## Abstract

We present a comparative study of the non-linear wave packet dynamics of two-mode coherent states of the Heisenberg-Weyl group, the SU(1,1) group and the SU(2) group under the action of a model anharmonic Hamiltonian. In each case, we find certain generic signatures of non-linear evolution such as quick onset of decoherence followed by Schrödinger cat formation and revival. We also report important differences in the evolution of coherent states belonging to different symmetry groups.

© 1999 Optical Society of America

## 1 Introduction

The quantum phenomena of revival and fractional revival (or the formation of Schrödinger cat and cat-like states) have been studied in many diverse systems and situations such as Rydberg atoms [1, 2, 3], the Jaynes-Cummings model [4], light propagation in Kerr media [5, 6] and even transient signals from multilevel quantum systems [7]. It is now apparent that these features are rather generic in that they are closely connected to the anharmonicity in the energy spectrum of the underlying Hamiltonian and to the periodicity of the time evolution operator.

The literature deals mostly with systems whose energy spectrum depends on a single quantum number with an associated revival time scale. However, there are many systems whose energy levels depend non-linearly on at least two quantum numbers [8, 9, 10]. In this paper, we focus on the non-linear wavepacket dynamics of coherent states under the action of a generic two-mode Hamiltonian.

Coherent states were originally defined for harmonic oscillators (H. O.) or the radiation field. From a group theoretic point of view, the H. O. coherent states arise in systems whose dynamical symmetry group is the Heisenberg-Weyl group. Coherent states of other symmetry groups also exist. Thus, for example, the much studied pair [11] and Perelomov [12, 13, 14] coherent states belong to the SU(1,1) group and are special cases of what may be called generalized SU(1,1) coherent states [10]. Coherent states of the SU(2) group have also been constructed [15, 16, 17, 18].

Our objective is to present a comparative study of how coherent states of various symmetry
groups evolve under the action of the *same* two-mode generic Hamiltonian.
Through a series of pictures and movies, we show how the initial coherent structure is lost due
to quantum dephasing and then regained later on to form spectacular and varied quasi-coherent
structures leading up to the formation of Schrödinger cats in some cases and to full
revival in all cases.

## 2 Coherent states

H. O. coherent states |*α, β*〉 are
superpositions of number states |*m, n*〉 with *m,
n*=0, 1, 2, …∞. One can also construct
‘even’ (+) and ‘odd’ (-) coherent states

SU(1,1) coherent states |*η, ξ, q*〉
are formed by superposing number states of the form
|*n*+*q, n*〉 where
*q*=*a*
^{†}
*a-b*
^{†}
*b*
is an integer constant and *n*=0, 1, 2, … ∞:

$$\times \sum _{n=0}^{\infty}\frac{{\xi}^{n}{\left(1-{\mid \eta \mid}^{2}\right)}^{n+\frac{q+1}{2}}\sqrt{\left(n+n\text{'}\right)!\left(n+n\text{'}+q\right)!}}{n!\left(n+q\right)!}\mid n+n\text{'}+q,n+n\text{'}\u3009,$$

where

For pair coherent states *η*→0 whereas for Perelomov
coherent states ξ→0. Since *η* and ξ
are continuous and (in general) complex parameters, infinitely many other cases of
|*η, ξ, q*〉 exist even for the same
value of *q*.

In the Schwinger representation of the SU(2) algebra [17], SU(2) coherent states |*θ, ξ,
N*〉 can be formed by the superposition of number states of the form
|*K, N-K*〉 where
*N*=*a*
^{†}
*a*+*b*
^{†}
*b*
is an integer constant and *K*=0, 1, 2, … *N*:

$$={\left(1+{\mid \tau \mid}^{2}\right)}^{-\frac{N}{2}}\sum _{K=0}^{N}{\left(\begin{array}{c}N\\ K\end{array}\right)}^{\frac{1}{2}}{\tau}^{K}\mid K,N-K\u3009.$$

The parameters *θ* and *ϕ* represent
angles in the range 0≤*θ,
ϕ*≤2*π*.

