In this lecture we will continue our discussion of general relativity.

We first introduce a
convention that allows us to drop the many factors of G and c that appear in formulae, then
talk in more detail about tensor manipulations. We then get into the business of deriving
important quantities in the Schwarzschild spacetime.

Geometrized units

You may have noticed that Newton’s constant G and the speed of light c appear a lot!
Partially because of laziness, and partially because it provides insight about fundamental
quantities, the convention is to use units in which G = c = 1. The disadvantage of this is
that it makes unit checking tougher. However, with a few conversions in the bag, it’s not
too bad.
In geometrized units, the mass M is used as the fundamental quantity. To convert to
a length, we use GM/c2 . To get a time, we use GM/c3 . Note, however, that there is no
combination of G and c that will allow you to convert a mass to a mass squared (or to
something dimensionless), a radius to a radius squared, or whatever. With these units, we
can for example write the Schwarzschild metric in its more common form:
ds2 = −(1 − 2M/r)dt2 + dr2 /(1 − 2M/r) + r 2 (dθ2 + sin2 θdφ2 ) . (1)

Tensor manipulations
The metric tensor is what allows you to raise and lower indices. That is, for example,
vα = gαβ v β , where again we use the summation convention. Similarly, v α = g αβ vβ , where
g αβ is the matrix inverse of gαβ : g αβ gβγ = δγα , where δ is the Kronecker delta (1 if α = γ,
0 otherwise). For a diagonal metric (such as Minkowski or Schwarzschild), it is particularly
simple: g tt = 1/gtt , g xx = 1/gxx , and so on (or g rr = 1/grr , etc.), and all off-diagonal
components are zero.

Specific energy and specific angular momentum

All that is great, but how does it help us? The most straightforward application emerges
from an examination of the components of the four-velocity, uα or uα . For this purpose, we
will define the concept of a test particle. This is useful in thinking about the effect of
spacetime on the motion of objects. A test particle is something that reacts to fields or
spacetime or whatever, but does not affect them in turn. In practice this is an excellent
approximation in GR whenever the objects of interest have much less mass than the mass
of the system.
The four-velocity of a particle with mass can be written as uα = dxα /dτ , where τ is
the proper time (specifically, dτ 2 = −ds2 in geometrized units). As we showed in an earlier
lecture, for any spacetime at all, and even in the presence of arbitrary forces, for a particle
with mass, the squared four-velocity is u2 = −1 (or −c2 in normal units), and for a photon
or other massless particle u2 = 0. By the way, one of the components of the four-velocity
may look odd: ut = dt/dτ . What the heck does that mean? It means the rate of change of
coordinate time per unit change in proper time, which you recall is the time as measured by
an observer riding along with the particle.
But what about the four-velocity with lowered indices? In Cartesian coordinates this
may not seem to be meaningful, because the metric is diagonal and has +1 for the space
components. For example,

ux = gαx uα = gtx ut + gxx ux + gyx uy + gzx uz = 0 + ux + 0 + 0 = ux . (2)

The same is true for the y and z components. It is true that ut = −ut , but that’s no big
deal.
In spherical polar coordinates, though, this is not the case. Consider the Minkowski line
element in such coordinates:

ds2 = −dt2 + dr2 + r2 (dθ2 + sin2 θdφ2 ) . (3)

The metric is still diagonal, but the components are not unity. We can understand this
by noting that although in our geometrized system time and distance have the same units,
angles are dimensionless and thus the metric coefficients can’t be unity. Let’s see how this
plays out if we consider orbits in the equatorial plane θ = π/2 (note that since the system
is spherically symmetric, we can always do this). Then, for example,

uφ = gαφ uα = gtφ ut + grφ ur + gθφ uθ + gφφ uφ = 0 + 0 + 0 + r 2 uφ . (4)

