Skip to main content

Full text of "Mesoscopic Quantum Superposition of Generalized Cat State: A Diffraction limit"

See other formats

arXiv: 1507.052lOvl [quant-ph] 18Jul2015 

Mesoscopic Quantum Superposition of Generalized Cat State: A Diffraction limit 

Suranjana Ghosh * 1 *, Raman Sharma 2 ^, and Prasanta K. Panigrahi 1 ^ 

1 Indian Institute of Science Education and Research Kolkata, Mohanpur - 741246, India 
2 Indian Institute of Technology Bombay, Powai, Mumbai - 400076, India 

The orthogonality of cat and displaced cat states, underlying Heisenberg limited measurement in 
quantum metrology, is studied in the limit of large number of states. The mesoscopic superposition 
of the generalized cat state is correlated with the corresponding state overlap function, controlled 
by the sub-Planck structures arising from phase space interference. The asymptotic expression of 
this overlap function is evaluated and the validity of large phase space support and distinguisha- 
bility of the constituent states, in which context the asymptotic limit is achieved, are discussed 
in detail. For large number of coherent states, uniformly located on a circle, the overlap function 
significantly matches with the diffraction pattern for a circular ring source with uniform angular 
strength. This is in accordance with the van Cittert-Zernike theorem, where the overlap function, 
similar to the mutual coherence function, matches with a diffraction pattern. The physical situation 
under consideration is delineated in phase space by utilizing Husimi-Q function. 


Cat states and their generalizations are known to 
achieve Heisenberg limited sensitivity in estimation of pa¬ 
rameters like coordinate/momentum displacements and 
phase space rotations [l|. A criterion to distinguish quan¬ 
tum states without classical counterparts, from those not 
possessing the same, are studied in [2, Q. For these non- 
classical states, subtle interference effects in the phase 
space pj lead to sub-Planck structures in their Wigner 
functions, which in turn allow precision measurement of 
quantum parameters, bettering the standard quantum 
limit. Recently, sub-Planck structures in different phys¬ 
ical systems have^been investigated IM1. It has been 
demonstrated BIO that, the sensitivity of the state 
used in quantum metrology is directly related to the area 


of the sub-Planck structures: p = with A being the 
action of the effective support of the Wigner function. 
The interference in phase space is a pure quantum phe¬ 
nomenon, arising due to the fact that these states are 
superposition of the coherent states (CSs), which them¬ 
selves are classical. The increase in the number of inter¬ 
fering coherent states in the phase space is akin to emer¬ 
gence of diffraction in classical optics, when the number 
of interfering sources becomes large with sufficient phase 
space support. 

Here, we analyze this diffraction limit of the small¬ 
est interference structures and find an exact asymptotic 
value of the displacement sensitivity. With the assump¬ 
tion of large phase-space support for the estimating state 
and smallness of the quantum parameters to be esti¬ 
mated, it is found that the asymptotic limit of the sen¬ 
sitivity reaches |i5| = ^y, where C is the first root of 

Jo, the 0 th order Bessel function. We explicitly show 


1 e-mail:ramansharma@cse. iitb. ac. in 

t e-mail:pprasanta@iiserkol. ac. in 

that this assumption is adequate for realistic values of 
the physical parameters; i.e., the average photon num¬ 
ber and the number of superposed CSs. The numerical 
analysis depicts how the asymptotic limit of exact over¬ 
lap function (OF) reaches to the 0 th order Bessel function 
for higher order mesoscopic superpositions. This limiting 
behavior in the phase space interference is found to be 
analogous to the van Cittert-Zernike theorem [Hi], relat¬ 
ing the mutual coherence in classical optics to diffraction. 
A phase space distribution (Q-function), having only pos¬ 
itive regions, reveals the actual physical situation at the 
point of resemblance between the two theories. 


