Lecture-2

January 21, 2025

Curved space-time can exist in absence of matter

▶ Curvature is sourced by the matter field. But we can have

non-trivial geometry even in absence of any matter.
▶ In previous lecture we have seen several such metrics which
solve the EoM without any matter stress tensor.
▶ Some of those solutions have a singularity ( a point where the
curvature diverges) - we call them Black Holes.
▶ In this lecture we shall explore the properties of such Black
Hole solutions.
Schwarzschild Black Hole
▶ The simplest of such Black Hole solutions is the Schwarzschild
Black Hole in 4 space-time dimensions.
▶ Despite of being the simplest it displays a rich and intricate
structure, some of which are generic to any black hole.
▶ The metric has the following structure in coordinates
{r , t, θ.ϕ}

dr 2
 
2 2M
dsBH = − 1− dt 2 + r 2 (dθ2 + sin2 θ dϕ2 )
1 − 2M r
r
where M is some arbitrary constant
(1)

▶ Note that the metric has singular behaviour at r = 2M and at

r =0
grr diverges at r = 2M and gtt diverges at r = 0.
Probing the space-time by light rays - path of photons
First a few physics facts, if you do not know it already please
accept them as facts at the moment.
▶ Information can travel fastest via light rays.
▶ This limits the region of space-time that could be accessed or
influenced by any observer at any arbitrary space-time point.
▶ It also says that studying the properties of light rays could be
one effective way to gain physical insight about a particular
space-time.
▶ Light rays travel in null curves - trajectories where the length
element vanishes.
▶ That is, if ds 2 is the length element between two
infinitesimally separated points {x a } and {x a + δx a } on the
trajectory of a photon, then

ds 2 |light = gab δx a δx b = 0
Radially ingoing light rays in Sch. Background

▶ Now we shall probe the Sch. Black hole space-time using light
rays.
▶ Points on such trajectories could be parametrized as
{r (λ), t(λ), θ(λ) = θ0 , ϕ(λ) = ϕ0 }, where λ is the parameter
along the trajectory and (θ0 , ϕ0 ) are angular constants.
▶ Light rays are null:
 2
dr 2 dt 2 dθ 2
+ gϕϕ dϕ




⇒ grr dλ + gtt dλ + gθθ dλ dλ =0
▶ Equations for radial light rays are simpler
 2  2
dr dt
⇒ grr + gtt =0 (2)
dλ dλ
 
dθ

dϕ
since dλ = dλ = 0 for radial paths.
Radially ingoing light rays in Sch. Background (contd)
▶ Equations for light rays:
 2  2
dr dt
⇒ grr + gtt =0
dt dλ
  s   (3)
dr gtt
=± −
dt grr

dθ

 dϕ 
since dλ = dλ = 0 for radial paths.
1 2M


▶ For Sch. Black Hole grr = and gtt = − 1 −
1− 2M
r
r
▶ Substituting and solving eq.(3) we get the following
 
∗ ∗ r − 2M
Lightrays : t±r = constant where r ≡ r +2M log
2M
▶ + sign corresponds to ingoing null rays: r decreases with time.
− sign corresponds to outgoing null rays: r increases with
time.
Radially ingoing light rays in Sch. Background (contd)

▶ Radial light rays in Sch. Black Hole background

 
∗ ∗ r − 2M
Lightrays : t±r = constant where r ≡ r +2M log
2M
▶ Here ‘ + ‘ sign corresponds to ingoing null rays: r decreases
with time.
‘ − ‘ sign corresponds to outgoing null rays: r increases with
time.
▶ Note that
▶ r ∗ (r ) is real only if r > 2M and

lim r ∗ (r ) = −∞
r →2M +

▶ Ingoing rays: t → ∞ as r → 2M +
Outgoing rays: t → −∞ as r → 2M +
Radially ingoing light rays in Sch. Background (Summary)