## 3 Wave packet dynamics

To study the wave packet dynamics of these two-mode coherent states, we look at the evolution
of their quadrature distributions under the action of a generic, phase insensitive, two-mode
Hamiltonian (we use *ħ*=1):

which can be readily diagonalised:

Recall that the quadrature distribution for a state vector
|*ψ*(*t*)〉 is defined as
|*ψ*(*x, y,
t*)|^{2}=|〈*x,
y*|*ψ*(*t*)〉|^{2},
where |*x, y*〉 is the eigenvector of
(*a*+*a*
^{†})/√2 and
(*b*+*b*
^{†})/√2 with
eigen-values x and y respectively. Details of analytical results on the evolution of H. O. and
SU(1,1) coherent states for this Hamiltonian have been described elsewhere [9, 10]. In this paper we include the evolution of SU(2) coherent states as well.
Our major results can be summarised as follows:

(a) The non-linearity of the Hamiltonian destroys the initial coherence of the wave packets. But the long-time evolution guarantees revival for all these states and at least one example each of cat formation by coherent states of all three dynamical groups considered.

(b) The details of their evolution are different. For H. O. coherent states, the evolution of
the initial wave packet depends on its symmetry and on the ratio of the two times scales
*T*
_{+} and *T*
_{-}. For SU(1,1)
coherent states, on the other hand,
*q*=*a*
^{†}
*a*-*b*
^{†}
*b*
being constant, it is clear from Eq. (6) that *T*
_{+} and hence the ratio
*T*
_{-}=*T*
_{+} does not play an
active role. However, the parity of *q* was found to be crucial in determining
the revival features of SU(1,1) coherent states. Likewise, for SU(2) coherent states,
*N*=*a*
^{†}
*a*+*b*
^{†}
*b*
is constant. In this case also only one of the time scales is significant, but the parity of
*N* decides the revival time.

#### 3.1 H. O. Coherent States

At *t*=0, the H. O. coherent states |*α,
β*〉 and |*α,
β*〉_{±} are represented by
*ψ*(*x, y,*
0)=*ϕ*
_{-}(*x, y*) and ${\psi}_{\pm}(x,y,0)=\left[{\varphi}_{-}(x,y)\pm {\varphi}_{+}(x,y)\right]\u2044\sqrt{2{P}_{\pm}}$ respectively, where
*ϕ*
_{±}(*x, y*) are Gaussians
centered at *x*=±*α*√2 and
*y*=±*β*√2:

The Gaussian pairs that represent the odd and even states at *t*=0 are well
separated for *α, β*>1 since the interference
between *ϕ*
_{-} and
*ϕ*
_{+} is negligible. However, as the wave
packets spread, interference between the components of the wave packets soon destroys the
initial coherent structure (**Movie 1**). But as shown in **Movie 2**, the
same interference will bring back some coherence in the structure at some later instants of
time for a given value of the ratio
*T*
_{+}/*T*
_{-}. For example, when
*T*
_{+}/*T*
_{-}=2/3, a six-way
fractional revival occurs for the odd state at *t*=*T*-/3. Note
that the odd state is *fully revived at t*=2*T*- for
*T*
_{+}/*T*
_{-}=2/3 whereas the even
state takes twice as long to revive. Although the evolution of the odd and even states may
appear identical in the initial stages (**Movie 1**), it is clear that these states
evolve quite differently at later times (**Movie 2**). Thus the symmetry of the
initial state plays a crucial role in the long term dynamics.

In **Movie 3**, we show that the ratio of *T*
_{+}
and *T*
_{-} is equally as important in determining the time evolution of
the wave packets. Snapshots of the quadrature distribution for
|*α, β*〉 are shown for
*T*
_{+}/*T*
_{-}=2/3 (left picture)
and 3/5 (right picture). Note that the revival time is 16*T*
_{-} in the
former case whereas it is only 6*T*
_{-} for the latter. Note also that
for the same value of time, a change in the ratio
*T*
_{+}/*T*
_{-} will drastically
alter the characteristics of fractional revival. It can be shown that at
*t*=4*T*
_{-} and
*T*
_{+}/*T*
_{-}=2/3,
*ψ*(*x, y,
t*)=[exp(-*iπ*/4)*ϕ*-(*x,
y*)+exp(*iπ*/4)*ϕ*+(*x,
y*)]/√2. The components of the wave function do not interfere with each
other as they are *π*/2 out of phase. The corresponding quadrature
distribution represents a Schrödinger cat state, a two-way fractional revival.