Cat states and their generalizations play a signifi¬ 
cant role in quantum optics and quantum computation 
0- A number of experimental schemes exist to produce 
cat states in laboratory conditions 0. These “pointer 
states” [lH often naturally manifest, when suitable quan¬ 
tum systems are coupled with decohering environment. 
It has been observed that the robustness of these states, 
made out of classical CSs, is a result of “quantum Dar¬ 
winism” 0- We consider a single oscillator, with the CS 
being an eigen state of a: a\a) = a\a), with annihilation 
and creation operator a and a f : [a,a^] = 1. 

The generalized cat state is composed of CSs, equally 
phase displaced on a circle: 

1 n ^ n 

\cat n a ) = —= \e~^a) = —= ^ D(e~^a) |0), (1) 

where, |a) = D(a)|0), with the displacement operator, 
D(a) = e“ a ^“ a and a|0) = 0. Here, it is worth men¬ 
tioning that the CSs are assumed to be distinguishable 
01. The displacements in the coordinate and momenta 
can be realized through an appropriately displaced cat 

state 0 : l c «4,a) = D(5)\cat n , a ). 

For checking the sensitivity of the estimating state 


| catn^o), one computes the overlap of the same with the displaced state and studies the orthogonality conditions, 


(cat n ^(x\cut n a ) — ^^^^(0|D(e n a)^D(S)D(e n a)|0) 

j —i fc=i 

1 Gh JL „ >27rfc , i2n(k-j) i27rfc 

— ( e " + e n ) + l“l e " ))( e “i l 5 + a ( e n -e " , 

j =1 fc=l 

n n 

= SEE' 

(2rcos( n ^ n k ^ ) sin{6— 7r ^J" fc ^ )1 q;| 2 sm( 2f ^ )) 

J=1 fc=l 

X g—^(|^| 2 +2 |q:| 2 (1— c os( 2?r ^ ))+4r sm( fc ^ )sm(fl — 7r ( J ^ l ~ fc ) )) 

( 2 ) 


where, r 
S = \6\e l 


= |a||<5| and 9 = (9s — 9 a ) with a = \a\e lffa 
9s . 

' W 




Distinguishable Cat-states • 









# Indistinguishable Cat-states 





20 40 60 80 100 


FIG. 1: Parameter domain of the phase space support a to 
maintain the condition of distinguishability. The shaded re¬ 
gion designates the domain for distinguishable cat-states. 

The phase space of the generalized cat state of Eq. © 
is composed of ‘n’ CSs, equally placed in a circle of ra¬ 
dius |a|, where large phase space support means the large 
magnitude of |a|. Now the natural question arises, how 
large should be the phase space support to maintain the 
distinguishability for a given mesoscopic superposition, 
i.e., domain of a for a given n. This is delineated in 
Fig. [TJ where we have taken upto a large value of n, ( e.g. 
n = 100). The upper region of the plot (shaded) de¬ 
picts the allowed parameter domain of a for mesoscopic 
superposition of CSs. Table-I reveals the values of al¬ 
pha (accurate upto first decimal place), above which the 
states are distinct. We further observe that the maxi¬ 
mum value of the ratio n/a to maintain distinguishabil¬ 
ity takes the average value v = 2.016. Thus one does 
not need a very large a to generate the said interference 
structures in phase space. In fact, increasing n is quite 
difficult in experiments, as it requires a large nonlinearity 

of the medium. On the contrary, the absolute value of a 
is directly related to the average photon number of the 
coherent state, which can be manipulated by controlling 
the laser beam. Hence, the allowed maximum order of 
mesoscopic superposition (‘n’) for a given a, conforming 
our result, is sufficiently large in reality. 

It is important to mention that, the OF between the 
initial and displaced cat states can also be represented 
by the phase space Wigner distribution: 

|(cat„ iQ ,|cat^ Q! )| 2 = J J W ca t„ ta (x,p)W cat s na (x,p)dx dp. 


