Professional Documents
Culture Documents
Selected Exercises From Susskind's General Relativity (TTM4)
Selected Exercises From Susskind's General Relativity (TTM4)
Selected Exercises From Susskind's General Relativity (TTM4)
*Exercise 2
Page 30
[Is it possible to find a curved surface and a lattice of rods arranged on it that cannot be flattened out, but can change shape? ]
Hint:
This exercise is answered in the text.
■
2. Tensor mathematics
Exercise 1
Page 65
[Prove that, in an orthonormal basis, equation (5) is equivalent to equation (6).
Hint: Do it in two dimensions. Then - it is slightly more involved - we encourage you to try to do it in any dimension.]
Solution:
Equation (5) is the geometric definition of dot product:
V · W = |V||W| cos θ
V · W = V 1W 1 + V 2W 2 + · · · + V N W N
In an orthonormal base:
V = V i ei
W = W j ej
V · W = V i W j ei · ej = V i W j δi j = V i W i
In an orthonormal base contravariant and covariant components are exactly the same. In plane old components, by geometric
definition:
V · ei = |V||ei | cos θi = |V| cos θi = V i
1
where θi is the angle between V and ei , so
V · W = V · (W i ei = (V · ei )W i = V i W i
*Exercise 2
Page 109
[Explain why the covariant derivative of the metric tensor is always zero.]
Hint: see discussion in the text.
■
Exercise 3
Page 109
[On Earth, with the polar coordinates θ for latitude and ϕ for longiture, find
1. the metric tensor gmn
2. its inverse g mn
3. the Christoffel symbols at point (θ, ϕ).]
Note:
This exercise can be conjoined with Exercise 1 on page 133, chapter 4, items 1. and 2..
Solution:
The ordinary distance on a 2-sphere of unitary radius is ds2 = dθ2 + cos2 θdϕ2 . We get:
1 0
gmn =
0 cos2 θ
1 0
g mn = 1
0 cos2 θ
dX m dX n
t · t = tm tn gmn = gmn =1
dS dS
2
■
On page 129 it is stated that "Parallel-transporting a vector ... preserves its length. It can be shown as a consequence of equation
(14)."
Solution:
Equation (14) (on page 127) reads dV n + Γnmr V r dX m = 0. Its covariant counterpart is dVn − Γrmn Vr dX m = 0. The ordinary
change in length is
d (V · V) = Vm dV m + V m dVm
Plugging in it dV m = −Γmnr V r dX n and dVm = Γrmn Vr dX n we get d (V · V) = 0
■
Exercise 1
Note:
Exercise 1 continues on pages 133–134 of chapter 1, items 3., 4. and 5..
[3. Show that the tangent vector to a meridian has everywhere components t1 = 1 and t2 = 0.
4. Show that the tensor that is the covariant derivative of this tangent vector is
1 0
0 − tan θ
5. Show that if we follow a meridian, the covariant change of the tangent vector is always zero.]
Solution:
3. Remember that tθ = dS dθ
, tϕ = dS
dϕ
. As t is tangent to a meridian ϕ = const. Hence dS = dθ, coming up with tθ = 1 and tϕ = 0.
4. From equation (8) on page 124
Dr tm = ∂r tm + Γrnm tn
and the solution of item 3. it follows that ∂r tm = 0, whence Dr tm = Γrθm .
5. From
Dtn = dtn + Γmrn tr dX m
and the solution of item 3. it follows that dtn = 0, whence Dtn = Γmθn dX m .
We get:
Dtθ = Γmθθ dX m = 0
as Γmθθ = 0;
Dtϕ = Γmθϕ dX m = 0
as Γθθϕ = 0; and Γϕθϕ dϕ = − tan θdϕ = 0 as dϕ = 0 on the meridian.
