Professional Documents
Culture Documents
Gravitational Physics
Yi-Zen Chu1,2 and Glenn D. Starkman
1
CERCA, Physics Department, Case Western Reserve University, Cleveland, OH 44106-7079, USA
Electromagnetic and gravitational radiation do not propagate solely on the null cone in a generic
curved spacetime. They develop tails, traveling at all speeds equal to and less than unity. If
sizeable, this off-the-null-cone effect could mean objects at cosmological distances, such as supernovae, appear dimmer than they really are. Their light curves may be distorted relative to their
flat spacetime counterparts. These in turn could affect how we infer the properties and evolution
of the universe or the objects it contains. Within the gravitational context, the tail effect induces
a self-force that causes a compact object orbiting a massive black hole to deviate from an otherwise geodesic path. This needs to be taken into account when modeling the gravitational waves
expected from such sources. Motivated by these considerations, we develop perturbation theory for
solving the massless scalar, photon and graviton retarded Greens functions in perturbed spacetimes
g = g + h , assuming these Greens functions are known in the background spacetime g .
In particular, we elaborate on the theory in perturbed Minkowski spacetime in significant detail;
and apply our techniques to compute the retarded Greens functions in the weak field limit of the
Kerr spacetime to first order in the black holes mass M and angular momentum S. Our methods
build on and generalizes work appearing in the literature on this topic to date, and lays the foundation for a thorough, first principles based, investigation of how light propagates over cosmological
distances, within a spatially flat inhomogeneous Friedmann-Lematre-Robertson-Walker (FLRW)
universe. This perturbative scheme applied to the graviton Greens function, when pushed to higher
orders, may provide approximate analytic (or semi-analytic) results for the self-force problem in the
weak field limits of the Schwarzschild and Kerr black hole geometries.
I.
2
text, semi-infinite) manifold. Let there be some field
producing source J present in the volume V , and nontrivial initial conditions
for the fields at t0 , for example,
00
0
00
00
[x = t ] and [x = t0 ]. Denote the scalar, photon, and graviton Greens functions as Gx,x0 , G 0 and
G0 0 respectively. Then the scalar field evaluated
at some point x lying on the t surface can be written
as3
Z
1
(1)
x =
dd x0 |g 0 | 2 Gx,x0 Jx0
V
Z
0
0
1
+
dd1 ~x0 |g 0 | 2 Gx,x0 0 x0 0 Gx,x0 x0
00
0
J[x]
x =t
(3)
x =t0
where P 0 0 (1/2)( + g g0 0 ).
From equations (1), (2) and (3), we see that the
physical solution of a linear field theory can be expressed as the sum of two integrals of the Greens function (and its gradient), one weighted by the sources
present in the system at hand and the other weighted
by the initial conditions of the fields themselves. In
particular, the d-dimensional volume integrals (with
respect to x0 ) of the Greens functions, weighted by the
field-producing Js, reaffirms the interpretation that
the Greens function is the field of a spacetime-point
unit charge because these volume integrals corresponds to calculating the field at x by superposing the
3
other is proportional to [t t0 ][x,x0 ], and describes
propagation on the interior of the future light cone of
x0 .
Gx,x0 =
[t t0 ]
(Ux,x0 [x,x0 ] + Vx,x0 [x,x0 ]) ,
4
(4)
0
[t t ]
(U 0 [x,x0 ] + V 0 [x,x0 ]) , (5)
4
0
[t t ]
U0 0 [x,x0 ]
=
4
+ V0 0 [x,x0 ] .
(6)
G 0 =
G0 0
This is barring special properties, such as the conformal symmetry enjoyed by the Maxwell action, which says that light is
blind to conformal factors of the metric: a2 g and g are
equivalent in its eyes. We will shortly elaborate on this point.
diag[1, 1, 1, 1],
(7)
where a tells us the relative size of the universe at various times along its evolution. Most of our inference
of the properties of the universe come from examining
light emanating from objects at cosmological or astrophysical distances, and furthermore our interpretation
of electromagnetic signals are based on the assumption that they travel on null geodesics. This statement
is precisely true when the metric is (7) because the
Maxwell action that governs the dynamics of photons
in vacuum, is insensitive to the conformal factor a2 .
Specifically, in 4 dimensional spacetime, SMaxwell []
and SMaxwell [a2 ] are exactly the same object; the conformal factor a2 drops out.
(d=4)
(d=4)
(8)
This means electromagnetic radiation in 4 dimensional spatially flat FLRW universes behaves no differently from how it does in 4 dimensional Minkowski
spacetimes. In particular, it travels only along null
geodesics. However, cosmological and astrophysical
observations have become so sensitive that it is no
longer sufficient to model our universe as the exactly
smooth and homogeneous spacetime in (7). Rather,
one needs to account for the metric perturbations,
g = a2 ( + h ) .
(9)
Because the a2 drops out of the Maxwell action, we recognize that a first principles theoretical investigation
of the propagation of light over cosmological distances
is equivalent to the same investigation in perturbed
Minkowski spacetime. Moreover, since the geometry is
now curved (albeit weakly so), light traveling over cosmological length scales should therefore develop tails.
As already mentioned in the abstract, if a significant
portion of light emitted from a supernova at cosmological distances leaks off the light cone, then the observer
4
on Earth may mistakenly infer that it is dimmer than
it actually is, as some of the light has not yet arrived.
