GENERAL ⎜ ARTICLE

An Introduction to the Raychaudhuri Equations

Sayan Kar

T h e R a y ch a u d h u ri e q u a tio n s a re ¯ rst in tro d u c e d

th ro u g h sim p le e x a m p le s a n d illu stra tio n s. T h e ir
u se a n d th e re su ltin g c o n se q u e n c e s in c o sm o lo g y
a re th e n b rie ° y o u tlin e d .
1 . P re a m b le : A n a lo g ie s a n d Illu stra tio n s
E q u ation s, id en tities a n d in eq u alities p rov id e a q u a n ti-
tative realisatio n o f scien ti¯ cally in terestin g p h en o m en a . Sayan Kar is on the faculty
of the Department of
T h e level o f g en era lity of an eq u ation (or id en tity, in -
Physics and Meteorology,
eq u a lity ) is u su a lly u n d ersto o d b y ask in g th e q u estion : IIT Kharagpur and is also
h ow u b iq u itou s is it? T a ke fo r ex am p le L ap lace's eq u a - associated with the Centre
tio n r 2 Á = 0 . It is w ell k n ow n th a t su ch a n eq u a tio n for Theoretical
arises in m a n y co n tex ts { tw o o f th e m ost p rom in en t Studies there. His main
research interest is in
on es b ein g electricity an d grav ity.
gravitational physics.

In p h y sics, if yo u lo o k carefu lly, th ere aren 't to o m a n y

su ch u b iq u itou s eq u ation s. T h o se w h ich are, h av e sto o d
th e test o f tim e a n d h ave b lo o m ed in m an y an u n ex -
p ected scen a rio .
In 1 955 , w h ile w ork in g on co sm ology, A m a l K u m a r R ay -
ch a u d h u ri d iscovered on e su ch eq u a tion . T o d ay, th is
eq u a tio n (w h ich b ea rs h is n a m e) h as fou n d its u se in
scen arios w h ich h e, p rob ab ly n ever im ag in ed . T h a t is
b ecau se, th o u gh cosm o lo gy is w h ere it all b eg an , th e
R ay ch au d h u ri eq u a tio n s are, a t a very b asic level, la rgely
geom etric/m ath em atica l statem en ts. H en ce, yo u can
im a gin e a p p ly in g th em in m an y d iv erse situ a tion s w h ich
are fa r rem ov ed fro m th e d o m a in s o f co sm o lo g y.
Keywords
L et u s lo o k a t so m e o f th ese situ a tio n s n ow . T h e ¯ rst o f Raychaudhuri equations, de-
th ese is n ot an ap p lica tion in th e tru e sen se, b u t a n ice formable media analysis, fluid
w ay o f in tro d u cin g w h at th e actu al eq u ation s in ten d to flows, Einstein’s general relativ-
tell u s. ity, geodesics, cosmology.

RESONANCE ⎜ April 2008 319

 GENERAL ⎜ ARTICLE

(a) (b)

Figure 1. Vector field of a . T h e T ea c h e r -in -th e -c la s sro o m A n a lo g y

look-at directions of the stu-
dents’ eyes for (a) a good Im a g in e a lectu re h a ll (say, a rea lly b ig o n e) fu ll o f stu -
lecture (b) a bad lecture. d en ts listen in g w ith ra p t a tten tio n to a n in terestin g ta lk .
T h in k o f little v ecto rs w h ich tell u s in w h ich direction a
stu d en t is lo o k in g , a t a n y m o m en t o f tim e. M a p th ese
v ecto rs o n sep a ra te sh eets o f p a p er (o n e fo r ea ch in -
sta n ce o f tim e d u rin g th e ta lk ), m a rk th e lo ca tio n o f
th e tea ch er a n d g iv e it to a cu rio u s frien d w h o w a n ts to
k n ow if th e ta lk w a s g o o d o r b a d . A t a ll in sta n ces, th e
frien d ¯ n d s th a t th ese a rrow s p o in t tow a rd a co m m o n
lo ca tio n (n ea r th e b la ck b o a rd )). H e co n clu d es th a t th e
ta lk m u st h av e b een in terestin g { o th erw ise w h y w o u ld
ev ery o n e k eep lo o k in g a t th e b la ck b o a rd co n tin u o u sly ?
In F igu re 1 (a ) a n d (b ), w e sh ow th e b eh av io u r o f th is
v ecto r ¯ eld o f look-at d irectio n s.
In (a ) w e h av e th e ca se w h en th ese a rrow s focu s to w a rd s
a p o in t (a g o o d tea ch er!).
In (b ) th ey seem h a p h a za rd ly d istrib u ted (th e u n in ter-
estin g o r b a d tea ch er!).
In su m m a ry, fo r a n in terestin g lectu re th e lo o k -a t d irec-
tio n s focu s tow a rd s th e b la ck b o a rd .
b . D e fo r m a b le M ed ia
O u r seco n d ex a m p le co n cern s d efo rm a b le m ed ia . S u p -
p o se y o u a re g iv en a ru b b er d isc. Y o u a re a sk ed to

