Example Sheet 3 Part III: General Relativity

Claude Warnick Michaelmas 2022

Exercise 3.1. a) A scalar field obeying the Klein–Gordon equation ra rb m2 = 0 has

energy-momentum tensor Tab = ra rb 1
g
2 ab (r c r c
+ m2 2
). Show that Tab is
conserved.

b) Show that the source free Maxwell equations imply that the energy-momentum tensor of the
electromagnetic field is conserved.

c) Show that the conservation of the energy-momentum tensor of a perfect fluid implies that

ua ra ⇢ + (⇢ + p)ra ua = 0, (⇢ + p)ub rb ua = (gab + ua ub )rb ⇢.

Exercise 3.2. Physically reasonable matter with energy-momentum tensor Tab is expected to
satisfy the weak energy condition: Tab ua ub > 0 for all timelike ua . Give a physical interpretation
of this condition.
You measure the components of T a b in some basis and determine its eigenvalues and
eigenvectors v a satisfying T a b v b = v a . You find that it has precisely one timelike eigenvector
with eigenvalue ⇢ and three spacelike eigenvectors with eigenvalues p(↵) . In terms of these
eigenvalues, find a necessary and sufficient condition for the weak energy condition to hold.

Exercise 3.3. Consider the energy-momentum tensor describing a point mass at the origin: in
“almost inertial” coordinates it is T00 (t, x) = M 3 (x), T0i = Tij = 0. Determine the linearised
gravitational field produced by this energy-momentum tensor, assuming it to be independent of t.
For what values of R = |x| is the linear approximation valid?

Exercise 3.4. In “almost inertial” coordinates the energy-momentum tensor of a straight cosmic
string aligned along the z-axis is

Tµ⌫ = µ (x) (y)diag(1, 0, 0, 1),

where µ is a small positive constant. Neglecting terms of order µ2 , show that the linearised
Einstein equation has a time independent solution for which h11 = h22 = p are the only non-zero
components of the perturbed metric tensor, where = 8µ log(r/r0 ), r = x2 + y 2 , and r0 is an
arbitrary length.
Show that the perturbed metric can be written in cylindrical polar coordinates as

ds2 = dt2 + dz 2 + (1 )(dr2 + r2 d 2

).

Introduce a new radial coordinate, r0 , by (1 )r2 = (1 8µ)r02 to obtain

ds2 = dt2 + dz 2 + dr02 + (1 8µ)r02 d 2

.

Finally, change the angular coordinate to obtain

ds2 = dt2 + dz 2 + dr02 + r02 d 02

.

Is this Minkowski space? Show intuitively how a distant object may give rise to double images.

Please send any corrections to [email protected]

2

Exercise 3.5. Consider a large thin spherical shell of mass M and radius R which rotates
slowly about the z-axis (in “almost inertial” coordinates) with angular velocity ⌦, so that
terms of O(R2 ⌦2 ) can be neglected. Introduce a shell density ⇢ = M (r R)/(4⇡R2 ), where
r2 = x2 + y 2 + z 2 , and a 4-velocity uµ = (1, ⌦y, ⌦x, 0). The energy-momentum tensor has
components T µ⌫ = ⇢uµ u⌫ . We can regard this source as a superposition of two sources, one for
which only T00 is non-zero and one for which only T0i is non-zero.

a) Solve the linearized Einstein equations sourced by T00 . Show that the result agrees with
Newtonian theory.

b) Consider the perturbation sourced by T0i . Argue that the only non-vanishing components are
h0i , which satisfy r2 h0i = 16⇡T0i . Consider the combination h = h01 + ih02 and work in
spherical polar coordinates. Show that r2 h = F (r) sin ✓ei for some F (r). Observing that
sin ✓ei is the spherical harmonic with l = m = 1, and recalling that a general solution to
Laplace’s equation can be expanded in spherical harmonics, deduce that h = f (r) sin ✓ei for
some f (r) which you should determine. Hence obtain the solution
⇢
!(y, x, 0), r<R
h0i = 3
! Rr3 (y, x, 0) r>R

where ! = 4M ⌦/(3M ). Note that this decays as 1/r2 for large r in contrast to the contribution
from part a) which decays as 1/r. This is a general result: the effect of rotation on hµ⌫ is
subleading to the effect arising from total mass.

c) Consider a free particle moving non-relativistically inside the shell. Working to first order in
the perturbation, show that the geodesic equation implies ẍ = 2! ⇥ ẋ, where ! = (0, 0, !)
and a dot denotes a derivative with respect to t. This is the same as the equation of a
slowly moving free particle in Minkowski spacetime, using a reference frame rotating with
angular velocity ! with respect to an inertial frame. Hence, inside the shell, the background
Minkowski frame is rotating with angular velocity ! relative to a local inertial frame. Thus
the local inertial frames are rotating with angular velocity ! with respect to the background
Minkowski frame. The local inertial frames are dragged around by the rotating shell. This is
the Lens-Thirring effect (1918).