This leakage may also modify the light curves of these
objects at cosmological distances. To our knowledge,
the size of the electromagnetic tail effect in cosmology
has not been examined before. Our development of
perturbation theory for the photon Greens function
(and confirmation of DeWitt and DeWitts first order
results [1]) in perturbed Minkowski, is therefore the
first step to a thorough, first principles based, understanding of the properties of light in the cosmological
context. This may in turn affect how we interpret cosmological and astronomical observations.7
JWKB Now, the JWKB approximation where
one assumes that the wavelength of the massless fields
are extremely small relative to the characteristic length
scales of the spacetime geometry (which, in term, usually amounts to neglecting all geometric terms relative
to the in the wave equation), is often used to justify
that null cone propagation is the dominant channel
of travel for massless fields in generic curved spacetimes. (See for example Misner, Thorne and Wheeler
[6].) Here, we caution that, even in cases where the
JWKB approximation yields exact results, it does not
imply that light travels solely on the light cone. Such a
counterexample is that of odd dimensional Minkowski
spacetimes, where the momentum vector k satisfies
the exact dispersion relation k k = 0, but the
Greens functions of massless fields develop power law
tails: for odd d, V [x, x0 ] ((tt0 )2 (~x ~x0 )2 )(d2)/2 .
(See Soodak and Tiersten [7] for a pedagogical discussion on tails of Greens functions in Minkowski spacetimes.) This tells us that, even for 4 dimensional flat
spacetime, the rigorous way to prove that light travels
on the null cone is by computing the photon Greens
function, since it is the Greens function (via the Kirchhoff representations in (1), (2) and (3)) that determines how physical signals propagate away from their
sources.
In the next section, we will review the general theory
of Greens functions and some geometrical constructs
related to them. Perturbation theory for Greens functions will then be delineated in the subsequent two sections; following that, we will apply the technology to
calculate the Greens functions in the Kerr black hole
spacetime, up to first order in its mass and angular momentum. We will conclude with thoughts on possible
future investigations.
II.
GENERAL THEORY
This section will summarize the key technical features of Greens functions we will need to understand
in the development of perturbation theory in the following two sections. We refer the reader to Poissons
review [4] for an in-depth discussion. We first examine the world function x,x0 , van Vleck determinant
x,x0 and the parallel propagator g 0 , which are geometrical objects needed for the formal construction
of the Greens functions themselves. We will record
the equations obeyed by the Greens functions, and
then describe the coefficients of [x,x0 ] and [x,x0 ]
in (4), (5) and (6). Finally we will compute the x,x0 ,
x,x0 and g 0 in Minkowski and perturbed Minkowski
spacetimes.
World Function The world function x,x0 defined in the introduction is half the square of the
geodesic distance between x and x0 . Assuming there
is a unique geodesic whose worldline has coordinates
{ []| [0, 1]; [0] = x0 , [1] = x }, it has the
integral representation
Z
1 1
g [] d
(10)
x,x0 =
2 0
with d/d.
van Vleck Determinant Closely related to x,x0
is the van Vleck determinant x,x0
x,x0 =
det[ 0 x,x0 ]
.
|gg 0 |1/2
(11)
Parallel Propagator
The parallel propagator
g 0 is formed by contracting two sets of orthonormal
basis tangent vector fields {A |A, = 0, 1, 2, 3, . . . , d
1}, one based at x and the other at x0 . (The A-index is
raised and lowered with AB and the -index is raised
and lowered with the metric.)
g 0 [x, x0 ] AB A [x] 0 B [x0 ],
(12)
g 0 [x, x] = g [x].
(13)
(14)
5
massless scalar,
x0 Gx,x0 = x Gx,x0 =
d [x x0 ]
|gg 0 |1/4
(15)
0 0
x0 G 0 R 0 G0 = x G 0 R G 0
= g 0
d [x x0 ]
.
|gg 0 |1/4
(16)
(17)
We will later note that our perturbative result satisfies (17). Proceeding to the de Donder gauge graviton
( = /2, with g ),
1
{ } g g ( R + 2) + 2R
2
+ R { }
g R g R G0 0
= ;0 0
d [x x0 ]
|gg 0 |1/4
(18)
where we have included a non-zero cosmological constant . The ;0 0 is built out of the parallel propagator g 0 [x, x0 ],
;0 0
1
(g0 g0 + g0 g0 ) .
2
(19)
Greens functions are bitensors. Coordinate transformations at x can be carried out independently from
x0 (and vice versa). Derivatives with respect to x are
independent of that with respect to x0 , so for instance,
G 0 = G 0 G 0 .
Hadamard form We now have sufficient vocabulary to describe the coefficients of [x,x0 ] and [x,x0 ]
in the Greens functions in equations (4), (5) and (6).
Assuming x and x0 lie in a region of spacetime where
( g
Our Christoffel symbol is
{ } g );
= (1/2)g
( ); the
Riemann tensor is R =
+
there is a unique geodesic joining them, in 4 dimensional spacetimes, the null cone pieces are built out of
the van Vleck determinant and the parallel propagators
p
Ux,x0 = x,x0
(20)
p
U 0 = x,x0 g 0
(21)
p
U0 0 = x,x0 P0 0
(22)
where
P0 0
1
(g0 g 0 + g 0 g0 g g0 0 ) .
