GR Brief Notes

Abhik Bhattacharjee
April 2023

1 Tensors, Curvature and Geodesics

For this section, I have mostly referred to

• Gravity by Hartle

• Spacetime and Geometry by Carroll

• Mathematical Methods for Physics and Engineering by Riley, Hobson, and Bence

1.1 Vectors, Dual Vectors and Tensors

A vector V can be written as V ≡ V α eα = V α ∂α (in coordinate basis). Under a coordinate
transformation, the components of V transform as

∂x′β α
V ′β = V . (1.1)
∂xα

A dual vector ω can be written as ω ≡ ωα eα , where {eα } are dual to {eα }: eα · eβ = δβα . Its
components transform as

∂xα
ωβ′ = ωα . (1.2)
∂x′β

Under a coordinate transformation, a mixed tensor T α β transforms as

∂x′α ∂xν µ
T ′α β = T ν. (1.3)
∂xµ ∂x′β

For mixed tensors, the horizontal position of indices is important: T α β , in general, is not the same
as Tβ α = gβµ (T µ ν )g να , where gµν is the metric tensor which is used to lower indices (its inverse
g µν is used to raise indices).

1
1.1.1 The Levi-Civita symbol
The Levi-Civita symbol is an object εαβγδ whose components change sign under the interchange of
any pair of indices and whose only non-zero components are ±1. We set ε0123 = +1. It transforms
not as a tensor, but a tensor density (of weight −1) under coordinate transformations x → x′ :

∂xα ∂xβ ∂xγ ∂xδ ∂x

′µ ′ν ′ρ ′σ
εαβγδ = ′
εµνρσ = J −1 εµνρσ , (1.4)
∂x ∂x ∂x ∂x ∂x

where J denotes the determinant of the Jacobian

p of the coordinate transformation. We can de-
fine the Levi-Civita pseudo-tensor ϵµνρσ ≡ |g|εµνρσ which transforms like a tensor but with a
multiplicative factor of sign(J).

A note on Special Relativity: Under a Lorentz transformation, a spacetime 4-vector trans-

forms as
∂x′µ ν
x′µ = x ≡ Λ µ ν xν (1.5)
∂xν
such that the line element

ds2 = ηµν dxµ dxν = −dt2 + dx2 + dy 2 + dz 2 (1.6)

remains invariant, which requires ηρσ = Λµ ρ Λν σ ηµν . Under a Lorentz transformation, the
differential operator (a co-vector) transforms as

∂µ′ = (Λ−1 )ν µ ∂ν . (1.7)

1.2 The Covariant Derivative

∇µ ϕ = ∂µ ϕ (scalar). (1.8)

∇ν V µ = ∂ν V µ + Γµνλ V λ or V;νµ = V,νµ + Γµνλ V λ (vector). (1.9)

∇µ (V ν ων ) = ∂µ (V ν ων ) =⇒ ωµ;ν = ωµ,ν − Γλµν ωλ (co-vector). (1.10)

T µ ν;ρ = T µ ν,ρ + Γµσρ T σ ν − Γσνρ T µ σ (mixed tensor). (1.11)

Here Γµνλ is the Christoffel symbol given by

1
Γρµν = g ρσ (gµσ,ν + gνσ,µ − gµν,σ ) . (1.12)
2

2
Under a coordinate transformation, the Christoffel symbol transforms as
∂xσ ∂xρ ∂x′λ τ ∂xσ ∂xρ ∂ 2 x′λ
Γ′λ
µν = Γ − (1.13)
∂x′ν ∂x′µ ∂xτ ρσ ∂x′ν ∂x′µ ∂xρ ∂xσ
∂xσ ∂xρ ∂x′λ τ ∂x′λ ∂ 2 x′τ
= Γ + . (1.14)
∂x′ν ∂x′µ ∂xτ ρσ ∂xτ ∂xµ ∂xν
The second line follows from the first upon using ∂λ′ (∂ρ x′µ ∂ν′ xρ ) = ∂λ′ (δνµ ) = 0.

A note on Lie derivative: Let the coordinates of a point P be xα and that of Q be x′α =
xα + dxα = xα + uα dλ. They are points on a curve γ, parameterized by λ, and let Aα be a
vector field in the neighbourhood of γ. Then the Lie derivative of Aα is defined as:

Aα (Q) − A′α (Q)

Lu Aα (P ) ≡
dλ


A + uβ ∂β Aα dλ
α

P
− Aα
+ ∂β uα Aβ dλ P
=
dλ (1.15)
β α α β
= u ∂β A (P ) − ∂β u A (P )
= uβ ∇β Aα (P ) − ∇β uα Aβ (P )
≡ [u, A]α (P ) (Lie bracket).