■
Exercise 2
Page 149
[In figure 13, what is the speed, relative to the stationary frame, of the observer who sees R, S, and T as simultaneous events? ]
Solution:
From equation (31) on page 146 the coordinates of the events R, S, and T, are:
XR = cosh ω TR = sinh ω
XS = 2 cosh ω TS = 2 sinh ω
XT = 3 cosh ω TT = 3 sinh ω
Lorentz transforming the coordinates:
X ′ = γ(X − vT ) T ′ = γ(T − vX)
′
XR = γ(cosh ω − v sinh ω) TR′ = γ(sinh ω − v cosh ω)
′
XS = 2γ(cosh ω − v sinh ω) TS′ = 2γ(sinh ω − v cosh ω)
XT′ = 3γ(cosh ω − v sinh ω) TT′ = 3γ(sinh ω − v cosh ω)
where v is the speed of the sought-after observer. Imposing the simultaneity condition TR′ −TS′ = TS′ −TT′ = γ(sinh ω−v cosh ω) =
0 we get
v = tanh ω
■
3
Exercises left to the reader
Let’s revert to page 142 in order to prove that from the rest frame of the first observer (R), the distance to the second observer
(S ) grows.
Solution:
From the inertial frame the segment RS at some time T is, by definition, of fixed length. Lorentz inverse-transforming from the
rest frame of R (primed coordinates) to the non-primed frame:
|RS| = XS − XR = γ(XS′ − XR
′
) + γv(TS′ − TR′ )
4
Passing to Cartesian coordinates the second derivative can be directly computed:
p
x = r 1 + (t/r)2
1 −3/2
ax = 1 + (t/r)2
r
Note:
Notice that for t = 0 the ordinary acceleration ax coincides with the magnitude of the proper acceleration |a|. In the book
it is remarked that the (proper) acceleration on a given trajectory in figure 14, for example the (proper) acceleration at point
N on the hyperbola R = 2, is the ordinary acceleration. And it is the constant (proper) acceleration we would experience all
along the trajectory. In other words, the observer at point R will experience a constant acceleration |a|. At point N the inertial
observer would determine an acceleration ax = |a| (that diminishes later on). Notice that |a|, the magnitude of the relativistic
(four-)acceleration, coincides with γ 3 ax in the case that the (ordinary) (three-) acceleration ax and the velocity have the same
direction (this is of no concern for uni-dimensional motion); a is constant and Lorentz-invariant (only the magnitude, not the
direction) in the particular case of an object accelerating parallel to its velocity; γ 3 a, where a is the ordinary (three-)acceleration,
2 i
is the proper acceleration. (Do not confuse with the spatial part of the four-acceleration ai = ddτx2 ). Let’s recall a few definitions:
i
proper velocity: v i = dxdτ
i
proper acceleration: αi = dv dt
2 i
(relativistic) (four-)acceleration: ai = ddτx2
2
(ordinary) (three-)acceleration: ax = ddt2x
2
spatial part of the four-acceleration: a1 = ddτx2 = 1r cosh τr
The magnitudes of the relevant quantities are recapitulated in the following tables:
hyperbolic motion
magnitudes formulas
(four-)acceleration (1) 1
r √
1−γ 2
spatial part of (1) a1 = γ
r (the time part is a0 = r )
proper acceleration 1
r
γ −3
ordinary acceleration r
ordinary velocity v = tanh τr
proper velocity γv = sinh τr
gamma 1
γ = √1−v 2
= cosh τr
parabolic motion
magnitudes formulas
ordinary acceleration a
proper acceleration γ3a
ordinary velocity at
proper velocity γat
space part of four-acceleration γ4a
magnitude of four-acceleration γ3a
gamma γ=√ 1 (t = γτ )
1−(at)2
constraints of parabolic motion a = const. for 0 ≤ t < a,
c
a(t) = c
t for
t ≥ ac .