2
(23)
(24)
Because x,x0 is proportional to the tangent vector at x (it points in the direction of greatest rate
of change in geodesic distance), x x,x0 describes the rate of change of the cross sectional area of
the congruence of geodesics (the expansion) through
the neighborhood of x, which via (24) is related to the
gradient of x,x0 along the geodesics. This expansion
scalar is related to the evolution of the angular diameter distance, which then in turn is related to the
luminosity distance relation. (See, for example. Visser
[8] and Flanagan et al. [9].) Along similar lines, initially parallel null rays from an extended source become deflected due to gravitational effects (weak lensing). Since the parallel propagator describes the parallel transport of an orthonormal reference frame along
these trajectories, namely
0
(see (12)),
(25)
they ought to contain physical content regarding polarization, rotation and shear of null bundles of photons.
To sum, the light cone part of the massless scalar and
6
photon Greens function should provide an alternate
means, from the standard ones in use by cosmologists
today, of getting at the physics of null light traveling
through the universe. This warrants more study.
Before moving on to develop our perturbation theory, let us take a few moments to calculate the world
function, van Vleck determinant and parallel propagator up to first order in h in perturbed Minkowski
spacetime. This will allow us to construct the null
cone piece of the scalar, photon and graviton Greens
function, and in turn, serve as a consistency check on
our first Born approximation results below.9 In fact,
this was how Kovacs and Thorne [2] constructed the
null cone piece of their Greens functions, by calculating separately the van Vleck determinant and Synges
world function. But we shall argue that this is not necessary. The Born series scheme we have devised gives
us a single coherent framework where all three geometric objects appearing in the null cone piece of the
Greens function are byproducts of the computation.
Specifically, the van Vleck determinant and the world
function can be read off the massless scalar Greens
function Gx,x0 , and the parallel propagator can be read
off the Lorenz gauge photon Greens function G 0 .
, and g 0 in Minkowski The geodesic equation in Minkowski spacetime is
g 0 = .
(31)
(32)
1
{ } .
P =
2
(35)
(33)
(34)
with
(36)
d
= 0,
d2
(26)
(27)
x,x0 =
1
,
2
(28)
(29)
which is not to be confused with the van Vleck determinant (we will always place the spacetime coordinates
as subscripts for the latter). Since
x,x0 = ,
0
x,x0 =
0
x,x0 =
(0)
x,x0
x,x0 + Ib
,
(37)
where
(30)
1
(0)
Ib
h []d.
(38)
0
(0)
7
(Tr denotes trace, and A1 is the inverse of A.) We
then deduce the square root of the van Vleck determinant is
p
1
1
x,x0 1 h h0 + Ib(0)
(40)
4
4
1 b(0)
0 b(0)
I 0 I .
+
2
(0)
Here, h0 h0 0 and Ib(0) Ib .
In [8], Visser developed perturbation theory for solving the van Vleck determinant.
p In particular, he
showed that the O[h] accurate x,x0 is given by (his
equation 61)
Z
p
1
(41)
x,x0 1 +
d(1 )(R|1) [],
2
0
4
4
4
0
This relation can be reached by recognizing
00 00 h = d2 h/2 , followed by integrating-byparts the d2 /d2 .
Equation (14) says the parallel propagator is parallel propagated along . If we write g 0 = + h 0
and keep only the O[h] terms in the Christoffel symbol
in h 0 = h 0 h 0 , (14) is then approximately equivalent to
d
h 0 [[], x0 ]
(44)
d
1
+ 00 h00 00 []
00 h00 00 []
[]
=
00 h00 00 []
2
dh00 00 []
= 00 h00 00 []
d
(45)
1
(h + h0 0 ) +
2
2
(46)
As can be checked explicitly,
= ( + 0 )h 00 00 .
00 h 00 00 []
(47)
(0)
We may thus re-write (46) in terms of Ib in (38),
g 0 +
1
(0)
(h + h0 0 ) + ([ + [0 )Ib] .
2
(48)
III.
PERTURBATION THEORY
8
Similarly, by replacing one of the fields in S [
g ] with
x00 ,x0 and the other with Gx,x00 , one obtains
G
x00 ,x0 ]
2S [
g ; Gx,x00 , G
(52)
Z
00
1
g = g h + . . .
1
1
1
|g| 2 = |
g | 2 1 + h + . . . ; h g h
2
(54)
(55)
to as high an order in h as desired. (We are now raising and lowering all indices with the background metric g .) To obtain the first Born approximation, the
O[h]-accurate result for Gx,x0 , one replaces the Gx,x00
x,x00 and
occurring within the integrals in (53) with G
00 1/2
00 00
only need to expand the |g |
and g
to first order. The result is
Z
1
x,x0 + dd x00 |
Gx,x0 G
g 00 | 2
(56)
1 00 00 00
00 00
00 Gx,x00
h g
h
00 Gx00 ,x0
2
00 00
with h00 g h00 00 . In perturbed Minkowski spacetime, we set g = , employ Cartesian coordinates,
and then use the spacetime translation symmetry re x,x0 for any d, namely
flected by the Greens function G
x,x0 = 0 G
x,x0 ,
G
(57)
x,x0
Gx,x0 G
(53)
Z
00 00
1
x,x00 00 Gx00 ,x0
= dd x00 |g 00 | 2 g 00 G
Z
00 00
1
00
x00 ,x0 .
dd x00 |
g 00 | 2 g Gx,x00 00 G
First Born Approximation Perturbation theory may
now be carried out by iterating (53) as many times as
one wishes (followed by dropping the remainder integral terms containing Gx,x0 ), and expanding
9
function is negative of ours.