This relation reveals the physical significance of the os¬ 
cillations of the OF in a particular direction in phase 
space and connects our result with the mesoscopic su¬ 
perposition structures. The oscillation of the OF is the 
signature of quantum interference structures of dimen¬ 
sion less that Planck’s constant, i.e., sub-Planck scale 
structures. Each zero of the OF signifies the orthogonal¬ 
ity of the original and displaced states, thereby implying 
the sensitivity limit of Heisenberg limited measurement. 

Now, it is intuitive as well as numerically verified by us 
that the entire contribution of the OF in Eq. © or © 
mainly originates from the adjacent components of the 
original and displaced cat states, i.e., j ~ k. Therefore, 
\j — k\ « n , cos(ir(j — k)/n) —> 1 and sin(Tr(j — k)/n) —> 
0. Then Eq. © takes the simpler form 

2r»m(0 - ” U + k) ) 


The off-diagonal terms in the above expression have 
negligible contribution. This assumption bears similar 
meaning of the classical situation, where an incoherent 
ring source is assumed, for which the cross-correlations 
between the different points of the source can be ne¬ 
glected. Now, with the assumption of sufficient phase- 
space support for the estimating state and smallness of 
quantum parameters to be estimated, one can consider 

/ i A \ 6 2 ' ' ^—> v—\ 

{cat n ^\cat n0i ) =-^ cos 


TABLE I: Estimate of the numerically obtained minimum value of phase space area (proportional to |a|), required for a 
mesoscopic superposition of n- CSs. 





























FIG. 2: Analogy between two-source interference: The solid 
lines show the maximum and dashed lines show the minimum 
intensity values. The crosses are the positions of coherent 
states and diamonds show the equivalent positions of sources 
of light which will produce the same pattern at a distance. 
The inset shows the equivalent position of sources for the 
state \cat2 a ^) = l a )+l™ e *) 

|<5| sin(0(5 — 9 a ) and <5|| = |<5| cos(0a — 9 a ). As is depicted 
in Fig. 0 it is interesting to observe that the above ex¬ 
pression is analogous to the double slit interference pat¬ 
tern, where the normalized intensity can be written as 
- y— = cos 2 (^) 0- The path difference between 
the two waves reaching at the observation point is yb/s, 
where b defines the distance between the two slits, s is 
the separation between the aperture and the screen, and 
y corresponds to the vertical coordinate of the detector. 
The above analogy can be mathematically established by 
taking A in the unit of s and redefining the commutation 
relation, [a, a'] = 7tA _1 : 

|(caf 2ja |caf 2 , Q }| 2 = cos 2 


|a|5 L7r 

( 8 ) 

where 2\a\ is the separation of the two coherent state 
sources. Use of the phase shifted cat state, \cat2 at< f,) = 

|a)+ [f e —-, would yield an interference pattern at an angle 
■| and fringe width , 2\a\ sin -f: 

only the diagonal terms and obtain 
e-5^1 2 71 

{cat n ^ a \cat° na ) 

'£ cos 

j =i 

2 rsin(6 — 




j= i 

2 rsin(9 -—) 


( 6 ) 

It needs to be mentioned that the state overlap depends 
only on 8a*, which leads to the conclusion that the sen¬ 
sitivity of estimating 5 is inversely proportional to |a|. 
It is easily checked that the OF, being of interferomet¬ 
ric origin, is only sensitive to the difference in phase : 
(cat^ a \cat^ a ) = e 1 ^ 1 (catn t a\cat^~ S2 ). The OF for n = 2, 

\{cat 2 ,a\cat s 2 a )\ 2 « cos 2 (2|a|<5j_)> 


\(cat2 a ^\cat2 5 a ^)| 2 = cos 2 (2|a| sin ^(8± sin ^+$\\ cos ^)) 


Introducing a phase between the constituent CSs of a 

cat state with n = 2 gives the state |caf2^} = 

The OF for this state is 

|(caf2^|cat2^ ,l5 )| 2 = cos 2 (2HAl) - 0), (10) 

akin to the phenomenon of “fringe shift ” observed in clas¬ 
sical optics. 