■
5
is equivalent to the definition of a geodesic given by equation (6), which says that the tangent vector to the trajectory in space-time
stays constant:
d2 X µ µ dX σ dX ρ
= −Γ σρ ]
dτ 2 dτ dτ
Solution: r
dX µ dX ν dτ
L = −m −gµν (X) = −m
dt dt dt
dX µ dX ν dX µ dτ dX ν dτ
∂L m ∂m gµν dt dt m ∂m gµν dτ dt dτ dt m dX µ dX ν dτ
m
= dτ
= dτ
= ∂m gµν
∂X 2 dt
2 dt
2 dτ dτ dt
dX µ dX ν
m gµm dt + gmν dt dX µ dX ν
∂L m
= dτ
= gµm + gmν
∂ Ẋ m 2 dt
2 dτ dτ
dX µ dX ν dX λ dX µ dX λ dX ν d2 X µ d2 X ν dτ
d ∂L m d m
= gµm + gmν = ∂λ gµm + ∂λ gmν + gµm + gmν
dt ∂ Ẋ m 2 dt dτ dτ 2 dτ dτ dτ dτ dτ 2 dτ 2 dt
dX µ dX ν dX λ dX µ dX λ dX ν d2 X µ d2 X ν
1 1
∂m gµν = ∂λ gµm + ∂λ gmν + gµm 2
+ gmν
2 dτ dτ 2 dτ dτ dτ dτ dτ dτ 2
dX λ dX µ dX λ dX ν dX µ dX ν d2 X ν
1
− ∂λ gµm + ∂λ gmν − ∂m gµν = gmν
2 dτ dτ dτ dτ dτ dτ dτ 2
dX ν dX µ dX µ dX ν dX µ dX ν d2 X ν
1
− ∂ν gµm + ∂µ gmν − ∂m gµν = gmν
2 dτ dτ dτ dτ dτ dτ dτ 2
1 dX µ dX ν d2 X ν
− ∂ν gµm + ∂µ gmν − ∂m gµν = gmν
2 dτ dτ dτ 2
1 dX σ dX ρ d2 X µ
− ∂ρ gσm + ∂σ gmρ − ∂m gσρ = gmµ
2 dτ dτ dτ 2
1 dX σ dX ρ d2 X µ
−g mµ · ∂ρ gσm + ∂σ gmρ − ∂m gσρ = g mµ gmµ
2 dτ dτ dτ 2
dX σ dX ρ d2 X µ
−Γµσρ =
dτ dτ dτ 2
■
The Hamiltonian is
H = pṙ − L
where
1
∂L 1− 2M G ṙ
r
p= =r
∂ ṙ 2M G
1
1− r − 1− 2M G ṙ2
r
6
i.e. the equation (36) (Note: in the book ṙ2 in equation (36) is misprinted as ṙ). Hence, solving for ṙ2 :
2 2M G 3
1−
2 2M G r
ṙ = 1− −
r E2
Exercise 2
Page 193
[Show that from equation (36) for the energy, and equation (37) for ṙ2 , it follows that
r
r − 2M G
ṙ ≈
2M G
as r → 2M G]
Note: q
r−2M G 2
[In the book the above equation is misprinted. Read it ṙ ≈ or ṙ ≈ r−2M
2M G .]
G
2M G
Solution:
Assuming a radial falling of an unitary mass from infinity (assuming at infinity ṙ∞ = 0), then from equation (36) E = 1, and
from equation (37), taking r = 2M G + ϵ, we get to the first order approximation in ϵ:
s 2 3 s 2
2M G 2M G r − 2M G 2M G r − 2M G
ṙ = 1− − 1− ≈ ≈
r r 2M G + ϵ 2M G + ϵ 2M G
where F = 1 − r .
2M G
Resolving for ϕ̇:
p
L F(r)
ϕ̇ = √
r2 L2 + m2 r4
Plugging it back into equation (15)
F(r)m
E=q
F(r) − r2 ϕ̇2
we get equation (17) for E as a function of r and given angular momentum:
q
L 2
p
F(r) r2 m + r4
E=m 2
r
■
On page 213 it is left to the reader to show that the point where E is stationary is at r = 3M G.
Solution:
To this aim we differentiate the RHS of equation (19), that is
√
F
E=L
r
7
Figure 1: Energy of a test particle as a function of the radial distance (arbitrary units).
Figure 2: Energy of a massive body as a function of the radial distance (arbitrary units).
8
Figure 3: Energy of a photon as a function of the radial distance (arbitrary units).
whence
MG
F=
r
i.e.