Photon
Next, we turn to the photon. The
Maxwell action in terms of electric and magnetic fields
F is
Z
00 00
00 00
1
1
SMaxwell =
dd x00 |g 00 | 2 g g F00 00 F00 00 .
4
(59)
We have already noted in the introduction, that this
action SMaxwell enjoys a conformal symmetry in 4 dimensions, namely, it evaluates to the same object in
both the metric g and the metric a2 g ; the conformal factor a2 drops out. Whenever there is such a
conformal factor, for instance, as in the context of a
spatially flat inhomogeneous FLRW universe described
by the metric in (9) we will choose the Lorenz gauge
with respect to g and not a2 g :
A
1
|g|
1
2
1
|g| 2 g A = 0
(60)
00 00
00 00
1
00 0
dd x00 |
g 00 | 2 g g 00 G00 00 G
00 00 G00 G
00 0 .
+R
G,
of the schematic form G(|1)
covariant differentiation of the Greens functions. The
(|1) is the first order in h variation of the Christoffel
symbol,
(|1) =
1
g
{ h} h .
2
(64)
0 + 1 G
0 h0 0 + 1 h G
0
G 0 G
2
2
Z
1 00 00 00 00 00
d 00 00 12
00 00 00 00
00 0
00
00
+ d x |
g | G
h g
g
h
g
00 G
2
1 00
00 00
00 0 1 G
00 h00 00 00 0 G
x00 ,x0
+ G
00 G
00 h
2
2
00
1
00 0 1 G
00 0
x,x00 h00 00 00 G
00 h00 00 00 G
+ G
2
2
10
00
1
x,x00 h00 00 G
00 0 + 1 00 G
00 h00 00 G
00 0
00 G
2
2
1 00
00 h00 00 G
00 h00 00 0 G
x00 ,x0
00 0 + 1 00 G
00 G
2
2
00
00 h00 00 R
00 + R
00 00 h00 00 G
00 0
+G
1 00 00 00
00 00
G00 0 .
+ G00 (R|1)
+ h R
2
In (65), we are again raising and lowering all indices with the background metric g . Here and
below, (R|n), (R|n) and (R|n) are the portion of the respective geometric tensors (built out of
g = g + h ) containing precisely n powers of the
perturbation h .
x,x0 +
G 0 G
Z
+
1
Gx,x0 (h0 0 + h )
2
1 00 00 00
d 00
00 00
x00 ,x0
G
h
h
d x 0 Gx,x00
2
1
00
(65)
(66)
metric g + 32GN into the Einstein-Hilbert action (67) and find a resulting infinite series in .
The quadratic piece, which will determine for us the
Greens function of the graviton, is
1
S [g] =
2
1
d x|g|
2
d
1
2
2R 2 R + 2 R
1
+ 2 (R 2) ,
(68)
2
where we have chosen the de Donder gauge =
with g . (The geometric tensors in
(68), such as R , are built out of g .) From (68)
and following the preceding analysis for the scalar and
photon, we may write down the integral equation involving the graviton Greens functions
1
2 ,
0 0
G0 0 G
Z
1 00 00 00 00
d 00 00 12
00 00
00 00 00 00
00
00
00
= d x |g | G g
g
g
g
g
00 G00 00 0 0
2
11
00 00
00 00
00 00 00 00
1 00 00 00 00
g g g g (R 2) 2R
2
00 00
00 00
00 00
00 00
00 00
00 00
00 00
00 00
R g R g + R g + R g G00 00 0 0
Z
00 00
00 00
00 00
1
1 00 00 00 00
00 00 0 0
dd x00 |
g 00 | 2 00 G00 00 g g g g g 00 G
2
00 00
00 00
1 00 00 00 00
2 2R
00 00 00 00
+ G00 00
g g g g
R
2
00 00
00 00
00 00
g R
g 00 00 + R
00 00 g00 00 + R
00 00 g00 00 G
00 00 0 0 .
R
00 00
+G
(69)
(71)
0
0
1
1
1
( + 0 )( + ) 0
2
2
2
1 2 1 02
1
0 2
( + )
2
2
2
(72)
We then arrive at
1
x00 ,x0
0
00
+ d x P Gx,x
h h
G
2
1
0
x,x00 h00 ] 00 G
x00 ,x0 + ([ + [0 )G
x,x00 h00 ] 00 G
x00 ,x0
+
([ + [0 )G
4
x,x00 h00 ] 00 G
x00 ,x0 + ([ + [0 )G
x,x00 h00 ] 00 G
x00 ,x0
+ ([ + )G
[ 0
(73)
12
One scattering approximation Let us examine
(58), (66) and (73). The terms that do not involve
any integrals can be viewed as the propagation of null
signals, modulated by the metric perturbations mul x,x0 . The terms involving integrals, go
tiplying the G
R
x,x00 h[x00 ]G
x00 ,x0 . Due
schematically as x x0 d4 x00 G
this can be interto the causal structure of the Gs,
x00 ,x0 tells us our
preted as a scattering process. The G
0
massless field begins at the source x and travels along
a null ray to x00 ; the h[x00 ] says it then scatters off the
metric perturbations (and its derivatives) at x00 ; and
x,x00 informs us that it then propagates along a
the G
null path from x00 to reach the observer at x. The full
(scattered) signal consists of integrating over all the
x00 from which the signal can scatter off. This is the
perturbative picture for the origin of tails of massless
fields in weakly curved spacetime.11 From this heuristic point of view, we can already anticipate that high
order perturbation theory will involve more than one
scattering events contributing to the tail effect. This
scattering picture may also help us estimate its size
without detailed calculations, and deserves some contemplation.
matrix integral
1
I
4
x,x00 h00 00 G
x00 ,x0
d4 x00 G
In this section we will restrict ourselves to 4 dimensions and analyze further the first order results for
the scalar (58), photon (66) and graviton (73) Greens
functions we have obtained in perturbed Minkowski
spacetime, and show that to O[h], concrete results for
the Greens functions can be gotten once a single matrix of integrals (involving h ) can be performed. We
will also decompose these scalar, photon and graviton
Greens functions into their null cone and tail pieces.
