, radial light rays propagate as in a two-
dimensional Minkowski metric according to
dv
2
= du
2
,
_
du
dv
_
2
= 1 , (10.29)
which shows that the light cones remain undeformed in the new
coordinates.
The Jacobian matrix of the transformation from the
Schwarzschild coordinates (t, r, , ) to the new coordinates
(v, u, , ) is
J
=
_
_
v
t
u
t
0 0
v
r
u
r
0 0
0 0 1 0
0 0 0 1
_
_
, (10.30)
where subscripts denote derivatives,
v
t
=
t
v , v
r
=
r
v (10.31)
and likewise for u. The metric g in the new coordinates,
g = diag(f
2
, f
2
, r
2
, r
2
sin
2
) , (10.32)
is transformed into the original Schwarzschild coordinates by
g = J gJ
T
(10.33)
=
_
_
f
2
_
v
2
t
u
2
t
_
f
2
(v
t
v
r
u
t
u
t
) 0 0
f
2
(v
t
v
r
u
t
u
t
) f
2
_
v
2
r
u
2
r
_
0 0
0 0 r
2
0
0 0 0 r
2
sin
2
_
,
which, by comparison with our requirement (10.28), yields the
three equations
_
1
2m
r
_
= f
2
_
v
2
t
u
2
t
_
,
1
1 2m/r
= f
2
_
v
2
r
u
2
r
_
,
0 = v
t
v
r
u
t
u
t
. (10.34)
For convenience, we now fall back to the radial coordinates r and
introduce the function
F( r)
1 2m/r
f
2
(r)
, (10.35)
assuming that f will turn out to depend on r only since any de-
pendence on time and on the angles and phi is forbidden in a
static, spherically-symmetric space-time. Then,
v
r
=
dr
d r
v
r
=
_
1
2m
r
_
v
r
(10.36)
and the same for u.
The equations (10.34) then transform to
F( r) = v
2
t
u
2
t
, F( r) = v
2
r
u
2
r
, v
t
v
r
u
t
u
r
= 0 . (10.37)
Now, we add the two equations containing F( r) and then add and
subtract from the result twice the third equation from (10.37).
This yields
(v
t
v
r
)
2
= (u
t
u
r
)
2
. (10.38)
Taking the square root of this equation, we can choose the signs.
The choice
v
t
+ v
r
= u
t
+ u
r
, v
t
v
r
= (u
t
u
r
) (10.39)
avoids that the Jacobian matrix could become singular, det J = 0.
Adding and subtracting the equations (10.39), we nd
v
t
= u
r
, u
t
= v
r
. (10.40)
Taking partial derivatives once with respect to t and once with
respect to r allows us to combine these equations to nd the wave
equations
v
tt
v
r r
= 0 , u
tt
u
r r
= 0 , (10.41)
which are solved by any two functions h
propagating with unit

velocity,
v = h
+
( r + t) + h
( r t) , u = h
+
( r + t) h
( r t) , (10.42)
where the signs were chosen such as to satisfy the sign choice in
(10.39).
Now, since
v
t
= h
+
h
, u
t
= h
+
+ h
,
v
r
= h
+
+ h
, u
r
= h
+
h
, (10.43)
where the primes denote derivatives with respect to the functions
arguments, we nd from (10.37)
F( r) =
_
h
+
h
_
2
_
h
+
+ h
_
2
= 4h
+
h
. (10.44)
We start from outside the Schwarzschild radius, assuming r >
2m, where also F( r) > 0 according to (10.35). The derivative of
(10.44) with respect to r yields
F
( r) = 4
_
h
+
h
+ h
+
h
_
(10.45)
or, with (10.44),
F
F
=
h
+
h
+
+
h
. (10.46)
the derivative of (10.44) with respect to time yields
0 = 4
_
h
+
h
+
h
_

h
+
h
= 0 . (10.47)
The sum of these two equations gives
(ln F)
= 2(ln h
+
)
. (10.48)
Now, the left-hand side depends on r, the right-hand side on the
independent variable r + t. Thus, the two sides of this equation
must equal the same constant, which we call 2C:
(ln F)
= 2C = 2(ln h
+
)
. (10.49)
The left of these equations yields
ln F = 2C r + const. F = const.e
2C r
, (10.50)
while the right equation gives
ln h
+
= C( r + t) + const. (10.51)
or
h
+
= const.e
C( r+t)
. (10.52)
For later convenience, we choose the remaining constants in
(10.51) and (10.52) such that
F( r) = C
2
e
2C r
, h
+
( r + t) =
1
2
e
C( r+t)
, (10.53)
and (10.47) gives
h
( r t) =
1
2
e
C( rt)
, (10.54)
where the negative sign must be chosen to satisfy both (10.44)
and F > 0.
Working our way back, we nd
u = h
+
( r + t) h
( r t) =
1
2
_
e
C( r+t)
+ e
C( rt)
_
= e
C r
cosh(Ct) =
_
r
2m
1
_
2mC
e
Cr
cosh(Ct) , (10.55)
using (10.12) for r. Similarly, we nd
v =
_
r
2m
1
_
2mC
e
Cr
sinh(Ct) , (10.56)
and the function f follows from (10.35),
f
2
=
1 2m/r
F
=
1 2m/r
C
2
e
2C r
=
1 2m/r
C
2
e
2Cr
exp
_
4mC ln
_
r
2m
1
__
=
2m
rC
2
_
r
2m
1
_
14mC
e
2Cr
. (10.57)
Now, since we want f to be non-zero and regular at r = 2m, we
must require 4mC = 1, which nally xes the Kruskal transfor-
mation of the Schwarzschild metric,
u =
_
r
2m
1 e
r/4m
cosh
_
t
4m
_
,
v =
_
r
2m
1 e
r/4m
sinh
_
t
4m
_
,
f =
32m
3
r
e
r/2m
. (10.58)
We have thus achieved our goal to replace the Schwarzschild
coordinates by others in which the Schwarzschild metric remains
prefectly regular at r = 2m.
10.2.2 Physical meaning of the Kruskal continuation
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
v
u
I
II
III
IV
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
v
u
I
II
III
IV
Illustration of the Kruskal contin-
uation in the u-v plane. The
Schwarzschild domain r > 2m is
shaded in blue, the forbidden region
r < 0 is shaded in red.
Since cosh
2
(x) sinh
2
(x) = 1, eqs. (10.58) imply
u
2
v
2
=
_
r
2m
1
_
e
r/2m
,
v
u
= tanh
t
4m
. (10.59)
This means u = v for r = 2m, which is reached for t .
Lines of constant coordinate time t are straight lines through the
origin in the (u, v) plane with slope tanh(t/4m), and lines of con-
stant radial coordinate r are hyperbolae.
The metric in Kruskal coordinates (10.28) is regular as long as
r(u, v) > 0, which is the case for
u
2
v
2
> 1 , (10.60)
as (10.59) shows. The hyperbola limiting the regular domain in
the Kruskal manifold is thus given by v
2
u
2
= 1. If (10.60) is
satised, r is uniquely dened, because the equation
(x) (x 1)e
x
= u
2
v
2
> 1 (10.61)
is monotonic for x > 0:
(x) = e
x
(x 1) + e
x
= xe
x
> 0 (x > 0) . (10.62)
The domain of the original Schwarzschild solution is restricted
to u > 0 and v < u, but this is a consequence of our choice for
the relative signs of h
in (10.54). We could as well have chosen

h
+
< 0 and h
> 0, which would correspond to the replacement

(u, v) (u, v).
The original Schwarzschild solution for r < 2m also satises Ein-
steins vacuum eld equations. There, the Schwarzschild metric
shows that r then behaves like a time coordinate because g
rr
< 0,
and t behaves like a spatial coordinate.
Looking at the denition of F( r) in (10.35), we see that r < 2m
corresponds to F < 0, which implies that h
+
and h
must have
the same (rather than opposite) signs because of (10.44). This
interchanges the functions u and v from (10.58), i.e. u v and
v u. Then, the condition v < u derived for r > 2m changes to
v > u.
In summary, the exterior of the Schwarzschild radius corresponds
to the domain u > 0, v < u, and its interior is bounded in the
(u, v) plane by the lines u > 0, v = u and v
2
u
2
= 1.
Radial light rays propagate according to ds
2
= 0 or dv = du,
i.e. they are straight diagonal lines in the (u, v) plane. This shows
that light rays can propagate freely into the region r < 2m, but
there is no causal connection from within r < 2m to the outside.
The Killing vector eld K =
t
for the Schwarzschild space-time
outside r = 2m becomes space-like for r < 2m, which means that
the space-time cannot be static any more inside the Schwarzschild
radius.
10.2.3 Eddington-Finkelstein coordinates
We now want to study the collapse of an object, e.g. a star. For
this purpose, coordinates originally introduced by Eddington and
re-discovered by Finkelstein are convenient, which are dened by
r = r
, =
, =
t = t
2mln
_
r
2m
1
_
(10.63)
in analogy to the radial coordinate r from (10.10), where the up-
per and lower signs in the second line are valid for r > 2m and
r < 2m, respectively.
Since
e
t/4m
= e
t
/4m
_
_
_
r
2m
1
_
1/2
(r > 2m)
_
r
2m
+ 1
_
1/2
(r < 2m)
, (10.64)
inserting these expressions into the Kruskal coordinates (10.58)
shows that they are related by
u =
e
r/4m
2
_
e
t
/4m
+
r 2m
2m
e
t
/4m
_
v =
e
r/4m
2
_
e
t
/4m
r 2m
2m
e
t
/4m
_
, (10.65)
such that
r 2m
2m
e
r/2m
= u
2
v
2
, e
t
/2m
=
r 2m
2m
u + v
u v
. (10.66)
Arthur S. Eddington (18821944)
David R. Finkelstein (1929)
The rst of these equations shows again that r can be uniquely
determined from u and v if u
2
v
2
> 1. The second equation
determines t
uniquely provided r > 2m and (u + v)/(u v) > 0,

or r < 2m and (u + v)/(u v) < 0. This is possible if v > u.
Using
dt = dt
2m
1
r/2m 1
dr
2m
= dt

2m
r
dr
1 2m/r
, (10.67)
we nd
_
1
2m
r
_
dt
2
=
_
1
2m
r
_
dt
2
+
4m
r
dt
dr
4m
2
r
2
dr
2
1 2m/r
,
(10.68)
and thus the metric in Eddington-Finkelstein coordinates reads
g =
_
1
2m
r
_
dt
2
+
_
1
4m
2
r
2
_
dr
2
1 2m/r
+
4m
r
dt
dr + r
2
d
2
(10.69)
=
_
1
2m
r
_
dt
2
+
_
1 +
2m
r
_
dr
2
+
4m
r
dt
dr + r
2
d
2
.
Thus, the components of the metric no longer depend on t
, but
we now have o-diagonal elements.
For radial light rays, d = 0 and ds
2
= 0, which implies from
(10.69)
_
1
2m
r
_
dt
2
_
1 +
2m
r
_
dr
2
4m
r
dt
dr = 0 , (10.70)
which can be factorised as
__
1
2m
r
_
dt
2
_
1 +
2m
r
_
dr
2
_
(dt
+ dr) = 0 . (10.71)
Thus, the light cones are dened either by
dr
dt
= 1 r = t
+ const. (10.72)
or by
dr
dt
=
r 2m
r + 2m
, (10.73)
which shows that dr/dt
1 for r 0, dr/dt
= 0 for r = 2m,
and dr/dt
= 1 for r . Due to the vanishing derivative

of r with respect to t
at r = 2m, geodesics cannot cross the

Schwarzschild radius from inside, but they can from outside be-
cause of (10.72).
10.2.4 Redshift approaching the Schwarzschild radius
Suppose a light-emitting source (e.g. an astronaut with a torch) is
falling towards a (Schwarzschild) black hole, what does a distant
observer see? let v and u be the four-velocities of the astronaut and
the observer, respectively. Then according to (4.48) the redshift
of the light from the torch as seen by the observer is
1 + z =
em
obs
=
(k, v)
(k, u)
, (10.74)
where k is the wave vector of the light.
We transform to the retarded time t
ret
t r, with r given by
(10.12). Then, (10.10) implies that
dt
ret
= dt
dr
1 2m/r
, (10.75)
thus
dt
2
= dt
2
ret
+
dr
2
(1 2m/r)
2
+
2dt
ret
dr
1 2m/r
(10.76)
and the Schwarzschild metric transforms to
g =
_
1
2m
r
_
dt
2
ret
2dt
ret
dr + r
2
d
2
. (10.77)
For radial light rays, d = 0, this means
0 =
_
1
2m
r
_
dt
2
ret
2dt
ret
dr , (10.78)
which is possible only if dt
ret
= 0. This shows that light rays must
propagate along r, or k
r
, which is of course a consequence of
our using the retarded time t
ret
.
For a distant observer at a xed distance r > 2m, the metric
(10.77) simplies to
g dt
2
ret
, (10.79)
which shows that t
ret
is also that distant observers proper time.
Expanding the astronauts velocity as
v =

t
ret
t
ret
+ r
r
, (10.80)
we nd
(k, v) = (
r
, v) = g
t
ret
r
t
ret
=
t
ret
(10.81)
because according to the metric (10.77) (
r
,
r
) = 0.
Similarly, we have for a xed distant observer u =

t
t
. Expanding
t
into
t
ret
and
r
, we nd
t
ret
= a
t
t
+ a
r
r
, a
t
= dt
ret
t
=
_
dt
dr
1 2m/r
_
t
= 1
a
r
= dt
ret
r
=
1
1 2m/r
(10.82)
using (10.75), and thus
(k, u) =

t
_
r
,
t
ret
+
r
1 2m/r
_
=
t 1 (10.83)
because of (10.79). This gives the redshift
1 + z

t
ret
=

t
r
1 2m/r
. (10.84)
When restricted to radial motion,

= 0 = L, the equation of
motion (9.15) is
r
2
+ 1
2m
r
= E
2
, (10.85)
where E was dened as
E =
t
_
1
2m
r
_
, (10.86)
see (9.12). Requiring that the astronauts proper time increases
with the coordinate time,

t > 0 and E < 0. Since r < 0 for the
infalling astronaut,
r = E
_
1

E
2
, (10.87)
with 1 2m/r.
The redshift (10.84) can now be written
1 + z =
E
_
1 +
_
1