This measures how the vector field Aα changes along the flow of the vector field uβ .

1.3 Vector Operators in Tensor Form

Gradient: ∇ϕ = (∂µ ϕ)eµ . (1.16)

p
1 ∂ |g| ν 1 ∂ p
Divergence: ∇µ V µ = ∂µ V µ + Γµνµ V ν = V,µµ + p V = p ( |g|V µ ). (1.17)
|g| ∂xν |g| ∂xµ
1 ∂ p
Laplacian: ∇µ ∇µ ϕ = p µ
( |g|g µν ∂ν ϕ). (1.18)
|g| ∂x

1.4 Local Inertial Frames

It follows from the equivalence principle that at each point P in a curved spacetime manifold
described by a metric gµν (x), it is possible to transform to a new basis {x} → {x′ } such that
′
gµν (x′P ) = ηµν . (1.19)
′
We can make the coordinate transformation such that ∂gµν /∂x′ρ |xP = 0. Then the coordinate
′
system {x } defines a local inertial frame (LIF) at point P . However, no coordinate transformation
exists which satisfies the above two conditions that makes all second derivatives vanish at xP .
Under the coordinate transformation, the metric tensor transforms as

′ ∂xρ ∂xσ
gµν (x′ ) = gρσ (x), (1.20)
∂x′µ ∂x′ν
3
which involves factors of ∂x/∂x′ . We can expand xα can be in a power series about xP :
 α
α ′β ∂x
α ′β
x (x ) = x (xP ) + ′β
(x′β − x′βP)
∂x xP
 2 α 
1 ∂ x
+ ′β ′γ
(x′β − x′β ′γ
P )(x − xP )
′γ
2 ∂x ∂x xP
∂ 3 xα
 
1
+ (x′β − x′β ′γ ′γ ′δ ′δ
P )(x − xP )(x − xP ) + · · · (1.21)
6 ∂x′β ∂x′γ ∂x′δ xP

Plugging this in the transformation equation above and comparing with

 ′ 
∂gµν
′ ′ ′ ′
gµν (x ) = gµν (xP ) + (x′ρ − x′ρ
P)
∂x′ρ xP
 2 ′ 
1 ∂ gµν
+ (x′ρ − x′ρ ′σ ′σ
P )(x − xP ) + · · · , (1.22)
2 ∂x′ρ ∂x′σ xP

we see that there are 16 numbers (∂xα /∂x′β )xP to adjust to make the transformed values of the
′ ′
metric gµν equal to ηµν . But gµν only has 10 independent components, so this leaves us with 6
numbers to spare – these correspond to the 6 d.o.f, 3 rotations and 3 Lorentz boosts, that keep
ηµν invariant. There are 10 × 4 = 40 first derivatives of the metric and 4 × (4 · 5/2) = 40 second
derivatives (∂ 2 xα /∂x′β ∂x′γ )xP which we can adjust to set all ∂gµν
′
/∂x′ρ to zero at xP . Now coming
to the second derivative of the metric, there are 10 × (4 · 5/2) = 100 independent components, but
only 4 × (4 · 5 · 6/6) = 80 independent components of (∂ 3 xα /∂x′β ∂x′γ ∂x′δ )xP . This leaves 20
non-vanishing components which turn out to be the components of the Riemann curvature tensor.

1.5 Absolute derivatives along curves and geodesics :

Consider the problem of calculating the derivative of a tensor along a curve r(s) characterized by
the parameter s. The term within the brackets after the last equality is called the absolute derivative
of v i along the curve r(s) and is often denoted by DV i /ds.

dV i dV i ∂ej dxk dV i dxk DV i

 
dV dei
= ei + V i = ei + V j k = + Γijk V j ei ≡ ei . (1.23)
ds ds ds ds ∂x ds ds ds ds