As a consistency check of our Born approximation, we
show that their null cone pieces matches the Hadamard
form described by equations (20), (21) and (22); this
generalizes the analysis carried out in Pfenning and
Poisson [3] to the case of arbitrary perturbations h .
In the scalar (58), photon (66) and graviton (73)
Greens functions results, we have to deal with derivatives (with respect to x or x0 ) acting on the following
11
We are
R being slightly 00inaccurate here, in that some of the
x,x00 h[x ]G
x00 ,x0 terms also contribute to null
x x0 d4 x00 G
propagation, as we will see in the next section. But we want
to introduce this scattering picture here, because it is easier
to see it from (58), (66) and (73), written in terms of the
than from (84), (86), and
Minkowski Greens function Gs,
(87) below, which are expressed in terms of [x,x0 ] and the
b
I-integrals
in (78).
(76)
In appendix (A) we show that I involves the integral of h (but in Euclidean 3-space) over the surface
generated by rotating the ellipse with foci at ~x and ~x0
and semi-major axis (t t0 )/2, about the line joining
~x and ~x0 . (This is the dashed oval in Fig. (2).)
I [x, x0 ] [t t0 ][
x,x0 ]Ib [x, x0 ]
(74)
(77)
with
Ib [x, x0 ]
(78)
"
#
Z
0
0
~
d
t+t
||
~x + ~x
1
00
h00 00
+
cos ,
+ ~x .
=
2 S2 4
2
2
2
The infinitesimal solid angle is d = d cos d, and the
Cartesian components of ~x00 are given by
!
r
r
0
0
0
x,x
x,x
~x00
sin cos ,
sin sin ,
cos .
2
2
2
(79)
To separate the light cone versus tail pieces of the
Greens functions, we now carry out the necessary
derivatives on (77) as they occur in (58), (66) and (73).
There is no need to differentiate the [t t0 ], because
that would give us [t t0 ] and its derivatives. Since
this would be multiplied by either [
x,x0 ] or possi~ 2 /2 < 0,
bly [
x,x0 ], 0 [
x,x0 ], etc., while
x,x0
0
these , , . . . terms can never be non-zero when t =
t0 . Schematically, therefore, the derivatives now read
b (where the two derivatives are both
[t t0 ]([
]I)
with respect to either x or x0 or one each), which in
turn would yield two types of terms. One is the tail
term, proportional to [
] Ib and the other the null
b [
b
cone ones, proportional to either [
]
I,
]
I,
b Following that, we would impose the
or 0 [
]
I.
constraint
x,x0 = 0 on the coefficients of the [
] and
13
0 [
] terms. This requires that we develop a power series in
x,x0 of Ib . Since there is at most one derivab however, we only need to do so up to
tive acting on I,
linear order. (Higher order terms would automatically
vanish once we put
= 0.) In appendix (A) we find
(0)
(1)
Ib = Ib +
x,x0 Ib + . . .
(80)
Z
1 1
x,x0
h00 00 []d
0
+ ....
= 1
2
2 0
R1
(0)
Gx,x0
(82)
h
i
0
0
[t t0 ]
1
1
1
(0)
(0)
(0)
x,x0 + Ib
1 h h0 + Ib(0) + Ib Ib
4
4
4
2
h
i 1
(0)
+
x,x0 + Ib
0 Ib 0 Ib
,
h h ; h0 h0 0 .
(84)
2
14
G 0
p
[t t0 ]
g 0 x,x0 [x,x0 ]
(86)
4
1 b
1
\
+ [x,x0 ]
0 I 0 Ib + ( 0 ) [ + [0 Ib] + (R|1)
,
2
2
G0 0
p
[t t0 ]
P0 0 x,x0 [x,x0 ]
4
1 b
0 I 0 Ib
+ [x,x0 ] P
2
0
1
+
([ + [0 )Ib00 ] 00 + ([ + [0 )Ib00 ] 00
4
+ ([ + [0 )Ib00 ] 00 + ([ + [0 )Ib00 ] 00
\ + (R|1)
[ 00 00 + (R|1)
[ 00 00 1 { (R|1)
[ 00 00 1 { (R|1)
[ 00 00
+ P (R|1)
}
}
2
2
[ 00 00 00 00
+ (R|1)
.
(87)
{ }
4
Z
(1/2) ( 0 )Ib . Hence, the rest of the tail terms
4 00
#
"
Geometry and tails Let us notice that it was all
Z
~
1
t + t0
d
||
~x + ~x0
00
the
differentiation that took place in our work on the
=
(R|1)00 00
+
cos ,
+ ~x .
2 S2 4
2
2
2
perturbative solution of the Greens functions, which
[
[
\
The geometric terms (R|1)
, (R|1) , and (R|1)
in (86) and (87) can be obtained by taking the corresponding linearized tensors in terms of the perturbation h , and replacing all the h with Ib and all
derivatives 00 with + .