320 RESONANCE ⎜ April 2008

 GENERAL ⎜ ARTICLE

d efo rm it w ith a n ex tern a l fo rce a n d th en leav e it a lo n e,

i.e., let th e d efo rm a tio n ev o lv e. D efo rm in g it is o f co u rse
ea sy, b u t h ow to q u a n tify th e w ay th e d efo rm a tio n ev o lv es
in tim e? T o q u a n tify, y o u n eed to lea rn w h a t a re th e
ty p es o f d efo rm a tio n s w h ich m ay co m b in e to g iv e y o u a
co llectiv e e® ect. F irstly, th in k o f th e situ a tio n w h en y o u
g iv e a ra d ia lly in w a rd / o u tw a rd fo rce a lo n g th e circu m -
feren ce o f th e d isc. T h e d isc w o u ld th en eith er iso tro p i-
ca lly ex p a n d o r co n tra ct. Its sh a p e w ill n o t ch a n g e. O n
th e o th er h a n d , y o u m ay g iv e a fo rce su ch th a t th ere is
a sh ea rin g o f th e d isc in to a n ellip tica l sh a p e. O r, th ere
co u ld b e a tw ist (a ro ta tio n ) o f th e d isc. In g en era l,
th ese a re th e o n ly th ree essen tia l w ay s in w h ich y o u ca n
d efo rm th e d isc in tia lly a n d leav e it a lo n e (let it ev o lv e).
T h u s, th e in itia l co n d itio n o n th e d efo rm a tio n m ay b e
a n in itia l ex p a n sio n , a n in itia l sh ea r o r a n in itia l ro ta -
tio n o r a co m b in a tio n o f a ll o r so m e. F igu re 2 illu stra tes
th e co n cep ts o f ex p a n sio n , sh ea r a n d ro ta tio n in th is
ex a m p le.
Isotropic expansion

A A >A
1 2 1

Area at O 1 Area at O2

Shear

A A

Area at O 1 Area at O 2

Rotation
Figure 2. The defining fea-
tures of expansion, shear
A A
and rotation of the disc as
mentioned in the second
Area at O Area at O 2
1 example.

RESONANCE ⎜ April 2008 321

 GENERAL ⎜ ARTICLE

c . F lo w in g F lu id s
Area enclosing fluid flow lines
F in a lly, let u s lo o k a t ° ow in g ° u id s. F low in g ° u id s a re
rep resen ted b y th e v elo city v ecto r ¯ eld d e¯ n ed a t ea ch
p o in t in th e ° u id . T h e v elo city ¯ eld is a lw ay s ta n g en -
tia l to th e so -ca lled ° u id ° ow lin es. C o n sid er th e cro ss-
sectio n a l a rea o rth o g o n a l to ° u id ° ow a t so m e va lu e o f
a p a ra m eter (say ¸ ) w ith w h ich p o in ts in th e ° ow lin es
Figure 3. A set of fluid flow
a re la b elled . T a k e a ¯ x ed n u m b er o f ° ow lin es p a ssin g
lines enclosing cross- th ro u g h th is a rea a n d lo o k a t th e sa m e set o f ° ow lin es
sectional areas at two dif- a t a n o th er lo ca tio n (a n o th er va lu e o f ¸ ) in th e ° u id .
ferent values of the param- D o es th e a rea ex p a n d / co n tra ct, d efo rm in sh a p e o r g et
eter λ. Note that the flow tw isted ? If it d o es, th en w e h av e a situ a tio n sim ila r to
has a positive expansion: th e ca se o f d efo rm a b le m ed ia . W e ca n th erefo re ch a r-
the area at λ1 is smaller a cterise th e ° ow in g ° u id b y lo o k in g a t h ow th e cro ss-
than that at λ2.
sectio n a l a rea ch a n g es w ith th e ° ow o f th e ° u id . F igu re
3 p rov id es a n illu stra tio n o f th e ° ow lin es a n d th e cro ss-
sectio n a l a rea s a t tw o d i® eren t va lu es o f ¸ . D i® eren tia l
eq u a tio n s g ov ern in g th e k in em a tic ev o lu tio n o f th e ex -
p a n sio n , sh ea r a n d ro ta tio n ca n th erefo re tell u s m o re
a b o u t th e n a tu re o f ° u id ° ow . In o th er w o rd s, if y o u
k n ow th e in itia l va lu es o f th e ex p a n sio n , sh ea r a n d ro ta -
tio n , y o u ca n d eterm in e th em a t a la ter sta g e b y so lv in g
th ese d i® eren tia l eq u a tio n s.
In a ll th e th ree ex a m p les g iv en a b ov e, w e n o te a co m m o n
fea tu re { n a m ely, th e n a tu re o f ev o lu tio n o f a v ecto r ¯ eld
a lo n g a p a ra m etrised d irectio n . T h e p a ra m eter co u ld b e
sp a tia l (a s in th e ¯ rst ex a m p le) o r it co u ld b e tim e (a s
in th e seco n d o n e) o r a n a rb itra ry p a ra m eter (a s in th e
la st ex a m p le). T h e eq u a tio n s g ov ern in g th e ev o lu tio n o f
The equations
th e q u a n tities su ch a s ex p a n sio n / co n tra ctio n , sh ea rin g
governing the
o r tw istin g o f su ch ° ow lin es (a s in th e ° u id ca se) a re
evolution of the
th e R ay ch a u d h u ri eq u a tio n s (in a so m ew h a t g en era lised
expansion, shear
sen se). W e sh a ll m a k e th ese id ea s m o re p recise b elow .
or rotation along
the flow lines are 2 . D e v e lo p in g th e D e fo rm a b le M e d ia A n a lo g y
the Raychaudhuri
L et u s n ow d ev elo p th e secon d scen a rio g iv en in th e
equations.

