SERIES ⎜ ARTICLE

Snippets of Physics
14. The Power of Nothing

T Padmanabhan

T h e v a c u u m sta te o f th e e le c tr o m a g n e tic ¯ e ld is
fa r fr o m tr iv ia l. A m o n g st o th e r th in g s it c a n
e x e rt fo r c e s th a t a r e m e a su r a b le in th e la b , in a
c u r io u s p h e n o m e n o n k n o w n a s C a sim ir e ® ec t.
W e a ll k n ow th a t cla ssica l electro m a g n etic ¯ eld s ca n ex -
T Padmanabhan works at
ert fo rces o n ch a rg ed p a rticles. T h e sta n d a rd ex p ressio n
IUCAA, Pune and is
fo r th e cla ssica l fo rce is q (E + v £ B ) w h ich , o f co u rse, interested in all areas
va n ish es w h en E = B = 0 . T h a t so u n d s em in en tly rea - of theoretical physics,
so n a b le. especially those which
have something to do with
B u t th en , w e k n ow th a t th e rea l w o rld is q u a n tu m m e- gravity.
ch a n ica l in ch a ra cter a n d n o t cla ssica l, w h ich im p lies
th a t w e n eed to trea t th e electro m a g n etic ¯ eld a s a q u a n -
tu m en tity. W h en w e d o th a t, p h o to n s em erg e a s th e
q u a n ta o f electro m a g n etic ¯ eld . A s in th e ca se o f a n y
o th er q u a n tu m sy stem , e.g ., th e h y d ro g en a to m , o n e ca n
d escrib e th e p h y sics in term s o f th e q u a n tu m sta tes o f
th e electro m a g n etic ¯ eld . In th is la n g u a g e, o n e ca n a lso
d e¯ n e a sta te w ith zero p h o to n s w h ich co u ld b e th o u g h t
o f a s th e va cu u m sta te o f th e electro m a g n etic ¯ eld . O n e
w o u ld h av e im a g in ed th a t, if th ere a re n o p h o to n s, th en
th ere w ill b e n o m ea su ra b le p h y sica l e® ects d u e to th e
electro m a g n etic ¯ eld . W h ile th is is m o re o r less tru e {
w h ich is ra th er rea ssu rin g { th ere a re in d eed in terestin g
situ a tio n s in w h ich it is n ot tru e! W e w ill d escrib e o n e
su ch co n tex t, ca lled C a sim ir e® ect, in th is in sta llm en t.
T h e sim p lest, th o u g h a b it id ea lized , d escrip tio n o f C a si-
m ir e® ect is th e fo llow in g . S u p p o se y o u k eep tw o p a ra l- Keywords
lel, p erfectly co n d u ctin g p la tes in th e o th erw ise em p ty Harmonic oscillator, quantum
sp a ce, sep a ra ted b y a d ista n ce L . T h en , th ey w ill a ttra ct mechanics, electromagnetism.

RESONANCE ⎜ February 2009 179

 SERIES ⎜ ARTICLE

Remember that
ea ch o th er w ith a fo rce
there are no net ¼ 2 ~c
F = ¡ A; (1 )
charges put on the 240 L 4
plates; we are not w h ere A is th e cro ss-sectio n a l a rea o f th e p la tes!! R e-
talking about a m em b er th a t th ere a re n o n et ch a rg es p u t o n th e p la tes;
charged parallel w e a re n o t ta lk in g a b o u t a charged p a ra llel p la te ca p a c-
plate capacitor. The ito r. T h e fo rce a cts b etw een tw o p la tes k ep t in th e va c-
force acts between u u m . T h is e® ect w a s p red icted [1 ] b y th e D u tch p h y si-
two plates kept in the cist H en d rick C a sim ir in 1 9 4 8 a n d h a s a ctu a lly b een
vacuum. m ea su red in th e la b [2 ]. O n e n ice w ay o f u n d ersta n d in g
th is resu lt is a s a ta n g ib le fo rce ex erted b y th e electro -
m a g n etic va cu u m . L et u s see h ow .
B efo re la u n ch in g in to m a th em a tics, let m e ex p la in th e
b a sic rea so n fo r th is p h en o m en o n in q u a lita tiv e term s.
C o n sid er th e fa m ilia r ex a m p le o f a h a rm o n ic o scilla to r,
w ith th e H a m ilto n ia n
1
H (p ;q ) = [p 2 + ! 2 q 2 ]: (2 )
2
W e h av e set th e m a ss o f th e p a rticle to u n ity fo r sim p lic-
ity. C la ssica lly, th e m in im u m en erg y fo r su ch a sy stem is
zero (E cla ss = 0 ), w h ich is a ch iev ed w h en q = p = 0 . W e
k n ow , h ow ev er, th a t th is is n o t p o ssib le in q u a n tu m th e-
o ry, essen tia lly b eca u se o f th e u n certa in ty p rin cip le. T o
m in im ize th e p o ten tia l en erg y, w e n eed to set q = 0 ; b u t
if w e k n ow th e p o sitio n to su ch in ¯ n ite p recisio n , th e
m o m en tu m w ill b e in ¯ n itely u n certa in a n d w e ca n n o t
g u a ra n tee a low va lu e fo r p 2 = 2 ! S o to m in im ize th e to ta l
en erg y, w e n eed to a llow fo r so m e a m o u n t o f ° u ctu a tio n
To minimize the total
in b o th q a n d p th a t is co m m en su ra te w ith th e u n cer-
energy, we need to ta in ty p rin cip le ¢ p ¢ q & ~. T h e resu ltin g g ro u n d sta te,
allow for some a s w e k n ow , h a s a n o n -zero en erg y E q u a n t = (1= 2 )~! .
amount of fluctuation
in both q and p that S u p p o se w e co n sid er a d i® eren t p h y sica l sy stem w ith th e
is commensurate H a m ilto n ia n H n ew = H (p ;q) ¡ (1 = 2 )~! w h ere H (p ;q) is
with the uncertainty g iv en b y (2 ). S in ce th e su b tra ctio n o f a co n sta n t fro m
principle Δp Δq &
th e H a m ilto n ia n d o es n o t ch a n g e th e eq u a tio n s o f m o -
~.
tio n , w e still a g a in h av e a h a rm o n ic o scilla to r b u t w ith