1
[
b
b
+
+
+
+
(88)
I
I
(R|1)
]
]
2 [
[ 1 + { + Ib
(R|1)
}
2
+ + Ib + + Ib
(89)
\ + + Ib Ib ,
(R|1)
(90)
15
can be traced to the operator, that gave us both
the terms in the arguments and coefficients of the functions in the scalar, photon, and graviton Greens
function. In turn, we have identified them as various
terms in the world function, the van Vleck determinant and the parallel propagator (in their perturbative guises). This re-affirms our assertion earlier that
it is the differential operator that is solely responsible for the behavior of massless radiation on the light
cone. On the other hand, because of (70), at the level
of the Born approximation, we see that the geometric
tensors contribute only to the tail piece of the Greens
function.
V.
(93)
where
h00
1
1+
M
2r
M
2r
!2
1
(94)
2
3
M
M
M
= 4
+8
12
+ ...,
2r
2r
2r
!
4
M
hij ij
1+
1
(95)
2r
!
2
3
M
M
M
= ij 4
+6
+4
+ ... ,
2r
2r
2r
h0i = 0.
(96)
p
Here r ij xi xj , M is the mass of the black hole,
and we have set Newtons constant to unity, GN = 1.
The power series expansion of h00 and hij can be subb
stituted into the I-integral
in (78). At order (M/2r)2
and beyond, the solution would of course receive contributions from more iterations and high order h terms
from the integral equations (53), (62), and (69), and
b
would likely involve two or more overlapping I-type
integrals. Here we will focus on the first Born approximation.
Within the one scattering approximation, the main
technical hurdle to overcome is therefore the class of
integrals
Z +1 Z 2
d(cos 00 )d00
bI(n) 1
00
00
4 1 0 |~x [, , 00 ] ~z[s, + , + ]|n
(97)
where n is a positive integer, ~x00 has Cartesian components defined in (79) (so that, in particular, = 0 =
t t0 ), and
~z
~x + ~x0
.
2
(98)
cos + =
r0 r
.
|~x ~x0 |
(100)
16
the Columb (electric) potential of a charged perfectly
conducting ellipsoid defined by ~x00 .12 By definition,
the conducting surface is an equipotential one. This
implies that the answer to bI(1) has to depend on the
s-coordinate of ~z only, for that would automatically be
a constant on the ellipsoidal surface. For ~z lying away
from the ellipsoidal surface, our integral must satisfy
Poissons equation g ij zi zj bI(1) [s] = 0 (with the inverse metric g ij of (A5)), which in turn is equivalent
to the ordinary differential equation
0 = (1
d2bI(1) []
2 )
d2
dbI(1) []
,
d
s/R. (101)
R
s
(103)
12
13
17
Starting from the Kerr metric written in BoyerLindquist coordinates (see equation 33.2 of [6]), we first
perform the following transformation on the r coordinate
2
M
.
(109)
r r 1+
2r
(This coordinate transformation would yield, when
S = 0, the Schwarzschild metric in isotropic coordinates.) Denoting the unit vector in the 3-direction as
eb3 and further define
~ Sb
S
e3 ,
(110)
(111)
1
0 i
J
M
{ }
|~x ~x0 |
0
r + r + |~x ~x0 |
ln
r + r0 |~x ~x0 |
World Function
The world function is x,x0
(0)
(t
t
)
+
(~
x
~
x
)
ln
,
|~x ~x0 |
r + r0 |~x ~x0 |
where
0 i
~
h [t, ~x] 2 M + {
} S
~x
In a Cartesian basis,
~
S
= S 2, 1,0 .
~x i
x x
i
i
x,x0
x,x0
1
.
r
(112)
(113)
0 i
The off diagonal nature of {
} implies that the first
order in mass Ib00 and Ibij for the Kerr black hole are
identical to that of the Schwarzschild case. As for Ib0i ,
by referring to (74), integrating by parts the spatial
gradient acting on r1 , and using (57) to pull the resulting two derivatives out of the integral, we observe
that it can be gotten by acting
~ +
Ji S
(114)
~x ~x0
i
(117)
1
1
+ 0
r
r
(119)
1
(0)
0
+
0 Ib
2
1
1
1
0
= (h[x] + h[x ]) = M
+ 0 . (120)
4
r
r
(115)
Altogether, to first order in mass M and angular momentum S, the Kerr spacetime hands us
0 i
Ib = M + {
} Ji bI(1) ,
(116)
14
18
Parallel Propagator
According to (46), the symmetric portion of the parallel propagator can be read
off the metric perturbations, namely
1
(g 0 + g0 )
(121)
2
1
1
= M
+ 0
r
r
i
i
1 ~
1 ~
0
0
+ {
}i
S
~
x
S
~
x
+
r3
r03
At this point, it is convenient to define
Vj
x
bj + x
b0j
.
rr0 (1 + x
bx
b0 )
(122)
i Vi = i Vi = 0
and it also satisfies
~
Vj = 0.
S
~x ~x0
(123)
The second is
(124)
(125)
1
(gjk0 gkj 0 ) = M [k (t t0 )J[k Vj] .
2
(127)
and
1
(A)
Ib
[+ Ib] ,
2
(130)
(131)
1
(g0j 0 gj00 ) = M (t t0 ) + i Ji Vj
2
Tails in Kerr
reads
(128)
M
(A)
Ibjk =
[r + r0 (t t0 )]Vj]
2 [k
+ J[k [r + r0 (t t0 )]Vj] .
