Controllability of pure states

for the Po schlTeller potential with a

dynamical group SU(2)
Shi-Hai Dong
a,
*
, Yu Tang
b
, Guo-Hua Sun
c
, F. Lara-Rosano
d
,
M. Lozada-Cassou
a
a
Programa de Ingenier a Molecular, Instituto Mexicano del Petro leo, La zaro Ca rdenas 152, 07730
Me xico, D. F., Mexico
b
Facultad de Ingenier a, UNAM, Me xico City 04510, D. F., Mexico
c
IIMAS of UNAM. A. P. 20-726, Del. Alvaro Obrego n, Me xico 01000, D. F., Mexico
d
Centro de Ciencias Aplicadas y Desarrollo Tecnolo gico, UNAM, Apartado Postal 70-186,
04510 Me xico, D. F., Mexico
Received 29 June 2004; accepted 17 September 2004
Available online 15 December 2004
Abstract
The controllability of a quantum system for the modied Po schlTeller (MPT) potential
with the discrete bound states is investigated. The creation and annihilation operators of this
potential are constructed directly from the normalized wave function with the factorization
method and associated to an su(2) algebra. It is shown that this quantum system with the non-
degenerate discrete bound states can, in principle, be strongly completely controllable, i.e., the
system eigenstates can be guided by the external eld to approach arbitrarily close to a target
state, which could be theoretically realized by the actions of the creation and annihilation
operators on the ground state.
2004 Elsevier Inc. All rights reserved.
PACS: 03. 65. Ca; 03. 65. Fd
Keywords: Controllability; SU(2) group; Modied Po schlTeller potential
www.elsevier.com/locate/aop
Annals of Physics 315 (2005) 566581
0003-4916/$ - see front matter 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.aop.2004.09.003
*
Corresponding author. Fax: +00525591756239.
E-mail address: [email protected] (S.-H. Dong).
1. Introduction
With the revolution of the microelectronic processing power, the study of the
quantum control theory for the manipulation of the quantum system state has been
the subject of interest [118]. Since the pioneering work on the quantum computa-
tion by DiVincenzo [2], the investigation of quantum control has attracted much
attention. Almost all of these studies are motivated by the recent advances in the eld
of the nuclear magnetic resonance and laser spectroscopy, e.g., the quantum control
by laser pulse can be realized experimentally [3]. Recently, Ramakrishna and Rabitz
[4] have shown how sequences of short laser pulses can be applied to control the state
of a molecule with a given Hamiltonian through a unitary operator representing
quantum evolution. Additionally, the nondegenerate quantum control problem
has been solved analytically [5]. The control of spin systems has been systematically
studied by the Cartan decomposition of SU(2
n
) [6]. At present, the quantum control
has been widely applied to the molecular and atomic physical problems, particularly
much attention has been paid to the problems of quantum systems with the non-
degenerate eigenstates.
Most of these works mentioned above are primarily focused on investigating the
controllability of the pure states for some quantum systems. However, due to the
importance of the some general mixed states in, for example, the control of temper-
ature distribution in systems and the need to include dissipative eects, the control-
lability of the mixed states has been studied by Schirmer et al. [18], in which the
criteria for reachability of quantum states and the constructive control of quantum
systems using factorization of unitary operators have been investigated. Neverthe-
less, the present work is conned to study the controllability of the pure states with
the group theoretical method as used in our previous work [17].
As we know, the basic problem for the quantum control is to control a quantum
system. Roughly speaking, given a quantum system the controllability refers to the
possibility of driving an initial state W(t = 0) = W
0
to any selected target state W(t)
at any chosen time under the guidance of an external eld. That is to say, the aim
of the quantum control is to steer a quantum system along a desired trajectory, for
example, to optimize the yield of a chemical reaction and to process quantum
information, etc. Until now, the quantum control theory has been developed well
for the closed quantum systems governed by the time-dependent Schro dinger equa-
tion [1,4,7,8,11,17], in which the state W(t) evolving from the initial state can be
described by the Schro dinger equation ihdWt=dt HWt. As we know, the state
W(t) may be realized by acting a unitary time-evolution operator Ut e
iHt=h
on
the initial state W
0
, i.e., W(t) = U(t) W
0
. Such an investigation has been carried out
in our recent work on the controllability of the quantum system for the Morse po-
tential [17]. It should be noted that we restrict our attention here to the closed
quantum system not for the open quantum system since it is possible to nd a
dynamical Lie algebra for the pure states of the closed quantum system. Once a
suitable dynamical algebra, e.g., the su(2) or the su(1,1), could be found in the
studied quantum system, we are able to use some well-known and established the-
orems of the control theory to investigate this system. Hence, this has to force us
S.-H. Dong et al. / Annals of Physics 315 (2005) 566581 567
to make use of some necessary and useful mathematical tools including the alge-
braic, group theoretical, and topological methods for studying the quantum con-
trol problem since the systems quantum Hamiltonian, sometimes, is associated
to the Lie algebras such as the compact Lie algebra su(2) or noncompact one
