2nd Reading

November 3, 2015 13:12 WSPC/S0218-2718 142-IJMPD 1544023

International Journal of Modern Physics D

Vol. 24, No. 12 (2015) 1544023 (6 pages)

c World Scientific Publishing Company
DOI: 10.1142/S021827181544023X

Gravitational waves: Some less discussed intriguing issues∗

C. Sivaram
Indian Institute of Astrophysics, Sarjapur Road,
Koramangala, Bangalore 560034, India
sivaram@iiap.res.in
Int. J. Mod. Phys. D 2015.24. Downloaded from www.worldscientific.com

Received 11 June 2015

by 187.134.87.192 on 02/16/16. For personal use only.

Accepted 5 October 2015

Published 5 November 2015

Attempts to detect gravitational waves is actively in progress with sophisticated devices

like LIGO setup across continents. Despite being predicted almost 100 years ago, there
has so far been no direct detection of these waves. In this work, we draw attention to
some of the less discussed but subtle aspects arising, for example, from high orbital
eccentricities, where emission near periastron could be millions of times more than that
in the distant parts of the orbit. The strong field nonlinear effects close to the compact
objects can substantially slow down and deflect the waves in the last (few) orbit(s) where
much of the intensity is expected. Spin–orbit and other forces could be significant. There
would also be plasma like resonant absorption (of kilohertz radiation) during the collapse.
Recent observation of supermassive black holes at high redshift implies cluster collapse,
where the gravitational wave intensity depends on very high powers of the mass. Any
unambiguous claim of detection should perhaps consider several of these effects.

Keywords: Black hole physics; gravitation; gravitational waves; relativistic processes;

supermassive black holes.

1. Description
Almost a 100 years ago, in 1918, Einstein established that the small perturbations
of the Minkowski metric tensor propagate at the speed of light and are generated by
masses undergoing acceleration. These perturbations, the gravitational waves (in
the linearized theory) describe transverse or shear deformation of space being asso-
ciated with the quadrupolar moment (moment of inertia) of the source. This was
the first definitive prediction of gravitational waves. In electromagnetism, the trans-
verse polarization is a vector generated in first approximation by dipolar moment of
source. As is well known, because of conservation of momentum, there are no dipole

∗ This
essay received an Honorable Mention in the 2015 Essay Competition of the Gravity Research
Foundation.

1544023-1
 2nd Reading
November 3, 2015 13:12 WSPC/S0218-2718 142-IJMPD 1544023

C. Sivaram

gravitational waves, the lowest order being quadrupolar1 (monopole radiation van-
ishes in both cases from conservation of charge and mass (energy), respectively).
Although Maxwell’s prediction of electromagnetic waves (of all frequencies) was ver-
ified experimentally very soon by Hertz, gravitational waves have yet to be directly
observed (the slowing down of the binary pulsars being consistent within a fraction
of a percent from the radiation loss predicted by the Einstein quadrupole formula
being the strongest indirect evidence). It is of course very clear why it is so difficult
to observe gravitational waves, apart from the vast difference in coupling strength,
the quadrupole involves the inverse fifth power of the velocity of light. A rod spin-
ning close to break up would generate less than an attowatt, the earth–moon binary
hardly some microwatts! The energy carried away by gravitational waves leads to a
decrease in the period of a binary system, this effect being conspicuously observed
Int. J. Mod. Phys. D 2015.24. Downloaded from www.worldscientific.com

first in the 8 h period binary pulsar.2,3 The many years of observation by Taylor
and Weisberg have enabled the slow down data to be consistent with that expected
from gravitational waves to within 0.4 of a percent.4 Binary supermassive black
by 187.134.87.192 on 02/16/16. For personal use only.