The absolute derivatives of vectors and rank-2 tensors can be written as

DV i dxk DVi dxk
= V;ki ; = Vi;k . (1.24)
ds ds ds ds
DT ij ij dx
k
DT i j i dxk DTij dxk
= T;k ; = T j;k ; = Tij;k . (1.25)
ds ds ds ds ds ds
If we identify the parameter s as the arc length and choose as our defining property for a geodesic
the path along which the absolute derivative of the tangent vector to the path vanishes, then

4
Dt/ds = 0.
If we now introduce an arbitrary coordinate system xi with basis vectors ei , then we find
k
 i k

Dt i dx dt i j dx
= t;k ei = + Γjk t ei = 0. (1.26)
ds ds ds ds

But, since tj = dxj /ds, it follows that the equation satisfied by a geodesic is

d2 xi j
i dx dx
k
+ Γ jk = 0. (1.27)
ds2 ds ds
A vector V is parallell transported along a curve parameterized by λ if DV j /dλ = ti ∇i V j = 0

i
(ti = dxi /dλ). For a tensor T i j , we have DT
dλ j
= tk ∇k T i j = 0.

1.5.1 The affine parameter

If a curve r(λ) = {xk (λ)} is parametrized by a scalar variable λ, then the length of the curve
between two points, A and B, is given by
Z B Z B Z Bp
L= ds ≡ f dλ = gij ẋi ẋj dλ, where ẋk = dxk /dλ. (1.28)
A A A

Using the calculus of variations, we can find the curve r ∗ (λ) that minimizes L:
d gik ẋi ∂k gij ẋi ẋj
   
d ∂f ∂f
= =⇒ =
dλ ∂ ẋk ∂xk dλ f 2f
f ẍi − ẋi f˙ ẋi j ∂k gij ẋi ẋj
=⇒ gik + ∂ j g ik ẋ =
f2 f 2f
i i ˙ i j
f ẍ − ẋ f ẋ ẋ ∂k gij ẋi ẋj
=⇒ gik + (∂ g
j ik + ∂ g
i jk ) =
f2 2f 2f
f˙ s̈
=⇒ gik ẍi + Γkij ẋi ẋj = gik ẋi = gik ẋi (∵ ṡ = f ), (1.29)
f ṡ
where in the last line I have used the definition of the Christoffel symbol in terms of the metric
tensor. Multiplying throughout with g kl , we arrive at
d2 xl i
l dx dx
j
s̈ dxl
+ Γ ij = . (1.30)
dλ2 dλ dλ ṡ dλ
If λ = as + b (a, b being constants), then ṡ = 1/a and s̈ = 0; further d/dλ = (1/a)d/ds. With this
substitution the above equation becomes
d2 xl i
l dx dx
j
+ Γij = 0, (1.31)
ds2 ds ds
which is the geodesic equation (1.27).
A parameter which, like λ, is a sum of a linear transformation of s and a translation is called an
affine parameter.

5
1.6 Linearized gravity
The Newtonian limit is defined by three requirements:
• The particles are moving slowly: dxi /dτ ≪ dt/dτ .
• The gravitational field is weak: gµν = ηµν + hµν with |hµν | ≪ 1.
• The field is static: hµν (x) = hµν (x).
The equation of motion is
1 i d 2 xi
∂ h00 = = −∂ i Φ =⇒ h00 = −2Φ. (1.32)
2 |{z} dt2 |{z}
geodesic eqn. Newton’s law

1.7 The Riemann Curvature Tensor

It is defined by
∇µ ∇ν V ρ − ∇ν ∇µ V ρ = ∂µ Γρνσ + Γρµλ Γλνσ − (µ ↔ ν) V σ ≡ Rρ σµν V σ .