▶ ▶ Equation of the light rays:

r −2M
t ± r ∗ = constant where r ∗ ≡ r + 2M log


2M
▶ Ingoing rays: t → ∞ as r → 2M +
Outgoing rays: t → −∞ as r → 2M +
▶ It follows that an observer, who uses t as her time
coordinates, will never see or influence any region beyond
r = 2M even if she survives from t = −∞ to t = ∞.
Exercise-1
1. Suppose instead of radial light rays we are studying the more
general onesthat also travel along the angular directions i.e.,
dθ

dϕ
dt ̸= 0, dt ̸= 0.
How will it affect the above conclusion that the region beyond
r = 2M is inaccessible by Sch. observer?
▶ Now if it is true that no observer can ever access the region
beyond r = 2M, then we might as well assume that region
does not exist. The space-time ends at r = 2M.
▶ But all observers need not use the Schwarzschild time t as
their time coordinates.
▶ There might exist new ‘time-like’ coordinate functions such
that null rays reaches r = 2M and go beyond it within finite
time, when expressed in terms of this new coordinate.
▶ If this happens, it simply says that there exist regions of
space-time that cannot be covered with the previous choice of
Schwarszchild coordinates {r , t, θ, ϕ}
Singularity as coordinate artefacts
▶ It turns out that we could remove the singularity at r = 2M
by a different choice of coordinates.
▶ Define a new coordinate as v = t + r ∗ = t +
R dr
2M
1− r
▶ It follows
!2 !
dr dv dr
dt 2 = dv 2 + −2 (4)
1 − 2M
r 1 − 2M
r

▶ Substituting (4) in (1) we find the metric in {r , v , θ, ϕ}

coordinates
 
2 2M
dsBH = 2dv dr − 1 − dv 2 + r 2 (dθ2 + sin2 θ dϕ2 )
r
(5)

▶ In this new coordinate system {r , v , θ, ϕ} none of the metric

components or its inverse diverges at r = 2M.
Radial null rays in new coordinate system
 
r −2M
New coordinate: v = t + r ∗ where r ∗ ≡ r + 2M log 2M
; the lines of constant v in (r − v ) plane are

actually the ingoing light rays

▶ v could be finite and real at all r , in particular, for r ≤ 2M

when r ∗ diverges or become imaginary.
▶ Choose a new ‘time’ coordinate τ = v − r
Note v cannot really be chosen as ‘time’ for any observer since surfaces of constant v are not like ‘space’

where distances between any two distinct points are always positive. For example, consider two distinct

points on a constant v surface with coordinates {r1 , θ0 , ϕ0 } and {r1 + δr , θ0 , ϕ0 } with δr ̸= 0. The

distance between these points vanishes.

▶ Metric and its inverse continue to remain finite in {r , τ, θ, ϕ}

coordinates.

2 2M
dsBH = dr 2 − dτ 2 + r 2 dΩ2 + (dτ + dr )2
r
▶ Ingoing null rays: τ + r = constant
r −2M


Outgoing null rays: τ − r − 4M log 2M = constant
Radial null rays in new coordinate system (contd)

▶ Ingoing null rays: τ + r = constant

Outgoing null rays: τ − r − 4M log r −2M


2M = constant
▶ As τ runs from large negative value to large positive value ,
the ingoing light ray starting from a very large r falls towards
r = 0 smoothly passing through the surface r = 2M
▶ However, the outgoing light ray has different behaviour. Here
τ → −∞ as r → 2M +
▶ So in summary:
If an observer uses {r , τ } coordinates, then she can send
signals beyond r = 2M, but cannot receive any signal that has
originated behind r = 2M even if we allow the signal to take
infinite time to reach the observer.
Outgoing null rays as new coordinates

Exercise:
1. Define another time coordinate τout = t − r ∗ + r
2. Express the metric, using this new time coordinate.
3. Show that both the metric and its inverse are finite at r = 2M
4. Determine the equations for ingoing and outgoing light rays.
5. What does it say about the observed universe using {τout , t}
as coordinates?
Summary so far
▶ We have defined two sets of coordinate systems, where the
metric becomes manifestly regular at r = 2M.
▶ In one of them, - {r , τ, θ, ϕ} coordinates, the ingoing radial
light rays smoothly go past the r = 2M surface and hits r = 0
within finite τ . However, the starting point of outgoing null
rays cannot be behind r = 2M surface, even if we go back
infinitely in past.
▶ If we complete the exercise, we see the exact opposite
situation when the same space-time is described in the other
set of regular coordinates, namely {r , τout , θ, ϕ}
▶ Note τ − τout = 4M log r −2M