322 RESONANCE ⎜ April 2008

 GENERAL ⎜ ARTICLE

p rev io u s sectio n , i.e., th e ca se o f d efo rm a b le m ed ia . A s

w e sa id b efo re, w e a re n o t in terested in a sk in g h ow th e
d efo rm a tio n o ccu red . T h a t is to say { w e a re in th e d o -
m a in o f k in em a tics. T h e o b v io u s n ex t q u estio n w o u ld
b e: W h a t a re th e k in em a tica l q u a n tities? W e h av e a l-
rea d y lea rn t a b o u t th e ex p a n sio n , sh ea r a n d ro ta tio n .
W e n ow d e¯ n e th em a b it m o re p recisely.
C o n sid er a tw o -d im en sio n a l m ed iu m (ex a m p le a sh eet
o f ru b b er). Im a g in e th a t u n d er so m e in itia l ex tern a l
stress th e sh eet is d efo rm ed a t a n a rb itra ry tim e t =
0 . T h is stress p ro d u ces th e in itia l d efo rm a tio n w h ich
a cts a s a set o f in itia l co n d itio n s. T h e stress is th en
rem ov ed a n d th e sy stem is a llow ed to ev o lv e in tim e.
W h a t h a p p en s su b seq u en tly ? T o a n sw er th is q u estio n
w e n eed to so lv e th e co rresp o n d in g ev o lu tio n eq u a tio n s
fo r th e d efo rm a tio n .
D en o te th e d efo rm a tio n b y a v ecto r » i (w h ere i = 1 ,2 ,
i.e., w e h av e a d efo rm a tio n v ecto r w ith co m p o n en ts » 1
a n d » 2 a lo n g th e d irectio n s 1 a n d 2 resp ectiv ely ). T h e
tim e ra te o f ch a n g e o f » i ca n b e a n a ly sed a s fo llow s: F o r
sm a ll tim e in terva ls w e m ay w rite

» i(t1 ) = » i(t0 ) + ¢ » i(t0 ); (1 )

w h ere
¢ » i = B ji (t0 )» j (t0 )¢ t + O ((¢ t)2 ); (2 )
a n d th e in d ex j is su m m ed ov er (eg . ¢ » 1 = B 11 » 1 + B 21 » 2 :
a n d , sim ila rly fo r ¢ » 2 ). T h erefo re, w e m ay w rite

d »i
= B ij (t)» j ; (3 )
dt
w h ere B ij (t) is a n a rb itra ry seco n d ra n k ten so r a n d o n ce
a g a in , th e rep ea ted in d ices (j , fo r in sta n ce, in th e a b ov e
1
eq u a tio n s) m ea n s w e h av e to su m o v er th em , eg . dd»t =
B 11 » 1 + B 12 » 2 a n d so o n . If y o u a re n o t fa m ilia r w ith
ten so rs, th in k o f th e B ij a s a 2 £ 2 m a trix w ith elem en ts

RESONANCE ⎜ April 2008 323

 GENERAL ⎜ ARTICLE

B 11 (t), B 12 (t), B 21 (t) a n d B 2j (t) ch a ra cterisin g th e tim e

ev o lu tio n o f th e d efo rm a tio n v ecto r.
O u r sch em e w o u ld n ow b e a s fo llow s:
S tep (i). D i® eren tia te th e a b ov e eq u a tio n o n ce m o re w rt
tim e.
i
S tep (ii). C o n v ert it to a n eq u a tio n fo r B j (t) a n d its
¯ rst tim e d eriva tiv e w ith »Äi o n its R H S .
S tep (iii). U se th e eq u a tio n o f m o tio n fo r » i a s p rescrib ed
la ter a n d a lso eq u a tio n (3 ) to a rriv e a t th e ¯ n a l resu lt.
Im p lem en tin g th e S tep s (i) a n d (ii) a b ov e (S tep (iii) w ill
b e ca rried o u t la ter) w e o b ta in :
µ i ¶
dB j
+ B k B j » j = »Äi;
i k
(4 )
dt
w h ere, a s b efo re, th e in d ex k is su m m ed ov er.
A . E x p a n sio n , R o ta tio n a n d S h ea r
B efo re w e w rite d ow n th e ev o lu tio n eq u a tio n s let u s ¯ rst
sta te th e q u a n tities o f in terest. T h e a rb itra ry seco n d
ra n k ten so r B ij ca n b e d eco m p o sed in to its tra ce, sy m -
m etric tra celess a n d a n tisy m m etric p a rts w h ich w ill co n -
stitu te th e iso tro p ic ex p a n sio n (sca la r), sh ea r (sy m m et-
ric tra celess ten so r) a n d ro ta tio n (a n tisy m m etric ten -
so r).
1
B ij = µ ± ij + ¾ ij + ! ij : (5 )
2
W e ca n ex p licitly w rite th e a b ov e ex p ressio n in term s o f
Any arbitrary th e fo llow in g 2 £ 2 m a trices:
second rank tensor µ 1 ¶ µ ¶ µ ¶
µ 0 ¾ ¾ £ 0 ¡ !
can be split into its £ ´ 2 ;§ ´ +
;− ´ ;
0 12 µ ¾£ ¡ ¾+ ! 0
trace, symmetric
(6 )
traceless and
w h ere w e h av e n a m ed th e sh ea r co m p o n en ts a s ¾ + a n d
antisymmetric
¾ £ a n d th e o n ly ro ta tio n co m p o n en t a s ! . N o te th a t
parts.
th e m a trix £ h a s tra ce µ , th e m a trix § is tra celess a n d