(1.33)
Alternatively, it can be defined through the parallel transport of a vector Aµ around a parallelogram
P QP ′ Q′ P spanned by two infinitesimal vectors aα and bβ (shown in Fig. 1a):
δAµ = dAµ (P QP ′ ) − dAµ (P Q′ P ′ ) = −Rµ λαβ Aλ aα bβ , (1.34)
where P Q = aα , P Q′ = bβ , and Q′ P ′ = (a + da)α = aα − Γαµν aµ bν , QP ′ = (b + db)β =
bβ − Γβµν aµ bν are obtained by parallel transporting aα and bβ respectively.
The Riemann tensor has the following properties:
(a) Rαβγδ = −Rβαγδ and Rαβγδ = −Rαβδγ .
(b) Rαβγδ = Rγδαβ .
(c) Rαβγδ + Rαδβγ + Rαγδβ = 0 (cyclicity).
(d) ∇λ Rαβγδ + ∇β Rλαγδ + ∇α Rβλγδ = 0 or equivalently ∇[λ Rαβ]γδ = 0 (Bianchi identity).
In D dimensions, it has D2 (D2 − 1)/12 independent components.
The Ricci tensor and the Ricci scalar are obtained by contracting the Riemann tensor:
Rµν = Rλ µλν ; R = g µν Rµν . (1.35)
The Ricci tensor and scalar contain all of the information about traces of the Riemann tensor,
leaving us with the trace-free parts, which are captured by the Weyl tensor
2
2
Cρσµν = Rρσµν − gρ[µ Rν]σ − (ρ ↔ σ) + gρ[µ gν]σ R, (D ≥ 4)
(D − 2) (D − 1)(D − 2)
(1.36)
The Ricci tensor and scalar are used to construct the Einstein tensor:
1
Gµν ≡ Rµν − gµν R. (1.37)
2
A useful resource : Decompositions of the Riemann tensor.

6
(a) The parallelogram used to define the Rie- (b) Evolution of two geodesics with separation
mann tensor through parallel transport. B µ in curved spacetime.

Figure 1: Definition of Riemann tensor through parallel transport [from A College Course on Rela-
tivity and Cosmology by Ta-Pei Cheng] and geodesic deviation [from Cosmology by D. Baumann].

1.8 Equation of Geodesic Deviation

The geodesic deviation equation is (see Fig. 1b)

D2 µ
B = ∇σ (∇ρ B µ U ρ )U σ = ∇σ (∇ρ U µ B ρ )U σ = Rµ νρσ U ν U ρ B σ (1.38)
dτ 2

Here B µ = ∂xµ /∂λ is the separation vector and U µ = ∂xµ /∂τ is the tangent vector, where τ is an
affine parameter along the geodesics and λ is a parameter labelling different geodesics.

A problem: Show that if tα = dxα /dλ obeys the geodesic equation in the form DtαR/dλ = κt
α ,
then uα = dxα /dµ satisfies Duα /dµ = 0 if µ and λ are related by dµ/dλ = exp κ(λ)dλ .
We have
Dtα d 2 xα α
α dx dx
γ
dxα
= + Γ βγ = κ(λ) (1.39)
dλ dλ2 dλ dλ dλ
α 2 α β γ
Du dx dx dx
= 2
+ Γαβγ =0 (1.40)
dµ dµ dµ dµ
Comparing these two equations we arrive at the following differential equation
2
d2 λ

dλ
+ κ(λ) = 0, (1.41)
dµ2 dµ

which is a differential equation for the variable v = dλ/dµ. This can be solved to get
 Z  Z 
dµ
v = exp − κ(λ)dλ and hence = exp κ(λ)dλ .
dλ

7
Tutorial I
PROBLEM 1

Prove the quotient theorem which states that if Aγ Pγαβ is a tensor for an arbitrary vector
Aγ , then Pγαβ is a tensor.

PROBLEM 2

Show that
√
1 1 ∂ −g
Γµνµ = ∂ν (ln |g|) = √ . (1.42)
2 −g ∂xν

PROBLEM 3

The action for a test particle of charge q and mass m in an EM field in curved spacetime is
Z Z
S = −m dτ + q dxµ Aµ . (1.43)

(a) Obtain the equation of motion, i.e. the Lorentz force law
Duµ q
= F µν uν , (1.44)
dτ m
where D/dτ is the absolute derivative, uµ = dxµ /dτ is the 4-velocity of the particle and
F µν is the EM field-strength tensor, and note that F 0i = E i and F ij = ϵijk Bk .

(b) Show that in flat space time it reduces to

dp
= q(E + v × B). (1.45)
dt

PROBLEM 4

Show that Maxwell’s equations in curved spacetime, viz. ∇µ F µν = −J ν follow from the
following action:

√
Z  
4 1 µν λ
S = − d x −g Fµν F + Aλ J . (1.46)
4

(a) Obtain the following wave equation (in the Lorenz gauge ∇µ Aµ = 0):

2Aν = −Jν + Rµ ν Aµ , (where 2 = ∇µ ∇µ ). (1.47)

(b) Establish charge conservation, i.e. ∇µ J µ = 0.