holes like the system OJ 287 also show conspicuous effects (clearly observed in
Ref. 5). Attempts at the detection of gravitational waves is now well known to
be an active area of research with huge detectors like LIGO and VIRGO (or in
the future LISA) interferometers having been built to detect them as they reach
Earth.6,7 One of the intriguing factors is the dependence of the gravitational wave
intensity on the binary orbital parameters, especially the eccentricity. It is well
recognized that eccentricity of orbits boosts relativistic effects substantially. The
presence of (1 − e2 ) in the denominator can result in several times the effect for
a circular orbit. The increase is for more dramatic for gravitational waves. For
instance, the ratio of the instantaneous power radiated at the periastron P and the
apoastron A is given by:
P (A) (1 + e)
= , (1)
P (P ) (1 − e)6
which in the case of the binary pulsar (e = 0.6) implies that this ratio is nearly
6000! The gravitational wave emission is 6000 times more at periastron than at
apoastron, so that most of the energy is radiated near about P , with the associated
travel time being hardly 3% of the orbital period. For the time averaged power, the
dependence on eccentricity scales inversely as (1 − e2 )3.5 (this being the lead term
in the formula8 ).
P = P c × f (e). (2)
P c for a circular orbit being:
P c = KGµ2 c−5 ω 6 a4 , (3)
(µ is the reduced mass, ω is the orbital frequency, a is the separation). For e = 0.9,
Eq. (1) predicts that the gravitational wave intensity is two million times more
intense at periastron within a travel time of hardly half a percent of the orbital

1544023-2
 2nd Reading
November 3, 2015 13:12 WSPC/S0218-2718 142-IJMPD 1544023

Gravitational waves

period. Accurate estimates of expected flux from several such systems should con-
sider these aspects in some detail. Again much of the gravitational wave radiation
is expected from the last or final few orbits (or even last orbit) before merger of the
compact objects, either neutron stars or more interestingly black holes. Millisecond
spikes or radiation are expected. This can give rise to several related effects. Higher
multipolar terms may not be insignificant. For n (multipolar) order, P scales as
P = P c × f (e) where
P cn = βGµ2 ω 2(n+1) R2n c−(2n+1) . (4)
6 8 −7
For n = 3, we have for example dependence scaling as R ω c , etc., but as for
the last orbit (or before merger), R = GM 2 3
c2 , we have (with ω R ≈ GM ), similar
contributions seem to arise from Pc2 and Pc3 , as ωR → c other terms could become
Int. J. Mod. Phys. D 2015.24. Downloaded from www.worldscientific.com

comparable and cannot be neglected as for larger separations so estimates to com-

pare with observed values (if and when observed!) should perhaps go into these
questions. Moreover at close separations, spin–orbit, spin–spin and other gravita-
by 187.134.87.192 on 02/16/16. For personal use only.

tional couplings become comparable to the static values and could affect the instan-
taneous orbital parameters.9,10 Gravitomagnetic effects, which are suppressed by
(v/c) could (for very close separations), become comparable to the other relativistic
effects.10 It is to be noted that we are dealing with strong gravitational fields in the
vicinity of compact objects like black holes and unlike electromagnetism, gravity is
a nonlinear field, the energy–momentum carried by the field (waves) adding nonlin-
early to the field intensities. The higher order self-couplings would become larger
and the waves would tend to be confined in the strong field (rather than propagate
freely!).11
Approximations carried over from the linear theory (inclusive of post-Newtonian
corrections) can no longer correctly describe the propagation of the waves when the
curvatures and fields are large close to the merging compact bodies (black holes).
The velocity of gravitational waves being light velocity (in vacuum) is only in the
linear theory where the wave equation (corresponding to small perturbations of flat
d2
space) is:
hab = KTab ,
= 2 − ( c12 dt2 ). (Again there are subtle issues like only

if the energy–momentum is traceless, it is possible to find a trace free haa = 0.)

Full nonlinear general relativity is not a field theory in Minkowskian (background)
spacetime but spacetime geometry itself. Moreover as is well known, notion of grav-
itational energy density is plagued with troubles. Problem is worse in the quantum
theory, as a hypothetical graviton carrying energy and momentum cannot be clas-
sically approximated by pp waves as the only value of the ADM energy–momentum
associated with the waves is zero. Gravitational waves are nonlinear and in strong
fields, their velocity is retarded by the gravity field (like light), the velocity cG now
being (in the vicinity of the source)


2GM
cG = cGO 1 − , (5)
rc2
where cGO = c the velocity of light. Moreover like light, gravitational waves would
also be subject to strong deflection in the region where most of the radiation is