2M
So if r ≤ 2M, then the space-time region where τ is real and
finite, τout is not and vice versa.
▶ These two different coordinate systems are uncovering two
different non-overlapping regions of space-time if r ≤ 2M
Kruskal coordinates - uncovering yet new regions in black
hole space-time

▶ As of now, we have seen three regions in the Sch. black hole

space-time.
1. Region I: r > 2M, covered by all the three coordinates
introduced.
2. Region II: r ≤ 2M and covered by coordinates based on
ingoing light rays.
3. Region III: r ≤ 2M and covered by coordinates based on
outgoing light rays.
▶ A natural question arises: Can we have coordinates that
covers all these three regions?
▶ It turns out that we can do this if we take the parameters
along the ingoing and outgoing light rays as coordinates and
define r implicitly in terms of them.
Kruskal coordinates (contd)

▶ We shall perform this coordinate transformation in two steps.

▶ Step -1:

dr
u = t − r ∗ ⇒ du = dt −
1 − 2Mr
∗ dr
v = t + r ⇒ dv = dt + (6)
1 − 2Mr
 
2 2M
⇒ dsBH =− 1− du dv + r 2 dΩ2
r

▶ Here r is implicitly defined as r ∗ = r + 2M log r −2M

v −u
2M = 2
▶ The above metric is not regular at r = 2M because of the
presence of the factor 1 − 2M


r , multiplying the du dv term.
Kruskal coordinates (contd)

▶ Step -2:

u U
U = −e − 4M ⇒ dU = − du
4M
v V
V = e 4M ⇒ dV = dv
4M

2M

UV −1
 (7)
2
⇒ dsBH = 1 − dU dV + r 2 dΩ2
r 16M 2
32M 3
 
r
= e − 2M dU dV + r 2 dΩ2
r
r

r
e 2M 1 − 2M


▶ Here r is implicitly defined as 2M r = UV
Different regions of the space-time

Sequence of coordinate transformations:

1. Sch. Coordinates: {r , t, θ, ϕ}
2. Ingoing and Outgoing light rays as coordinates:
u = t − r ∗ , v = t + r ∗ , r ∗ = r + 2M log 1 − 2M


r
u v
3. Kruskal coordinates: U = −e − 4M , V = e 4M
▶ Naively r ∗ becomes imaginary when r < 2M.
   
2M 2M
If r < 2M, log 1 − = log − 1 + iπ
r r
▶ Still we could have well defined real u or real v by analytically
continuing t in complex plane.
▶ But note that both u and v cannot not be real simultaneously
if r < 2M.
Different regions of the space-time in Kruskal coordinates
Sequence of coordinate transformations:

1. Sch. Coordinates: {r , t, θ, ϕ}
 
2. Ingoing and Outgoing light rays as coordinates: u = t − r ∗ , v = t + r ∗, r ∗ = r + 2M log 1 − 2M
r

− u v
3. Kruskal coordinates: U = −e 4M , V = e 4M

1. Region-I:
(r > 2M) This is the region covered by Sch. coordinates.
In this region both u and v are real and therefore
U < 0, V > 0.
2. Region-II:
(r < 2M) and only v is real but u is not; We uncovered this
region once using coordinates defined in terms of ingoing light
rays. Here U > 0, V > 0
3. Region-III:
(r < 2M) and only u is real but v is not; Here coordinates are
defined in terms of outgoing light rays. Here U < 0, V < 0
A new region in Kruskal coordinates
Sch. Coordinates: {r , t, θ, ϕ}
 
Ingoing and Outgoing light rays as coordinates: u = t − r ∗ , v = t + r ∗, r ∗ = r + 2M log 1 − 2M
r
− u v
Kruskal coordinates: U = −e 4M , V = e 4M