180 RESONANCE ⎜ February 2009

 SERIES ⎜ ARTICLE

a sh ift in th e en erg y. C la ssica lly, th e m in im u m en erg y Quantum mechanics

sta te w ill still co rresp o n d to q = p = 0 b u t w ith en - allows you to have a
erg y E cla ss = ¡ (1= 2 )~! . B u t q u a n tu m m ech a n ica lly, state for the harmonic
th e g ro u n d sta te w ill ex h ib it ° u ctu a tio n s in q a n d p b u t oscillator with the
w ill h av e zero en erg y ; E q u a n t = 0 ! T h is is th e cru cia l Hamiltonian Hnew (p,q)
p o in t. Q u a n tu m m ech a n ics a llow s y o u to h av e a sta te fo r
such that Equant= 0
th e h a rm o n ic o scilla to r w ith th e H a m ilto n ia n H n ew (p ;q)
which can host
su ch th a t E q u a n t = 0 w h ich ca n h o st ° u ctu a tio n s in th e
fluctuations in the
d y n a m ica l va ria b les q a n d p .
dynamical variables q
S o m eth in g v ery a n a lo g o u s h a p p en s in th e ca se o f a n elec- and p.
tro m a g n etic ¯ eld . A s w e sh a ll see th e electro m a g n etic
¯ eld ca n b e th o u g h t o f a s a b u n ch o f h a rm o n ic o scilla -
to rs. T h e g ro u n d sta te w ill co rresp o n d to a sta te o f zero
p h o to n s a n d o n e ca n a rra n g e m a tters su ch th a t it h a s
zero en erg y. B u t th e electric a n d m a g n etic ¯ eld s w ill
p lay th e ro le a n a lo g o u s to p a n d q o f th e o scilla to r a n d
th ey w ill ex h ib it ° u ctu a tio n s { u su a lly ca lled va cu u m
° u ctu a tio n s { in th e g ro u n d sta te. T h erefo re o n e ca n n o t
rea lly say th a t th e electro m a g n etic ¯ eld s va n ish in th e
va cu u m sta te ev en th o u g h w e ca n m a k e its en erg y va n -
ish . O n ce w e reco g n ize th is fa ct, it is n o t su rp risin g th a t
th e electro m a g n etic va cu u m ca n ex ert fo rces o n b o d ies.
A ctu a lly th e situ a tio n is a little b it m o re co m p lica ted
b eca u se th e p ro ced u re a n a lo g o u s to th e su b tra ctio n o f
(1= 2 )~! is m o re n o n triv ia l in th is ca se b u t th e essen tia l
id ea is th e sa m e.
L et u s n ow try to u n d ersta n d th is in a m o re m a th e-
m a tica l la n g u a g e a n d in a so m ew h a t b ro a d er co n tex t.
A s it tu rn s o u t, th e essen tia l id ea ca n b e illu stra ted b y
ig n o rin g tw o co m p lica tio n s o f th e rea l w o rld . F irst is
One cannot really say
th e v ecto r n a tu re o f electro m a g n etism a n d th e seco n d is
that the
th e fa ct th a t sp a ce is th ree d im en sio n a l. W e w ill w o rk
o u t ¯ rst a sim p ler p ictu re u sin g ju st a sca la r ¯ eld w ith electromagnetic fields
o n e d eg ree o f freed o m (ra th er th a n w ith th e electro m a g - vanish in the vacuum
n etic ¯ eld ) a n d a lso ig n o rin g th e tw o tra n sv erse d irec- state even though we
tio n s a n d trea tin g sp a ce a s o n e-d im en sio n a l. A fter w e can make its energy
w o rk o u t th e sim p li¯ ed p ictu re, w e w ill d escrib e h ow it vanish.