Exercise 3.6 (Optional). Show that the second order terms in the expansion of the Ricci tensor
around Minkowski spacetime are

(2) 1 ⇢ 1
Rµ⌫ [h] = h @ µ @ ⌫ h⇢ h⇢ @⇢ @(µ h⌫) + @µ h⇢ @⌫ h⇢ + @ h⇢ ⌫ @[ h⇢]µ
2 4 ✓ ◆
1 ⇢ 1 ⇢ 1 ⇢
+ @ (h @⇢ hµ⌫ ) @ h@⇢ hµ⌫ @ h⇢ @ h @(µ h⌫)⇢
2 4 2

Exercise 3.7. a) Use the linearised Einstein equation to show that, in vacuum,
D E
⌘ µ⌫ Rµ⌫
(2)
[h] = 0.

D ⇢ ⇢
E
b) Show that htµ⌫ i = 1
32⇡ @ µ h⇢ @ ⌫ h 1
2 @µ h@⌫ h 2@ h @(µ h⌫)⇢

c) Show that htµ⌫ i is gauge invariant i) to first order in ⇠µ ; ii) (optional, lengthy) to second
order in ⇠µ .
 3

Exercise 3.8. Two bodies, with masses m1 , m2 , move in an elliptical Newtonian orbit with
semi-major axis a and eccentricity e. Choosing coordinates so that the orbit lies in the (x, y)
plane with the centre of mass at the origin, the bodies have positions x1 = (r1 cos , r1 sin , 0)
and x2 = ( r2 cos , r2 sin , 0), where
✓ ◆ ✓ ◆
m2 m1
r1 = r, r2 = r.
m1 + m2 m1 + m2
The equation of the ellipse and the angular velocity of the orbit are
p
a(1 e2 ) ˙ (m1 + m2 )a(1 e2 )
r(t) = , (t) =
1 + e cos (t) r(t)2
In this question, you should consider the case of a circular orbit e = 0 with equal masses m1 = m2 .
If you want a challenge consider the extension to non-circular orbits and unequal masses.
Treating the bodies as non-relativistic point masses (in the sense of Exercise 3.3), compute
the corresponding energy-momentum tensor, the second moment of the energy distribution Iij ,
and the metric perturbation hij . Show that the time average of the total power radiated in
gravitational waves is
32 m21 m22 (m1 + m2 ) 1 + (73/24)e2 + (37/96)e4
hP i = f (e), f (e) = 7
5 a5 (1 e2 ) 2
Note that f (e) increases rapidly as e ! 1, so highly eccentric orbits emit more gravitational
radiation than circular ones. [This calculation is taken from P. C. Peters and J. Mathews, Physical
Review 131, p435 (1963)]
Exercise 3.9. As the above system emits gravitational radiation, it loses energy and angular
momentum which causes the shape of the orbit to change gradually. We can model this by
allowing a and e to be slowly varying functions of time. The energy of a Newtonian orbit is
E = m1 m2 /(2a), so a decreases over time. Setting dE/dt = hP i gives an expression for da/dt.
For a circular orbit (e = 0), use this to show that a(t) reaches zero at a time
5a(0)4
T = .
256m1 m2 (m1 + m2 )
Consider two black holes in a circular orbit, each with mass 30M , which gives m1 = m2 ⇡ 45 km.
What is the time to merger is the initial distance between them is one astronomical unit (AU)?
(1AU is the Earth-Sun distance: 1.5 ⇥ 108 km.) For comparison, the age of the universe is
about 14 ⇥ 109 years. How far apart are the black holes when the time to merger is 1 year? For
comparison, the radius of the sun is about 7 ⇥ 105 km.
[The general case of an elliptic orbit with evolving a and e is considered in P. C. Peters,
Physical Review 136, p1224 (1964). Emission of gravitational radiation causes e to decrease so
the orbit becomes more circular over time.]
Exercise 3.10. Consider the above system after emission of gravitational radiation has caused
the orbit to become circular. How is the frequency f of the gravitational waves related to ˙ ? Use
your expression for da/dt from the previous question to derive an expression for ¨, the rate of
change of the angular velocity due to emission of gravitational radiation. Eliminate a to obtain
an expression relating ˙ , ¨, m1 and m2 . Hence show that by measuring f and f˙ it is possible to
determine the chirp mass
3
(m1 m2 ) 5
Mchirp = 1
(m1 + m2 ) 5
[This gives an estimate of the total mass of the system. One can then estimate the distance to
the binary using the magnitude of the gravitational radiation.]