E
2
_
_

2E
(10.88)
near the Schwarzschild radius, where 0+.
The change in the astronauts radius with the observers proper
time is
dr
dt
ret
=
r
t
ret
=
r
1 + z

E
2E
=
2
. (10.89)
Since, by denition,

2
=
1
2
_
1
2m
r
_

r 2m
4m
, (10.90)
we have
d(r 2m) = dr =
2
dt
ret

r 2m
4m
d , (10.91)
thus
d(r 2m)
r 2m

dt
ret
4m
r 2m e
t
ret
/4m
(10.92)
and

e
t
ret
/4m
r

e
t
ret
/4m
4m
, (10.93)
showing that the astronauts redshift
1 + z
2E
8mEe
t
ret
/4m
(10.94)
grows exponentially as he approaches the Schwarzschild radius.
Chapter 11
Charged, Rotating Black Holes
11.1 Metric and electromagnetic eld
11.1.1 The Reissner-Nordstrm solution
The Schwarzschild solution is a very important exact solution of
Einsteins vacuum equations, but we expect that real objects col-
lapsing to become black holes may be charged and rotating. We
shall now generalise the Schwarzschild solution in these two di-
rections.
First, we consider a static, axially-symmetric solution of Ein-
steins equations in presence of an electromagnetic charge q at
the origin of the Schwarzschild coordinates, i.e. at r = 0. The
electromagnetic eld will then also be static and axially symmet-
ric.
Hans J. Reissner (18741967)
Expressing the eld tensor in the Schwarzschild tetrad, we thus
expect the Faraday 2-form (5.71) to be
F =
q
r
2
dt dr =
q
r
2
e
ab
0

1
. (11.1)
We shall verify below that a = b also for a Schwarzschild solu-
tion with charge, so that the exponential factor will become unity
later.
Gunnar Nordstrm (18811923)
The electromagnetic energy-momentum tensor
T
=
1
4
_
F

1
4
g
_
(11.2)
is now easily evaluated. Since the only non-vanishing compo-
nent of F
is F
01
and the metric is diagonal in the Schwarzschild
tetrad, g = diag(1, 1, 1, 1), we have
F
= F
01
F
01
+ F
10
F
10
= 2F
2
01
. (11.3)
132
CHAPTER 11. CHARGED, ROTATING BLACK HOLES 133
Using this, we nd the components of the energy-momentum ten-
sor
T
00
=
1
4
_
F
01
F
0
1

1
2
F
2
01
_
=
1
8
F
2
01
=
q
2
8r
4
e
2(a+b)
,
T
11
=
1
4
_
F
10
F
1
0
+
1
2
F
2
01
_
=
q
2
8r
4
e
2(a+b)
= T
00
,
T
22
= F
2
01
=
q
2
8r
4
e
2(a+b)
= T
33
. (11.4)
Inserting these expressions instead of zero into the right-hand side
of Einsteins eld equations yields, with (8.60),
G
00
=
1
r
2
e
2b
_
1
r
2

2b
r
_
=
8G
c
4
T
00
=
Gq
2
c
4
r
4
e
2(a+b)
,
G
11
=
1
r
2
+ e
2b
_
1
r
2
+
2a
r
_
=
8G
c
4
T
00
= G
00
. (11.5)
Adding these two equations, we nd a
+ b
= 0, which implies
a + b = 0 because the functions have to tend to zero at innity.
This conrms that we can identify dt dr =
0

1
and write
F
01
= 1/r
2
.
Analogous to (8.63), we note that the rst of equations (11.5) with
a = b is equivalent to
_
re
2b
_
= 1
Gq
2
c
4
r
2
, (11.6)
which gives
e
2b
= e
2a
= 1
2m
r
+
Gq
2
c
4
r
2
, (11.7)
if we use 2m as the integration constant as for the neutral
Schwarzschild solution.
Dening
r
2
2mr +
Gq
2
c
4
, (11.8)
we thus obtain the metric of a charged Schwarzschild black hole,
g =
r
2
dt
2
+
r
2
dr
2
+ r
2
_
d
2
+ sin
2
d
_
. (11.9)
This is the Reissner-Nordstrm solution. Of course, for q = 0
it returns to the Schwarzschild solution. Recall that the unit of
the electric charge in the cgs system is g
1/2
cm
3/2
s
1
, so that the
quantity Gq
2
/c
4
has the dimension of a squared length.
Before we proceed, we should verify that Maxwells equations
are satised. First, we note that the Faraday 2-form (11.1) is exact
because it is the exterior derivative of the 1-form
A =
e
r
dt , dA =
e
r
2
dr dt =
e
r
2
dt dr =
e
r
2

0
1
. (11.10)
Thus, by Poincar es lemma, dF = d
2
A = 0, so that the homoge-
neous Maxwell equations are satised.
Moreover, we notice that
F =
e
r
2

2

3
, (11.11)
which is easily veried using (5.66),
(
0

1
) =
1
2
g
00
g
11
01

(11.12)
=
1
2
_
2

3

3

2
_
=
2

3
.
Inserting the Schwarzschild tetrad from (8.40) yields
F = e sin d d = d (e cos d) , (11.13)
which shows, again by Poincar es lemma, that d(F) = 0, hence
also (d)F = 0 and F = 0, so that also the inhomogeneous
Maxwell equations (in vacuum!) are satised.
11.1.2 The Kerr-Newman solution
Roy Kerr (1934)
The formal derivation of the metric of a rotating black hole is
a formidable task which we cannot possibly demonstrate during
this lecture. We thus start with general remarks on the expected
form of the metric and then immediately quote the metric coe-
cients without deriving them.
In presence of angular momentum, we expect the spherical sym-
metry of the Schwarzschild solution to be broken. Instead, we
expect that the solution must be axisymmetric, with the axis xed
by the angular momentum. Moreover, we seek to nd a stationary
solution.
Then, the group R SO(2) must be an isometry of the met-
ric, where R represents the stationarity and SO(2) the (two-
dimensional) rotations about the symmetry axis. Expressing these
symmetries, there must be a time-like Killing vector eld k and
another Killing vector eld m which is tangential to the orbits of
SO(2).
These two Killing vector elds span the tangent spaces of the
two-dimensional submanifolds which are the orbits of RSO(2),
i.e. cylinders.
We can choose adapted coordinates t and such that k =
t
and
m =
. Then, the metric can be decomposed as

g = g
ab
(x
i
)dx
a
dx
b
+ g
i j
(x
k
)dx
i
dx
j
, (11.14)
where indices a, b = 0, 1 indicate the coordinates on the orbits
of R SO(2), and indices i, j, k = 2, 3 the others. Note that, due
to the symmetry imposed, the remaining metric coecients can
only depend on the coordinates x
i
.
Astationary, axi-symmetric space-time (M, g) can thus be foliated
into M = , where is dieomorphic to the orbits of R
SO(2), and the metric coecients in adapted coordinates can only
depend on the coordinates of . We write
(4)
g = + g (11.15)
and have
=
ab
(x
i
)dx
a
dx
b
. (11.16)
The coecients
ab
are scalar products of the two Killing vector
elds k and m,
(
ab
) =
_
(k, k) (k, m)
(k, m) (m, m)
_
, (11.17)
and we abbreviate the determinant of by

det =
_
(k, k)(m, m) + (k, m)
2
. (11.18)
Without proof, we now quote the metric of a stationary, axially-
symmetric solution of Einsteins eld equations for either vacuum
or an electro-magnetic eld. Dening
= r
2
2mr + Q
2
+ a
2
,
2
= r
2
+ a
2
cos
2
,
2
= (r
2
+ a
2
)
2
a
2
sin
2
, (11.19)
we can write the coecients of the metric as
g
tt
= 1 +
2mr Q
2
2
=
a
2
sin
2

2
,
g
t
=
2mr Q
2
2
a sin
2
=
r
2
+ a
2

2
a sin
2
,
g
rr
=
, g
=
2
, g
2
sin
2
. (11.20)
The quantities Q and a abbreviate the charge q and the angular
momentum L of the black hole. They are dened by
Q
2
:=
Gq
2
c
4
, a :=
L
Mc
=
GL
mc
3
(11.21)
and both have the dimension of a length.
Evidently, for a = 0 = Q, = r, = r
2
2mr and = r
2
and we
return to the Schwarzschild solution (8.66). For a = 0, we still
have = r and = r
2
, but = r
2
2mr + Q
2
as in (11.8), and
we return to the Reissner-Nordstrm solution (11.9). For Q = 0,
we obtain the Kerr solution for a rotating, uncharged black hole,
and for a 0 and Q 0, the solution is called Kerr-Newman
solution.
Also without derivation, we quote that the vector potential of the
rotating, charged black hole is given by the 1-form
A =
qr
2
_
dt a sin
2
d
_
, (11.22)
from which we obtain the Faraday 2-form
F = dA =
_
q
2
2qr
2
4
dr +
qr
4
2a
2
sin cos d
_
_
dt a sin
2
d
_
+
qr
2
2a sin cos d d
=
q(r
2
a
2
cos
2
)
4
dr
_
dt a sin
2
d
_
+
2qar
4
sin cos d
__
2
+ a
2
_
d adt
_
. (11.23)
For a = 0, this trivially returns to the eld (11.1) for the Reissner-
Nordstrm solution.
The eld components can now be read o the result (11.23). We
nd
E
r
= F
rt
=
q(r
2
a
2
cos
2
)
4

q
r
2
,
E
=
F
t
r
=
2qa
2
4
sin cos 0 ,
E
=
F
t
r sin
= 0 (11.24)
for the electric eld and
B
r
=
F
r
2
sin
=
2qa
r
4
cos
_
2
+ a
2
_

2qa
r
3
cos ,
B
=
F
r
r sin
=
aq(r
2
a
2
cos
2
)
r
4
sin
aq
r
3
sin ,
B
=
F
r
r
= 0 (11.25)
for the magnetic eld. In the limit of large r, the electric eld thus
becomes that of a point charge q at the origin, and the magnetic
eld has a characteristic dipolar structure.
It is known from electrodynamics that the magnetic dipole mo-
ment of a sphere with charge q, mass M and angular momentum
L is
= g
q
L
2Mc
, (11.26)
where g is the gyromagnetic moment. A comparison of the mag-
netic eld in (11.25) with the magnetic dipole eld from electro-
dynamics,
B =
3( e
r
)e
r

r
3
, (11.27)
shows that the magnetic dipole moment of a charged, rotating
black holes has the modulus
= qa =
qL
Mc
= 2
qL
2Mc
, (11.28)
showing that its gyromagnetic moment is g = 2.
11.1.3 Schwarzschild horizon, ergosphere and Killing
horizon
By construction, the Kerr-Newman metric (11.20) has the two
Killing vector elds k =
t
, expressing the stationarity of the so-
lution, and m =
, which expresses its axial symmetry.