#### 3.2 SU(1,1) Coherent States

The SU(1,1) coherent states |*η, ξ,
q*〉 qi will have different initial quadrature distributions for different
values of *η, ξ* and *q*. The initial
distributions are coherent structures. As the system evolves, this coherence is lost rather
quickly but is restored partially (fully) at times of fractional (total) revival much as in the
case of H. O. coherent states. The details are different. For odd values of *q*,
|*ψ*(*x, y,
T*
_{-})|^{2}=|*ψ*(*x,
y*, 0)|^{2}. This would be the case for even values of
*q* also provided both *η* and
*ξ* are pure imaginary, or one of them is zero while the other is
pure imaginary. Otherwise, |*ψ*(*x, y*,
2*T*-)|^{2}=|*ψ*(*x,
y*, 0)|^{2} for even values of *q*. We illustrate
these features in **Movies 4–5**.

In **Movie 4**, the system is a Perelomov coherent state
(*ξ*=0, *η*=-*i* tanh
π/4). In this case, the initial quadrature distribution is a Gaussian for
*q*=0 and a vortex [19, 20] for *q*=1. But the revival time is
*T*
_{-} for *all* values of *q*. Note
that the quadrature distribution for *q*=0 develops hyperbolic dark fringes at
*t=T*
_{-}/2 whose origin can be traced analytically [10]. In **Movie 5**, the system is a pair coherent state
(*ξ*=3, *η*=0). Since both
*ξ* and *η* are real, the revival time is
2*T*
_{-} when *q*=0 and *T*
_{-}
when *q*=1. Furthermore, this system allows the formation of
Schrödinger cats at *t*=*T*
_{-}/2 for
*q*=0. At *t=T*
_{-} the initial distribution rotates by
*π*/2. At *t=T*
_{-}/2, these two
distributions add incoherently as the components of the wave function, being
*π*/2 out of phase, do not interfere with each other.

#### 3.3 SU(2) Coherent States

The SU(2) coherent states given by Eq. (4) are represented by the wave function

Clearly the corresponding quadrature distribution will depend not only on *N*
but also on the amplitude and phase of the parameter *τ*, that is, on
the angles *θ* and *ϕ*. The distribution,
which is a Gaussian modulated by the square of a Hermite polynomial, will have dark fringes at
the nodes of the latter and will be lined up at an angle tan^{-1}
(1/*τ*) with respect to the positive *x*- axis. In
the limit *τ*→0 (or
*τ*→∞), all the *N* bosons are
in one mode only (see also Eq. 4) and the quadrature distribution is aligned vertically (or
horizontally). For τ=±1, the distribution is along the diagonals
*x*=∓*y*. An altogether different pattern arises in
the limiting case *τ*→±*i*. One
obtains a wave function with vortex structure [19, 20]:

The quadrature distributions corresponding to *τ*=1 and
*τ*=*i* are shown in Fig. (6) for *N*=11.

The remarkable transition from one pattern to another can be effected by changing the phase
of *τ* while keeping
|*τ*|=1. This is shown in **Movie 6**.

For *N*≫1, the quadrature distributions of
|*τ, N*〉 and |*τ,
N*+1〉 will have similar initial patterns and short-time
evolutions. But their long-time evolutions and revival features will depend critically on the
parity of *N*. In fact, one can show that under the action of the Hamiltonian
*H* given by Eqs (5) and (6), |*τ, N*〉 revives (but
for an over-all phase factor) at *all integer* multiples of
*T*
_{+} if *N* is odd, and at
*even* multiples of *T*
_{+} if
*N* is even. If *τ* is pure imaginary, then the
revival time is the same for *all* values of *N*.