RESONANCE ⎜ February 2009 181

 SERIES ⎜ ARTICLE

g en era lizes to th e rea l life.

O n ce w e ig n o re th e v ecto r n a tu re o f th e electro m a g n etic
¯ eld , w e ca n w o rk w ith a sin g le sca la r ¯ eld Á (t;x ) w h ich
is a fu n ctio n o f o n e sp a ce d im en sio n a n d tim e. (If y o u
w a n t, y o u ca n th in k o f it a s a n a lo g o u s to a n y o n e co m -
p o n en t o f th e electro m a g n etic ¯ eld .) In th e a b sen ce o f
so u rces, w e k n ow th a t ea ch co m p o n en t o f th e electro -
m a g n etic ¯ eld sa tis¯ es th e w av e eq u a tio n ; so w e w ill
a ssu m e th a t o u r sca la r ¯ eld sa tis¯ es th e eq u a tio n :

@ 2Á @ 2Á
¡ = 0: (3 )
@ t2 @x2
(W e h av e ch o sen u n its w ith c = 1 . In 3 -d im en sio n s,
th e seco n d term w o u ld h av e b een ¡ r 2 Á w h ich b eco m es
¡ (@ 2 Á = @ x 2 ) w h en w e ig n o re tw o sp a tia l co o rd in a tes.)
T h is eq u a tio n ca n b e sim p li¯ ed b y in tro d u cin g th e sp a -
tia l F o u rier tra n sfo rm Q k o f Á (t;x ) b y
Z1
dk
Á (t;x ) = Q k (t) ex p (ik x ): (4 )
¡1 (2¼ )

S u b stitu tin g th is in (3 ) w e ¯ n d th a t Q k (t) sa tis¯ es th e

eq u a tio n QÄ k + k 2 Q k = 0: T h e ¯ eld Á (t;x ) is co m p letely
sp eci¯ ed b y th e fu n ctio n Q k (t) so th a t w e ca n th in k o f
Q k (t) a s th e d y n a m ica l va ria b les d escrib in g o u r sy stem .
T h e fa ct th a t w e a re d ea lin g w ith th e ¯ eld tra n sla tes
in to th e fa ct th a t w e n ow h av e a n in ¯ n ite n u m b er o f
d y n a m ica l va ria b les, o n e fo r ea ch va lu e o f k . O th er th a n
th a t, w e ca n w o rk d irectly w ith Q k (t) in stea d o f th e
The field φ (t,x) is o rig in a l ¯ eld Á (t;x ).
completely specified
O n e m in o r p ro b lem w ith Q k (t) is th a t it is a co m p lex
by the function Qk(t) n u m b er (sin ce Á (t;x ) is rea l) a n d w e w o u ld lik e to w o rk
so that we can think w ith d y n a m ica l va ria b les th a t a re rea l. T h is is ea sily
of Qk(t) as the ta k en ca re o f. A s Q k is co m p lex , w e h a v e tw o d eg rees o f
dynamical variables freed o m co rresp o n d in g to th e rea l a n d im a g in a ry p a rts
describing our o f Q k fo r ea ch k w ith th e co n stra in t Q ¤k = Q ¡k . If w e
system. w rite Q k = (A k + iB k ), th en , sin ce Á is a rea l sca la r

182 RESONANCE ⎜ February 2009

 SERIES ⎜ ARTICLE

¯ eld , w e ca n rela te th e va ria b les fo r k to th a t fo r ¡ k In particular, we can

a s A k = A ¡k a n d B k = ¡ B ¡ k . E v id en tly, o n ly h a lf quantize the
th e m o d es co n stitu te in d ep en d en t d eg rees o f freed o m . field by quantizing
T h erefo re, w e ca n w o rk w ith a n ew set o f rea l m o d es q k , each of the harmonic
d e¯ n ed fo r a ll va lu es o f k w ith a su ita b le red e¯ n itio n , oscillators qk(t). (In
say, b y ta k in g q k = A k fo r o n e h a lf o f k a n d q ¡k = B k fact, that is the
fo r th e o th er h a lf. T h is w ill, o f co u rse, lea d to th e sa m e essence of quantum
eq u a tio n b u t fo r th e real va ria b le q k (t).
field theory of non-

qÄk + k 2 q k = 0: (5 ) interacting fields; rest

is just detail.)