Since the metric coecients in adapted coordinates satisfy
g
tt
= (k, k) , g
= (m, m) , g
t
= (k, m) , (11.29)
they have an invariant meaning which will now be claried.
Let us consider an observer moving with r = const. and =
const. with uniform angular velocity . If her four-velocity is u,
then
=
d
dt
=
t
=
u
u
t
(11.30)
for a static observer at innity, whose proper time can be identi-
ed with the coordinate time t. Correspondingly, we can expand
the four-velocity as
u = u
t
t
+ u
= u
t
_
t
+
_
= u
t
(k + m) , (11.31)
inserting the Killing vector elds. Let
k + m ((k + m, k + m))
1/2
(11.32)
dene the norm of k + m, then the four-velocity is
u =
k + m
k + m
. (11.33)
Obviously, k + m is a time-like Killing vector eld, at least at
suciently large distances from the black hole. Since then
(k + m, k + m) = (k, k) + (m, m) + 2(k, m)
= g
tt
+ g
+ 2
2
g
t
< 0 , (11.34)
k + m becomes light-like for angular velocities
=
g
t

_
g
2
t
g
tt
g
. (11.35)
If we dene

g
t
g
=
(k, m)
(m, m)
, (11.36)
we can write (11.35) as
=
_
g
tt
g
. (11.37)
For an interpretation of , we note that freely-falling test particles
on radial orbits have zero angular momentum and thus (u, m) = 0.
By (11.33), this implies
0 = (k + m, m) = g
t
+ g
(11.38)
and thus
=
g
t
g
= (11.39)
according to the denition (11.36). This shows that is the
angular velocity of a test particle falling freely towards the black
hole on a radial orbit.
The minimum angular velocity
from (11.37) vanishes if and

only if g
tt
= (k, k) = 0, i.e. if the Killing vector eld k turns
light-like. With (11.20), this is so where
0 = a
2
sin
2
= 2mr r
2
Q
2
a
2
cos
2
, (11.40)
i.e. at the radius
r
0
= m +
_
m
2
Q
2
a
2
cos
2
. (11.41)
This is the static limit: for an observer at this radius to remain
static with respect to observers at innity (i.e. with respect to the
xed stars), she would have to move with the speed of light. At
smaller radii, observers cannot remain static against the drag of
the rotating black hole.
We have seen in (4.48) that the light emitted by a source with
four-velocity u
s
is seen by an observer with four-velocity u
o
with
a redshift
s
=
(
k, u
o
)
(
k, u
s
)
, (11.42)
where

k is the wave vector of the light.
Observers at rest in a stationary space-time have four-velocities
proportional to the Killing vector eld k,
u =
k
(k, k)
, hence k =
_
(k, k) u . (11.43)
Moreover, the projection of the wave vector

k on the Killing vec-
tor eld k does not change along the light ray, since
k
(
k, k) = (
k
k, k) + (
k,
k
k)
=

k
= 0 , (11.44)
because light rays are null geodesics,
k
k = 0, and because the

Killing equation (5.34) implies that the last expression is the con-
traction of the symmetric tensor eld

k
with the antisymmetric

tensor eld
.
Using this in a combination of (11.42) and (11.43), we obtain
s
=
(
k, k)
o
(
k, k)
s
(k, k)
s
(k, k)
o
=
(k, k)
s
(k, k)
o
. (11.45)
For an observer at rest far away from the black hole, (k, k)
o
1,
and the redshift becomes
1 + z =
(k, k)
s
, (11.46)
which tends to innity as the source approaches the static limit.
The minimum and maximum angular velocities
from (11.37)
both become equal to for
2
=
_
g
t
g
_
2
=
g
tt
g
g
2
t
g
tt
g
= 0 . (11.47)
This equation means that the Killing eld k +m turns light-
like,
(, ) = (k, k) + 2(k, m) +
2
(m, m)
= g
tt
+ 2g
t
+
2
g
= g
tt
2
g
2
t
g
+
g
2
t
g
=
g
tt
g
g
2
t
g
= 0 . (11.48)
Interestingly, writing the expression from (11.47) with the metric
coecients (11.20) leads to the simple result
g
2
t
g
tt
g
= sin
2
, (11.49)
so that the condition (11.47) is equivalent to
0 = = r
2
2mr + q
2
+ a
2
, (11.50)
which describes a spherical hypersurface with radius
r
H
= m +
_
m
2
q
2
a
2
. (11.51)
By its denition (11.36), the angular velocity on this hypersur-
face H can be written as
H
=
g
t
g
H
=
a(2mr q
2
)
H
=
a(2mr
H
q
2
)
(r
2
H
+ a
2
)
2
, (11.52)
since
2
= (r
2
+ a
2
)
2
because of = 0 at r
H
. Because of (11.50),
the numerator is a(r
2
H
+ a
2
), and we nd
H
=
a
r
2
H
+ a
2
. (11.53)
This means remarkably that the hypersurface H is rotating at con-
stant angular velocity like a solid body.
Since the hypersurface H is dened by the condition = 0, its
normal vectors are given by
grad = d
. d = 2(r m)dr (11.54)

Thus, the norm of the normal vectors is
(grad, grad) = 4g
rr
(r m)
2
, (11.55)
now, according to (11.20), g
rr
= 0 on the hypersurface, show-
ing that H is a null hypersurface.
Because of this fact, the tangent space to the null hypersurface H
at any of its points is orthogonal to a null vector, and hence it does
not contain time-like vectors. H is called a Killing horizon.
The hypersurface dened by the static limit is time-like, which
means that it can be crossed in both directions, in contrast to the
horizon H. The region in between the static limit and the Killing
horizon is the ergosphere, in which k is space-like and no observer
can be prevented from following the rotation of the black hole.
Formally, the Kerr solution is singular where = 0, but this sin-
gularity can be lifted by a transformation to coordinates similar to
the Eddington-Finkelstein coordinates for a Schwarzschild black
hole.
11.2 Motion near a Kerr black hole
11.2.1 Keplers third law
We shall now assume q = 0 and consider motion on a circular
orbit in the equatorial plane. Thus r = 0 and = /2, and
= r
2
2mr + a
2
and = r , (11.56)
further
2
= (r
2
+ a
2
)
2
a
2
= r
4
+ a
2
r
2
+ 2ma
2
r , (11.57)
and the coecients of the metric (11.20) become
g
tt
= 1 +
2m
r
, g
t
=
2ma
r
,
g
rr
=
r
2
, g
= r
2
,
g
2
r
2
= r
2
+ a
2
+
2ma
2
r
. (11.58)
Since

= 0 and r = 0, the Lagrangian reduces to
2[ =
_
1
2m
r
_

t
2
4ma
r

t
+
_
r
2
+ a
2
+
2ma
2
r
_

2
. (11.59)
By the Euler-Lagrange equation for r and due to r = 0, we have
d
dt
[
r
= 0 =
[
r
, (11.60)
which yields, after multiplying with r
2
/
t
2
,
m + 2ma + (r
3
ma
2
)
2
= 0 (11.61)
where we have introduced the angular frequency according to
(11.30). Noticing that
r
3
ma
2
= (r
3/2
m
1/2
a)(r
3/2
+ m
1/2
a) , (11.62)
we can write the solutions as
=
m
1/2
r
3/2
m
1/2
a
. (11.63)
This is Keplers third law for a Kerr black hole: The angular ve-
locity of a test particle depends on whether it is co-rotating with
or counter-rotating against the black hole.
11.2.2 Accretion ow onto a Kerr black hole
We now consider a stationary, axially-symmetric ow of a perfect
uid onto a Kerr black hole. Because of the symmetry constraints,
the Lie derivatives of all physical quantities in the direction of the
Killing vector elds k =
t
and m =
need to vanish.
As in (11.31), the four-velocity of the ow is
u = u
t
(k + m) . (11.64)
We introduce
e (u, k) = u
t
, j (u, m) = u
, l
j
e
=
u
u
t
(11.65)
and use
u
= g
t
u
t
+ g
= u
t
_
g
t
+ g
_
u
t
= g
tt
u
t
+ g
t
u
= u
t
_
g
tt
+ g
t
_
(11.66)
to see that
l =
g
t
+ g
g
tt
+ g
t
=
g
t
+ lg
tt
g
+ lg
t
. (11.67)
Moreover, by the denition of l and , and using 1 = (u, u) =
u
t
u
t
+ u
, we see that
l =
u
u
t
u
t
u
t
u
t
=
1
1 l
. (11.68)
Finally, using (11.66) and (11.68), we have
u
2
t
= u
t
u
t
(g
tt
+ g
t
) =
g
tt
+ g
t
1 l
. (11.69)
If we substitute from (11.67) here, we obtain after a short cal-
culation
u
2
t
= e
2
=
g
2
t
g
tt
g
+ 2lg
t
+ l
2
g
tt
. (11.70)
It was shown on problem sheet 5 that the relativistic Euler equa-
tion reads
( + p)
u
u = gradp u
u
p , (11.71)
where and p are the density and the pressure of the ideal uid,
further
gradp = dp
, (11.72)
where dp
is the vector belonging to the dual vector dp.

We rst observe that, since the ow is stationary,
0 = [
u
p = up = u
p = u
p
u
p = 0 , (11.73)
thus the second term on the right-hand side of (11.71) vanishes.
Next, we introduce the dual vector u
belonging to the four-

velocity u. In components, u
= g
= u
. Then, from (5.32),

_
[
u
u
= u
+ u
= u
+ u
, (11.74)
where we have employed the symmetry of the connection . This
shows that
[
u
u
=
u
u
. (11.75)
Then, we introduce f 1/u
t
and compute [
f u
u
in two dierent
ways. First, according to (5.32), we can write
_
[
f u
u
= f u
+ u
( f u
) = f
_
u
+ u
f
(11.76)
hence
[
f u
u
= f [
u
u
d f = f
u
u
d f , (11.77)
where (11.75) was inserted.
On the other hand, f u = u/u
t
= k +m because of (11.64), which
allows us to write
[
f u
u
= [
k
u
.,,.
=0
+[
m
u
. (11.78)
Now, in components,
_
[
m
u
= m
+ u
(m
)
=
_
m
+ u
_
+ u
. (11.79)
But the Lie derivative of u
in the direction m must vanish because

of the axisymmetry, thus
0 =
_
[
m
u
= m
+ u
, (11.80)
and the rst term on the right-hand side of (11.79) must vanish.
Therefore,
[
f u
u
= [
m
u
= (u, m)d = jd . (11.81)

Equating this to (11.77) gives
f
u
u
= df + jd . (11.82)
However, from (11.68),
f =
1
u
t
= u
t
(1 l) = e(1 l) . (11.83)
When inserted into (11.82), this yields
e(1 l)
u
u
= (1 l)de eld edl + jd

= (1 l)de edl , (11.84)
where el = j was used. Thus,
u
u
= d ln e
dl
1 l
. (11.85)
Inserting this into Eulers equation (11.71), we obtain
dp
+ p
= d ln e +
dl
1 l
, (11.86)
which shows that surfaces of constant pressure are given by
ln e
_
dl
1 l
= const. . (11.87)
Setting l = 0 makes the second term on the left-hand side vanish,
and we nd from (11.70)
g
+ 2lg
t
+ l
2
g
tt
g
2
t
g
tt
g
=
1
e
2
= const. . (11.88)
If a = 0, we obtain the isobaric surfaces of the accretion ow onto
a Schwarzschild black hole. With
g
tt
= 1 +
2m
r
, g
= r
2
sin
2
, g
t
= 0 , (11.89)
we nd
r
r 2m

l
2
r
2
sin
2
= const. (11.90)
This describes toroidal surfaces around the black hole, the so-
called accretion torus.
11.3 Entropy and temperature of a black
hole
It was realised by Hawking, Penrose and Christodoulou that the
area of a possibly charged and rotating black hole, dened by
A := 4 := 4
_
r
2
+
+ a
2
_
(11.91)
cannot shrink. Here, r
+
is the positive branch of the two solutions
of (11.50),
r
= m
_
m
2
Q
2
a
2
. (11.92)
This led Bekenstein (1973) to the following consideration. If A
cannot shrink, it reminds of the entropy as the only other quan-
tity known in physics that cannot shrink. Could the area A have
anything to do with an entropy that could be assigned to a black
hole? In fact, this is much more plausible than it may appear at
rst sight. Suppose radiation disappears in a black hole. Without
accounting for a possible entropy of the black hole, its entropy
would be gone, violating the second law of thermodynamics. The
same holds for gas accreted by the black hole: Its entropy would
be removed from the outside world, leaving the entropy there
lower than before.
If, however, the increased mass of the black hole led to a suitably
increased entropy of the black hole itself, this violation of the
second law could be remedied. Any mass and angular momentum
swallowed by a black hole leads to an increase of the area (11.91),
which makes it appear plausible that the area of a black hole might
be related to its entropy. Following Bekenstein (1973), we shall
now work out this relation.
First, (11.92) implies
r
+
+ r
= 2m , r
+
r
= Q
2
+ a
2
, r
+
r
= 2
_
m
2
Q
2
a
2
,
(11.93)
thus
2mr
+
= (r
+
+ r
)r
+
= r
2
+
+ Q
2
+ a
2
(11.94)
and

A
4
= 2mr
+
Q
2
. (11.95)
The dierential of is
d = 2r
+
dr
+
+ 2ada , (11.96)
which leads to
d =
4
r
dm
4r
+
Q
r
dQ
4a d
L
r
, (11.97)
where r := r
+
r
and the relation (11.21) between a and the