8
Tutorial II
PROBLEM 1

Consider the following metric in 3 dimensions

dr2
ds2 = + r2 (dθ2 + sin2 θdϕ2 ). (1.48)
1 − (r/a)2

Calculate the circumference around the equator, area, volume, and distance from center-to-
surface of a sphere of coordinate radius R centered on r = 0 in this space.

PROBLEM 2

Consider the following metric in (t, r) coordinates

dr2
 
2 2 2 2
ds = L −(r − 1)dt + 2 (L = const.). (1.49)
r −1

Consider also a tensor field Sµν with components S00 = a(r2 − 1), S11 = −a/(r2 − 1), and
S01 = S10 = 0 for some constant a > 0.

(a) Compute all the Christoffel symbols in the (t, r) coordinates.

(b) Compute S µν and ∇µ S µν .

(c) Perform a coordinate transformation (t, r) → (θ, η) with θ = at and r = cosh η.

Compute Sµν in this new coordinate system.

PROBLEM 3

Consider the metric on the surface of 2-sphere of radius a

ds2 = a2 (dθ2 + sin2 θdϕ2 ). (1.50)

(a) Find the Christoffel symbols.

(b) Find the Riemann tensor.

(c) Show that the Ricci scalar R ≡ g µν Rµν = 2/a2 and the Kretschmann scalar K ≡
Rµνρσ Rµνρσ = 1/a4 .

(d) Consider the vector A0 = eθ at (θ, ϕ) = (θ0 , 0) on the surface of the 2-sphere. The
vector is parallel transported all the way around the latitude circle θ = θ0 . What is the
resulting vector A? What is the magnitude (A · A)1/2 ?

PROBLEM 4

9
A sphere can be projected onto a plane using stereographic projection. Consider again the
metric on a 2-sphere (from the previous problem). Perform the coordinate transformation

x = 2a tan (θ/2) cos ϕ, y = 2a tan (θ/2) sin ϕ, (1.51)

and write the metric in terms of the x and y coordinates.

(a) Compute the Christoffel symbols for the resulting metric.

(b) What is the condition satisfied by a 4-vector ξ for the quantity V α ξα (V being the
tangent vector) to be conserved along any geodesic?

(c) Show that ξ = −yex + xey satisfies this condition.

Figure 2: Stereographic projection of a unit sphere on a plane [from here].

10
 Home Assignment I
PROBLEM 1 [3 points]
Consider a second rank tensor R with components Rβ γ . We define
Tαβ γ = ∂α Rβ γ . (1)
(a) Show that Tαβ γ transforms as a (1, 2) tensor under Lorentz transformation but is not a
tensor under general coordinate transformations.
(b) Find terms that you need to add to Tαβ γ such that the resulting expression transforms as
a tensor under general coordinate transformations.
PROBLEM 2 [3 points]
A tangent vector in the Cartesian coordinates is written as V = V x î+V y ĵ+V z k̂. Determine
the unit vectors r̂, θ̂, and ϕ̂ in the spherical polar coordinate system.
PROBLEM 3 [7 points]
Consider a conformally flat metric gαβ (x) = ω(x)ηαβ = e2ϕ(x) ηαβ (in 4 dimensions).
(a) Show that the Christoffel symbols are given by
1
Γγαβ = (δαγ ∂β ω + δβγ ∂α ω − η γδ ηαβ ∂δ ω). (2)
2ω(x)
(b) Show that the Riemann tensor is given by

α 1 α ∂γ ω
R βγδ = (δδ ∂β ∂γ ω − ηβδ η αλ ∂λ ∂γ ω) − 2 (δβα ∂δ ω + δδα ∂β ω − ηβδ η αλ ∂λ ω)
2ω 2ω
 
1 α αλ α α α 2
+ 2 δγ ∂β ω∂δ ω − ηγβ η ∂λ ω∂δ ω + δγ ∂δ ω∂β ω + δδ ∂γ ω∂β ω − ηβδ δγ (∂ω)
4ω
− [γ ↔ δ]. (3)
(c) Finally show that the Ricci tensor is given by
Rαβ = −2ϕ,αβ + 2ϕ,α ϕ,β − ηαβ [∂ 2 ϕ + 2(∂ϕ)2 ]. (4)
PROBLEM 4 [7 points]
The line element of flat spacetime in a frame (t, x, y, z) rotating with an angular velocity Ω
about the z-axis of an inertial frame is
ds2 = −[1 − Ω2 (x2 + y 2 )]dt2 + 2Ω(y dx − x dy)dt + dx2 + dy 2 + dz 2 . (5)
(a) Verify this by transforming to polar coordinates and checking that the line element is
ds2 = −dt2 + dr2 + r2 dθ2 + r2 sin2 θdϕ2 (6)
with the substitution ϕ → ϕ − Ωt.
(b) Find the geodesic equations for x, y, and z in the rotating frame.
(c) Show that the above equations can be reduced to the usual equations of Newtonian
mechanics for a free particle in a rotating frame.
2 Gravitation
I have referred to Spacetime and Geometry by Carrol for this section.