We now derive the state overlap and sensitivity in pa¬ 
rameter estimation for very high order of mesoscopic su¬ 
perpositions. For convenience, we assume n is even: 

2 . 2 , 

(cat n , a \cat s na ) = - cos 
U 3=1 

2r sin(# — —^) 


which matches with the known result [?], with <5_l = 

In the asymptotic limit of n, one writes 


lim {catn t a\catn a) 

>oo, n/oi<v 

lim — cos(2rsin(0--)) 

n—> oo, nla.<.v 71 ' ^ 71 

~ 3 =1 

1 ^ 2 7 

lim — cos(2rsin(0--)) 

n—too, n/ot<.v 71 ' ^ 71 

3 =1 

/»1 -j^ /»27T 

/ cos(2r sin(^ — 27rx))dx = — / cos(2r sin(z))dz 

Jo ^7T Jq 

= Jo(2|a||5|). 

( 12 ) 

0.0 0.2 0.4 0.6 0.8 1.0 


FIG. 3: Deviation (Dev) of the overlap function (Eq. |[3|) from 
the zeroth order Bessel function for a = 15: n = 8 (dashed 
line), n = 16 (dotted line), and n = 30 (solid line). For larger 
value of ’n\ the OF almost coincides with the zeroth order 
Bessel function. 

This proves our assertion that states can be discrim¬ 
inated for | S | = 2 ^y due to orthogonality, where C is a 
root of the Bessel function (of first kind) of order zero, 
i.e., Jq- In Eq. E2 we evaluate the overlap function for 
very high order of mesoscopic superposition. Theoreti¬ 
cally, a limit n —> oo is taken, provided/implied that the 
states are still distinguishable. It is worth mentioning 
that theoretically the parameter a also does not have any 
upper limit and in principle can go upto infinity. Thus, 
the Bessel function is obtained without any further re¬ 
striction. In reality, neither n nor a can go upto infinity 
and thus one has to take the practical quantitative esti¬ 
mate of these two physical parameters, as described in 
Fig. [T] and Table-I. The distinguishability condition im¬ 
plies n/a < v, where the average value of v is 2.016. 

In Fig. [3] we have plotted the deviation (Dev = 
OF — Bessel Function) with respect to the displace¬ 
ment in phase space, for three different values of n with 
a = 15. The best result is obtained for n = 30 or for 
n/a ~ v, beyond which the constituent CSs become in¬ 

distinguishable. Thus, for higher order mesoscopic su¬ 
perposition, our result fits very well with the condition 
when the phase space support is sufficient enough. The 
result is verified by a three dimensional plotting of the 
functions in Fig. [4] The first plot corresponds to the 
bessel function for a = 15 (a) with respect to the real 
and imaginary components of the displacement parame¬ 
ter. The same is also performed for the overlap function. 
The OF for n = 10 (b) does not match with the Bessel 
function, whereas the oscillations find a remarkable sim¬ 
ilarity with the Bessel function for n = 30 (c). 

Husimi-Q-Function: So far, we have been discussing 
about the OF, which is a result of superposition of n-CSs. 
However, it becomes a natural question to ask: what is 
the physical situation of the CSs on a circle for the critical 
ratio n/a = vl We have tried to explore the answer by 
calculating the Husimi-Q-function. Q-function is a phase 
space distribution, which is always positive and does not 
include any interference structures. Hence, plotting the 
Q-function for a higher order superposition is less time 
consuming. The Q-function is calculated and then plot¬ 
ted in Fig.[5]for (a) a = 9.9, n = 20; (b) a = 15.9, n = 32; 
and (c) a = 19.9, n = 40. These cases are evaluated for 
n/a = v and shows a striking support to our intuition 
that the CSs are actually just touching each other and 
starts becoming indistinguishable; thereby creating a ex¬ 
tended source of ring-shaped light of radius \/2a. This 
established the required physical situation in the classical 
optics to obtain ring shaped light source with constant 
angular source strength. 