1544023-3
 2nd Reading
November 3, 2015 13:12 WSPC/S0218-2718 142-IJMPD 1544023

C. Sivaram

emitted (the last orbit of the merger). The deflection ∆ = 4GM rc2 + higher order,
would be large near r = 2M , and for rotating merging (compact) objects, the
angular momentum also contributes to the deflection ∆ = cG3 rJ2 (which for neutron
stars is about a radian and more for extremal black holes). Thus, gravitational
waves would no longer propagate direct from the source but could be deflected
considerably, slowed down and most of the energy may even be trapped. A strong
self-interacting (attractive) field tends to confine the waves not allowing them to
leave the source.11,12 An illustrative example is given by a self-interacting (nonlin-
ear) field, where the field energy density (φ)2 acts as a source of the field thus:
2 φ = K(φ)2 . (6)

This has a solution of φ = K
ln r i.e. a confining potential. This also holds in

Int. J. Mod. Phys. D 2015.24. Downloaded from www.worldscientific.com

curved space! Higher orders again change the propagation features near the source,
including the velocity.12 So gravity not only deflects and refracts electromagnetic
by 187.134.87.192 on 02/16/16. For personal use only.

waves but also gravitational waves in the strong field vicinity of the source, where
the effects are nontrivial and large. So exploring the possibility of detecting the
gravitational waves from such sources should consider these aspects carefully. Again
the belief that the interaction between matter and gravitational waves are very weak
and negligible may not be true as is elaborated in a recent work,13 where it was
shown that stars vibrating at the same frequency as the gravitational waves passing
through them can absorb large amount of energy from the ripples. The effects of
gravitational wave can be detected thus by looking at groups of stars. Population
of stars near a system of merging black holes pounding the stellar population with
gravitational waves make the more massive stars light up first. A similar suggestion
regarding such a resonance absorption was made much earlier14,15 where it was
indicated that the gravitational plasma frequency given by

ωGP ≈ (αGnmn )1/2 , (7)

which for a neutron star (n the neutron number density being 1038 , mn the neutron
mass, G the gravitational constant) becomes ωGP = 104 Hz which is same as the
frequency of the bulk of the gravitational waves emitted by such a collapsing com-
pact object. This could enable the resonance absorption of a large fraction of the
gravitational waves. A more detailed analysis gives a energy dependent damping
length scaling as:
ω2 c
ld ∼ · . (8)
Gnma v 3
The formula was modified for a relativistic gas. Recently, in connection with the
propagation of highest energy cosmic rays and energy dependent time delays in
gamma ray bursts, it was suggested that there could be minimal velocity or minimal
velocity change at the quantum level.16 This corresponds to ∆v ≈ 10−11 cm/s if this
concept is applied to detection of gravitational waves this would imply a minimal
constraint on the dimensionless strain ∼ ∆l
l invoked in gravitational wave detection.

1544023-4
 2nd Reading
November 3, 2015 13:12 WSPC/S0218-2718 142-IJMPD 1544023

Gravitational waves

Thus, if the relative velocity of two particles fluctuate under influence of the wave
and they collide during the wave cycle, their velocity fluctuation is
∆v ∼ hν. (9)
For v ∼ c, ∆v the value obtained in Ref. 14, we have h ∼ 10−22 close to the limit of
LIGO, etc. A method invoking Mossbauer effect to measure minute velocity changes
to possibly detect gravity waves was suggested which involves ∆v v ∼ 10
−21 15
. The
above value of h also interestingly corresponds to that invoked in the LIGO setup,
where7

1/2
t

m
hmin ≈ , (10)
Larm
Int. J. Mod. Phys. D 2015.24. Downloaded from www.worldscientific.com

which for t = 1 ms, Larm = 1 km, M = 103 kg implies hmin = 10−22 , same as
Eq. (9). The flux from gravitational waves, scales as h−2 and as the square of the
by 187.134.87.192 on 02/16/16. For personal use only.