Does this remind us of anything in particular? Remember that uφ = dφ/dτ is the

angular velocity. Then r 2 uφ = r(ruφ ) = rv, where v is the linear speed in the azimuthal
direction. Thus, uφ is the specific angular momentum, where in this context and others
“specific” means “per unit mass”. Similarly, as it turns out, e = −ut is the specific energy,
where the minus sign comes from our choice of metric signature. Both of these identifications
also hold true in Schwarzschild coordinates.
For example, consider a particle in circular motion, although not necessarily Keplerian.
Then there is no r or θ motion and u2 = −1 gives us ut ut + uφ uφ = −1. We can put this
into a more convenient form by writing (g tα uα )ut + (g φα uα )uφ = −1. The Schwarzschild
spacetime is diagonal, so this becomes simply g tt (ut )2 + g φφ (uφ )2 = −1. Consulting the line
element, we find g tt = 1/(1 − 2M/r) and g φφ = 1/r2 , so the specific energy is
q
e = (1 − 2M/r)(1 + u2φ /r2 ) . (5)
 p
For example, for a slowly rotating star with uφ ≈ 0, the energy is e = 1 − 2M/r. This
means that a particle of mass m originally at infinity will release a total energy m(1 − e) if it
finally comes to rest on the star’s surface. Let’s check if this makes sense in the Newtonian
limit r À M . Then e ≈ 1 − M/r, so the energy released is mM/r, which is the Newtonian
form when we put back in the units: GmM/r.
As another example, let’s go back to the Minkowski metric and consider a particle that
is moving purely in the radial direction. What is its specific energy? Suppose that the radial
speed is ur . Then since uθ = uφ = 0, u2 = −1 gives us
ut ut + u r ur = −1
g tt u2t + grr (ur )2 = −1
(6)
−u2t + (ur )2 = −1
1/2
e = −ut = [1 + (ur )2 ] .
Does this make sense? For ur ¿ 1 (i.e., a speed much less than the speed of light), an
expansion gives
1
e ≈ 1 + (ur )2 . (7)
2
Recalling that e is in units of mc2 , this is correct, as it gives the rest mass (the 1 part) plus
the kinetic energy. What about very close to the speed of light?
Here’s where we have to be careful. In the “lab” frame, being close to the speed of light
means that v = dr/dt is nearly 1. However, ur = dr/dτ , not dr/dt. We therefore have to go
the long way around:
dr dr/dτ ur ur ur
v= = = t = tt = . (8)
dt dt/dτ u g ut −ut
Therefore, ur = −vut . We therefore can rewrite the equation for the specific energy as
1/2
−ut = (1 + v 2 u2t )
u2t = 1 + v 2 u2t
(9)
(1 − v 2 )u2t =1
e = −ut = (1 − v 2 )−1/2 = γ .
This is also correct. Note how we had to be precise with our manipulations!

Orthonormal Tetrads
Let’s now return to a subject we’ve mentioned a few times: shifting to a locally Minkowski
frame. In general, you want to take a metric that looks like gαβ and shift into a frame such
that locally the metric is ηαβ = (−1, 1, 1, 1). It is conventional to represent the new coordi-
nates with hats (e.g., t̂, r̂, θ̂, φ̂), so that
ds2 = −dt̂2 + dr̂2 + dθ̂2 + dφ̂2 . (10)
The transformation from the local to the global coordinates is done with the transformation
matrices eα̂β and eαβ̂ . For example, uα = eαβ̂ uβ̂ . The components of the transformation
matrices come from the transformation of the metric tensor:

ηα̂β̂ = eµα̂ eνβ̂ gµν . (11)

This is especially easy for the Schwarzschild metric, because the metric is diagonal. Then,
for example, et t̂ = (1 − 2M/r)−1/2 and eφφ̂ = r−1 . Note that even after having transformed
into a reference frame in which the spacetime is as Minkowski as possible (i.e., first but not
second derivatives vanish), there is still freedom to choose the coordinates. Also, remember
that there is always freedom to have Lorentz boosts; that is, having found a frame in which
the spacetime looks flat, another frame moving at a constant velocity relative to the first
also sees flat spacetime. This means that your choice of frame (“orthonormal tetrad”) is
based to some extent on convenience. Around a spherical star, a good frame is often the
static frame, unmoving with respect to infinity. For a visualization of some of the effects on
space and time near a gravitating object, see Figure 1.