▶ In terms of t and r coordinates, the key feature of the

region-II and region-III is that here t is extended to complex
values in a way so that either u or v is real.
▶ If we do not extend t to complex values but still keep r < 2M,
we uncover a new region where both u and v are complex in a
way such that U and V are real and U > 0, V < 0.
▶ This is the fourth quadrant in {U, V } coordinates, not
covered by region I, II or III.
▶ So once we use the {U, V } coordinates, we uncovered another
new region of BH space-time - the region IV.
Curvature singularity vs. coordinate artefacts
− r
 
2
Metric in Kruskal coordinates: dsBH = 32M 3 e 2M dU dV + r 2 dΩ2
r

▶ We have seen that divergence of a metric component (or its

inverse) does not necessarily imply a singularity in the
space-time geometry since a component of a tensor might
diverge if we choose wrong coordinates.
▶ On the other hand, if some curvature invariants diverges at
some point in space-time, then it is a genuine singularity since
the value of a scalar or invariant cannot be changed under any
coordinate transformation.
▶ It turns out if we evaluate the scalar R abcd Rabcd on the
Schwarzschild metric and take the limit r → 0 limit, it diverges
and no coordinate transformation can remove this singularity.
▶ Therefore, in every coordinate systems the point r = 0 is
always singular and it is an irremovable singularity.
Can there be a fifth region

▶ So far the main technique we used to uncover new region is

1. Extend an existing coordinate to complex values
2. But the extension is done in a way so that the newly defined
coordinates are real.
▶ But the question remains:
Can we now extend the U and V coordinates to complex
values and generate a new region by further redefinition?
▶ The answer turns out to be ‘no’.
▶ The space-time in {U, V } coordinates with both coordinates
ranging from −∞ to +∞ is the maximal possible extension of
the Sch. black hole space-time.
Maximal extension of space-time from physics perspective

▶ A physicist’s description of a singularity (and therefore a

breakdown of the existing theory) :
For an observer, approaching singularity, some measurement
should lead to catastrophic results within finite time in her
clock.
▶ There is no point in trying to describe the ’catastrophe’
(which could be a prediction for infinite force, or imaginary
answers for real measurements), since we believe that it really
does not exist in nature, rather it indicates a breakdown of the
theory we are using to describe the nature.
▶ What is important is the fact that the catastrophe must
happen in ‘finite time’.
Which ‘time’ is measured by an observer
▶ We can express space-time in many different coordinate
systems, using different ‘time’ coordinates.
▶ But when an observer measures ‘time’ in her clock, we need
to know which of these infinite choices is being measured.
▶ An ’observer’s choice of coordinate system’ is the one where
1. she is always at the same position in space and moving only in
time
2. she sees the space-time in her immediate neighbourhood to be
just a flat space.
▶ An observer follows a curve or trajectory in the space-time -
called ‘worldline’. It follows that the length element along this
curve could be expressed as

ds 2 = −dτ 2 where τ is the time measured by her

▶ The τ is a parameter along the curve - called ‘proper time’.

The maximal extension in terms of observer’s worldline
▶ A ‘finite time’ in observer’s clock implies the finite proper
time along observer’s world line.
▶ A space-time point is non-singular if every possible observer’s
worldline go past the point smoothly (i.e., without any
catastrophe) in its finite proper-time.
▶ This simply means that around that point there exist
coordinate systems (the ones used by the observers) in which
the point is regular.
▶ However, if the singularity is such that it cannot be removed
by any coordinate redefinition (e.g. the curvature singularity
at r = 0 is Sch. space-time) then it will remain singular
(predicting catastrophic measurement) in every observer’s
worldline.
▶ The ‘maximally extended space-time’ are the ones where all
observer’s worldlines could be extended from −∞ to ∞ unless
it encounters an irremovable singularity.
The geodesics - the free falling observer
▶ The definition of maximal extension or space-time singularity
uses any observer’s worldline.
▶ However, if we are interested in the structure of the
space-time only, we should use those worldlines which are
acted upon only by gravity and nothing else.
▶ Observers moving under the influence of gravitational force
only are actually ‘free particles’ moving in a curved geometry.
▶ In Flat space we know such ‘free’ objects move in a straight
line - the path with minimum length.
▶ Extending this intuition to curved space-time, free particle
trajectory is the extrema of the path-length.
Z final point
Action for free particles ≡ Afree ∝ ds
initial point

▶ Extrema of the path-length is called geodesic’.