su(1,1) for one-dimensional quantum system. For example, the Morse and MPT
potentials widely used in molecular physics to describe the stretching and bending
vibrations, respectively, have been identied as a compact Lie algebra su(2) for the
pure states. A noncompact Lie algebra su(1,1) is identied for the Morse potential
by the potential group approach, which corresponds to a xed energy but dierent
potential strengths [19]. That a quantum system is completely controllable or not
depends on the knowledge of the Lie algebra generated by the systems quantum
Hamiltonian and the corresponding properties of the Lie group structure [7]. This
arises from the fact that every control algorithm for the studied quantum system
has to consider the nature of the underlying Lie algebra since the knowledge of
the structure of the Lie algebra may help in understanding the dynamics and
designing the control algorithms. For many quantum systems, however, the Lie
algebra is not known. The determination of the Lie algebra has become one of
the main questions in the quantum control for the quantum systems. How is the
system taken to the corresponding Lie algebra has been addressed in [20,21] for
some basic Lie algebras su(2) and su(1,1) in some detail since they have been
widely used in the various branches of physics and chemistry. Generally speaking,
it is possible to constitute the dynamical Lie algebra for the studied quantum sys-
tem by the factorization method or supersymmetric quantum mechanics approach.
We suggest the reader refer to [20,21] for more information. To our knowledge,
there are two kinds of dierent dynamical Lie algebras widely and popularly used
in physics and chemistry, e.g., the compact Lie algebra su(2) or su(n) (n being the
dimension of Hilbert space) and the noncompact one su(1,1). The control theories
are very dierent for them.
Recently, the controllability of the quantum system with the dynamical Lie alge-
bra has been studied [1,4,7,8,11,16,18,2224]. Therefore, it is possible to establish the
relation between the quantum system and the controllability through the dynamical
Lie algebra. That is to say, once the dynamical Lie algebra such as the su(2) or the
su(1,1) is found for the studied quantum system, we are able to study the controlla-
bility of this system through some well-known theorems on the control theory
[1,4,7,8,11,18,2224], which have supplied answers to questions arising in the quan-
tum control and given the basic results for the classical Lie algebra theory for the
type of controllability known as pure state controllability. It should be pointed out
that the conditions on the control depend on the quantum system being closed or
open, the quantum state being pure or mixed, the dynamical Lie algebra being com-
pact or noncompact. Moreover, a necessary and sucient condition for complete
controllability is that the dynamical Lie algebra of the quantum system is su(n), as
addressed by Schirmer et al. [18] (see Theorem 3). With this spirit, for example, Tarn
at al. [16] have carried out the controllability of the quantum system for the MPT
potential with the continuous spectra in terms of the noncompact Lie algebra
su(1,1). On the other hand, we have studied the controllability of the quantum
568 S.-H. Dong et al. / Annals of Physics 315 (2005) 566581
system for the Morse potential with the discrete bound states by the compact dynam-
ical Lie algebra su(2) [17].
The reasons why the great eorts have been made in quantum control are follow-
ing. First, it is possible to understand the basic physics associated with the quantum
structures such as the quantum uctuation and other quantum eects and to explore
their controllability and design a new species of useful devices. Therefore, the inves-
tigation in this recently sprung up eld will be excited and potential. Second, it is
known that the quantum control is a kind of underlying technology which makes
quantum computation possible. In turn, the quantum information processing may
provide the quantum control with many novel techniques for performing feedback
and state estimation, which are the fundamental problems in quantum information
processing, with possible applications to algorithmics, quantum cryptographic proto-
cols, quantum interrogation or potential applications in nanotechnology and nano-
science, etc. Third, the study in this eld is possibly used to investigate the problem
of controlling a two-level quantum system in implementations of the quantum com-
puters because the system like the spin 1/2 particle in electro-magnetic eld has been
used to perform logic operation in implementation of the quantum computers [2].
The purpose of this work is twofold. The rst is to review the basic control theory
to study the controllability of the quantum system for the MPT potential. The second
is to nd which type of dynamical Lie algebra exists for this quantum system with the
MPT potential. Once the dynamical Lie algebra is found, we are able to use the well-
known theorems of the control theory to investigate the controllability of this quan-
tum system following the similar approach to that of Dong et al. [17].
This paper is organized as follows. Section 2 is devoted to reviewing the prelimin-
ary on the control theory. In Section 3, the properties of the MPT potential are
briey studied. The creation and annihilation operators are constructed directly
from the normalized wave functions with the factorization method [25]. We nd that
these operators
^
L