(132)
The third and final building blocks are the geometric curvature terms. The non-zero components of the
Riemann terms are
M
[
(R|1)
{i+ [r + r0 (t t0 )]Vj} , (133)
0i0j =
4
1
0
0
[
(R|1)
0ijk = [j + Jk] ([r + r (t t )]Vi ) , (134)
2
M
[
(R|1)
[k+ l]i [r + r0 (t t0 )]Vj
ijkl =
2
(i j)
(135)
where the (i j) means one has to take the preceding
term and swap the indices i and j. Performing the
appropriate contractions and utilizing (123) yields the
Ricci tensor and scalar terms
[t t0 (r + r0 )]
0 i
[
(R|1)
=
M
SJ
i
{ }
rr0
(136)
19
0
0
\ = 2M [t t (r + r )]
(R|1)
0
rr
(137)
ij
i
{ }
Z
d00 (3) ~x + ~x0
1
+ ~x00 ,
2
4
2
(the 4
[t t0 ]
[x,x0 ] + [x,x0 ] Ib(S)
4
(138)
[t t0 ]
g 0 [x,x0 ]
(139)
4
(A)
[
+ [x,x0 ] Ib(S) + Ib
+ (R|1)
,
[t t0 ]
1 b(A)
(A)
(A)
(S)
(A)
b
b
b
b
G0 0
I + I + I + I
P0 0 [x,x0 ] + [x,x0 ] P I +
4
2
1
1
[
[
[
\
[
[
.
+ P (R|1) + (R|1)00 00 + (R|1)00 00 { (R|1)00 }00 { (R|1)00 }00 + (R|1)00 {00 00 }00
2
2
20
t-t'
VI.
r'
r'
In this paper, we have developed a general Born series expansion for solving the minimally coupled massless scalar, photon, and graviton Greens function in
perturbed spacetimes described by the metric g =
g + h . The key starting points are the integral
equations for the scalar (53), photon (62) and graviton (69) cases, which were gotten from the quadratic
portions of the actions of the respective field theories.
From these, one performs a power series in the perturbation h and iterate these equations (followed by
dropping the remainder terms) however many times
necessary to achieve the desired accuracy. We derived
a first order integral representation for the scalar (56)
and photon (65) Greens functions in generic backgrounds, and for scalar (58), photon (66) and graviton (73) in a Minkowski background. Furthermore, in
(84), (86), and (87), we decomposed these perturbed
Minkowski results into their light cone and tail pieces,
showing their consistency with the Hadamard form.
We reiterate that, at first order in metric perturbations, the solution of the scalar, photon and graviton
Greens functions is reduced to the evaluation of the
single matrix integral in (78); the remaining work is
mere differentiation. Even though we have applied our
perturbation theory only to massless scalars, photons
and gravitons, because all we have exploited are the
quadratic actions of the field theories involved, our
methods should in fact apply to any field theory whose
quadratic action is hermitian.
As a concrete application of our formalism, we have
calculated the Greens functions of the massless scalar
(138), photon (139), and graviton (140) in the weak
field limit of the Kerr black hole geometry, to first order in its mass M and angular momentum S. A subset
of these weak field results for the Schwarzschild case
have previously been obtained by DeWitt and DeWitt
[1], and Pfenning and Poisson [3]. Our Kerr calculation shows that, to first order in angular momentum
S, there will be rotation-induced corrections to these
Schwarzschild Greens functions, only on and near the
null cone, namely |~x ~x0 | t t0 r + r0 (where
r |~x| and r0 |~x0 |). Beyond that, t t0 > r + r0 ,
the behavior of the Greens functions changes abruptly
and is governed solely by the mass of the black hole.
Of the previous approaches we have studied DeWitt and DeWitt [1], Kovacs and Thorne [2] and Pfenning and Poisson [3] DeWitt and DeWitts seems to
be the most general. They utilized Julian Schwingers
perspective that the Greens function is an operator
in a fictitious Hilbert space, for example, Gx,x0 =
b 0 i, from which they found its variation. Howhx|G|x
ever, on the level of classical field theory, the main
concern of this paper, our methods do not require any
21
additional structure than the quadratic action of the
field theory at hand. Hence, we hope it is accessible
to a wider audience.15 Our null cone versus tail decomposition was modeled after the work of Pfenning
and Poisson [3] (except we generalized it to arbitrary
metric perturbations), who in turn state that their
work was based on calculations by Kovacs and Thorne
[2]. In Pfenning and Poissons work, they wrote down
a perturbative version of the differential equations in
(15), (16) and (18) for a weakly curved spacetime with
only scalar perturbations , and derived integral representations of the solutions using the flat spacetime
x,x0 ; their methods can very likely
Greens function G
be generalized to arbitrary perturbations. However,
repeated (and un-necessary) use was made of the equations obeyed by the gravitational potential . We feel
this obscures the fact that the solution of the Greens
function of some field theory depends on the geometry
but not on the underlying dynamics of the geometry
itself.