Maximal extension and geodesic completeness

▶ We interpret maximal extension of a space-time with geodesic

completeness.
▶ A space-time is ‘maximally extended’ if every geodesic could
be extended to arbitrary value of its proper time (affine
parameter) unless it terminates on a irremovable singularity.
▶ It turns out that the Sch. Black hole with all the four regions
as revealed in Kruskal coordinates {U, V , θ, ϕ} has this
property of maximal extension i.e., any geodesic here
1. Either ends in the curvature singularity at r = 0 in finite
proper time
or
2. it could be extended to all values of the proper time from −∞
to ∞.
▶ We shall come back to the description of Sch. space-time in
Kruskal coordinates.
▶ But before that we shall develop some technical skill to deal
with geodesics since they are the key instruments to be used
to probe the structure of the space-time.
Trajectory of a particle moving only under gravitational
force

▶ Trajectory of any point-like object can again be derived by

extremizing some ‘Action’ that is a functional of the path or
trajectory. .
▶ Free particle trajectory is the extrema of the path-length.
Z final point
Action for free particles ≡ Afree ∝ ds
initial pont
Geodesic equation in arbitrary curved space-time
▶ Coordinates of the points on a curve can have a parametric
representation of the form {x µ (λ)} where λ is a parameter
that monotonically varies along the curve.
▶ Suppose {x µ (λ)} and {x µ (λ + dλ)} are two infinitesimally
separated points on the curve.

Afree [{x µ (λ)}] ∝ Total path length

s
Z λfinal  µ ν 
dx dx (8)
= dλ gµν (x)
λinitial dλ dλ

▶ Extremizing Afree w.r.t the path {x µ (λ)} we get the geodesic

equation
 2 µ  α β  µ
d x µ dx dx dx
2
+ Γαβ ∝ (9)
dλ dλ dλ dλ
Few simplifying properties of the geodesic equation
It turns out that
▶ Afree is invariant under reparametrization of the curve i.e.,
λ → σ = σ(λ)
Equivalent to coordinate redefinition invariance along the one-dimensional manifold - the trajectory

▶ It is always possible to choose the parameter σ in a way (affine

parametrization) so that the geodesic equation simplifies to
 2 µ  α β
d x µ dx dx
2
+ Γαβ =0 (10)
dσ dσ dσ
▶ From equation (10) it follows that along the trajectory
µ
dx ν

L(σ) ≡ gµν (x) dx
dσ dσ is conserved i.e.,
  µ ν 
d dx dx
gµν (x) =0
dσ dσ dσ
Time-like, space-like or null geodesic
 
dL(σ) dx µ dx ν



▶ dσ = 0, ⇒ gµν (x) dσ dσ = constant
▶ Note we can change the magnitude of the constant by a
constant rescaling of σ, but we cannot alter its sign.
▶ So there could be three distinct classes of geodesics.
1. Time-like: The constant is chosen to be (-1).
2. Space-like: The constant is chosen to be (+1).
3. Null: The constant is chosen to be (0).
▶ It turns out that no particle’s trajectory could ever be
space-like.
We shall discuss this point further at some later point

▶ So if we want to explore the black hole space-time by probe

particle’s trajectory, we have to study null or time-like
geodesic.
Exercise

1. Derive the geodesic equation (9) by extremizing the total

path-length as given in equation (8) with respect to the
{x µ (λ)}. Fix the proportionality factor.
2. So that we could always reparametrize the curve so that the
geodesic equation simplifies to equation (10).
3. Using equation (10) show that
  µ   ν 
dx dx
gµν (x) = c = constant ∀ σ (11)
dσ dσ