angular momentum

L was used. Solving this equation for dm
yields
dm = d + dQ +

d
L (11.98)
with the denitions
:=
r
4
, :=
r
+
Q
,

:=
a
. (11.99)
This reminds of the rst law of thermodynamics if we tentatively
identify m with the internal energy, with the entropy and the
remaining terms with external work.
Let us now see whether a linear relation between entropy S and
area will lead to consistent results. Thus, assume S = with
some constant . Then, a change will lead to a change S =
in the entropy.
Bekenstein showed that the minimal change of the eective area
is twice the squared Planck length (1.5), thus
=
2G
c
3
. (11.100)
On the other hand, he identied the minimal entropy change of
the black hole with the minimal change of the Shannon entropy,
which is derived from information theory and is
S = ln 2 . (11.101)
This could e.g. correspond to the minimal information loss when
a single particle disappears in a black hole. Requiring
ln 2 = S = =
2G
c
3
(11.102)
xes ,
=
ln 2
2
c
3
G
, (11.103)
and thus the Bekenstein entropy becomes
S =
ln 2
8
c
3
k
B
G
A , (11.104)
where A is the area of the black hole and the Boltzmann constant
k
B
was inserted to arrive at conventional units for the entropy.
The quantity from (11.99) must then correspond to the temper-
ature of the black hole. If this should be consistent, it must agree
with the thermodynamic denition of temperature,
1
T
=
_
S
E
_
V
. (11.105)
For E, we can use the mass or rather
E = Mc
2
=
mc
4
G
. (11.106)
Then,
S
E
=
G
c
4
S
m
=
ln 2
8
k
B
c
A
m
. (11.107)
Taking A from (11.91) yields
A
m
=
4
(11.108)
and therefore
T =
2
ln 2
c
k
B
=
2
ln 2
c
k
B
r
A
. (11.109)
This result leads to a remarkable conclusion. If black holes
have a temperature, they will radiate and thus lose energy or its
mass equivalent. They can therefore evaporate. By the Stefan-
Boltzmann law, the luminosity radiated by a black body of area A
and temperature T is
L = AT
4
, =
2
k
4
B
60
3
c
2
. (11.110)
For an uncharged and non-rotating black hole, r = 2m and A =
16m
2
, thus its temperature is
T =
1
4 ln 2
c
k
B
m
=
1
4 ln 2
c
3
k
B
GM
=
1
4 ln 2
m
2
Pl
c
2
k
B
M
(11.111)
with the Planck mass m
Pl
= 2.2 10
5
g from (1.4). For a black
hole of solar mass,
T = 5.7 10
7
K . (11.112)
Chapter 12
Homogeneous, Isotropic
Cosmology
12.1 The generalised Birkho theorem
12.1.1 Spherically-symmetric space-times
Generally, a space-time (M, g) is called spherically symmetric if
it admits the group SO(3) as an isometry such that the groups
orbits are two-dimensional, space-like surfaces.
For any point p M, we can then select the orbit (p) of SO(3)
through p. In other words, we construct the spatial two-sphere
containing p which is compatible with the spherical symmetry.
Next, we construct the set of all geodesics N(p) through p which
are orthogonal to (p). Locally, N(p) forms a two-dimensional
surface which we also call N(p). Repeating this construction for
all p M yields the surfaces N.
We can now introduce coordinates (r, t) on N and (, ) on ,
i.e. such that the group orbits are given by (r, t) = const. and
the surfaces N by (, ) = const.
Then, the metric of the space-time M can be written in the form
g = g + R
2
(t, r)
_
d
2
+ sin
2
d
2
_
, (12.1)
where g is an indenite metric in the coordinates (t, r) on N.
Without loss of generality, we can now choose t and r such that g
is diagonal, which we denote by
g = e
2a(t,r)
dt
2
+ e
2b(t,r)
dr
2
, (12.2)
where a(t, r) and b(t, r) are functions to be determined.
148
CHAPTER 12. HOMOGENEOUS, ISOTROPIC COSMOLOGY 149
We introduce the dual basis
0
= e
a
dt ,
1
= e
b
dr ,
2
= Rd ,
3
= Rsin d (12.3)
and nd its exterior derivatives
d
0
= a
e
b
1

0
, d
1
=

be
a
0

1
, (12.4)
d
2
=
R
R
e
a
0

2
+
R
R
e
b
1

2
,
d
3
=
R
R
e
a
0

3
+
R
R
e
b
1

3
+
cot
R

2

3
,
where the overdots and primes denote derivatives with respect to
t and r, respectively.
In the dual basis (12.3), the metric is g = diag(1, 1, 1, 1), thus
dg = 0, and Cartans rst structure equation (8.13) implies
i
j

j
= d
i
(12.5)
for the connection 1-forms
i
j
. This yields
0
1
=
1
0
= a
e
b
0
+

be
a
1
,
0
2
=
2
0
=
R
R
e
a
2
,
0
3
=
3
0
=
R
R
e
a
3
,
1
2
=
2
1
=
R
R
e
b
1
,
1
3
=
3
1
=
R
R
e
b
3
,
2
3
=
3
2
=
cot
R

3
.(12.6)
Cartans second structure equation (8.13) then yields the curva-
ture 2-forms
i
j
,
0
1
= d
0
1
E
0

1
,
0
2
= d
0
2
+
0
1

1
2

E
0

2
+ H
1

2
,
0
3
= d
0
3
+
0
1

1
3
+
0
2

2
3
=

E
0

3
+ H
1

3
,
1
2
= d
1
2
+
1
0

0
2
H
0

2
+

F
1

2
,
1
3
= d
1
3
+
1
0

0
3
+
1
2

2
3
=

F
1

3
H
0

3
,
2
3
= d
2
3
+
2
0

0
3
+
2
1

1
3
F
2

3
, (12.7)
where the functions
E = e
2a
_
b a
b +

b
2
_
e
2b
_
a
+ a
2
_
,
E =
e
2a
R
_
R a
R
_
e
2b
R
a
,
H =
e
ab
R
_

R

bR
_
,
F =
1
R
2
_
1 R
2
e
2b
+

R
2
e
2a
_
,
F =
e
2a
R
R +
e
2b
R
_
b
_
(12.8)
were dened.
According to (8.20), the curvature forms imply the components
R
(e
, e
) (12.9)
of the Ricci tensor, for which we obtain
R
00
= E 2

E , R
01
= 2H , R
02
= 0 = R
03
,
R
11
= E + 2

F , R
12
= 0 = R
13
R
22
=

E +

F + F = R
33
, R
23
= 0 (12.10)
and the Ricci scalar
R = (E + 2

E) + (E + 2

F) + 2(

E +

F + F)
= 2(E + F) + 4(

E +

F) , (12.11)
and nally for the Einstein tensor
G
= R

R
2
g
(12.12)
the components
G
00
= F + 2

F , G
01
= 2H , G
02
= 0 = G
03
,
G
11
= 2

E F , G
12
= 0 = G
13
,
G
22
= E

E

F = G
33
, G
23
= 0 . (12.13)
12.1.2 Birkhos theorem
We can now state and prove Birkhos theorem in its general
form: Every C
2
solution of Einsteins vacuum equations which
is spherically symmetric in an open subset U is locally isometric
to a domain of the Schwarzschild-Kruskal solution.
The proof proceeds in four steps:
1. If the surfaces R(t, r) = const. are time-like in U and dR
0, we can choose R(t, r) = r, thus

R = 0 and R
= 1. Since
H =
be
ab
/R then, the requirement G
01
= 0 implies

b =
0. The sum G
00
+ G
11
= 2(

F

E) must also vanish, thus
e
2b
R
_
b
+ a
_
= 0 , (12.14)
which means a(t, r) = b(r) + f (t). By a suitable choice
of a new time coordinate, a can therefore be made time-
independent as well. Moreover, we see that
0 = G
00
= F + 2

F =
1 e
2b
R
2
+
2b
e
2b
R
(12.15)
is identical to the condition (8.62) for the function b in the
Schwarzschild space-time. Thus, we have e
2b
= 1 2m/r
as there, further a(r) = b(r), and the metric turns into the
Schwarzschild metric.
2. If the surfaces R(t, r) = const. are space-like in U and dR
0, we can choose R(t, r) = t and proceed in an analogous
way. Then,

R = 1 and R
= 0, thus H = a
e
ab
/R, hence
G
01
= 0 implies a
= 0 and, again through G

00
+G
11
= 0, the
condition a +

b = 0 or b(t, r) = a(t) + f (r). This allows us
to change the radial coordinate appropriately so that b(t, r)
also becomes independent of r. Then, G
00
= 0 implies
0 = G
00
=
1
R
2
_
1 + e
2a
_
2 a
e
2a
R
, (12.16)
where

b = a was used. Since R = t, this is equivalent to
(te
2a
) = 1 e
2a
= e
2b
=
2m
t
1 , (12.17)
with t < 2m. This is the Schwarzschild solution for r <
2m because r and t change roles inside the Schwarzschild
horizon.
3. If R(t, r) = const. are space-like in some part of U and
time-like in another, we obtain the respective dierent do-
mains of the Schwarzschild space-time.
4. Assume nally (dR, dR) = 0 on U. If R is constant in U,
G
00
= R
2
= 0 implies R = . Therefore, suppose dR is not
zero, but light-like. Then, r and t can be chosen such that
R = t r and dR = dt dr. For dR to be light-like,
(dR, dR) = g(dR, dR) = e
2a
+ e
2b
= 0 , (12.18)
we require a = b. Then, G
00
+ G
11
= 0 or
e
2a
R
_
a +

b a
_
= 0 , (12.19)
implies a = a
, which again leads to R = through G

00
= 0.
This shows that the metric reduces to the Schwarzschild metric in
all relevant cases.
It is a corollary to Birkhos theorem that a spherical cavity
in a spherically-symmetric spacetime has the Minkowski met-
ric. Indeed, Birkhos theorem says that the cavity must have
a Schwarzschild metric with mass zero, which is the Minkowski
metric.
12.2 Cosmological solutions
12.2.1 Homogeneity and isotropy
There are good reasons to believe that the Universe at large is
isotropic around our position. The most convincing observational
data are provided by the cosmic microwave background, which
is a sea of blackbody radiation at a temperature of 2.726 K whose
intensity is almost exactly independent of the direction into which
it is observed.
There is furthermore no good reason to believe that our position
in the Universe is in any sense prefered compared to others. We
must therefore conclude that any observer sees the cosmic mi-
crowave background as an isotropic source such as we do. Then,
the Universe must also be homogeneous.
We are thus lead to the expectation that our Universe at large may
be described by a homogeneous and isotropic space-time. Let us
now give these terms a precise mathematical meaning.
A space-time (M, g) is called spatially homogeneous if there ex-
ists a one-parameter family of space-like hypersurfaces
t
that
foliate the space-time such that for each t and any two points
p, q
t
, there exists an isometry of g which takes p into q.
Before we can dene isotropy, we have to note that isotropy re-
quires that the state of motion of the observer needs to be spec-
ied rst because two observers moving with dierent velocities
through a given point in space-time will generally observe dier-
ent redshifts in dierent directions.
Therefore, we dene a space-time (M, g) as spatially isotropic
about a point p if there exists a congruence of time-like geodesics
through p with tangents u such that for any two vectors v
1
, v
2
V
p
orthogonal to u, there exists an isometry of g taking v
1
into v
2
but
leaving u and p invariant. In other words, if the space-time is
spatially isotropic, no prefered spatial direction orthogonal to u
can be identied.
Isotropy thus identies a special class of observers, with four-
velocities u, who cannot identify a prefered spatial direction. The
spatial hypersurfaces
t
must then be orthogonal to u because oth-
erwise a prefered direction could be identied through the mis-
alignment of the normal direction to
t
and u, breaking isotropy.
We thus arrive at the following conclusions: a homogeneous and
isotropic space-time (M, g) is foliated into space-like hypersur-
faces
t
on which g induces a metric h. There must be isometries
of h carrying any point p
t
into any other point q
t
. Be-
cause of isotropy, it must furthermore be impossible to identify
prefered spatial directions on
t
. These are very restrictive re-
quirements which we shall now exploit.
12.2.2 Spaces of constant curvature
Consider now the curvature tensor
(3)
R induced on
t
(i.e. the cur-
vature tensor belonging to the induced metric h). We shall write it
in components with its rst two indices lowered and the following
two indices raised,
(3)
R =
(3)
R
kl
i j
. (12.20)
In this way,
(3)
R represents a linear map from the vector space
of 2-forms
_
2
into
_
2
, because of the antisymmetry of
(3)
R with
respect to permutations of the rst and the second pairs of indices.
Thus, it denes an endomorphism
L :
_
2
_
2
, (L)
i j
=
(3)
R
kl
i j

kl
. (12.21)
Due to the symmetry (3.77) of
(3)
R upon swapping the rst with
the second pair of indices the endomorphism L is self-adjoint. In
fact, for any pair of 2-forms ,
_
2
,
(, L) =
(3)
R
kl
i j

i j
kl
=
(3)
R
i jkl
i j
kl
=
(3)
R
kli j
i j
kl
=
(3)
R
i j
kl

i j
kl
= (, L) , (12.22)
which denes a self-adjoint endomorphism.
We can now use the theorem stating that the eigenvectors of a
self-adjoint endomorphism provide an orthonormal basis for the
vector space it is operating on. Isotropy now requires us to con-
clude that the eigenvalues of these eigenvectors need to be equal
because we could otherwise dene a prefered direction (e.g. by
the eigenvector belonging to the largest eigenvalue). Then, how-
ever, the endomorphism L must be proportional to the identical
map
L = 2kI , (12.23)
with some k R.
By the denition (12.21) of L, this implies for the coecients of
the curvature tensor
(3)
R
kl
i j
= k
_
k
i
l
j

k
j
l
i
_
(12.24)
because
(3)
R must be antisymmetrised. Lowering the indices by
means of the induced metric h yields
(3)
R
i jkl
= k
_
h
ik
h
jl
h
jk
h
il
_
. (12.25)
The Ricci tensor is
(3)
R
jl
=
(3)
R
i
jil
= kh
is
_
h
si
h
jl
h
ji
h
sl
_
= k
_
3h
jl
h
jl
_
= 2kh
jl
, (12.26)
and the Ricci scalar becomes
(3)
R =
(3)
R
j
j
= 6k . (12.27)
In the coordinate-free representation, the curvature is
R(x, y)v = k ((x, w)y (y, w)x) . (12.28)
from (8.18) and (12.25), we nd the curvature forms
i
j
=
1
2
(3)
R
i
jkl

k

l
=
k
2
h
is
_
h
sk
h
jl
h
jk
h
sl
_

k

l
= k
i

j
(12.29)
in a so far arbitrary dual basis
i
.
The curvature parameter k must be (spatially) constant because
of homogeneity. Space-times with constant curvature can be
shown to be conformally at, which means that coordinates can
be introduced in which the metric h reads
h =
1
2
3
i=1
(dx
i
)
2
, (12.30)
with a yet unknown arbitrary function . This leads us to intro-
duce the dual basis
dx
i
, (12.31)
from which we nd
d
i
=
2
dx
j
dx
i
= (
j
)
i

j
. (12.32)
In this basis, the metric h can be represented by h = diag(1, 1, 1).
Therefore, we do not need to distinguish between raised and low-
ered indices, and dh = 0. Hence Cartans rst structure equation
(8.13) implies the connection forms
i j
= (
i
)
j
(
j
)
i
. (12.33)
According to Cartans second structure equation, the curvature
forms are
i j
= d
i j
+
ik