Cosmology We close with some thoughts on applying our work to cosmological physics. We have already shown that the classical theory of light in a spatially flat inhomogeneous FLRW universe is equivalent
to that in a perturbed Minkowski spacetime. Consider
a source of photons that turns on for a finite duration
of time, say a gamma ray burst at redshift z = 6. We
display in Fig. (4) that not only would these photons
sweep out a null cone of finite thickness proportional to
the duration of the burst, but they will also fill its interior. If t is the present time, the dark oval represents
the light that has leaked off the light cone. From our
calculation in (86), the tail part of the Greens function
(tail)
begins at O[h].
and hence the vector potential A
Because the components of the stress energy tensor of
the electromagnetic fields in an orthonormal frame Tbb
(which is what an observer can measure) is quadratic
in the derivatives of the potential, Tbb a4 (A)2 ,
this means deep in the interior of the null cone Tbb
itself must be quadratic in the metric perturbations
h .16 In cosmology, because the metric perturbations
h are believed to be sourced by quantum fluctuations of fields in the very early universe, the h at a
15
16
Tail
22
in (2) and the complicated terms in (86), however, this
is a difficult calculation. We hope to report on this line
of investigation in a future publication.
VII.
ACKNOWLEDGEMENTS
x,x0 = 0. A similar discourse may be found in Pfenning and Poisson [3], but ours is more general because
we performed it for arbitrary metric perturbations.
Let us first display the integral in its most explicit
form, using the second equality of (32):
Z
1
I =
d4 x00 h00 00 [t00 , ~x00 ]
4
[t t00 |~x ~x00 |][t00 t0 |~x00 ~x0 |]
. (A1)
|~x00 ~x0 ||~x ~x00 |
We may integrate over t00 immediately, so that t00 =
t |~x ~x00 | = t0 + |~x00 ~x0 |. This in turns yields
the constraint that, viewed in Euclidean 3-space, the
observer at ~x and the emitter at ~x0 form the foci of a
prolate ellipsoid, with semi-major axis 0 /2, defined
by
t t0 = |~x ~x00 | + |~x00 ~x0 |.
(A2)
(A3)
x,x00 h00 00 G
x00 ,x0 , due to the causal structure
product G
of the Greens function, is non-zero if and only if the
x00 lie both on the future null cone of x0 and on the past
null cone of x. This can be satisfied if and only if x0 lies
on or within the past light cone of x or equivalently, if
and only if x lies on or within the future light cone of
x0 .
If we now assume that (A3) holds, then it is the surface of the ellipsoid in (A2) that we need to integrate
over, weighted by h00 00 [t00 , ~x00 ]. To see this, let us employ ellipsoidal coordinates centered at (1/2)(~x + ~x0 ),
i.e. put ~x00 (1/2)(~x0 + ~x) + ~x000 , with
p
000
(s/2)2 (R/2)2 sin cos ,
~x [s, , ] =
p
(s/2)2 (R/2)2 sin sin ,
(s/2) cos .
(A4)
These coordinates fix the foci to be at ~x and ~x0 but
allow the size of the ellipse to vary
p with s. (The 1and 2-components of ~x000 tell us (s/2)2 (R/2)2 act
as the radial coordinate in the 12-plane, and hence we
shall require s R. This means
R all volume integrals
involve s would have limits R ds.) The Euclidean
spatial metric in 3 dimensions goes from gij = ij for
Cartesian coordinates to
(s/2)2 (R/2)2 cos2
gij = diag
,
s2 R2
2
s R,
(A5)
s R
cos[],
2
2
|~x00 ~x0 | =
s R
+ cos[].
2
2
23
= [t t0 ][
x,x0 ]
(A7)
Z
0
0
d
t+t
R
~x + ~x
1
000
00
00
h
+ cos[],
+ ~x ,
2 S2 4
2
2
2
where now ~x000 = ~x000 [t t0 , , ].
Small
x,x0 expansion For small
x,x0 , we may
b
develop I as a series expansion in powers of
x,x0 .
Right on the null cone tt0 = |~x~x0 |, and if we lie ~x~x0
along the positive 3-axis, the spacetime arguments of
h take on the Cartesian components
0
x3 + x03
R
t + t0
+
cos , 0, 0,
+ cos
2
2
2
2
(A8)
x,x0
+
cos
(A9)
2
2
2
and the 3-component (1/2)(x3 + x03 + 0 cos ) as
s
2
R
x3 + x03
x,x0
+
cos .
(A10)
+
2
2
2
Perform a change of variables in (A7) cos 21 and
Taylor expand h in powers of
0 in the time and
x,x
3-components and in powers of
x,x0 in the remaining
orthogonal directions. One would find it is necessary
to expand the orthogonal directions up to second order
to achieve a non-zero result. With = x0 + (x x0 ),
x,x0
4
x,x0
4
dh00 00
(A11)
0
Z 1
2 1
2 1
d
000 h00 00
300 h00 00
R
t t0
0
Z 1
d(1 ) 1200 + 2200 h00 00 + . . .
0
2 1
2 1
Z 00 h00 00
t00 h00 00
d
R
t t0
0
Z
x,x0 1
x,x0
4
(1 )
2 h00 00 []
h00 00 []
=
,
x x0
(A13)
x,x0
b
I 1
0
h00 00 []d .
2
2 0
(A14)
May 1993
[8] M. Visser, Phys. Rev. D 47, 2395 (1993) [arXiv:hepth/9303020].
[9] E. E. Flanagan, E. Rosenthal and I. M. Wasserman,
Phys. Rev. D 79, 044032 (2009) [arXiv:0810.0535
[astro-ph]].
[10] J. S. Schwinger, J. Math. Phys. 2, 407 (1961).
[11] C. R. Galley and M. Tiglio, Phys. Rev. D 79, 124027
(2009) [arXiv:0903.1122 [gr-qc]].
[12] E. Poisson, Phys. Rev. D 66, 044008 (2002) [arXiv:grqc/0205018].