Thus, the overlap function [Eq. (fl2l) ] is the result of 
superposition of n CSs situated in a ring of radius \/2a. 
Hence, the superposition is analogous to the diffrac¬ 
tion pattern generated when light passes through the 
thin r ing shaped opening. The van Cittert-Zernike the¬ 
orem [13( states that diffraction problem is identical to 
the coherence problem and the two problems result in 
the same mathematical formalism through the quantity 
called complex degree of coherence. In this work, n co¬ 
herent states, symmetrically situated in circle produce 
interference ripples in the center. This phase space in¬ 
terference structures is known to manifest in the overlap 
function between the original and displaced cat-states. 
We have shown that this interference term is analogous 


FIG. 5: Q-function in phase space for different a: (a) a = 9.9, n = 20; (b) a = 15.9, n = 32; and (c) a = 19.9, n = 40. We 
have plotted, maintaining the critical ratio: n/a = v, where the CSs are just touching each other. 

to the diffraction pattern resulting from the equivalent 
optical sources after proper scaling of the parameters. 
Thus, the fact that the overlap between the cat states and 
their shifted forms, is of the same form as the diffraction 
pattern centered at one of the states, bears strong re¬ 
semblance to the van Cittert-Zernike theorem. Here the 
normalized mutual coherence function 712 ( 0 ), for a ring 
shaped opening with constant angular source strength, 
can be written explicitly as, 

equation should be compared with the OF for large n 
(Eq. [12]), for unit distance from the screen to the opening 
(R = 1 ) and for [a, a 1 '] = 7 rA _1 : 

(cat n , Q |cat£ >Q ) = J 0 . ( 14 ) 


712 ( 0 ) 

(■ Ei{t)E 2 {t)*)r 

^(E 1 (t)E 1 (t)*) T (E 2 (t)E 2 (t)*) T 

f 2nr 0 \fl - E 2 \\ 

J ° ( \R ) ■ 


712 ( 0 ) actually signifies the complex degree of spatial co¬ 
herence of the two points at the same instant in time, 
when fields arriving at the observation screen being E\ (t) 
and E 2 (t) respectively, ro is the radius of ring, R is the 
distance of the screen from the opening and |r"i — r^j is 
the path difference between the points. 

The suffix T in the expectation value signifies the time 
average according to the ergodic hypothesis. The above 

In conclusion, the sensitivity of cat-like states to quan¬ 
tum parameter estimation is studied for large number of 
constituent CSs. The phase space support is explored 
for accessible parameter ranges in realistic situation. We 
provide a quantitative estimate of the phase space sup¬ 
port for a given superposition. In the large n limit, the 
state OF, determining the orthogonality of cat and dis¬ 
placed cat states, approaches the Bessel function. Ac¬ 
cording to the van Cittert-Zernike theorem, the coher¬ 
ence problem is mathematically identical with the diffrac¬ 
tion problem by complex degree of coherence. The fact 
that the OF is having the same form as the diffraction 
pattern results on the same expression of normalized mu- 


tual coherence function for large n. This is similar to the 
mutual coherence function of a circular ring, which yields 
Bessel function of order zero, matching with the theorem 
of van Cittert-Zernike. In addition, this work opens up 
several scope for future studies: i) Mesoscopic superposi¬ 
tion is a purely quantum phenomenon and overlap func¬ 
tion plays a very crucial role in quantum parameter esti¬ 
mation. The overlap function, without asymptotic limit, 
can be investigated for different relative phases (9 a and 
9g) of the CS and displacement parameters, which will 
provide information in various direction in phase space; 
ii) Studying mesoscopic superposition in realistic quan¬ 
tum systems, which are not modelled by harmonic oscil¬ 
lator CSs are quite nontrivial in general. Hence, it has 
become a frequent practice to study harmonic oscillator 
system and use that knowledge to investigate other solv¬ 
able quantum mechanical potentials; iii) Quantum sen¬ 

sitivity has become a very fascinating area of research, 
where it is known that increasing the value of the CS- 
parameter will make the system more and more sensi¬ 
tive. However, it was completely unknown that the OF 
would saturate to the Bessel function. This novel fact 
can be further utilized for investigating various physical 
situations of interest. 