k
j
(12.34)
=
_
k

j

j
k

i
_
(
k
k
)
i

j
,
but at the same time we must satisfy (12.29). This immediately
implies
k
= 0 (i k) , (12.35)
thus has to be of the form
=
3
k=1
f
k
(x
k
) (12.36)
because otherwise the mixed derivatives could not vanish.
Inserting this result into (12.35) shows
i j
=
_
f
i
+ f
j

f
k
f
k
_

i

j
. (12.37)
In order to satisfy (12.29), we must have
f
i
+ f
j
=
k + f
k
f
k
. (12.38)
Since the right-hand side does not depend on x
i
or x
j
, the second
derivatives f
i
and f
j
must all be equal and constant, and thus
the f
i
must be quadratic in x
i
with a coecient of (x
i
)
2
which is
independent of x
i
. Therefore, we can write
= 1 +
k
4
3
i=1
(x
i
)
2
(12.39)
because, if the linear term is non-zero, it can be made zero by
translating the coordinate origin, and a constant factor on is
irrelevant because it simply scales the coordinates.
12.2.3 Form of the metric and Friedmanns equations
Alexander Friemann (18881925)
According to the preceding discussion, the homogeneous and
isotropic spatial hypersurfaces
t
must have a metric of the form
h =
_
3
i=1
(dx
i
)
2
(1 + kr
2
/4)
2
, r
2
i=1
(x
i
)
2
. (12.40)
By a suitable choice of the time coordinate t, the metric of a spa-
tially homogeneous and isotropic space-time can then be written
as
g = c
2
dt
2
+ a
2
(t)h , (12.41)
because the scaling function a(t) must not depend on the x
i
in
order to preserve isotropy and homogeneity.
We choose the appropriate dual basis
0
= cdt ,
i
=
a(t) dx
i
1 + kr
2
/4
, (12.42)
in terms of which the metric coecients are g = diag(1, 1, 1, 1).
The exterior derivatives of the dual basis are
d
0
= 0 ,
d
i
=
a dt dx
i
1 + kr
2
/4

a
(1 + kr
2
/4)
2
k
2
x
j
dx
j
dx
i
=
a
ca

0

i
+
kx
j
2a

i

j
. (12.43)
Since the exterior derivative of the metric is dg = 0, Cartans rst
structure equation (8.13) implies
i
j

j
= d
i
, (12.44)
suggesting the curvature forms
0
i
=
i
0
=
a
ca

i
,
i
j
=
j
i
=
k
2a
_
x
i

j
x
j

i
_
, (12.45)
which evidently satisfy (12.44).
Their exterior derivatives are
d
0
i
=
aa a
2
c
2
a
2

0
i
+
a
ca
d
i
=
a
c
2
a

0
i
+
k ax
j
2ca
2

i
j
(12.46)
and
d
i
j
=
k a
2a
2

0
_
x
i
j
x
j
i
_
(12.47)
+
k
2a
_
dx
i

j
dx
j

i
+ x
i
d
j
x
j
d
i
_
=
k
a
2
_
1 +
k
4
r
2
_

i

j
+
k
2
4a
2
_
x
i
x
k

j

k
x
j
x
k

i

k
_
Cartans second structure equation (8.13) then gives the curvature
forms
0
i
= d
0
i
+
0
k

k
i
=
a
c
2
a

0

i
, (12.48)
i
j
= d
i
j
+
i
0

0
j
+
i
k

k
j
=
k + a
2
/c
2
a
2

i

j
,
from which we obtain the components of the Ricci tensor
R
jk
= R
i
jik
=
i
j
(e
i
, e
k
) (12.49)
as
R
00
= 3
a
c
2
a
, R
11
= R
22
= R
33
=
a
c
2
a
+ 2
k + a
2
/c
2
a
2
. (12.50)
The Ricci scalar is then
R = R
i
i
= 6
_
a
c
2
a
+
k + a
2
/c
2
a
2
_
, (12.51)
and the Einstein tensor gets the components
G
00
= 3
k + a
2
/c
2
a
2
, G
11
= G
22
= G
33
=
2 a
c
2
a

k + a
2
/c
2
a
2
.
(12.52)
For Einsteins eld equations to be satised, the energy-
momentum tensor must be diagonal, and its components must not
depend on the spatial coordinates in order to preserve isotropy and
homogeneity. We set T
00
= c
2
, which is the total energy density,
and T
i j
= p
i j
, where p is the pressure.
This corresponds to the energy-momentum tensor of an ideal
uid,
T
= (c
2
+ p)u
+ pg
(12.53)
as seen by a fundamental observer (i.e. an observer for whom
the spatial hypersurfaces are isotropic). For such an observer,
u
0
, and the components of T
are as given.
Then, Einsteins eld equations in the form (6.80) with the cos-
mological constant reduce to
3
k + a
2
/c
2
a
2
=
8G
c
2
+ ,
2 a
c
2
a

k + a
2
/c
2
a
2
=
8G
c
4
p . (12.54)
Adding a third of the rst equation to the second, and re-writing
the rst equation, we nd Friedmanns equations
a
2
a
2
=
8G
3
+
c
2
3

kc
2
a
2
a
a
=
4G
3
_
+
3p
c
2
_
+
c
2
3
. (12.55)
12.2.4 Density evolution
After multiplication with 3a
2
and dierentiation with respect to t,
Friedmanns rst equation gives
6 a a = 8G( a
2
+ 2a a) + 2c
2
a a . (12.56)
If we eliminate
6 a = 8Ga
_
+
3p
c
2
_
+ 2c
2
a (12.57)
by means of Friedmanns second equation, we nd
+ 3
a
a
_
+
p
c
2
_
= 0 (12.58)
for the evolution of the density with time.
This equation has a very intuitive meaning. To see it, let us con-
sider the energy contained in a volume V
0
, which changes over
time in proportion to V
0
a
3
, and employ the rst law of thermody-
namics,
d(c
2
V
0
a
3
) + pd(V
0
a
3
) = 0 d(c
2
a
3
) + pd(a
3
) = 0 .
(12.59)
We can use the rst law of thermodynamics here because isotropy
forbids any energy currents, thus no energy can ow into or out
of the volume a
3
.
Equation (12.59) yields
a
3
+ 3a
2
a +
3p
c
2
a
2
a = 0 , (12.60)
which is identical to (12.58). This demonstrates that (12.58) sim-
ply expresses energy-momentum conservation. Consequently,
one can show that it also follows from the contracted Bianchi
identity, T = 0.
Two limits are typically considered for (12.58). First, if matter
moves non-relativistically, p < c
2
, and we can assume p 0.
Then,

= 3
a
a
, (12.61)
which implies
=
0
a
3
(12.62)
if
0
is the density when a = 1.
Second, relativistic matter has p = c
2
/3, with which we obtain

= 4
a
a
(12.63)
and thus
=
0
a
4
. (12.64)
This shows that the density of non-relativistic matter drops as ex-
pected in proportion to the inverse volume, but the density of rel-
ativistic matter drops faster by one order of the scale factor. an
explanation will be given below.
12.2.5 Cosmological redshift
We can write the metric (12.41) in the form
g = c
2
dt
2
+ a
2
(t)d
2
, (12.65)
where d
2
is the metric of a three-space with constant curvature
k. Since light propagates on null geodesics, (12.65) implies
cdt = a(t)d . (12.66)
Suppose a light signal leaves the source at the coordinate time t
0
and reaches the observer at t
1
, then (12.66) shows that the coordi-
nate time satises the equation
_
t
1
t
0
cdt
a(t)
=
_
observer
source
d , (12.67)
whose right-hand side is time-independent. Thus, for another
light signal leaving the source at t
0
+dt
0
and reaching the observer
at t
1
+ dt
1
, we have
_
t
1
t
0
cdt
a(t)
=
_
t
1
+dt
1
t
0
+dt
0
cdt
a(t)
. (12.68)
Since this implies
_
t
0
+dt
0
t
0
dt
a(t)
=
_
t
1
+dt
1
t
1
dt
a(t)
, (12.69)
we nd for suciently small dt
0,1
that
dt
0
a(t
0
)
=
dt
1
a(t
1
)
. (12.70)
We can now identify the time intervals dt
0,1
with the inverse fre-
quencies of the emitted and observed light, dt
i
=
1
i
for i = 0, 1.
This shows that the emitted and observed frequencies are related
by
1
=
a(t
1
)
a(t
0
)
. (12.71)
Since the redshift z is dened in terms of the wavelengths as
z =
1

0
0
=
0

1
1
, (12.72)
we nd that light emitted at t
0
and observed at t
1
is redshifted by
1 + z =
0
=
a(t
1
)
a(t
0
)
. (12.73)
Thus, the wavelength of light is increased or decreased in the
same proportion as the universe itself expands or contracts.
We can now interpret the result (12.64) that the density of rela-
tivistic matter drops by one power of a more than expected by
mere dilution: as the universe expands, relativistic particles are
redshifted by another factor a and thus loose energy in addition
to their dilution.
12.2.6 Alternative forms of the metric
Before we proceed, we bring the spatial metric (12.40) into a dif-
ferent form. We rst write it in terms of spherical polar coordi-
nates as
h =
dr
2
+ r
2
(d
2
+ sin
2
d
2
)
(1 + kr
2
/4)
2
(12.74)
and introduce a new radial coordinate u dened by
u =
r
1 + kr
2
/4
. (12.75)
Requiring that r u for small r and u, we can uniquely solve
(12.75) to nd
r =
2
ku
_
1
1 ku
2
_
, (12.76)
which gives the dierential
dr =
r
u
1 ku
2
du . (12.77)
by (12.75), we have
1
1 + kr
2
/4
=
u
r
, (12.78)
and thus
dr
1 + kr
2
/4
=
du
1 ku
2
. (12.79)
In terms of the new radial coordinate u, we can thus write the
spatial part of the metric in the frequently used form
h =
du
2
1 ku
2
+ u
2
d
2
, (12.80)
where d abbreviates the solid-angle element. The constant k
can be positive, negative or zero, but its absolute value does not
matter since it merely scales the coordinates. Therefore, we can
normalise the coordinates such that k = 0, 1.
Yet another form of the metric is found by introducing a radial
coordinate w such that
dw =
du
1 ku
2
. (12.81)
Integrating both sides, we nd that this is satised if
u = f
k
(w)
_
_
sin w (k = 1)
w (k = 0)
sinh w (k = 1)
. (12.82)
We thus nd that the homogeneous and isotropic class of cos-
mological models based on Einsteins eld equations are charac-
terised by the metric
g = c
2
dt
2
+ a
2
(t)
_
dw
2
+ f
2
k
(w)
_
d
2
+ sin
2
d
2
__
(12.83)
which is equivalent to
g = c
2
dt
2
+ a
2
(t)
_
du
2
1 ku
2
+ u
2
_
d
2
+ sin
2
d
2
_
_
(12.84)
with u related to w by (12.82), and the scale factor a(t) satises
the Friedmann equations (12.55).
Metrics of the form (12.83) or (12.84) are called Robertson-
Walker metrics, and Friedmann-Lematre-Robertson-Walker met-
rics if their scale factor satises Friedmanns equations.
Chapter 13
Two Examples of Relativistic
Astrophysics
13.1 Light Propagation
13.1.1 Arbitrary spacetimes
We had seen in (6.14) that the separation vector n between two
geodesics out of a congruence evolves in a way determined by
the equation of geodesic deviation or Jacobi equation
2
u
n = R(u, n)u , (13.1)
where u is the tangent vector to the geodesics and R(u, n)u is the
curvature as dened in (3.47).
We apply this now to a light bundle, i.e. a congruence of light
rays or null geodesics propagating from a source moving with a
four-velocity u. Let now k be the wave vector of the light rays,
then the frequency of the light at the source is
s
= (k, u) , (13.2)
and we introduce the normalised wave vector

k = k/
s
which
satises (
k, u) = 1. Since k is a null vector, so is

k.
Next, we introduce a screen perpendicular to k and to u. It thus
falls into the local three-space of the source, where it is per-
pendicular to the light rays. Since it is two-dimensional, it can
be spanned by two orthonormal vectors E
1,2
, which are parallel-
transported along the light bundle such that
k
E
i
= 0 =
k
E
i
. (13.3)
162
CHAPTER13. TWOEXAMPLES OF RELATIVISTICASTROPHYSICS163
Notice that the parallel transport along a null geodesic implies
that the E
i
remain perpendicular to

k,
k
(
k, E
i
) = (
k
k, E
i
) + (
k,
k
E
i
) = 0 . (13.4)
In a coordinate basis e
and its conjugate dual basis

i
, they
can be written as
E
i
= E
i
e
, E
i
=
(E
i
) . (13.5)
The separation vector n between rays of the bundle can now be
expanded into the E
1,2
,
n = n
= N
i
E
i
, (13.6)
showing that its components n
in the basis e
are
n
(n) =
(N
i
E
i
) = N
i
E
i
. (13.7)
Substituting the normalised wave vector

k for the four-velocity u
in the Jacobi equation (13.1), we rst have
k
n = R(
k, n)
k . (13.8)
Writing n = N
i
E
i
and using (13.3), we nd
k
n = E
i
k
N
i
,
2
k
n = E
i
k
N
i
, (13.9)
and thus
E
i
k
N
i
= R(
, n
. (13.10)
Next, we apply
to the vectors on either side of this equation

and nd
E
i

2
k
N
i
= R
= R
j
N
j
, (13.11)
where we have used the components of the curvature tensor for
the arbitrary basis e
as given in (8.20). Finally, we multiply

this equation with E
i
from the left and use the orthonormality of
the vectors E
i
,
E
i
E
i
= 1 , (13.12)
to nd the equation
k
N
i
= R
E
i
j
N
j
= R
i

k
j
N
j
(13.13)
describing how the perpendicular cross section of a light bundle
changes along the bundle.
It will turn out convenient to introduce the Weyl tensor C, whose
components are determined by
R
= C
+ g
[
R
]
g
[
R
]