and
^
L
0
satisfy the commutation relations of an su(2) algebra. We
analyze the controllability of this quantum system with the compact SU(2) group in
Section 4. This work ends with some concluding remarks in Section 5.
2. Preliminary on the control theory
Generally speaking, the quantum system, whose state W(t) evolves from the initial
state W(t = 0) = W
0
, can be described by the time-dependent Schro dinger equation as
ih
d
dt
Wt HWt H
0

m
i1
H
i
u
i
t
_ _
Wt; 1
where u
i
(t) are control functions of the quantum system. Through adjusting these
control functions, it is possible to guide the time development of the state W(t) so
as to attain a specied objective. Generally speaking, the state vector could be
expressed by a unitary transformation W(t) = U(t)W
0
on the initial state W
0
, where
U(t) is a unitary time-evolution operator with the form
S.-H. Dong et al. / Annals of Physics 315 (2005) 566581 569
Ut expiHt=h exp
^
Ht; h 1; 2
where we have denoted the skew-Hermitian counterpart iH of H by
^
H as done by
Clark et al. [22].
1
The
^
H
0
describes the Hamiltonian of free evolution for the given quantum
system to be controlled (e.g., a molecule), while the
^
H
i
(with i = 1,2, . . . , m)
can be interpreted as the eect of the external perturbations, which are su-
ciently weak in comparison with that of the
^
H
0
. The energy levels thus are
not changed signicantly by the perturbations. Additionally from the point of
view of quantum theory, imposing unit norm at the initial time t = 0, we have
W(t)|W(t) = 1 "t, which means that the system evolves on the unit sphere S
H
in
the Hilbert space H.
To address the existence issue systematically, it is necessary to precise the reach-
able sets and controllability. As mentioned above, the purpose of this work is to
study the controllability of the quantum system for the MPT potential in terms of
the well-known and established theorems for the control theory on Lie groups
[1,23,24], that is to say, what we are interested in is which type of dynamical group
is hidden in the studied quantum system and we then determine the controllability of
this system is strongly completely controllable or not. For simplicity, we recall some
most fundamental and necessary denitions and theorems, which are useful for
studying the quantum system for the MPT potential with the discrete bound states
only. The reader is advised to refer to [1,23,24] for more information and a rigorous
treatment on the subject.
Notation. Throughout this work, G will denote a compact semi-simple Lie group.
Let G be the Lie algebra of right invariant vector elds on G. The creation operator
and the annihilation one are denoted by
^
J