324 RESONANCE ⎜ April 2008

 GENERAL ⎜ ARTICLE

sy m m etric, w h ile th e m a trix − is a n tisy m m etric. T h u s, The Raychaudhuri

th e fo u r q u a n tities µ , ¾ + , ¾ £ a n d ! ch a ra cterise a n y equations
d efo rm a tio n o f th e tw o -d im en sio n a l d efo rm a b le m ed iu m . constitute a
T h e fu llB m a trix w ith elem en ts B ij ca n b e w ritten d ow n coupled, first
b y ju st a d d in g th e a b ov e th ree m a trices (i.e., m a trix
order, nonlinear
B = £ + § + − ).
system of ordinary
E x e rc is e 1 . C o n sid er a circle o f ra d iu s a . A n y p o in t o n differential
th e circle ca n b e p a ra m etrised a s (a co s Á ;a sin Á ). F in d equations.
o u t h ow ex p a n sio n , sh ea r a n d ro ta tio n ca n d efo rm th e
circu la r b o u n d a ry. Y o u h av e to ¯ n d d » i in ea ch ca se { d o
it sep a ra tely, o n ce w ith th e ex p a n sio n m a trix , th en w ith
th e sh ea r m a trix a n d ¯ n a lly w ith th e ro ta tio n m a trix .
Ig n o re o rd er (¢ t)2 term s.
B . T h e E v o lu tio n (R a y c h a u d h u r i) E q u a tio n s
W e n ow u se th e in p u ts fro m th e p rev io u s su b sectio n to
rew rite th e ev o lu tio n eq u a tio n fo r B ij a s fo u r co u p led
n o n lin ea r ¯ rst o rd er eq u a tio n s in v o lv in g th e d ep en d en t
va ria b les µ , ¾ + , ¾ £ a n d ! . B efo re th a t, o f co u rse, w e
n eed to w rite d ow n th e eq u a tio n o f m o tio n w e a ssu m e
fo r Ȁi. T h is, in a lin ea r a p p rox im a tio n , is ta k en a s
»Äi = ¡ K i
j » j ¡ ¯ »_i ; (7 )
w h ere ¯ a n d K ij d en o te d a m p in g a n d sti® n ess co e± -
cien ts resp ectiv ely a n d th e m a trix K (w h o se elem en ts
a re th e K ij ) is a ssu m ed to b e o f th e fo rm :
µ ¶
k + k+ k£
K ´ : (8 )
k£ k ¡ k+
T h e ¯ rst term in (7 ) is n o th in g m o re th a n th e w ell-
k n ow n H o o k e's L aw 1 . It lo o k s a b it d i® eren t b eca u se 1
Hooke’s law of elasticity states
w e h av e in clu d ed th e p o ssib ility o f a n iso tro p ic stresses that the amount by which a body
is deformed is proportional to
(»Ä1 n o t ju st eq u a l to ¡ K 11 » 1 ¡ ¯ »_1 b u t, ca n b e o f th e the force causing the deforma-
fo rm »Ä1 = ¡ K 11 » 1 ¡ K 12 » 2 ¡ ¯ »_1 , w h ere K 12 en co d es th e tion.
e® ect o f a n iso tro p y ).
F in a lly, im p lem en tin g S tep (iii) m en tio n ed ea rlier (i.e.,
th e u se o f th e eq u a tio n o f m o tio n (7 ) fo r » i a n d (3 ) in (4 ),

RESONANCE ⎜ April 2008 325

 GENERAL ⎜ ARTICLE

an d th e elim in a tio n of » i from b oth sid es), th e eq u a tio n s

tu rn o u t to b e:
1
µ_ + µ 2 + ¯ µ + 2 k + 2 (¾ +2 + ¾ £2 ¡ ! 2 ) = 0 : (9)
2
¾_+ + (¯ + µ ) ¾ + + k + = 0 : (1 0)
¾_£ + (¯ + µ ) ¾ £ + k £ = 0 : (1 1)
!_ + (¯ + µ ) ! = 0 : (1 2)
W e n ote th a t th e eq u ation s for ¾ + , ¾ £ an d ! are stru c-
tu rally sim ilar w h erea s th e equ atio n for µ is k n ow n in
th e literatu re as a R iccati d i® eren tial eq u a tio n . F u rth er,
th e eq u a tio n s con stitu te a cou p led , ¯ rst ord er, n o n lin ear
sy stem .
E x e rc is e 2 . O b tain eq u ation s (9){ (12 ) b y u sin g eq u a -
tio n s (3 ), (4), (7) a n d (8). T h e w ay to d o it is to rew rite
th e eq u ation (4 ) u sin g eq u a tion s (7) an d (3 ) ¯ rst. T h is
dB i
w ill tu rn ou t to b e d t j + B ik B kj = ¡ K ij ¡ ¯ B ij a fter
elim in atin g » i fro m b oth sid es. T h en , ex p licitly w rite
dB 1 dB 1 dB 2 dB 2
d ow n eq u ation s for d t 1 , d t 2 , d t 1 an d d t 2 . F in a lly,
u se eq u ation (6) an d th e com b in a tio n s ddt (B 11 § B 22 ) a n d
d
dt
(B 12 § B 21 ) to get th e fo u r eq u a tio n s (9 ){(12 ).
S olv in g th e eq u ation s (9 ){(12), o n e ca n d eterm in e h ow
th e q u an tities µ , ¾ + , ¾ £ a n d ! evo lv e in tim e. T h ou gh
com p licated , th e eq u ation s ca n in d eed b e solved . A gen -
eralisa tio n to th ree d im en sion s ca n also b e w o rked ou t.
T h e in terested read er can lo ok u p th e seco n d referen ce
for fu rth er d eta ils.
E x e rc is e 3 . A ssu m e k + an d k £ a s zero . A lso take
¯ = 0 . S h ow th a t if on e d e¯ n es a q u an tity I = ¾ +2 +
¾ £2 ¡ ! 2 , th en th e fou r eq u ation s can b e reca st in to tw o
eq u a tio n s in vo lv in g I a n d µ . T ry to so lv e th ese eq u a tio n s
an d o b tain th e evolu tion of th e ex p an sion , sh ea r a n d
rota tio n .