2.1 Einstein’s equation

1 8πG
Gµν ≡ Rµν − gµν R = 4 Tµν . (2.1)
2 c

2.2 Lagrangian Formulation

Einstein-Hilbert action:
√
Z Z
4
SEH = d xL = d4 x −gR. (2.2)

This leads to Einstein’s equation when varied w.r.t g µν :

R √ R √
• (δS)1 = d4 x −gg µν δRµν = d4 x −g∇σ [gµν ∇σ (δg µν ) − ∇λ (δg σλ )], which can be set
to 0 by Stoke’s theorem.1
R √
• (δS)2 = d4 x −gRµν δg µν .
Note1. δgµν = −gµρ gνσ δg ρσ .
R √ R √
• (δS)3 = d4 xRδ −g = d4 xR(− g(gµν /2)δg µν ).
Note2. Variation of the identity ln (det M ) = Tr(ln M ) yields δ(det M )/ det M = T r(M −1 δM ) →
δg = −ggµν δg µν .
In order to derive the full Einstein equation, we consider the action
c4
S= SEH + SM , (2.3)
16πG
and we define the energy-momentum tensor to be
−2 δSM
Tµν ≡ √ , (2.4)
−g δg µν
which is automatically symmetric since g µν is symmetric.

2.3 Symmetries and Killing vectors

If gµν , a solution of Einstein’s equations, is independent of xσ∗ (σ∗ denoting a particular coordi-
nate), then we define a vector K = ∂σ∗ or K µ = δσ∗ µ
. We say that the vector K µ generates an
isometry (symmetry of the metric). Then constancy of the quantity2 pσ∗ ≡ K · p (a scalar) along
a geodesic implies Dpσ∗ /dτ = 0 or Dpσ∗ /dλ = 0:
pµ ∇µ (K · p) = 0 =⇒ pµ pν ∇µ Kν = 0 =⇒ ∇(µ Kν) = 0 (Killing’s eqn.). (2.5)
p p
If V µ is a vector field over a region Σ with boundary ∂Σ, then Σ ∇ · V |g|dD x = ∂Σ n · V |γ|dD−1 x, where
1
R R

nµ is normal to ∂Σ and γij is the induced metric on ∂Σ.

2
The 4-momentum is pµ = mdxµ /dτ for massive particles and pµ = dxµ /dλ for massless particles.

12
 • If K and L are 2 Killing vectors, P = aK + bL is also a Killing vector. Their commutator
M = [K, L] is also a Killing vector:

M ≡ M α ∂α = K α ∂α Lβ ∂β − Lβ ∂β K α ∂α = K β ∇β Lα − Lβ ∇β K α ∂α ;


(2.6)

∇α Mβ + ∇β Mα = Rβδαγ K γ Lδ − Rαδβγ K δ Lγ − Rβδαγ K δ Lγ + Rαδβγ K δ Lγ = 0. (2.7)

• The derivatives of Killing vectors can be related to the Riemann tensor: ∇µ ∇σ K ρ = Rρ σµν K ν .

• The directional derivative of the Ricci scalar along a Killing vector field vanishes: K α ∇α R =
0.

For a maximally symmetric spacetime of dimension D, the number of linearly independent Killing
vectors is D(D + 1)/2. A maximally symmetric spacetime has constant curvature:

Rµνρσ = K(gµρ gνσ − gµσ gνρ ) =⇒ R = KD(D − 1). (2.8)

For flat spacetime, K = 0 and for positively/negatively curved spacetime, K ≷ 0.