R
3
g
[
g
]
(13.14)
= C
+
1
2
_
g
+ g
R
6
_
g
_
.
Inserting this into (13.13) gives
k
N
i
= C
i

k
j
N
j
+
1
2
_
(
k, E
i
) R
j
N
j
(E
i
, E
j
) R
N
j
(
k,

k) R
i
E
j
N
j
+ (
k, E
j
) R
i

k
N
j
_
R
6
_
(
k, E
i
)(
k, E
j
) (
k,

k)(E
i
, E
j
)
_
.
The second, the fth and the sixth terms on the right-hand side
vanish because the E
i
are perpendicular to

k, and the fourth and
the last terms vanishes because

k is a null vector. Using the or-
thonormality of the E
i
, we thus obtain
k
N
i
=
1
2

i
j
R
N
j
+ C
i

k
j
N
j
=
1
2
R
N
i
+ C
i

k
j
N
j
. (13.15)
The evolution of the bundles perpendicular cross section can thus
be described by a matrix 7,
k
_
N
1
N
2
_
= 7
_
N
1
N
2
_
, (13.16)
which can be written in the form
7 =
1
2
R
_
1 0
0 1
_
+ C
_
E
1
E
1
E
1
E
2
E
2
E
1
E
2
E
2
_
.
(13.17)
Further insight can be gained by extracting the trace-free part
from 7. Since the trace is
tr 7 = R
+ C
_
E
1
E
1
+ E
2
E
2
_
, (13.18)
the trace-free part of 7 is
7
1
2
tr 7 J =
1
2
C
_
E
1
E
1
E
2
E
2
2E
1
E
2
2E
2
E
1
E
2
E
2
E
1
E
1
_
.(13.19)
The Weyl tensor has the same symmetries as the curvature tensor,
thus we can transform
C
i

k
j
= C
i

k
j
= C
i

k
j
= C
j

k
i
, (13.20)
where we have renamed the summation indices in the last step.
Thus, the trace-free part of 7 is symmetric.
In a last step, we introduce the complex vector
E
1
+ iE
2
, (13.21)
in terms of which we can write
E
1
E
1
+ E
2
E
2
=
1
E
1
E
2
E
2
= T
_
_
,
2E
1
E
2
= 0
_
_
. (13.22)
Summarising, we dene the two scalars
F
1
2
R
+
1
2
C
,
/
1
2
C
, (13.23)
in terms of which the matrix 7 can be brought into the form
7 =
_
F + T(/) 0(/)
0(/) F T(/)
_
. (13.24)
this is called the optical tidal matrix.
The eect of the tidal matrix becomes obvious if we start with a
light bundle with circular cross section, for which the components
N
i
of the distance vector can be written as
_
N
1
N
2
_
=
_
cos
sin
_
, (13.25)
where is the polar angle on the screen spanned by the vectors
E
1,2
. Before we apply the optical tidal matrix, we rotate it into its
principal-axis frame,
_
F + T(/) 0(/)
0(/) F T(/)
_

_
F + / 0
0 F /
_
, (13.26)
which shows that it maps the circle onto a curve outlined by the
vector
_
x
y
_

_
(F + /) cos
(F /) sin
_
. (13.27)
this is an ellipse with semi-major axis F + / and semi-minor
axis F /, because obviously
x
2
(F + /)
2
+
y
2
(F /)
2
= cos
2
+ sin
2
= 1 . (13.28)
Thus, for / = 0, the originally circular cross section remains
circular, with F being responsible for isotropically expanding or
shrinking the it, while the light bundle is elliptically deformed if
/ 0.
13.1.2 Specialisation to homogeneous and isotropic
spacetimes
In an isotropic spacetime, it must be impossible to single out pre-
ferred directions. This implies that / = 0 then, because oth-
erwise the principal-axis frame of the optical tidal matrix would
break isotropy. If the spacetime is homogeneous, this must hold
everywhere, so that we can specialise
7 =
_
F 0
0 F
_
, (13.29)
with F dened in (13.23).
Moreover, we see that
G
=
_
R

R
2
g
_

k
= R
(13.30)
because

k is a null vector. Thus, we can put
F =
1
2
G
=
4G
c
4
T
, (13.31)
using Einsteins eld equations in the second step.
Next, we can insert the energy-momentum tensor (12.53) for a
perfect uid,
T
= (c
2
+ p)u
+ pg
, (13.32)
and use the fact that fundamental observers (i.e. observers for
whom the universe appears isotropic) have u
= u
=
t
.
Since the frequency measured by an observer moving with four-
velocity u is (k, u), due to our denition of

k = k/
s
, and because
of the cosmological redshift (12.73), we can write
u
=
1
s
(k, u) =
s
= 1 + z , (13.33)
where
o
is the frequency measured by the observer, and z is the
redshift of the source relative to the observer.
Thus, we nd
F =
4G
c
2
_
+
p
c
2
_
(1 + z)
2
=
4G
c
2
+ p/c
2
a
2
, (13.34)
where a is the scale factor of the metric.
If p < c
2
, the density scales like a
3
as shown in (12.62), and
then
F =
4G
c
2

0
a
5
. (13.35)
We are still free to choose a suitable parameter along the ducial
light ray. Since the tangent vector

k is given by
k =
dx
d
(13.36)
and (
k, u) = 1 + z as shown above, we must have

_
dx
d
, u
_
=
dx
0
d
=
cdt
d
= 1 + z =
1
a
, (13.37)
where u =
t
was used in the second step. Thus, we must have
d = cadt. Then, observing that da = adt, we nd
d = cadt =
cada
a
=
cda
a/a
. (13.38)
With this result, we can rewrite
k
N
i
=

k
N
i
=
dx
N
i
=
dN
i
d
, (13.39)
and the optical tidal matrix becomes
d
2
N
i
d
2
=
4G
c
2

0
a
5
N
i
. (13.40)
This equation can be substantially simplied to reveal its very
intuitive meaning. From the metric in the form (12.83), we see
that radially propagating light rays must satisfy
cdt = adw , (13.41)
where w is the radial distance coordinate dened in (12.81). The
sign can be chosen depending on whether the distance should
grow with increaing time (i.e. into the future) or with decreasing
time (i.e. into the past), but it is irrelevant for our consideration.
We choose cdt = adw and therefore, with (13.38),
d = cadt = a
2
dw . (13.42)
Due to the isotropy of the light bundles deformation, we can re-
place its two diameters N
1,2
by a single diameter D. Moreover,
we consider the comoving diameter D/a. substituting dw for d
as a measure of length, we rst see that
d
2
dw
2
_
D
a
_
= a
2
d
d
_
a
2
d
d
_
D
a
_
_
= a
2
d
d
(aD
D)
= a
2
(aD
D) , (13.43)
where the prime denotes the derivative with respect to the param-
eter .
Next, we use (13.38) to write
da
d
=
1
c
_
a
a
_
da
da
=
1
c
2
_
a
a
_
d( a/a)
da
=
1
2c
2
d( a/a)
2
da
, (13.44)
which enables us to insert Friedmanns equation (12.55) in the
form
_
a
a
_
2
=
8G
3
0
a
3
+
c
2
3

kc
2
a
2
(13.45)
to nd
a
=
da
d
=
4G
c
2

0
a
4
+ ka
3
. (13.46)
Now, we substitute D
= (N
i
)
from (13.40) and a
from (13.46)
into (13.43) and obtain
d
2
dw
2
_
D
a
_
=
4G
c
2

0
a
2
D +
4G
c
2

0
a
2
D kDa
1
= k
_
D
a
_
, (13.47)
which is a simple oscillator equation for the comoving bundle di-
ameter D.
Equation (13.47) is now easily solved. We set the boundary con-
ditions that the bundle emerges from a source point, hence D = 0
at the source, and that it initially expands linearly with the radial
distance w, hence d(D/a)/dw = 1 there. Then, the solution of
(13.47) is
D = af
k
(w) = a
_
_
1
k
sin(
kw) (k > 0)
w (k = 0)
1
k
sinh(
kw) (k < 0)
, (13.48)
with f
k
(w) dened in (12.82).
This shows that the diameter of the bundle increases linearly if
space is at, diverges hyperbolically if space is negatively curved,
and expands and shrinks as a sine if space is positively curved.
13.2 Relativistic Stellar Structure Equations
13.2.1 The Tolman-Oppenheimer-Volko equation
We now consider an axially symmetric, static solution of Ein-
steins eld equations in presence of matter. As usual for an
axisymmetric solution, we can work in the Schwarzschild tetrad
(8.40), in which the energy-momentum tensor of a perfect uid,
T = T
with T
= (c
2
+ p)u
+ pg
, (13.49)
simplies to
T
= diag(c
2
, p, p, p) (13.50)
because u = u
0
e
0
= e
0
in the static situation we are considering.
It had been shown on problem sheet 5 that the relativistic Euler
equation is
(c
2
+ p)
u
u = gradp u
u
p , (13.51)
which had been derived by contracting the local conservation
equation
= 0 (13.52)
with the projection tensor h = g + u u.
Richard C. Tolman (18811948)
J. Robert Oppenheimer (1904
1967)
George M. Volko (19142000)
Specialising (13.51) to our situation, we use (8.9) to see that
u
u = (du
+ u
, u)e
= (u
0
1
0
, u)e
1
= a
e
b
e
1
, (13.53)
because the only non-vanishing of the connection forms
0
for a
static, axially symmetric spacetime is
1
0
= a
e
b
0
, (13.54)
as shown in (8.50). Moreover, in the static situation,
u
p = 0.
The gradient gradp was dened as
gradp = dp
= p
dr
= p
e
b
(
1
)
= p
e
b
e
1
. (13.55)
Substituting (13.53) and (13.55) into (13.51) yields
(c
2
+ p)a
= p
=
p
c
2
+ p
, (13.56)
which is the relativistic hydrostatic equation.
With the components of the Einstein tensor given in (8.60) and
the energy-momentum tensor (13.50), the two independent eld
equations read
1
r
2
+ e
2b
_
1
r
2

2b
r
_
=
8G
c
2

1
r
2
+ e
2b
_
1
r
2
+
2a
r
_
=
8G
c
4
p . (13.57)
The rst of these equations is equivalent to
_
re
2b
_
= 1
8G
c
2
r
2
. (13.58)
Integrating, and using the mass
M(r) = 4
_
r
0
(r
)r
2
dr
, (13.59)
shows that the function b is determined by
e
2b
= 1
2GM
rc
2
. (13.60)
If we subtract the rst from the second eld equation (13.57), we
nd
2e
2b
r
(a
+ b
) =
8G
c
4
(c
2
+ p) (13.61)
or
a
= b
+
4G
c
4
e
2b
(c
2
+ p)r . (13.62)
On the other hand, (13.60) gives
2b
e
2b
=
2GM
r
2
c
2

2GM
rc
2
=
2GM
r
2
c
2

8G
c
2
r , (13.63)
or
b
=
_
4G
c
2
r
GM
r
2
c
2
_
e
2b
, (13.64)
which allows us to write (13.62) as
a
=
_
GM
r
2
c
2
+
4G
c
4
pr
_
e
2b
=
G(M + 4pr
3
/c
2
)
rc
2
(r 2GM)
. (13.65)
But the hydrostatic equation demands (13.56), which we combine
with (13.65) to nd
p
=
G(c
2
+ p)(M + 4pr
3
/c
2
)
rc
2
(r 2GM)
. (13.66)
This is the Tolman-Oppenheimer-Volko equation for the pres-
sure gradient in a relativistic star.
This equation generalises the hydrostatic Euler equation in New-
tonian physics, which reads for a spherically-symmetric congu-
ration
p
=
GM
r
2
. (13.67)
This shows that gravity acts on c
2
+ p instead of alone, the
pressure itself adds to the source of gravity, and gravity increases
more strongly than r
2
towards the centre of the star.
13.2.2 The mass of non-rotating neutron stars
Neutron stars are a possible end product of the evolution of mas-
sive stars. When such stars explode as supernovae, they may leave
behind an object with a density so high that protons and electrons
combine to neutrons in the process of inverse decay. Objects
thus form which consist of matter with nuclear density.
Let us assume that we can describe such objects with an energy-
momentum tensor of the form (13.50) in which anisotropic
stresses are unimportant. The pressure p is related to the density
by an equation-of-state p().
Furthermore, let the density be positive, > 0, the pressure be a
monotonically increasing function of ,
dp
d
0 , (13.68)
and the equation-of-state be known for densities
0
, where
0
reaches nuclear densities.
We must satisfy the Tolman-Oppenheimer-Volko equation
(13.66), equation (13.65) for the function a(r) in the metric, and
the equation
M
= 4r
2
(13.69)
determining the mass.
The boundary conditions are
p(r = 0) = p(
c
) p
c
, M(0) = 0
e
a(R)
= 1
2GM
Rc
2
, (13.70)
where
c
is the central density. The latter equation is valid at the
stellar radius R and follows from Birkhos theorem.
In order to prevent innite pressure gradient forces, we require
2GM
c
2
= 2m < r (13.71)
due to the Tolman-Oppenheimer-Volko equation. Assuming
that the core of a neutron star reaches nuclear densities,
0
5 10
14
g cm
3
, (13.72)
and that the core has radius r
0
, the core mass is limited by
m
0
<
r
0
2
. (13.73)
By the assumption that the density does not increase outward,