T h a t is, th e d y n a m ica l va ria b le q k (t) sa tis¯ es th e h a r-

m o n ic o scilla to r eq u a tio n w ith freq u en cy ! = jk j, fo r
ea ch va lu e o f k . O u r ¯ eld is m a th em a tica lly th e sa m e a s
a n in ¯ n ite n u m b er o f h a rm o n ic o scilla to rs, o n e fo r ea ch
k . It fo llow s th a t ev ery th in g w e k n ow a b o u t h a rm o n ic
o scilla to rs ca n n ow b e a p p lied to th is sy stem . In p a rtic-
u la r, w e ca n q u a n tize th e ¯ eld b y q u a n tizin g ea ch o f th e
h a rm o n ic o scilla to rs q k (t). (In fa ct, th a t is th e essen ce
o f q u a n tu m ¯ eld th eo ry o f n o n -in tera ctin g ¯ eld s; rest is
ju st d eta il.)
C la ssica lly, w e ca n n ow co n stru ct th e g ro u n d sta te b y
ta k in g q k = 0 fo r a ll va lu es o f k . T h is w ill, o f co u rse,
m a k e th e ¯ eld va n ish a lo n g w ith its en erg y, a s to b e
ex p ected fro m a sen sib le g ro u n d sta te. B u t, a s w e d is-
cu ssed ea rlier, th is d o es n o t h o ld fo r th e q u a n tu m g ro u n d
sta te. T h e g ro u n d sta te o f th e h a rm o n ic o scilla to r fo r a
g iv en va lu e o f k is d escrib ed b y th e g ro u n d sta te en erg y
eig en fu n ctio n
³! ´1 = 4 µ ¶
k 1 2
à (q k ) = ex p ¡ ! k q k : (6 )
¼ 2

W e a re u sin g u n its w ith ~ = 1 fo r sim p lify in g th e ex -

p ressio n s. T h e g ro u n d sta te w av e fu n ctio n fo r th e fu ll
sy stem , m a d e o f a b u n ch o f in d ep en d en t o scilla to rs, ca n
b e d escrib ed b y th e p ro d u ct o f th e g ro u n d sta te w av e

RESONANCE ⎜ February 2009 183

 SERIES ⎜ ARTICLE

This expression fu n ctio n s o f ea ch o f th e o scilla to rs:

can be interpreted Y ³! k ´1 = 4 µ ¶
1 2
along similar lines ª [Á (x )] = ex p ¡ ! k q k
¼ 2
k
as we interpret a · Z1 ¸
1 d k 2
harmonic oscillator ´ N¹ ex p ¡ ! k qk : (7 )
wave function in 2 ¡1 (2¼ )
usual quantum
T h is ex p ressio n ca n b e in terp reted a lo n g sim ila r lin es a s
mechanics. w e in terp ret a h a rm o n ic o scilla to r w av e fu n ctio n in u su a l
q u a n tu m m ech a n ics. S u p p o se w e h av e a h a rm o n ic o scil-
la to r in th e g ro u n d sta te a n d w e m ea su re th e p o sitio n
q . T h en th e rela tiv e p ro b a b ility th a t w e w ill g et a va lu e
q = a co m p a red to a va lu e q = b is g iv en b y

jà (a )j2 ¡ 2 2
¢
R = = ex p ¡ ! [a ¡ b ] : (8 )
jà (b)j2
N ow su p p o se w e h av e a qu an tu m ¯ eld w h ich is in th e
g ro u n d sta te a n d w e m ea su re th e ¯ eld ev ery w h ere a t,
say, t = 0 . T h en , th ere is so m e p ro b a b ility th a t w e
w ill g et a ¯ eld co n ¯ g u ra tio n d escrib ed b y th e fu n ctio n
Á (0;x ) = f 1 (x ) a n d so m e o th er p ro b a b ility th a t ¯ eld
co n ¯ g u ra tio n is d escrib ed b y th e fu n ctio n Á (0;x ) = f 2 (x ).
J u st a s in th e p rev io u s ca se, w e w a n t to k n ow th e rela -
tiv e p ro b a b ility o f g ettin g o n e co n ¯ g u ra tio n co m p a red to
a n o th er. T o ¯ n d th is, w e ¯ rst o b ta in th e sp a tia l F o u rier
tra n sfo rm s o f f 1 (x ) a n d f 2 (x ) a n d ca ll th em a k a n d b k .
T h en th e rela tiv e p ro b a b ility is g iv en b y
µ Z1 ¶
jª (f 1 (x ))j2 dk 2 2
For any choice of f1(x) R = = ex p ¡ ! k [a k ¡ b k ] : (9 )
jª (f 2 (x ))j2 ¡1 (2¼ )
and f2(x) the number R
can be computed, F o r a n y ch o ice o f f 1 (x ) a n d f 2 (x ) th e a b ov e n u m b er ca n
allowing us to b e co m p u ted , a llow in g u s to d eterm in e th e p ro b a b ilities
determine the o f d i® eren t ¯ eld co n ¯ g u ra tio n s in th e va cu u m sta te.
probabilities of
different field
Y o u w o u ld h av e n o ticed th a t w e sw itch ed to rela tiv e
configurations in the
p ro b a b ilities ra th er th a n a b so lu te p ro b a b ilities in th is
d iscu ssio n . F o r a sin g le h a rm o n ic o scilla to r, o n e co u ld
vacuum state.
h av e sa id th a t jà (q )j2 d q g iv es th e a b so lu te p ro b a b ility