326 RESONANCE ⎜ April 2008

 GENERAL ⎜ ARTICLE

3 . F lu id F lo w s The behaviour of
A s m en tio n ed in th e ¯ rst sectio n , th e k in em a tics o f ° u id the expansion
° ow ca n b e u n d ersto o d u sin g a sim ila r fo rm a lism . In along the flow
o rd er to m a k e co n ta ct w ith co sm o lo g y w e n ow b rin g in would tell us
th e rela tiv istic fo rm a lism . T h e th in g s n eed ed to k n ow whether the flow
a b o u t su ch a fo rm a lism a re listed in B ox 1 . lines come
towards each other
C o m p a rin g w ith th e ca se o f d efo rm a b le m ed ia , w e n o -
or move away.
tice th a t th e d efo rm a tio n v ecto r » i is n ow rep la ced b y
th e v elo city v ecto r v i, w h ere i h ere ru n s fro m 0 to 3 .
A lso , w e a ssu m e v i to b e tim e-lik e (see B ox 1 fo r th e
d e¯ n itio n o f tim e-lik e).' W e co n sid er o n ly th o se ° ow
lin es w h ich a re g eo d esic, in th e sen se th a t v i a re ta n -
g en ts to g eo d esics. T h e g ra d ien t o f th e v elo city v ecto r
i
is o f co u rse @@ xv j . W h en w e h av e a cu rv ed sp a ce-tim e (re-
ca ll w h a t E in stein ta u g h t u s in G en era l R ela tiv ity : th e
g rav ita tio n a l ¯ eld is eq u iva len t to a cu rv ed sp a ce-tim e,
see B ox 1 fo r fu rth er d eta ils), th e u su a l p a rtia l d eriv -
a tiv e is rep la ced b y a cova ria n t d eriva tiv e, d en o ted b y
r j . T h u s, in su m m a ry, th e q u a n tity r j v i is a seco n d
ra n k ten so r a n d ca n b e sp lit, a s b efo re in to th e tra ce,
sy m m etric tra celess a n d a n tisy m m etric p a rts. T h ese a re
o n ce a g a in , th e ex p a n sio n , sh ea r a n d ro ta tio n .
S o , th e eq u a tio n s o f ev o lu tio n o f th e ex p a n sio n , sh ea r
a n d ro ta tio n a lo n g th e ° ow lin es ca n ea sily b e o b ta in ed
b y fu rth er d i® eren tia tin g r j v i, u sin g th e fa ct th a t th e
tra jecto ries a re g eo d esic a n d th en ex tra ctin g th e tra ce,
sy m m etric tra celess a n d a n tisy m m etric p a rts o n b o th
sid es (reca ll, th ese a re ex a ctly th e sa m e a s S tep s (i),
(ii) a n d (iii) m en tio n ed ea rlier, fo r d efo rm a b le m ed ia ).
T h is is a b it co m p lica ted to w o rk th ro u g h , so w e m en tio n
w h a t th e en d resu lt is a n d th en d iscu ss th e co n seq u en ces.
W e lo o k a t o n ly th e eq u a tio n fo r th e ex p a n sio n µ , b e-
ca u se th e o th er eq u a tio n s a re m o re co m p lica ted a n d a lso
b eca u se R ay ch a u d h u ri, in h is o rig in a l p a p er, d id ch o o se
to d iscu ss th is eq u a tio n in g rea ter d eta il.

RESONANCE ⎜ April 2008 327

 GENERAL ⎜ ARTICLE

B ox 1.

S p a c e -tim e : In special relativity, time t is elevated to the status of another coordinate.

tx y z constitutes a four-dimensional space-time continuum. One talks about space-time
and not space and time. The in¯nitesimal distance between two points in space-time is
given as ds 2 = ¡ c 2 dt2 + dx 2 + dy 2 + dz 2 (Minkowski distance), which can be positive
(space-like), negative (time-like) or zero (null). This is unlike the usual distance we are
familiar with, i.e., ds 2 = dx 2 + dy 2 + dz 2 which is always positive.

T ra n sfo r m a tio n s: To go from one inertial frame tx y z to another t0x 0y 0z 0, one has to
use the Lorentz transformations and n o t the Galilean transformations. The Minkowski
distance mentioned above remains unchanged under a Lorentz transformation of the
coordinates.

F o u r v e c to r : Like space-time co-ordinates, vectors also have a fourth component, hence

the name four vector. For instance, the q velocity four vector now has a fourth component
2
which is nothing but ° c (with ° = 1= 1 ¡ vc 2 ). Four vectors will transform via Lorentz
transformations and can have length as positive (space-like), negative (time-like) or zero
(null).