Now let’s see some examples of this in action. Suppose a particle moves along a circular
arc with a linear velocity in the φ direction v φ̂ as seen by a static observer at Schwarzschild
φ̂ φ̂ t̂
radius hr. What i is the angular velocity as seen at infinity? v = dφ̂/dt̂ = u /u . But this is
eφ̂φ uφ / et̂ t ut . Since Ω = dφ/dt = uφ /ut , then
à !
v φ̂
Ω = (et̂ t /eφ̂φ )v φ̂ = (1 − 2M/r)1/2 . (12)
r

This makes sense; it’s just the same as one would calculate in the Newtonian limit, except
that the frequency is less because of redshifting.
With this under our belts, let’s do a problem that points out some of the strange things
about black holes. Consider a particle of nonzero rest mass that is released from rest a
long way from an uncharged, nonspinning black hole (which thus can be described using the
Schwarzschild spacetime). At a distance r from the origin, as measured using Schwarzschild
coordinates, we want to know (a) what is the proper radial speed, (b) what is the radial
speed as measured at infinity, and (c) what is the radial speed as measured by a local static
observer?
To start off, we note that because the particle starts at rest at a large distance from the
hole, the total energy of the particle is just mc2 , hence −ut = 1. This is a conserved quantity.
We also note that because the particle is just falling radially, it means that u θ = uφ = 0,
and uθ = uφ = 0 as well. We can then work from conservation of the squared four-velocity
Fig. 1.— Effects on distance and time near a gravitating object. From
http://abyss.uoregon.edu/∼js/images/spacetime dilates.gif
to find
u2 = −1
ut ut + ur ur = −1
g tt u2t + grr (ur )2 = −1
−(1 − 2M/r)−1 (1) + (1 − 2M/r)−1 (ur )2 = −1 (13)
(1 − 2M/r)−1 (ur )2 = −1 + (1 − 2M/r)−1
(ur )2 = −(1 − 2M/r) + 1
ur = −(2M/r)1/2 .
The negative sign is because the particle is falling inward. If we were to put the units back
in, this would be −(2GM/r)1/2 .
To interpret this, we need to recall the meaning of proper speed, etc.
(a) Remember that “proper” means “as measured by a comoving observer”. Therefore, the
proper radial speed is dr/dτ . This is just ur , so the answer to the first part of our question
is that the proper radial speed is ur = −(2M/r)1/2 .
(b) What is the radial speed as measured at infinity? Recall from the definition of Schwarzschild
coordinates that the coordinate time t is the time as measured at infinity. Therefore, the
radial speed as measured at infinity is dr/dt. To relate this to the proper radial speed we
need to do the following:
dr/dt = (dr/dτ )/(dt/dτ )
= ur /ut
= ur /(g tt ut )
(14)
= ur /[−(1 − 2M/r)−1 (−1)]
= (1 − 2M/r)ur
= −(1 − 2M/r)(2M/r)1/2 .

(c) What is the radial speed as measured by a local static observer at r? For this, we need to
transform between the global four-velocity components (e.g., ur , ut ) to the local four-velocity
components in an orthonormal tetrad (e.g., ur̂ , ut̂ ). Remember the rule: the transformation
matrices take your metric and turn it into a Minkowski metric. For example, consider the
tt part of the Schwarzschild line element: gtt = −(1 − 2M/r). To turn this into gt̂t̂ = −1,
we need to multiply this by et t̂ et t̂ , which tells us that et t̂ = (1 − 2M/r)−1/2 . Similarly, to go
back, it must be that et̂ t = (1 − 2M/r)1/2 . Armed with this information, we then find
dr̂/dt̂ = ur̂ /ut̂
= er̂r ur /[et̂ t ut ]
= [er̂r /et̂ t ]ur /ut
(15)
= [(1 − 2M/r)−1/2 /(1 − 2M/r)1/2 ]dr/dt
= (1 − 2M/r)−1 [−(1 − 2M/r)(2M/r)1/2 ]
= −(2M/r)1/2 .
 Intuition Builder

In the last problem we did, there are several apparent anomalies. Indeed,
these anomalies are the source of much confusion and much spewing by crack-
pots. They are: (a) the proper radial speed seems to become greater than
1, i.e., greater than the speed of light, when r < 2M , (b) the radial speed
measured at infinity seems to go to zero as r → 2M , and (c) the radial speed
measured by a local static observer appears to become greater than the speed
of light when r < 2M . What is going on in each case?