184 RESONANCE ⎜ February 2009

 SERIES ⎜ ARTICLE

o f ¯ n d in g th e p a rticle in th e in terva l (q ;q + d q ). W h en We get around this by

w e h av e in ¯ n ite n u m b er o f o scilla to rs, th e n o rm a liza tio n talking about relative
fa cto r N¹ in (7 ) in v o lv es a n in ¯ n ite p ro d u ct w h ich is h a rd probabilities in which
to d e¯ n e rig o ro u sly. W e g et a ro u n d th is b y ta lk in g a b o u t the normalization
rela tiv e p ro b a b ilities in w h ich th e n o rm a liza tio n fa cto r factor cancels out.
ca n cels o u t.
B efo re w e p ro ceed fu rth er, let m e m en tio n th e co rre-
sp o n d in g resu lt in th ree sp a tia l d im en sio n s. In th is ca se
(7 ) h a s th e o b v io u s g en era liza tio n to :
Y ³! k ´1 = 4 µ ¶
1 2
ª [Á (x )] = ex p ¡ ! k jq k j
k
¼ 2
· Z 3 ¸
1 d k 2
= N ex p ¡
¹ ! k jq k j : (1 0 )
2 (2¼ )3

In fa ct, in th is ca se it is n icer to ex h ib it th e resu lt in

term s o f th e ¯ eld co n ¯ g u ra tio n itself b y u sin g ! k = jk j
a n d ! k jq k j2 = k 2 jq k j2 = jk j. S in ce ik q k is essen tia lly th e
F o u rier sp a tia l tra n sfo rm o f r Á , w e ca n ea sily o b ta in
Z 3 Z 3
d k 2 d k jk j2 jq k j2
! jq
k k j =
(2 ¼ 3 ) (2¼ 3 ) jk j

Z Z ½ ¾
1 3 3 r x Á ¢r y Á
= d x d y : (1 1 )
2¼ 2 jx ¡ y j2

(P rov e th is!) S u b stitu tin g th is in to (1 0 ) a n d ta k in g

th e m o d u lu s, w e g et th e p ro b a b ility d istrib u tio n in th e
Once again,
g ro u n d sta te to b e
expression (12)
P [Á (x )] = jª [Á (x )]j2 shows clearly that
the vacuum state of
½ ZZ ¾ the field can host,
1 3 3 r x Á ¢r y Á
= N ex p ¡ d x d y ; (1 2 ) what is usually
2¼ 2 jx ¡ y j2 called, zero point
fluctuations of the
w ith N = jN¹ j2 . O n ce a g a in , th is ex p ressio n sh o w s
field variable φ.
clea rly th a t th e va cu u m sta te o f th e ¯ eld ca n h o st, w h a t