The author, R. Sharma acknowledges the support by 
National Initiative on Undergraduate Science (NIUS) un¬ 
dertaken by HBCSE-TIFR, Mumbai, India. Discussions 
with Prof. Vijay. A. Singh of HBCSE are gratefully ac¬ 

[1] W. H. Zurek, Nature (London) 412, 712 (2001). 

[2] W. Vogel, Phys. Rev. Lett. 84 , 1849 (2000). 

[3] Th. Richter and W. Vogel, Phys. Rev. Lett. 89, 283601 

( 2002 ). 

[4] W. Schleich and J.A. Wheeler, Nature (London) 326, 574 
(1987); W. Schleich and J. A. Wheeler, J. Opt. Soc. Am. 
B 4 , 1715 (1987); W. Schleich, D. F. Walls, and J. A. 
Wheeler, Phys. Rev. A 38, 1177 (1988); W. P. Schle¬ 
ich, Quantum Optics in Phase Space (Wiley-VCH, Berlin, 
2001) and references therein. 

[5] G. S. Agarwal and P. K. Pathak, Phys. Rev. A 70, 053813 

[6] S. Ghosh, A. Chiruvelli, J. Banerji, and P. K. Panigrahi, 
Phys. Rev. A 73 , 013411 (2006). 

[7] F. Toscano, D. A. R. Dalvit, L. Davidovich, and W. H. 
Zurek, Phys. Rev. A 73 , 023803 (2006). 

[81 J. R. Bhatt, P. K. Panigrahi, and M. Vyas, Phys. Rev. 
A 78 , 034101 (2008). 

[91 S. Ghosh, U. Roy, C. Genes, and D. Vitali, Phys. Rev. A 
79 , 052104 (2009). 

[10] U. Roy, S. Ghosh, P. K. Panigrahi, and D. Vitali, Phys. 
Rev. A 80 , 052115 (2009). 

[11] S. Ghosh and I. Marzoli, Int. J. Quant. Inf. 9, 1519 

( 2011 ). 

[12] S. Ghosh, Int. J. Quant. Inf. 10 , 1250014 (2012). 

[13] M. Born and E. Wolf, Principles of Optics, 7th (ex¬ 
panded) ed., (Cambridge U. Press, Cambridge, 1995). 

[14] D. F. Walls and G. J. Milburn, Quantum Optics 2nd 
edition (Springer-Verlag, Berlin, Heidelberg, 2008) and 
references therein. 

[15] M. Brune, S. Haroche, J.M. Raimond, L. Davidovich, 
and N. Zagury, Phys. Rev. A 45 , 5193 (1992). 

[16] D. A. R. Dalvit, J. Dziarmaga, and W.H. Zurek, Phys. 
Rev. A 72, 062101 (2005). 

[17] W. Zurek, Nature Physics 5, 181 (2009). 

[18] Distingushibility is one of the essential criteria for ob¬ 
serving the interference structures of nonlocal superpo¬ 
sition of localized wave-packets. We have considered su¬ 
perposition of distinguishable CSs. In some other con¬ 
text, if one needs to consider indistinguishable CSs, which 
them selves overlapped each other, the normalization wil l 
be ^1/E" Efc Exp ]-M 2 (l - Exp[—(2in(j - k))/n\)}. 

For distingushible states j = k and it becomes 1 /s/n, 
where n is the total number of CSs. 

[19] E. Hecht, Optics fth Edition (Pearson Education, 2002). 

[20] G. S. Agarwal, G. Gbur, and E. Wolf, Optics Letters 29 , 
459 (2004).