2M G MG
1− =
r r
so r = 3M G, that does not depend on L. The calculation of the second derivative is pretty tedious but straightforward; the
second derivative is negative at r = 3M G, i.e. E reaches a maximum at r = 3M G (also cf. figure 7 on page 213).
On page 215 it is recommended to compare the calculations done for the photons with the corresponding Newtonian calculations.
Solution:
Let’s start with the radial falling of a test mass. The newtonian Lagrangian is
1 2 M Gm
L = mṙ − −
2 r
1 2 M Gm
H = pṙ − L = mṙ −
2 r
Resolving for ṙ: r
2E 2M G
ṙ = +
m r
For a radial light ray (m → 0) beamed from point r the latter expression diverges. For an orbiting particle the Lagrangian is
1 M Gm
L = m ṙ2 + r2 Ω̇2 − −
2 r
9
where dΩ2 = dθ2 + cos2 θdϕ2 . In the plane θ = 0 (the symmetry of the system allows this choice without loss of generality),
dθ = 0 and dΩ reduces to dϕ. The linear momentum is, as before, pr = mṙ and the angular momentum L = ∂L ∂ ϕ̇
= mr2 ϕ̇.
Resolving for ϕ̇:
L
ϕ̇ =
mr2
The Hamiltonian is
1 2 L2 M Gm
H = pr ṙ + Lϕ̇ − L = mṙ + −
2 2mr2 r
Resolving for ṙ: r
2E L2 2M G
ṙ = − 2 2+
m m r r
Again, the above expression diverges for an orbiting light ray (m → 0). As for a circular orbit ṙ = 0 light rays are not allowed
to follow circular orbits. For a circular orbit (ṙ = 0) of a massive particle:
L2 M Gm
E= 2
−
2mr r
Given L, the (effective radial potential) energy is a function of r: E = E(r). Differentiating the RHS of the above expression
w.r.t. r and setting the derivative equal to zero we get the condition
dE L2 M Gm
=− 3 + =0
dr mr r2
2
The equilibrium position is r = M LGm2 , that depends on L, i.e. a massive particle is can follow circular orbits at any distance
L2
depending on L. It is easy to show that the second derivative is mr 4 > 0 at the distance of the orbital radius r, i.e. the (effective
radial potential) energy reaches a minimum (stable equilibrium position) at the distance of the circular orbit.
Note:
Alternatively the energy equation can be solved w.r.t. r:
L2 M Gm
E= 2
−
2mr r
2mr2 E + 2M Gm2 r − L2 = 0
√
−2M Gm2 ± ∆
r=
4mE
2 2 3 2
and imposing ∆ = 4M 2 G2 m4 + 8mEL2 = 0, i.e. E = − M 2LG m
2 , we get the radius of the circular orbit r = M L
Gm2 .
Note:
For comparison with the Newtonian calculations in the general case ṙ ̸= 0 we need the formulas in chapter 6. So let’s skip to
chapter 6.
■
6. Black Holes
*Exercise 1
Page 217
[Explain why a light ray emitted from inside the photon sphere can escape, but a light ray cannot enter the photon sphere and
come out again.]
Hint: refer to figure 7 on page 213.
Note:
Let’s continue the Newtonian calculations in chapter 5 in the general case ṙ ̸= 0 starting from the equations (11) and (14) in
this chapter, that is:
mr2 ϕ̇
L= √
g
mF
E= √
g
10
Figure 4: Effective radial potential of a test particle in the Newtonian approximation (arbitrary units).