RESONANCE ⎜ February 2009 185

 SERIES ⎜ ARTICLE

is u su a lly ca lled , zero p o in t ° u ctu a tio n s o f th e ¯ eld va ri-

Let us now ask
a b le Á . T h e p ro b a b ility th a t o n e d etects a p a rticu la r
what happens if ¯ eld co n ¯ g u ra tio n Á (x ) w h en th e ¯ eld is in th e va c-
we introduce two u u m sta te ca n b e o b ta in ed b y eva lu a tin g th e va lu e o f
perfectly P fo r th is p a rticu la r fu n ctio n a l fo rm Á (x ). T h e resu lt
conducting, is in d ep en d en t o f tim e b eca u se o f th e sta tio n a rity o f
parallel plates th e va cu u m sta te. G iv en th e a m b ig u ity in th e ov era ll
into the vacuum. n o rm a liza tio n fa cto r N th is p ro b a b ility sh o u ld a g a in b e
in terp reted a s a rela tiv e p ro b a b ility. T h a t is, th e ra tio
P 1 = P 2 w ill g iv e th e rela tiv e p ro b a b ility b etw een tw o ¯ eld
co n ¯ g u ra tio n s ch a ra cterized b y th e fu n ctio n s Á 1 (x ) a n d
Á 2 (x ).
L et u s n ow a sk w h a t h a p p en s if w e in tro d u ce tw o p er-
fectly co n d u ctin g , p a ra llel p la tes in to th e va cu u m . T h e
fa ct th a t th e p la tes a re p erfectly co n d u ctin g im p lies th a t
th e electro m a g n etic ¯ eld { fo r w h ich o u r Á (t;x ) is a
p rox y { m u st sa tisfy so m e n o n -triv ia l b o u n d a ry co n -
d itio n s a t x = 0 a n d x = L w h ere th e p la tes a re lo -
ca ted . F o r th e sca la r ¯ eld , w e ca n ta k e th e b o u n d a ry
co n d itio n to b e th a t th e ¯ eld va n ish es a t th e p la tes:
Á (t;0 ) = Á (t;L ) = 0 in o n e sp a tia l d im en sio n . Y o u ca n -
n o t d escrib e a ¯ eld sa tisfy in g su ch a b o u n d a ry co n d itio n
u sin g th e F o u rier in teg ra l in (4 ) w ith k ta k in g a ll p o s-
sib le va lu es in ¡ 1 < k < 1 . In stea d w e ca n restrict
it to a d iscrete, th o u g h in ¯ n ite, set o f va lu es g iv en b y
k = n (¼ = L ) w ith n = 1;2;:::: a n d w rite

X1 h ¼x i
Á (t;x ) = q n (t) sin n (1 3 )
n= 1
L

so th a t th e b o u n d a ry co n d itio n s a t x = 0 a n d x = L a re
sa tis¯ ed . W e still h av e to d ea l w ith a n in ¯ n ite n u m b er o f
o scilla to rs b u t th eir freq u en cies a re n ow g iv en b y ! n =
k n = n (¼ = L ).
If w e n ow w o rk o u t th e co rresp o n d in g g ro u n d sta te, it
w ill clea rly b e d i® eren t fro m th e o n e d escrib ed b y (7 )
b eca u se th e in teg ra l ov er k w ill b e rep la ced b y th e su m

186 RESONANCE ⎜ February 2009

 SERIES ⎜ ARTICLE

ov er n . T h is is n eed ed b eca u se, o u r b o u n d a ry co n d itio n The introduction

tells u s th a t th e g ro u n d sta te sh o u ld n ow h av e zero p ro b - of the plates,
a b ility fo r ¯ eld co n ¯ g u ra tio n s w h ich d o n o t va n ish a t th e through changing
p la tes. T h e in tro d u ctio n o f th e p la tes, th ro u g h ch a n g in g the boundary
th e b o u n d a ry co n d itio n , h a s ch a n g ed th e g ro u n d sta te. condition, has
W h a t a b o u t th e en erg y o f th e g ro u n d sta te w ith a n d changed the
w ith o u t th e p la tes? T h ey a re a lso d i® eren t. In th e a b - ground state.
sen ce o f p la tes, ea ch h a rm o n ic o scilla to r co n trib u tes a n
en erg y (1= 2 )~! = (1= 2 )~jk j. S o th e to ta l g ro u n d sta te
en erg y, p er u n it len g th o f sp a ce, is g iv en b y a n in teg ra l
ov er a ll k o f (1 = 2 )~jk j. T h erefo re, th e en erg y in a reg io n
o f len g th L w ill b e:
Z1 Z1
L 1 L
E0= d k ~jk j = d k ~k : (1 4 )
(2¼ ) ¡1 2 (2 ¼ ) 0
T h is is m a n ifested ly in ¯ n ite, essen tia lly b eca u se th ere
a re a n in ¯ n ite n u m b er o f h a rm o n ic o scilla to rs. W h a t
a b o u t th e g ro u n d sta te en erg y in th e p resen ce o f th e
p la tes? T h is is g iv en b y th e su m

1X 1X
1 1
0
E 0 = ~! n = ~(n ¼ = L ) (1 5 )
2 n= 0 2 n= 0

w h ich is a lso in ¯ n ite, essen tia lly b ein g th e su m o f a ll

p o sitiv e in teg ers.
T h ese in ¯ n ities a re b a d n ew s b u t th ere is a trick to g et
a ro u n d th em . A s w e sa id b efo re, th e eq u a tio n o f m o tio n
fo r th e k -th o scilla to r w ill n o t ch a n g e if w e su b stra ct
fro m th e H a m ilto n ia n (1= 2 )~! k b u t it w ill `reg u la rize
th e g ro u n d sta te en erg y to zero . T h is is eq u iva len t to
lo o k in g a t th e d i® eren ce (E 00 ¡ E 0 ) a s th e p h y sica lly
releva n t q u a n tity. T o stu d y th is, it is co n v en ien t to in -
tro d u ce in (1 4 ) a co n tin u o u s va ria b le n v ia th e eq u a tio n
k = (¼ = L )n . T h en w e g et fro m (1 4 ) a n d (1 5 ):
"1 Z1 #
~¼ X
(E 00 ¡ E 0 ) = n ¡ dn n : (1 6 )
2L n = 0 0