A problem: Show that the following metric satisfies Einstein’s equation with a negative cos-
mological constant

R2
ds2 = (−dt2 + dx2 + dy 2 + dz 2 ). (2.9)
z2
The EH action with a cosmological constant is
√
Z
δSEH
SEH = d4 x −g(R − 2Λ) → = 0 =⇒ Gµν = −Λgµν . (2.10)
δgµν

The non-zero Christoffel symbols for the metric are

1 1
Γttz = Γxxz = Γyyz = Γzzz = − ; Γztt = Γzxx = Γzyy = (2.11)
z z
The Ricci tensor is given by (from (1.33))

Rσν = ∂ρ Γρνσ + Γρρλ Γλνσ − ∂ν Γρρσ − Γρνλ Γλρσ .

Given the Christoffel symbols, this yields the following non-zero components of the Ricci
tensor
3 3
−Rtt = Rxx = Ryy = Rzz = − 2
=⇒ Rµν = − 2 gµν . (2.12)
z R
Therefore, the Ricci scalar is R = −12/L2 . Hence the Einstein tensor is
3 6 3 compare with (2.10) 3
Gµν = − gµν + g µν = g µν −−− − − − − − − → Λ = − . (2.13)
R2 R2 R2 R2

13
Tutorial III
PROBLEM 1

Consider the following action describing the minimal coupling of gravity to a real scalar
field
√
Z  
4 1 µν 1 µν
S = d x −g 2
g R µν + L where L = − g ∂µ ϕ∂ν ϕ (κ2 = 8πG).
2κ 2
(2.14)
(a) Show that the equation of motion of the scalar field is
√ √
∂µ ( −gg µν ∂ν ϕ) = −gg µν ∇µ ∂ν ϕ = 0, (2.15)
and the Einstein equation is Gµν ≡ Rµν − gµν R/2 = κ2 Tµν where
Tµν = ∂µ ϕ∂ν ϕ + gµν L. (2.16)

(b) Derive the Ricci form of the Einstein field equations,

 
2 1 ρ
Rµν = κ Tµν − gµν T ρ . (2.17)
2

PROBLEM 2

Consider the following spherically-symmetric metric

ds2 = −e2α(r,t) dt2 + e2β(r,t) dr2 + r2 (dθ2 + sin2 θ dϕ2 ), (2.18)
The non-zero Christoffel symbols for this metric are (compute these)
Γttt = ∂t α, Γttr = Γtrt = ∂r α, Γtrr = e2(β−α) ∂t β,
Γrtr = Γrrt = ∂t β, Γrtt = e2(α−β) ∂r α, Γrrr = ∂r β,
1 1
Γθrθ = Γθθr = , Γrθθ = −re−2β , Γϕrϕ = Γϕϕr = ,
r r
Γrϕϕ = −re−2β sin2 θ, Γθϕϕ = − sin θ cos θ, Γϕθϕ = Γϕϕθ = cot θ,
(a) Show that the components of the Ricci tensor are
 
 2 2
 2(α−β) 2 2 2
Rtt = ∂t β + (∂t β) − ∂t α∂t β + e ∂r α + (∂r α) − ∂r α∂r β + ∂r β
r
(2.19)
 
2
Rrr = − ∂r2 α + (∂r α)2 − ∂r α∂r β − ∂r β + e2(β−α) ∂t2 β + (∂t β)2 − ∂t α∂t β
 
r
(2.20)
2
Rtr = Rrt = ∂t β, Rθθ = e−2β [r(∂r β − ∂r α) − 1] + 1, Rϕϕ = sin2 θRθθ .
r
(2.21)

14
(b) Use Einstein’s equations in vacuum to find α(t, r) and β(t, r). Substitute your results in
the metric to obtain
 rS  2  rS −1 2
ds2 = − 1 − dt + 1 − dr + r2 (dθ2 + sin2 θ dϕ2 ). (2.22)
r r

The fact that this metric is independent of t indicates that any spherically symmetric vacuum
metric possesses a timelike Killing vector K α = (1, 0, 0, 0).
Comparing the r ≫ rS limit of this metric with the Newtonian weak field metric yields
rS = 2GM . We have thus arrived at the Schwarzschild metric.

PROBLEM 3*

Find all the Killing vectors of the Schwarzschild metric:

   −1
2 2GM 2 2GM
ds = − 1 − dt + 1 − dr2 + r2 (dθ2 + sin2 θ dϕ2 ). (2.23)
r r

There will be four of them: K = ∂t (mentioned above), R, S, and T . By evaluating their

commutators show that R, S, and T form a closed algebra.

15