C u rv e d sp a c e -tim e : General Relativity goes further in order to incorporate gravity via

the Equivalence Principle. Locally inertial frames (locally like those of Special Relativity)
encode the global e®ects of gravity through a nontrivial, curved (as opposed to °at,
which is the case with the geometry de¯ned with the Minkowski distance mentioned at
the beginning of this box) space-time geometry. Between two such locally inertial frames
one transforms using completely general coordinate transformations.

G e o d e sic s: Geodesics are trajectories of test particles in a given gravitational ¯eld.

Since such ¯elds are equivalent to a curved space-time, the trajectories are those of
extremum distance in a given metric geometry (geometry is speci¯ed through the distance
function, e.g., on a two-dimensional sphere of radius a , the distance (in¯nitesimal) is
ds 2 = a 2 (dµ 2 + sin 2 µ dÁ 2 ) = g µ µ dµ 2 + g Á Á dÁ 2 ! the metric functions, g µ µ = a 2 and
g Á Á = a 2 sin2 µ ) . The tangent to a geodesic curve, at a speci¯c point on it, is a four
vector.

C u rv a tu r e : A measure of the curvature of a geometry is the Riemann curvature, which

involves second derivatives of the metric functions. Among derived quantities is the
Ricci tensor (the R ij is the Ricci tensor which appears in the Raychaudhuri equation).
Einstein's equations of gravity (the equations which replace the Newtonian equation
r 2 Á = 4¼ G ½ m ) may be written non-technically as G eo m etry = M a tter, where the geom-
etry part involves quantities constructed out of the Ricci tensor and the metric functions.

T h e eq u a tio n fo r th e ex p a n sio n is in d eed ra th er sim -

p le. W e h av e a lrea d y w ritten a sim ila r eq u a tio n (lo o k

328 RESONANCE ⎜ April 2008

 GENERAL ⎜ ARTICLE

a t eq u a tio n (9 ) w ith ¯ = 0 ) w h ile d iscu ssin g a b o u t d e-

The good teacher/
fo rm a b le m ed ia . H ere it is a g a in :
talk is like gravity
dµ 1 and the fact that
+ µ 2 + ¾ 2 ¡ ! 2 = ¡ R ij v iv j ; (1 3 )
d¸ 3 students’ eyes
focus towards the
w h ere ¾ 2 = ¾ ij ¾ ij a n d ! 2 = ! ij ! ij . T h e q u a n tity board is like
th a t a p p ea rs o n th e R H S is, essen tia lly, `g eo m etric' (o r, focusing of
eq u iva len tly fo llow in g E in stein , is rela ted to th e `m a t-
trajectories.
ter', w h ich cu rv es g eo m etry ). N o te th a t th is term ca n
b e th o u g h t o f a s th e p a ra llel o f 2 k in eq u a tio n (9 ). T h u s,
ju st a s k in eq u a tio n (9 ) ca n b e a fu n ctio n o f t (tim e-
d ep en d en t sp rin g co n sta n t), th e R ij v iv j term is, in g en -
era l, a fu n ctio n o f th e p a ra m eter ¸ . T h e ex p a n sio n , fo l-
low in g th e ° u id ° ow scen a rio , w o u ld tell u s w h eth er th e
° ow lin es co m e to g eth er, o r g o a p a rt fro m ea ch o th er.
¸ is a p a ra m eter th a t la b els p o in ts o n th e ° ow lin es.
S o , if th e ° ow lin es co m e clo ser, th e q u estio n is `w h a t
is ca u sin g th is to h a p p en ?'. T h ere is n o th in g else w h ich
ca n d o it o th er th a t th e term o n th e R H S a n d th e th ird
a n d fo u rth term s o n th e L H S in eq u a tio n (1 3 ). If, fo r
th e tim e b ein g w e set th e ro ta tio n to zero , w e w ill sh ow
b elow th a t it is th e a ttra ctiv e n a tu re o f g rav ity w h ich
ca u ses th e lin es to fo cu s. S o , g o in g b a ck to th e ¯ rst
a n a lo g y, w e m ig h t say th a t th e id ea o f a `g o o d tea ch er'
is rep la ced b y `g rav ity ' a n d th e fa ct `th e stu d en ts ey es
fo cu s to th e b la ck b o a rd ' is rep la ced b y `° u id ° ow lin es fo -
cu s tow a rd s a p o in t'. T h is is w h a t R ay ch a u d h u ri ta u g h t
u s { in d eed a v ery sim p le fa ct { th a t g rav ity, b y v irtu e
o f b ein g a ttra ctiv e, ca u ses g eo d esics (° ow lin es) to fo cu s
tow a rd s a p o in t. L et u s n ow see w h eth er w e ca n m a k e
th is a little m o re q u a n tita tiv e, b y a n a ly sin g th e eq u a tio n
(1 3 ).
A s w e sa id b efo re, w e w ill set th e ro ta tio n term to zero .
W e a lso m a k e a tra n sfo rm a tio n
3 dF
µ= : (1 4 )
F d¸