RESONANCE ⎜ February 2009 187

 SERIES ⎜ ARTICLE

Both the expressions

Y o u m ay th in k th a t th is is n o t m u ch h elp b eca u se th is
in (17) as well as their
is o f th e fo rm (1 ¡ 1 ) w h ich d o es n o t h av e a p re-
cise m ea n in g . T h a t is tru e b u t th ere a re w ay s o f g iv in g
difference are now
m ea n in g to su ch ex p ressio n s in a fa irly sy stem a tic m a n -
finite and the idea is to
n er. T h e sim p lest p ro ced u re is to co n sid er, in stea d o f
first compute the
th e ex p ressio n in (1 6 ), th e ex p ressio n :
difference as a
function of λ and then E 00(¸ ) ¡ E 0 (¸ )
take the limit of "1 Z1 #
λ → 0 hoping for the ~¼ X
´ n ex p (¡ n ¸ ) ¡ d n n ex p (¡ n ¸ ) : (1 7 )
best. 2L n = 0 0

H ere w e h a v e m u ltip lied b o th th e ex p ressio n s b y a `reg -

u la to r fu n ctio n ' ex p (¡ n ¸ ) w h ere ¸ is ju st a p a ra m eter.
B o th th e ex p ressio n s a s w ell a s th eir d i® eren ce a re n ow
¯ n ite a n d th e id ea is to ¯ rst co m p u te th e d i® eren ce a s
a fu n ctio n o f ¸ a n d th en ta k e th e lim it o f ¸ ! 0 h o p in g
fo r th e b est. T h a t is, w e in terp ret th e ex p ressio n in (1 6 )
a s th e lim it o f th e ex p ressio n in (1 7 ) w h en ¸ ! 0 . I w ill
let y o u w o rk o u t th e ex p ressio n s. Y o u sh o u ld ¯ rst g et:
X1 e ¡¸ 1 1 ¸2
n ex p (¡ n ¸ ) = = ¡ + + O (¸ 4 )
n= 0
(1 ¡ e ¡¸ )2 ¸2 12 240
(1 8 )

w h ich d iv erg es w h en ¸ ! 0 a s to b e ex p ected . S im ila rly

Z1
1
d n n ex p (¡ n ¸ ) = 2 (1 9 )
0 ¸

w h ich a lso d iv erg es w h en ¸ ! 0 . B u t, a sto n ish in g ly

en o u g h , th e d i® eren ce b etw een (1 8 ) a n d (1 9 ) rem a in s
¯ n ite a s ¸ ! 0 :
X1 Z1
n ex p (¡ n ¸ ) ¡ d n n ex p (¡ n ¸ )
n= 0 0