These findings are amply demonstrated in **Movies 7–8**. We consider two
SU(2) coherent states with the same value of *τ* but with
*N* values differing by unity. In movies (7) and (8) we compare the evolution of
their quadrature distributions under the action of the Hamiltonian (5). In both movies,
*N*=10 and 11 whereas *τ*=1 in **Movie 7**
and *τ*=*i* in **Movie 8**.

Fractional revivals (if any) will occur at times
*t*=(*r/s*)*T*
_{+}, where
*r* and *s* are mutually prime with
*r*<*s*. Defining
*U*(*t*)=exp(-*iHt*) to be the time evolution
operator, we can follow the method adopted earlier [10] to show that

$$\times \sum _{j=0}^{l-1}{\alpha}_{j}^{(r,s)}\mid \tau \mathrm{exp}\left(-i\pi [2j\u2044l-rN\u2044s]\right),N\u3009,$$

where

The coeficients ${\alpha}_{j}^{(r\mathit{,}s)}$ are given by

and can be evaluated analytically [21].

As an illustration of fractional revival, we will now explain the patterns obtained at
*t*=*T*
_{+}/2. From Eq. (10), one obtains

For Movie 7, we get

and

Note that the wave functions corresponding to |±1,
*N*〉 are real. In Eq. (14), these state vectors are added with a
*π*/2 phase difference. The resulting state vector represents a cat
state whose quadrature distribution is an incoherent sum of the distributions for
|1, 10〉 and |-1, 10〉:

$\frac{{e}^{-\left({x}^{2}+{y}^{2}\right)}}{\pi {2}^{N+1}N!}\left[{H}_{N}^{2}\left(\frac{y-x}{\sqrt{2}}\right)+{H}_{N}^{2}\left(\frac{y+x}{\sqrt{2}}\right)\right],N=10.$.

In Eq.(15), on the other hand, the state vectors
|±*i, N*〉 are added with a
*π*/2 phase difference. But the wave functions for
|±*i, N*〉 are complex giving rise to strong
interference. Using the wave functions for |±*i,
N*〉, the quadrature distribution for
*U*(*T*+/2)|1, 11〉 is found to
be

(1-sin 2*ηN*)× the quadrature distribution for
|±*i, N*〉, *N*=11.

Dark fringes will appear whenever

where *m* is an integer. Since
0≤*η*≤2*π*, the
total number of such fringes will be 2*N*. Although there will also be
2*N* bright patches, the phenomenon is not a 2*N*-way fractional
revival. Rather, it is more like ‘slicing a doughnut’ radially in
2*N* equal parts.

For **Movie 8**, we would get

$u({T}_{+}\u20442)\mid i,N\u3009\to \{\begin{array}{cc}\mid -i,N\u3009+i|i,N\u3009,& \mathrm{if}N=10;\\ \mid 1,N\u3009+i\mid -1,N\u3009,& \mathrm{if}N=11.\end{array}$.

The corresponding quadrature distributions can be explained in a similar fashion.

## 4 Conclusion

In conclusion, we have studied the non-linear wave packet dynamics of two-mode (a) H. O.
coherent states, (b) SU(1,1) coherent states and (c) SU(2) coherent states. The Hamiltonian used
was generic, phase insensitive and quadratic in the photon number operators. Although the
Hamiltonian was the same in each case, it was found that the states (a), (b) and (c) evolve
quite differently which can be traced to their belonging to different symmetry groups. However,
certain similarities in their evolution also emerged. Thus in *all cases*, the
initial coherent structure is quickly lost but is regained later on to experience the quantum
phenomenon of revival and the formation of Schrödinger cats.