frequency and with the above values, corresponds to a value of

2
FG = 2 × 10−5 W/m . (11)
It may be noted that the optical flux from Jupiter (on Earth) is 100 times less, at
2 × 10−7 W/m2 . This underlies the vast difference underlying the nature and detec-
tion of gravitational waves, which has defied any direct detection so far. Another
phenomenon to be considered is the formation of supermassive black holes espe-
cially in the early universe. While it is still unclear how such large black holes can
form,17 recent discoveries have uncovered several supermassive black holes residing
at the cores of ultra luminous quasars at high redshift. A very recent discovery18
shows the presence of a 12 billion solar mass black hole which formed less than a
billion years after the Big Bang expansion and challenges theories of how they can
grow so rapidly. A possible scenario19 is the collapse of a large stellar cluster (along
with other constituents). The rapid asymmetric shrinking of such a large assembly
(mass M and radius R) should be accompanied by release of gravitational waves,
the power scaling as a very high power of M . Briefly,17,19 the angular momentum
is quasi conserved, so that the relation,
M 3 R = const. (12)
holds. The two body relaxation time scales as

3 1/2
R
∼ M −5 , (13)
GM
and from Ref. 18, it follows that cluster density increases as M 9 and from Eqs. (2)–
(4), it follows that the gravitational wave power scales as M 20 ! For M ≈ 106 M ,
the frequency is ∼1 Hz and for 109 M is a milliHertz. LISA should detect several
such high redshift super collapse events if one goes by the number of black holes
being seen (at high redshifts). Higher multipole radiation could be dominant just
before the multimergers as discussed from Eq. (4). In short, considering the current

1544023-5
 2nd Reading
November 3, 2015 13:12 WSPC/S0218-2718 142-IJMPD 1544023

C. Sivaram

interest in the direct detection of gravitational waves the several questions addressed
above such as the quantitatively very different powers emitted in different parts
of orbits, the propagation deflection and trapping of waves in strong fields (just
before final collapse where most of the energy is emitted), are intriguing aspects to
be discussed. Again spin–orbit, spin–spin, gravitomagnetic effects would add to the
complexity and are by no means trivial. The inherent nonlinearities (in the vicinity
of strong fields) have to be considered before any unambiguous detection is claimed.

References
1. C. W. Misner, K. S. Thorne and J. A. Wheeler, Gravitation (Freeman, 1973).
2. J. H. Taylor and J. M. Weisberg, Astrophys. J. 253 (1982) 908.
3. J. H. Taylor and J. Weisberg, Astrophys. J. 345 (1989) 434.
Int. J. Mod. Phys. D 2015.24. Downloaded from www.worldscientific.com

4. J. Taylor and T. Damour, Phys. Rev. D 45 (1992) 1840.

5. C. Sivaram, Int. J. Mod. Phys. D 18 (2009) 2167.
by 187.134.87.192 on 02/16/16. For personal use only.

6. www.ligo.caltech.edu.
7. www.virgo.infn.in; lisa.nasa.gov/.
8. C. Peters and J. Mathews, Phys. Rev. 131 (1963) 435.
9. C. Sivaram, Lectures on black hole physics, in Proc. NATO Advances Study Institute,
ed. Z. Zhang (Kluwer Academic, Boston, 1991).
10. V. de Sabbata and C. Sivaram, Spin in Gravitation, 2nd edn. (World Scientific and
Imperial College Press, 2002).
11. C. Sivaram and K. Sinha, Phys. Rep. 51 (1979) 111.
12. C. Sivaram, Mod. Phys. Lett. A 14 (1999) 2363. arXiv:1402.5071.
13. B. McKernam, K. E. S. Ford, B. Kocsis and Z. Haiman, MNRAS 445 (2014) L74.
14. C. Sivaram, Basic Plasma Processes on the Sun, eds. E. R. Priest and V. Krishan,
Vol. 142 (Kluwer Academic Publishers, 1990), p. 62.
15. B. Bertotti and C. Sivaram, Nuovo Cimento B Serie 106 (1991) 1299.
16. B. Sreenath et al., Astrophys. Space Sci. 345 (2013) 209.
17. C. Sivaram and K. Arun, Adv. High Energy Phys. 2014 (2014) 8, doi:10.1155/2014/
924848.
18. X. B. Wu, F. Wang, X. Fan et al., Nature 518 (2015) 512.
19. C. Sivaram, Quasars, IAU Symposium, eds. G. Swarup and V. K. Kapahi, Vol. 119
(1986).

1544023-6