RESONANCE ⎜ April 2008 329

 GENERAL ⎜ ARTICLE

W ith th is tra n sfo rm a tio n th e eq u a tio n b eco m es

The divergence of
an initially negative d2F 1£ i j ¤
θ to negative + R ij v v + ¾ 2 F = 0 : (1 5 )
d¸ 2 3
infinity within a
finite value of λ is T h e co n ten ts in th e sq u a re b ra ck ets ca n b e term ed a s
focusing.
so m e fu n ctio n o f ¸ . N o te th a t th e v elo city v ecto r is
o b v io u sly a fu n ctio n o f ¸ a n d th e elem en ts o f R ij a re th e
fu n ctio n s o f th e co o rd in a tes w h ich a re, in tu rn , fu n ctio n s
of ¸ .
S o , th e eq u a tio n s resem b le so m eth in g v ery fa m ilia r to
a ll o f u s { a h a rm o n ic o scilla to r eq u a tio n w ith a tim e
d ep en d en t sp rin g co n sta n t:

d2F
+ k (¸ )F = 0: (1 6 )
d¸ 2
H ow ev er, it is a h a rm o n ic o scilla to r ty p e eq u a tio n o n ly
if k is a p o sitiv e q u a n tity. T h is m ea n s R ij v iv j + ¾ 2 ¸ 0 .
O r, w ith ¾ 2 ¸ 0 , w e h av e th e req u irem en t R ij v iv j ¸ 0 .
N ow , o n e m ay a sk { w h a t d o w e g et o u t o f th is h a rm o n ic
o scilla to r-lik e eq u a tio n ? T o u n d ersta n d th is, w e n eed to
g o b a ck to th e d e¯ n itio n o f µ = F3 dd F¸ . If F is o scilla to ry,
th en F w ill h av e zero s a t ¯ n ite va lu es o f ¸ , (rem em b er,
fu n ctio n s lik e th e sin e o r co sin e h av e zero s a t ¯ n ite va lu es
o f ¸ ). T h is w o u ld im p ly th a t µ ca n d iv erg e to n eg a tiv e
in ¯ n ity if, in itia lly (a t say, ¸ = 0 ), µ is n eg a tiv e. N ow
reca ll th a t µ is a q u a n ti¯ er o f th e iso tro p ic ex p a n sio n .
T h erefo re w e m ay w rite µ a s

A (¸ 2 ) ¡ A (¸ 1 )
µ = ; (1 7 )
A (¸ 2 )

w h ere A (¸ ) is th e cro ss-sectio n a l a rea en clo sin g a ¯ x ed

n u m b er o f ° ow lin es. T h u s, if A (¸ 2 ) g o es to zero (i.e.,
th e co n g ru en ce co lla p ses to a p o in t), th en µ g o es to n eg -
a tiv e in ¯ n ity. T h is is g eo d esic fo cu sin g a n d it h a p p en s
b eca u se g rav ity is a ttra ctiv e. T h e co n d itio n R ij v iv j ¸ 0
ca n b e sh ow n , u sin g E in stein 's eq u a tio n s o f G en era l

330 RESONANCE ⎜ April 2008

 GENERAL ⎜ ARTICLE

R ela tiv ity a s rela ted to `g rav ita tin g m a tter', w h ich is
a ttra ctiv e in n a tu re. L o o sely sp ea k in g , R ij v iv j is so m e-
w h a t lik e r 2 Á (w ith Á a s th e g rav ita tio n a l p o ten tia l).
T h erefo re, u sin g P o isso n 's eq u a tio n fo r N ew to n ia n g rav -
ity, r 2 Á is p ro p o rtio n a l to th e m a ss d en sity ½ m , w h ich
a s w e a ll k n ow , is a lw ay s ¸ 0 .
E x e rc is e 4 . A ssu m in g ! ij = 0 a n d R ij v iv j ¸ 0 , o n e
ca n reca st th e R ay ch a u d h u ri eq u a tio n a s a n in eq u a lity
dµ
d¸
+ 13 µ 2 · 0 . A n a ly se th is in eq u a lity to sh ow th a t if µ
a t ¸ 0 is n eg a tiv e, th en µ m u st ten d to n eg a tiv e in ¯ n ity
w ith in a ¯ n ite va lu e o f ¸ .
Y o u m ig h t b e cu rio u s, a t th is sta g e, a s to w h y w e a re
a lw ay s settin g th e ro ta tio n eq u a l to zero ? If y o u d o n 't,
y o u w ill g et a n o p p o site e® ect, b eca u se th e sig n o f th a t
term in eq u a tio n (1 3 ) is o p p o site to th a t o f th e ¾ 2 a n d
th e R ij v iv j term s. T h u s ro ta tio n ca n a ctu a lly lea d to
a defocu sin g (g eo d esics m ov in g aw ay fro m ea ch o th er)
e® ect.
4 . C o sm o lo g y
A s m en tio n ed rig h t a t th e b eg in n in g , th e eq u a tio n s o rig -
in a lly a ro se in th e co n tex t o f co sm o lo g y. In a m o d el o f
th e u n iv erse w h ich w e ca ll a co sm o lo g ica l m o d el, th e
u n iv erse is ta k en a s h o m o g en eo u s a n d iso tro p ic (i.e., n o
p referred lo ca tio n s o r d irectio n s) a t sca les in w h ich a
g a la x y is trea ted a s a p o in t. S u ch g a la x ies m a k e u p
th e co sm o lo g ica l ° u id w h ich ° ow s a lo n g g eo d esics in th e
g iv en g eo m etry. T h e ex p a n sio n o f th e u n iv erse is q u a n -
ti¯ ed b y a sca le fa cto r w h ich w e d en o te a s a (t). T h e a (t)
ca n b e o b ta in ed b y so lv in g th e E in stein eq u a tio n s w ith
a n a p p ro p ria te so u rce, u su a lly a p erfect ° u id . S o , it ca n
b e seen , h ow ° u id ° ow en ters th e p ictu re. O b v io u sly,
o n ce w e h a v e a ° ow , w e m u st h av e ex p a n sio n , sh ea r a n d
ro ta tio n . It tu rn s o u t th a t th e ex p a n sio n is g iv en q u ite
sim p ly :
a_ 1 d
µ = 3 = 3 a3: (1 8 )
a a dt

RESONANCE ⎜ April 2008 331

 GENERAL ⎜ ARTICLE

B ox 2.