2 2 !
mc2 2 mc2 3 L2
1 2 E E
mṙ 2
= F 2
− F − F3
2 mc 2 mc 2 2mr2
2 2
L2
1 2 mc ε mM G mc mM G
mṙ ≈ 1+2 2 −2 + ··· − −3 + ··· − F3
2 2 mc r 2 r 2mr2
In the limit c → ∞ F = 1 in the last term in the RHS. Rearranging we get the Newtonian approximation:
1 2 mM G L2
ε≈ mṙ − +
2 r 2mr2
Note:
In the case ε = 12 mv∞
2
> 0 the (hyperbolic) Newtonian trajectory does not make it to the singularity (r = 0) unless
L = 0. If it
L2
were you would get limr→0 ṙ = 0 (as the particle would make it to the singularity) and limr→0 v∞ r = limr→0 −2M Gr + m ;
2 2
2
2
but 0 ̸= limr→0 −2M Gr + m L
2 unless L = 0 (cf. figure 1 on page 199 of the book).
Note:
The non-approximated calculation can be done starting from equation (17) for ṙ = 0
q
L 2
p
F(r) r2 m + r4
E=m
r2
11
Figure 5: Effective radial potential of a massive body in the Newtonian approximation (arbitrary units).
12
Figure 7: Effective radial potential of a massive body (arbitrary units).
E2 L2
2M G 1 2M G
−1=− + 2 −
m2 r m r2 r3
The RHS behaves as an effective radial potential energy Vef f (r). We can find the roots (positions of the maximum and minimum)
and the minimum conditions under which there will be a minimum by deriving the LHS and setting the derivative of Vef f (r) to
zero. This is equivalent to setting the derivative of the energy to zero ( dE
dr = 0):
d E2 L2
2E dE 2M G 2 6M G
− 1 = · = − − =0
dr m2 m2 dr r2 m2 r 3 r4
L2 L2
r2 − 2
r+3 2 =0
M Gm m
s 2
L2 L2 L2
r= 2
± 2
−3 2
2M Gm 2M Gm m
2
2
L2
Depending on ∆ = 2MLGm2 − 3 m 2 we can have two real roots (a maximum and a minimum) if ∆ > 0 and no real roots if
2
2
L2 L2
∆ < 0. For ∆ = 2MLGm2 − 3 m 2 = 0 we have a minimum stable circular orbit for a given L: r = 2M Gm2 . Observing that
2
2
L2 L2 L2
∆ = 2MLGm2 − 3 m 2 = r 2M Gm2 − 3 m2 = 0, we get r = 6M G, independent of L.
13
8. Formation of a Black Hole
*Exercise 1
Page 291
[Suppose that there is a spherical mirror at some distance farther than the Schwarzschild radius, and that when the in-falling
radiations hit the mirror, they are reflected back outward.
1. In figure 23, draw the trajectory of the shell of radiations.
2. Discuss in which region people are doomed.]
Hint: refer to figure 23 on page 289. [The trajectory of the rays reflected off the mirror is slanted at +45° and make it to scry+;
and that’s all, no people doomed.]
■
where:
∂f
dX = dX ′ + dX ′m
∂X ′m
∂g
dY = dY ′ + dX ′m
∂X ′m
Solution:
In the appropriate approximations:
∂f ∂f ∂f ∂f ∂f
dX = dX ′ + dX ′ + dY ′ = 1+ dX ′ + dY ′ ≈ dX ′ + dY ′
∂X ′ ∂Y ′ ∂X ′ ∂Y ′ ∂Y ′
∂g ∂g ∂f ∂g ∂g
dY = dY ′ + dY ′
+ dX ′
= 1 + dY ′ + dX ′ ≈ dY ′ + dX ′
∂Y ′ ∂X ′ ∂Y ′ ∂X ′ ∂X ′
2
∂f 2 ∂f
dX 2 ≈ dX ′ + dY ′
≈ (dX ′ ) + (dX ′ dY ′ + dY ′ dX ′ )
∂Y ′ ∂Y ′
2
′ ∂g ′ 2 ∂g
2
dY ≈ dY + dX ≈ (dY ′ ) + (dX ′ dY ′ + dY ′ dX ′ )
∂X ′ ∂X ′
2 2 ∂f ∂g
dX 2 + dY 2 = (dX ′ ) + (dY ′ ) + + dX ′m dX ′n
∂Y ′ ∂X ′
So we get
∂f ∂g
hmn = +
∂Y ′ ∂X ′
where the terms of a higher order of magnitude in smallness have been omitted.
14