1 1
= ¡ + O (¸ 2 ) = ¡ (2 0 )
12 12

188 RESONANCE ⎜ February 2009

 SERIES ⎜ ARTICLE

w h en ¸ ! 0 . T h is a llow s u s to o b ta in th e fo llow in g
rem a rka b le resu lt:
¼~
E (L ) ´ (E 00 ¡ E 0 ) ´ lim (E 00(¸ ) ¡ E 0 (¸ )) = ¡ :
¸! 0 24L
(2 1 )
S o w e see th a t th e g ro u n d sta te en erg y o f th e sy stem
w ith th e p la tes { w h en reg u la rized b y su b tra ctin g aw ay
th e en erg y in th e a b sen ce o f th e p la tes { is a n eg a tiv e
n u m b er1 a n d is in v ersely p ro p o rtio n a l to th e sep a ra tio n
1
b etw een th e p la tes! C lea rly, th is w ill lea d to a n a ttra c- With this `regularization',
quantum ¯eld theorists of-
tiv e fo rce F = ¡ (d E = d L ) / L ¡2 b etw een th e p la tes, ten conclude that the sum
sin ce red u cin g th e sep a ra tio n b etw een th e p la tes lea d s of all positive integers is
to th e low erin g o f th e en erg y. A m o re p h y sica l w ay o f not only ¯nite but is a neg-
th in k in g a b o u t th is resu lt is a s fo llow s. If w e ch a n g e th e ative fraction (¡ 1= 12)! If
sep a ra tio n b etw een th e p la tes b y a n a m o u n t ¢ L , th e you are familiar with Rie-
mann-zeta
P1 function, ³ (x )
en erg y o f th e co n ¯ g u ra tio n w ill ch a n g e b y (d E = d L )¢ L = n = 1 n ¡ x , you will reco-
w h ich m u st b e a cco u n ted b y th e w o rk d o n e b y th e a g en cy gnize that the sum of all
sep a ra tin g th e p la tes a ctin g a g a in st th e a ttra ctiv e fo rce positive integers is formally
F . E q u a tin g it to ¡ F ¢ L , w e ¯ n d th a t F = ¡ d E = d L . the same as ³ (¡ 1). One
can de¯ne this quantity by
In th e m y th ica l w o rld o f o n e sp a tia l d im en sio n , th e p la tes analytic continuation in
a re zero -d im en sio n a l p o in ts w h ich is n o t o f m u ch u se. the complex plane and one
does recover the result
T h e co rresp o n d in g ca lcu la tio n fo r electro m a g n etic ¯ eld ³ (¡ 1) = ¡ 1= 12. Of course,
in 3 -d im en sio n s is m o re co m p lica ted a lg eb ra ica lly b u t this does not make one any
a ll th e co n cep ts rem a in th e sa m e. T h e ¯ n a l resu lt in wiser as to what is going
th is ca se is a n ex p ressio n fo r en erg y p er u n it tra n sv erse on.
a rea o f th e p la tes, g iv en b y :
(E 00 ¡ E 0 ) ¼ 2 ~c
= ¡ ; (2 2 )
A 720 L 3
w h ere w e h av e re-in tro d u ced th e c fa cto r. T h e fo rce p er
u n it a rea a ctin g b etw een th e p la tes is g iv en b y
F d (E 00 ¡ E 0 ) ¼ 2 ~c
= ¡ = ¡ : (2 3 )
A dL A 240 L 4
T h is tin y fo rce h a s a ctu a lly b een m ea su red in th e la b !
N o te th a t, th o u g h th e resu lt is electro m a g n etic b y n a -
tu re, it is in d ep en d en t o f th e electro n ic ch a rg e e. T h e

RESONANCE ⎜ February 2009 189

 SERIES ⎜ ARTICLE

electro m a g n etism o n ly en ters th ro u g h th e b o u n d a ry co n -

The essential lesson
d itio n o n th e p erfect co n d u cto rs, w h ich is o n e rea so n w e
is that the pattern of
co u ld m im ic it w ith a sca la r ¯ eld .
quantum fluctuations
is sensitive to the T h e w h o le p h en o m en o n is q u ite b ew ild erin g a n d if y o u
boundary conditions a re sh a k in g y o u r h ea d in d isb elief, I w ill n o t b la m e y o u !
we impose, both B u t th e rea lity o f th is e® ect is b ey o n d d isp u te a n d it
mathematically and h a s b een d eriv ed fro m sev era l d i® eren t p ersp ectiv es ov er
practically. y ea rs. T h e essen tia l lesso n is th a t th e p a ttern o f q u a n -
tu m ° u ctu a tio n s is sen sitiv e to th e b o u n d a ry co n d itio n s
w e im p o se, b o th m a th em a tica lly a n d p ra ctica lly. T h e
g ro u n d sta te o f th e electro m a g n etic ¯ eld in th e p resen ce
o f tw o p a ra llel, co n d u ctin g , p la tes is q u ite d i® eren t fro m
th e g ro u n d sta te in th e a b sen ce o f th e p la tes. T h is m u ch
a lo n e is ea sy to u n d ersta n d b eca u se th e g ro u n d sta te in
th e p resen ce o f th e p la tes m u st en su re th a t, th e ¯ eld co n -
¯ g u ra tio n s w h ich d o n o t sa tisfy th e b o u n d a ry co n d itio n s
a t th e p la tes, h av e zero p ro b a b ility fo r th eir ex isten ce.
B u t w h a t is ra th er cu rio u s is th a t th is g ro u n d sta te h a s
a n en erg y w h ich d i® ers fro m th a t in th e a b sen ce o f th e
p la tes b y a ¯ n ite a m o u n t. T h ere is n o sim p le ex p la -
Address for Correspondence
T Padmanabhan
n a tio n fo r th is fa ct, w h ich m a k es C a sim ir e® ect a ll th e
IUCAA, Post Bag 4 m o re fa scin a tin g .
Pune University Campus
Ganeshkhind
Suggested Reading
Pune 411 007, India.
Email:
[1] H B G Casimir, Proc. Kon. Ned. Akad. Wetensch., Vol.B51, p.793, 1948.
paddy@iucaa.ernet.in
[2] S K Lamoreaux, Phys. Rev. Lett., Vol.78, p.58 1997.
nabhan@iucaa.ernet.in
G Bressi, G Carugno, R Onofrio, G Ruoso, Phys. Rev. Lett., Vol.88,
p.041804, 2002.

190 RESONANCE ⎜ February 2009