0
m
0

4G
3c
2
r
3
0
0
, (13.74)
which yields the bound
r
0

_
3m
0
c
2
4G
0
_
1/3
. (13.75)
Combining this with (13.73), we nd
m
0
<
1
2
_
3m
0
c
2
4G
0
_
1/3
, (13.76)
which gives
m
0
<
1
2
_
3c
2
8G
0
_
1/2
, r
0
<
_
3c
2
8G
0
_
1/2
. (13.77)
Inserting the density (13.72) yields
r
0
< 18 km , M
0
=
m
0
c
2
G
< 6M
,
. (13.78)
Masses exceeding this limit will collapse into black holes.
Appendix A
Electrodynamics
A.1 Electromagnetic eld tensor
Electric and magnetic elds are components of the antisymmetric
eld tensor
F
(A.1)
formed from the four-potential
A
=
_

A
_
. (A.2)
The eld tensor can be conveniently summarised as
(F
) =
_
0

E
E J
_
(A.3)
with
J
i j
=
i ja
B
a
. (A.4)
Given the signature (, +, +, +) of the Minkowski metric, its as-
sociated rank-(0, 2) tensor has the components
(F
) =
_
0
E J
_
. (A.5)
A.2 Maxwells equations
The homogeneous Maxwell equations read
[
F
]
= 0 . (A.6)
For = 0, (, ) = (1, 2), (1, 3) and (2, 3), this gives
B + c

E = 0 , (A.7)
173
APPENDIX A. ELECTRODYNAMICS 174
and for = 1, (, ) = (2, 3), we nd

B = 0 . (A.8)
The inhomogeneous Maxwell equations are
=
4
c
j
, (A.9)
where
j
=
_
c
j
_
(A.10)
is the four-current density. For = 0 and = i, (A.9) gives

E = 4 , c
E = 4
j , (A.11)
respectively.
With the denition (A.1) and the Lorenz gauge condition
=
0, the inhomogeneous equations (A.9) can be written as
A
=
4
c
j
, (A.12)
where =
2
0
+

2
is the dAlembert operator. The particular
solution of the homogeneous equation is given by the convolution
of the source with the retarded Greens function
G(t, t
, x, x
) =
1
x x
_
t t
x x
c
_
_
, (A.13)
i.e. by
A
(t, x) =
1
c
_
d
3
x
_
dt
G(t, t
, x, x
) j
(t, x
) . (A.14)
A.3 Lagrange density and energy-
momentum tensor
The Lagrange density of the electromagnetic eld coupled to mat-
ter is
[ =
1
16
F

1
c
A
, (A.15)
from which Maxwells equations follow by the Euler-Lagrange
equations,
[
(
)

[
A
= 0 . (A.16)
APPENDIX A. ELECTRODYNAMICS 175
From the Lagrange density of the free electromagnetic eld,
[ =
1
16
F
, (A.17)
we nd the energy-momentum tensor
T
= 2
[
g
+ g
[ =
1
4
_
F

1
4
g
_
. (A.18)
From (A.3), we nd rst
F
= 2(
E
2

B
2
) , (A.19)
and the energy-momentum tensor can be written as
T
=
1
8
_

E
2
+

B
2
2
EJ
T
2J
E
2

B
2
+ JJ
T
_
. (A.20)
This yields the energy density
T
00
=
E
2
+

B
2
8
(A.21)
of the electromagnetic eld and the Poynting vector
cT
0i
=
c
4
E

B . (A.22)
Appendix B
Summary of Dierential
Geometry
B.1 Manifold
An n-dimensional manifold M is a suitably well-behaved space
that is locally homeomorphic to R
n
, i.e. that locally looks like
R
n
.
A chart h, or a coordinate system, is a homeomorphism from D
M to U R
n
,
h : D U , p h(p) = (x
1
, . . . , x
n
) , (B.1)
i.e. it assigns an n-tupel of coordinates x
i
to a point p D.
An atlas is a collection of charts whose domains cover the entire
manifold. If all coordinate changes between charts of the atlas
with overlapping domains are dierentiable, the manifold and the
atlas themselves are called dierentiable.
B.2 Tangent and Dual Spaces
The tangent space V
p
at a point p M is the vector space of all
derivations. A derivation v is a map from the space /
p
of C
functions in p into the real numbers,

v : /
p
R , f v( f ) . (B.2)
A derivation is a linear map which satises the Leibniz rule,
v(f + g) = v( f ) + v(g) , v( f g) = v( f )g + f v(g) . (B.3)
176
APPENDIX B. SUMMARY OF DIFFERENTIAL GEOMETRY 177
Tangent vectors generalise directional derivatives of functions.
A coordinate basis of the tangent vector space is given by the
partial derivatives
i
. Tangent vectors can then be expanded in
this basis,
v = v
i
i
, v( f ) = v
i
i
f . (B.4)
A dual vector w is a linear map assigning a real number to a vec-
tor,
w : V R , v w(v) . (B.5)
The space of dual vectors to a tangent vector space V is the dual
space V
.
Specically, the dierential of a function f / is a dual vector
dened by
d f : V R , v d f (v) = v( f ) . (B.6)
Accordingly, the dierentials of the coordinate functions x
i
form
a basis dx
i
of the dual space which is orthonormal to the coor-
dinate basis
i
of the tangent space,
dx
i
(
j
) =
j
(x
i
) =
i
j
. (B.7)
B.3 Tensors
A tensor t 7
r
s
of rank (r, s) is a multilinear mapping of r dual
vectors and s vectors into the real numbers. For example, a ten-
sor of rank (0, 2) is a bilinear mapping of 2 vectors into the real
numbers,
t : V V R , (x, y) t(x, y) . (B.8)
The tensor product is dened component-wise. For example, two
dual vectors w
1,2
can be multiplied to form a rank-(0, 2) tensor
w
1
w
2
(v
1
, v
2
) (w
1
w
2
)(v
1
, v
2
) = w
1
(v
1
) w
2
(v
2
) . (B.9)
A basis for tensors of arbitrary rank is obtained by the tensor
product of suitably many elements of the bases
i
of the tan-
gent space and dx
j
of the dual space. For example, a tensor
t 7
0
2
can be expanded as
t = t
i j
dx
i
dx
j
. (B.10)
If applied to two vectors x = x
k
k
and y = y
l
l
, the result is
t(x, y) = t
i j
dx
i
(x
k
k
) dx
j
(y
l
l
) = t
i j
x
k
y
l
i
k
j
l
= t
i j
x
i
y
j
. (B.11)
The contraction of a tensor t 7
r
s
is dened by
C : 7
r
s
7
r1
s1
, t Ct (B.12)
such that one of the dual vector arguments and one of the vec-
tor arguments are lled with pairs of basis elements and summed
over all pairs. For example, the contraction of a tensor t 7
1
1
is
Ct = t(dx
k
,
k
) = (t
i
j

i
dx
j
)(dx
k
,
k
) = t
i
j
k
i
j
k
= t
k
k
. (B.13)
The metric g 7
0
2
is a symmetric, non-degenerate tensor eld of
rank (0, 2), i.e. it satises
g(x, y) = g(y, x) , g(x, y) = 0 y x = 0 . (B.14)
The metric denes the scalar product between two vectors,
(x, y) = g(x, y) . (B.15)
B.4 Covariant Derivative
The covariant derivative or a connection linearly maps a pair of
vectors to a vector,
: V V V , (x, y)
x
y (B.16)
such that for a function f /
f x
y = f
x
y ,
x
( f y) = f
x
y + x( f )y . (B.17)
The covariant derivative of a function f is its dierential,
v
f = v f = d f (v) . (B.18)
Due to the linearity, it is completely specied by the covariant
derivatives of the basis vectors,
j
=
k
i j
k
. (B.19)
The functions
k
i j
are called connection coecients or Christoel
symbols. They are not tensors.
By means of the exponential map, so-called normal coordinates
can always be introduced locally in which the Christoel symbols
all vanish.
The covariant derivative y of a vector y is a rank-(1, 1) tensor
eld dened by
y : V
V R , y(w, v) = w(
v
y) . (B.20)
In components,
(y)
i
j
= y(dx
i
,
j
) = dx
i
(
j
y
k
k
) =
j
y
i
+
i
jk
y
k
. (B.21)
The covariant derivative of a tensor eld is dened to obey the
Leibniz rule and to commute with contractions. Specically, the
covariant derivative of a dual vector eld w V
is a tensor of
rank (0, 2) with components
(w)
i j
=
j
w
i

k
i j
w
k
. (B.22)
B.5 Parallel Transport and Geodesics
A curve is dened as a map from some interval I R to the
manifold,
: I M , t (t) . (B.23)
Its tangent vector is (t).
A vector v is said to be parallel transported along if

v = 0 . (B.24)
A geodesic curve is dened as a curve whose tangent vector is
parallel transported along ,

= 0 . (B.25)
In coordinates, let u = be the tangent vector to and with
components x
i
= u
i
, then
x
k
+
k
i j
x
i
x
j
= 0 . (B.26)
B.6 Torsion and Curvature
The torsion of a connection is dened by
T : VV V , (x, y) T(x, y) =
x
y
y
x[x, y] . (B.27)
It vanishes if and only if the connection is symmetric.
On a manifold M with a metric g, a symmetric connection can
always be uniquely denedby requiring that g = 0. This is the
Riemannian connection, whose Christoel symbols are
i
jk
=
1
2
g
ia
_
j
g
ak
+
k
g
ja

a
g
jk
_
. (B.28)
From now on, we shall assume that we are working with the Rie-
mannian connection whose torsion vanishes.
The curvature is dened by
R : V V V V ,
(x, y, z) R(x, y)z =
_
y

y
x

[x,y]
_
z . (B.29)
The curvature or Riemann tensor R 7
1
3
is given by
R : V
V V V R , (w, x, y, z) w[R(x, y)z] . (B.30)

Its components are
R
i
jkl
= dx
i
[R(
k
,
l
)
j
] =
k
i
jl

l
i
jk
+
a
jl
i
ak

a
jk
i
al
. (B.31)
The Riemann tensor obeys three important symmetries,
R
i jkl
= R
jikl
= R
jilk
, R
i jkl
= R
kli j
, (B.32)
which reduce its 4
4
= 256 components in four dimensions to 21.
In addition, the Bianchi identities hold,
(x,y,z)
R(x, y)z = 0 ,
(x,y,z)
x
R(y, z) = 0 , (B.33)
where the sums extend over all cyclic permutations of x, y, z. The
rst Bianchi identity reduces the number of independent compo-
nents of the Riemann tensor to 20. In components, the second
Bianchi identity can be written
R
i
j[kl;m]
= 0 , (B.34)
where the indices in brackets need to be antisymmetrised.
The Ricci tensor is the contraction of the Riemann tensor over its
rst and third indices, thus its components are
R
jl
= R
i
jil
= R
l j
. (B.35)
A further contraction yields the Ricci scalar,
R = R
i
i
. (B.36)
The Einstein tensor is the combination
G
i j
= R
i j

R
2
g
i j
. (B.37)
Contracting the second Bianchi identity, we nd the contracted
Bianchi identity,
j
G
i j
= 0 . (B.38)
B.7 Pull-Back, Lie Derivative and Killing
Vector Fields
A dierentiable curve
t
(p) dened at every point p M denes
a dieomorphic map
t
: M M. If
t
= v for a vector eld
v V,
t
is called the ow of v.
The pull-back of a function f dened on the target manifold at
t
(p) is given by
(
t
f )(p) = ( f
t
)(p) . (B.39)
This allows vectors dened at p to be pushed forward to
t
(p) by
(
t
v)( f ) = v(
t
f ) = v( f
t
) . (B.40)
Dual vectors w can then be pulled back by
(
t
w)(v) = w(
t
v) . (B.41)
For dieomorphisms
t
, the pull-back and the push-forward are
inverse,
t
=
1
t
.
As for vectors and dual vectors, the pull-back and the push-
forward can also be dened for tensors of arbitrary rank.
The Lie derivative of a tensor eld T into direction v is given by
the limit
[
v
T = lim
t0
t
T T
t
, (B.42)
where
t
is the ow of v. The Lie derivative quanties how a
tensor changes as the manifold is transformed by the ow of a
vector eld.
The Lie derivative is linear and obeys the Leibniz rule,
[
x
(y+z) = [
x
y+[
x
z , [
x
(yz) = [
x
yz +y[
x
z . (B.43)
It commutes with the contraction. Further important properties
are
[
x+y
= [
x
+ [
y
, [
x
= [
x
, [
[x,y]
= [[
x
, [
y
] . (B.44)
The Lie derivative of a function f is the ordinary dierential
[
v
f = v( f ) = d f (v) . (B.45)
The Lie derivative and the dierential commute,
[
v
d f = d[
v
f . (B.46)
The Lie derivative of a vector x is the commutator
[
v
x = [v, x] . (B.47)
By its commutation with contractions and the Leibniz rule, the
Lie derivative of a dual vector w turns out to be
([
x
w)(v) = x[w(v)] w([x, v]) . (B.48)
Lie derivatives of arbitrary tensors can be similarly derived. For
example, if g 7
0
2
, we nd
([
x
g)(v
1
, v
2
) = x[g(v
1
, v
2
)] g([x, v
1
], v
2
) g(v
1
, [x, v
2
]) (B.49)
with v
1,2
V.
Killing vector elds K dene isometries of the metric, i.e. the
metric does not change under the ow of K. This implies the
Killing equation
[
K
g = 0
i
K
j
+
j
K
i
= 0 . (B.50)
B.8 Dierential Forms
Dierential p-forms
_
p
are totally antisymmetric tensor
elds of rank (0, p). Their components satisfy
i
1
...i
p
=
[i
1
...i
p
]
. (B.51)
The exterior product is dened by
:
_
p
_
q
_
p+q
, (, ) =
(p + q)!
p!q!
7() ,
(B.52)
where 7is the alternation operator
(7t)(v
1
, . . . , v
p
) =
1
p!
sgn()t(v
(1)
, . . . , v
(p)
) . (B.53)
On the vector space
_
of dierential forms, the wedge product
denes an associative, skew-commutative Grassmann algebra,
= (1)
pq
, (B.54)
with
_
p
and
_
q
.
A basis for the p-forms is
dx
i
1
. . . dx
i
p
, (B.55)
which shows that the dimension of
_
p
is
dim
_
p
=
_
n
p
_
. (B.56)
The interior product i
v
is dened by
i : V
_
p
_
p1
, (v, ) i
v
() = (v, . . .) . (B.57)
In components, the interior product is given by
(i
v
)
i
2
...i
p
= v
j
ji
2
...i
p
. (B.58)
The exterior derivative turns p-forms into (p + 1)-forms d,
d :
_
p
_
p+1
, d = (p + 1)7() . (B.59)
Due to the symmetry of the Riemannian connection, its compo-
nents are given by partial derivatives,
(d)
i
1
,...,i
p+1
= (p + 1)
[i
1
i
2
...i
p+1
]
(B.60)
A dierential form is called exact if a dierential form exists
such that = d. It is called closed if d = 0.
B.9 Cartans Structure Equations
Let e
i
be an arbitrary basis and
i
its dual basis such that
(
i
, e
j
) =
i
j
. (B.61)
The connection forms
i
j