## References and links

**1. **I. Sh. Averbukh and N. F. Perelman, “Fractional revival: Universality in the long-term evolution of quantum wave packets beyond the correspondence principle dynamics,” Phys. Lett. A **139**, 449 (1989). [CrossRef]

**2. **Z. D. Gaeta and C. R. Stroud Jr., “Classical and quantum-mechanical dynamics of a quasiclassical state of the hydrogen atom,” Phys. Rev. A **42**, 6308 (1990). [CrossRef] [PubMed]

**3. **X. Chen and J. A. Yeazell, “Phase-conjugate picture of a wave-packet interference design for arbitrary target states,” Phys. Rev. A **59**, 3782 (1999). [CrossRef]

**4. **J. H. Eberly, N. B. Narozhny, and J. J. Sanchez-Mondragon, “Periodic spontaneous collapse and revival in a simple quantum model,” Phys. Rev. Lett. **44**, 1323 (1980). [CrossRef]

**5. **B. Yurke and D. Stoler, “Generating quantum mechanical superpositions of macroscopically distinguishable states via amplitude dispersion,” Phys. Rev. Lett. **57**, 13 (1986). [CrossRef] [PubMed]

**6. **K. Tara, G. S. Agarwal, and S. Chaturvedi, “Production of Schrödinger macroscopic quantum-superposition states in a Kerr medium,” Phys. Rev. A **47**, 5024 (1993). [CrossRef] [PubMed]

**7. **C. Leichtle, I. Sh. Averbukh, and W. P. Schleich, “Multilevel quantum beats: an analytical approach,” Phys. Rev. A **54**, 5299 (1996). [CrossRef] [PubMed]

**8. **R. Bluhm, V. A. Kostelecký, and B. Tudose, “Wave-packet revivals for quantum systems with nondegenerate energies,” Phys. Lett. A **222**, 220 (1996). [CrossRef]

**9. **G. S. Agarwal and J. Banerji, “Fractional revivals in systems with two time scales,” Phys. Rev. A **57**, 3880 (1998). [CrossRef]

**10. **J. Banerji and G. S. Agarwal, “Revival and fractional revival in the quantum dynamics of SU(1,1) coherent states,” Phys. Rev. A **59**, 4777 (1999). [CrossRef]

**11. **G. S. Agarwal, “Nonclassical statistics of fields in pair coherent states,” J. Opt. Soc. Am. B **5**, 1940 (1988). [CrossRef]

**12. **A. M. Perelomov, *Generalized Coherent States and Their Applications* (Springer, Berlin1986). [CrossRef]

**13. ***Coherent states*, edited by J. R. Klauder and B. S. Skagerstam, (World Scientific, Singapore, 1985).

**14. **C. M. Caves and B. L. Shumaker, “New formalism for two-photon quantum optics. I. Quadrature phases and squeezed states,” Phys. Rev. A **31**, 3068 (1985). [CrossRef] [PubMed]

**15. **F. T. Arecchi, E. Courtens, R. Gilmore, and H. Thomas, “Atomic coherent states in quantum optics,” Phys. Rev. A **6**, 2211 (1972). [CrossRef]

**16. **K. Wódkiewicz and J. H. Eberly, “Coherent states, squeezed fluctuations, and the SU(2) and SU(1,1,) groups in quantum-optics applications,” J. Opt. Soc. Am B **2**, 458 (1985). [CrossRef]

**17. **V. Bužek and T. Quang, “Generalized coherent state for bosonic realization of SU(2) Lie algebra,” J. Opt. Soc. Am B **6**, 2447 (1989). [CrossRef]

**18. **G. S. Agarwal, R. R. Puri, and R. P. Singh, “Atomic Schrödinger cat states,” Phys. Rev. A **56**, 2249 (1997). [CrossRef]

**19. **G. S. Agarwal, R. R. Puri, and R. P. Singh, “Vortex states for the quantized radiation field,” Phys. Rev. A **56**, 4207 (1997). [CrossRef]

**20. **L. Allen, M. J. Padgett, and M. Babiker, “The orbital angular momentum of light” in *Progress in Optics*, edited by E. Wolf, (North Holland, Amsterdam), 39, 294 (1999).

**21. **J. H. Hannay and M. V. Berry, “Quantization of linear maps on a torus-Fresnel diffraction by a periodic grating,” Physica , **1D**, 267 (1980).