What is a sin gu la rity? Loosely speaking, it is a location where `things' are ill-de¯ned.
Recall Coulomb's law or Newton's law of gravity and ask yourself { what happens when
you are just at the location of the point charge or the point mass { the potential and
the ¯eld diverges to in¯nity. Einstein taught us that gravity is manifest in the curvature
of space-time { so when the curvature becomes very large, gravity must be very strong
too. More speci¯cally, when certain co-ordinate invariant quantities (scalars) diverge at
some point, we say that we have a cu rva tu re sin gu la rity. The Big Bang and the black
hole are two examples, though they are di®erent in nature. For instance, in the case of
Big Bang, the scalars just mentioned about, vary as inverse powers of the scale factor
a (t) { so when a (t) ! 0 (the universe shrunk to a point in its past, as it is believed to
have been) these quantities diverge. However, note there is another characteristic of a
singularity { geodesics (trajectories) seem to end there, they cannot be continued further
beyond and this happens at a ¯nite value of the parameter that labels points on them.
Such situations are described technically by the term geod esic in co m p leten ess. But it
might happen that geodesic incompleteness can occur without there being any curva-
ture singularity, i.e., those scalars may not diverge at that point beyond which geodesics
cannot be continued. So, its a o n e-w a y statement which we can make { curvature sin-
gularities are locations where curvature diverges and geodesics end (focus) but geodesics
can focus in a completely benign way too (without encountering a curvature singular-
ity). In a nutshell, curvature singularities imply geodesics focusing there, but, geodesic
focusing does not necessarily imply a curvature singularity. Thus, you realise how im-
portant the Raychaudhuri equations must be in understanding these singularities, which,
despite Einstein's utter dislike of them, are inevitably there in his theory of gravity. The
details about singularities (their de¯nition, properties, the conditions under which they
are bound to appear etc.,) is what Hawking and Penrose worked on in the 1960s, early
1970s and one of their tools (among many others) were the Raychaudhuri equations.

If w e ta k e a (t) » tº , th en µ » t¡1 . G o in g b a ck w a rd s
in tim e (i.e., tow a rd s t ! 0 fro m to d ay ), it is th erefo re
ea sy to see th a t w e en d u p w ith a n ex p a n sio n w h ich b e-
co m es n eg a tiv e in ¯ n ity. T h a t's focu sin g { a n d w h ere d o
th ese g eo d esics fo cu s to ? { th e b ig b a n g ! S o , in ev ita b ly,
th e ex isten ce o f a sin g u la rity (see B ox 2 to g et a n id ea
o f a sin g u la rity ) b eco m es o b v io u s irresp ectiv e o f w h a t º
is. F u rth er, lo o k in g a t th e seco n d ex p ressio n fo r µ (i.e.,
µ = a13 ddt a 3 ), o n e ca n ea sily co n clu d e th a t µ is a q u a n tity
w h ich ch a ra cterises th e fra ctio n a l ch a n g e in th e v o lu m e
o f th e th ree-d im en sio n a l sp a ce (v o lu m e = a 3 ). T h ere-
fo re, it is a lso ca lled th e volu m e ex p a n sio n . In fa ct th e
R ay ch a u d h u ri eq u a tio n a n d g eo d esic fo cu sin g g av e th e

332 RESONANCE ⎜ April 2008

 GENERAL ⎜ ARTICLE

¯ rst h in t th a t sin g u la rities a re in ev ita b le in E in stein 's Acknowledgements

G en era l R ela tiv ity. L a ter o n in th e la te 1 9 6 0 s a n d ea rly
1 9 7 0 s, H aw k in g a n d P en ro se p rov ed th e so -ca lled sin g u - The author thanks A Das-
la rity th eo rem s o n th e b a sis o f R ay ch a u d h u ri eq u a tio n s! gupta and B Nath for their
comments.
E x e rc is e 5 : If y o u k n ow h ow to ca lcu la te th e R icci
ten so r R ij th en , a ssu m in g sh ea r a n d ro ta tio n a s zero a n d
v i ´ (1;0;0 ;0 ), sh ow th a t th e µ g iv en in eq u a tio n (1 8 )
sa tis¯ es th e R ay ch a u d h u ri eq u a tio n fo r th e ex p a n sio n .

Suggested Reading Address for Correspondence

Sayan Kar
[1] E Poisson, Relativist’s toolkit: The mathematics of black hole mechanics, Department of Physics and
Cambridge University Press, Ch. 1 and 2, 2004. Centre for Theoretical Studies
[2] A Dasgupta, H Nandan and S Kar, Kinematics of deformable media, Indian Institute of Technology
Annals of Physics, 2008, (to appear). Kharagpur 721 302, India.
[3] S Kar and S Sen Gupta, The Raychaudhuri equations: A brief review, Email:
Pramana – J. Phys., Vol.69, p.49, 2007. sayan@phy.iitkgp.ernet.in
[4] J B Hartle, Gravity: an introduction to Einstein’s General Relativity, sayan@cts.iitkgp.ernet.in
Pearson Education, 2003.

We couldn’t wait for the apple to drop...

RESONANCE ⎜ April 2008 333