_
1
are dened by
v
e
i
=
j
i
(v) e
j
. (B.62)
In terms of Christoel symbols, they can be expressed as
i
j
=
i
k j

k
. (B.63)
They satisfy the antisymmetry relation
dg
i j
=
i j
+
ji
. (B.64)
The covariant derivative of a dual basis vector is
i
=
i
j
(v)
j
. (B.65)
Covariant derivatives of arbitrary vectors x and dual vectors are
then given by
v
x = (dx
i
+ x
j
i
j
, v) e
i
,
v
= (d
i

j
j
i
, v)
i
(B.66)
or
x = e
i
(dx
i
+ x
j
j
i
) , =
i
(d
i

j
j
i
) . (B.67)
Torsion and curvature are expressed by the torsion 2-form
i
_
2
and the curvature 2-form
i
j

_
2
as
T(x, y) =
i
(x, y) e
i
, R(x, y)e
i
=
j
i
(x, y) e
j
. (B.68)
The torsion and curvature forms are related to the connection
forms and the dual basis vectors by Cartans structure equations
i
= d
i
+
i
k

k
,
i
j
= d
i
j
+
i
k

k
j
. (B.69)
The components of the torsion and curvature tensors are deter-
mined by
i
= T
i
jk

j

k
,
i
j
= R
i
jkl

k

l
. (B.70)
B.10 Dierential Operators and Integration
The Hodge star operator turns a p-form into an (n p)-form,
:
_
p
_
np
, . (B.71)
If e
i
is an orthonormal basis of the dual space, the Hodge star
operator is uniquely dened by
(e
1
. . . e
i
p
) = e
i
p+1
. . . e
i
n
, (B.72)
where the indices i
1
. . . i
n
appear in their natural order or a cyclic
permutation thereof. For example, the coordinate dierentials
dx
i
are an orthonormal dual basis in R
3
, and
dx
1
= dx
2
dx
3
, dx
2
= dx
3
dx
1
, dx
3
= dx
1
dx
2
.
(B.73)
The codierential is a dierentiation lowering the order of a p-
form by one,
:
_
p
_
p1
, , (B.74)
which is dened by
= sgn(g) (1)
n(p+1)
d . (B.75)
It generalises the divergence of a vector eld and thus has the
components
()
i
2
...i
p
=
1
_
g
i
1
(
_
g
i
1
i
2
...i
p
) . (B.76)
The Laplace-de Rham operator
d + d (B.77)
generalises the Laplace operator.
The canonical volume form is an n-form given by
=
_
g dx
1
. . . dx
n
. (B.78)
The integration of n-forms = f (x
1
, . . . , x
n
) dx
1
. . . dx
n
over
domains D M is dened by
_
D
=
_
D
f (x
1
, . . . , x
n
) dx
1
. . . dx
n
. (B.79)
Functions f are integrated by means of the canonical volume
form,
_
D
f =
_
D
f
_
g dx
1
. . . dx
n
. (B.80)
The theorems of Stokes and Gauss can be expressed as
_
D
d =
_
D
,
_
D
=
_
D
, (B.81)
where is an (n 1)-form, is a 1-form.
Index
accretion torus, 144
action
of a free particle, 38
adapted coordinates, 97
algebra, 59
alternation operator, 58
atlas, 13
dierentiable, 14
Bianchi identities
for Riemannian connection,
36
Bianchi identity
contracted, 37
rst, 33
second, 33
Birkhos theorem, 105
generalised, 150
Cartans structure equations, 94
chart, 13
Christoel symbols, 26
for linearised gravity, 79
of a Riemannian connection,
35
transformation, 26
codierential, 63
coordinate expression, 63
cofactor, 45
commutator, 20
connection
linear, 26
metric, 34
Riemannian or Levi-Civita, 35
symmetric, 27
connection forms, 93
contraction, 22
cosmological constant, 72
and energy-momentum tensor,
76
cosmological light propagation,
168
covariant derivative, 27
of a dual vector, 29
of a tensor, 31
covariant dierentiation, 26
Cramers rule, 45, 73
curvature, 32
and tidal eld, 70
curvature tensor, 32
in arbitrary basis, 96
symmetries, 36
curve, 19
derivation, 16
dieomorphisms
one-parameter group of, 19
dierential forms, 58
exact and closed, 61
dierential of a function, 21
Doppler shift, 6
dual vector, 21
Eddington-Finkelstein coordi-
nates, 127
eikonal equation, 42
Einstein elevator, 6
Einstein tensor, 37
Einsteins eld equations, 71
energy non-conservation, 48
energy-momentum tensor, 76
of electromagnetic eld, 76
equivalence principle
Einsteins, 4
weak, 4
events
simultaneous, 5
exponential map, 28
exterior derivative, 60
coordinate expression, 60
186
INDEX 187
exterior product, 58
Faraday 2-form, 63
Fermats principle, 83
eld equation
linearised in Hilbert gauge, 82
frames
freely-falling, 6
inertial, 5
Friedmanns equations, 157
Friedmann-Lematre-Robertson-
Walker metric, 161
Frobenius condition, 97
gauge transformation, 81
Gauss theorem, 65
generator of a transformation, 20
geodesic, 28
and free fall, 39
and light rays, 43
geodesic deviation
equation of, 68
geodesic equation in Newtonian
limit, 48
geodesics
congruence of, 66
geometrical optics and change of
amplitude, 43
Grassmann algebra, 59
gravitational wave
equation, 87
gauge-invariant polarisation
tensor, 89
polarisation tensor, 87
gravitomagnetic eld
equation of motion in, 86
Lense-Thirring eect, 86
gravity
weakness, 3
Hilbert gauge, 81
Hilbert-Palatini action, 76
Hodge operator, 62
index of refraction in weak gravi-
tational eld, 83
inhomogeneous wave equation for
electromagnetic eld, 41
integration of an n-form, 64
interior product, 59
isometry, 54
Jacobi equation, 68
Jacobian matrix, 18
Kerr solution, 136
Kerr spacetime
angular velocity, 138
ergosphere, 140
Keplers third law, 141
Killing horizon, 140
redshift approaching static
limit, 139
static limit, 138
Kerr-Newman solution, 136
electromagnetic eld, 137
gyromagnetic moment, 137
Killing
equation, 57
vector eld, 57
vector elds and conservation
laws, 57
Koszul formula, 35
Kruskal continuation, 126
Lense-Thirring eect, 86
Lie derivative, 54
and commutator, 55
coordinate expressions, 56
Lie transport, 67
light deection, 7
in Newtonian limit, 49
light speed in weak gravitational
eld, 83
Lorentz force, 40
Lorentzian spacetime, 24
Lovelocks theorem, 72
manifold, 13
dierentiable, 14
Maxwells equations
in general relativity, 40
in special relativity, 39
metric
Riemannian and pseudo-
Riemannian, 24
INDEX 188
metric tensor, 23
neutron stars
mass limit, 172
optical tidal matrix, 165
Palatini identity, 73
parallel transport, 27
perihelion shift
scalar theory of gravity, 12
Planck
length, 3
mass, 3
time, 3
polarisation
and parallel transport, 44
vector, 42
product manifolds, 15
pull-back, 53
push-forward, 53
quadrupole formula, Einsteins, 92
redshift
and Doppler shift, 44
approaching Schwarzschild
radius, 131
cosmological, 160
gravitational, 7, 8
Reissner-Nordstrm solution, 133
Ricci identity, 34
Ricci scalar
Ricci tensor, 33
Robertson-Walker metric, 161
scalar product, 23
Schwarzschild
metric, 103
Schwarzschild spacetime
Eddington-Finkelstein coordi-
nates, 127
eective potential, 108
equation of motion in, 109
Kruskal continuation, 126
Lagrangian of, 106
last stable orbit, 111
light deection, 114
perihelion shift, 112
spin precession, 117
Schwarzschild tetrad, 99
curvature forms of, 101
Einstein tensor of, 102
Ricci scalar of, 102
Ricci tensor of, 102
Shapiro delay, 84
spacetimes with constant curva-
ture
metric, 154
Ricci scalar, 154
Ricci tensor, 154
spatially homogeneous spacetime,
152
spatially isotropic spacetime, 152
spherically symmetric spacetime,
148
connection forms, 149
curvature forms, 149
Einstein tensor, 150
Ricci scalar, 150
Ricci tensor, 150
static spacetime, 97
metric, 97
stationary spacetime, 96
axially symmetric, 134
Stokes theorem, 65
structure equations, Cartans, 94
tangent space, 16
basis, 18
tensor, 21
transformation of, 22
tensor eld, 21
Tolman-Oppenheimer-Volko
equation, 170
torsion, 31
torsion forms, 94
torsion tensor, 32
in arbitrary basis, 96
two-sphere, 13
atlas, 14
variational principle
of general relativity, 74, 76
vector eld, 19
INDEX 189
autoparallel, 27
ow of, 52
global ow of, 52
integral curve of, 52
vector space, 16
vectors and innitesimal transfor-
mations, 20
volume form, 61
canonical, 62
Weyl tensor, 164

General Relativity: Matthias Bartelmann Institut F Ur Theoretische Astrophysik Universit at Heidelberg

Uploaded by

Copyright:

Available Formats

General Relativity: Matthias Bartelmann Institut F Ur Theoretische Astrophysik Universit at Heidelberg

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

General Relativity: Matthias Bartelmann Institut F Ur Theoretische Astrophysik Universit at Heidelberg

Uploaded by

Copyright:

Available Formats

General Relativity

h is the change in from the oor to the ceil-

. In such a theory, how ever it may look

is called an atlas of M if the domains of the

be two charts, and U

exists and denes a map between two open sets in R

, for which it is clear from ad-

, i.e. their derivatives of all or-

around its z axis. Let us

functions f from the

R dened on an open neighbourhood U

such that h(q) = x and h(p) = 0, i.e. x = h

instead of h, we obtain of course a dierent coordinate basis e

function f , its result (v( f ))(p) is a number

and their mul-

obtains the struc-

are called dual vectors.

function on M and v V an arbitrary tangent

and thus a dual vector. Choosing a coordinate rep-

instead, we arrive at the double-dual vector space V

is isomorphic to V and can thus be identied with V.

and can thus

V R. If we x a vector v V, T(, v) is a linear

R and thus an element of V

, which can be identied

) called the outer

by simply multiplying their results on the

because it linearly maps vectors into

vector elds another vector eld

V V R , (w, x, y) w[T(x, y)] (3.45)

and x, y V is a tensor of rank (1, 2) called the

V V V R , (w, x, y, v) = w[R(x, y)v] (3.51)

and x, y, v V denes a tensor of rank (1, 3) called

, we dene the covariant derivative of the torsion T such

= diag(1, 1, 1, 1) and use the notation (2.46)

is the current four-vector. The homogeneous equations

and by lowering them with g

does not vanish despite the Lorentz

. We further introduce the scalar am-

, and assigning frequencies

is antisymmetric, the second term on the right-

is contracted with the antisymmetric tensor A

is symmetric, however, this nal term remains, with important

. Thus, the result must

k) is the wave four-vector which satises

f )(p) = ( f )(p) = f [(p)] . (5.5)

pulls functions f on N back to M and is thus

f ) = v( f ). This denes a vector from

and the push-forward

can now be extended

is a dieomeorphism, which implies in partic-

its inverse, i.e.

is a symmetry transformation of T because T stays

. If the tensor eld

, n-form exists on M which vanishes nowhere. This

is the 1-form belonging

(y) = g(x, y) (5.97)

with (u, u) = 1 . (6.1)

(v, u))u , (6.6)