^
L

and
^
J

^
L

, respectively. M denotes
the differential manifold.
Denition 1. For the initial state W
0
and the nal state W
f
with the properties W
0
,
W
f
2 M, where M represents the nite or innite-dimensional differential manifold.
We say that the state W
f
is reachable from the initial state W
0
at time t
f
if there exists
a control u(t) such that W(t = t
f
|u,W
0
) = W
f
. The corresponding states are denoted
as R
t
f
W
0
. The set of states reachable from the initial one W
0
at some positive time
t are R(W
0
) =

t > 0
R
t
(W
0
).
It should be addressed that this concept has become one of the cornerstones of
mathematical system theory. It has also become a discipline widely accepted both in
classical dynamics and in quantum control theory. What this denition cares about
is that the nal state W
f
can be obtained from the initial state W
0
at time t
f
through a
control u(t).
1
On the other hand, it should be mentioned that the control system (1) divided by ih h 1 was rst
studied by Jurdjevic and Sussmann [23] in their classical work, where the system (1) was denoted by (X, U),
i.e.,
dxt
dt
X
0
xt

m
i1
u
i
tX
i
xt where X
0
, . . . , X
m
are the right-invariant vector elds on the Lie
group G. Likewise, the u
i
(t) are the control functions. To our understanding, the so-called right-invariant
vector elds on the Lie group G are nothing but the state space H of the studied system.
570 S.-H. Dong et al. / Annals of Physics 315 (2005) 566581
Denition 2. We say that the studied control system is strongly complete controlla-
bility if R
t
(W
0
) = M holds for all times t > 0 and all W
0
2 M. The system is com-
pletely controllable if R(W
0
) = M holds for all W
0
2 M.
Theorem 1. If dim Gm d holds for all m 2 M, then quantum system (1) is
strongly completely controllable, where G is the Lie algebra and it is assumed dim
M = d < 1.
This theorem was rst given by Kunita [24] and then applied by Huang et al.
[1]. What this theorem is concerned is that the dimension d of the Lie algebra gen-
erated by the studied quantum system is nite. It is known that a quantum system
is completely controllable or not depends on the knowledge of the Lie algebra gen-
erated by the systems quantum Hamiltonian as well as the property of the Lie
group structure. For a given quantum system, whose state spaces evolve on the real
manifold M, which resides in a nite or innite-dimensional space. If the Lie alge-
bra generated by the systems quantum Hamiltonian is compact (e.g., the state
spaces of the Lie algebra su(2) are nite), the dimension d of this Lie algebra is
nite, from which we can conclude that the studied quantum system is strongly
completely controllable.
Another theorem is the JurdjevicSussmann theorem proposed in 1972 [23]. In
their classical work, they have studied the control systems on Lie groups and found
the necessary and sucient conditions for the system to be controllable. The neces-
sary condition is that the studied system has the accessibility property. They have
shown that this condition is also sucient if the Lie algebra G is connected or com-
pact. Otherwise, the accessibility is not sucient for controllability. We now summa-
rize this theorem below.
Theorem 2. A necessary condition for system (1)
^
H; u to be controllable is that the
Lie group G connected. If Lie group G is compact, or if the system is homogeneous, the
condition is also sufcient.
This theorem given by Jurdjevic and Sussmann [23] implies whether the quantum
systemis controllable or not depends onone of twoconditions. The rst is the Lie group
G generated by the quantum system is connected and the second is that G is compact.
Theorem 3. (see [18]) A necessary and sufcient condition for a quantum system
governed by Eq. (1) and dynamical Lie algebra G to be: (1) completely controllable is
G un, (2) pure state controllability is G sun, or if N is even, G spn=2.
This theorem can be used in practice to decide the degree of the controllability of
the studied quantum system by constituting a suitable dynamical Lie algebra for the
studied quantum system and then to determine whether the Lie algebra is isomorphic
to one of the algebras mentioned above.
On basis of these theorems, it is possible to study the controllability of the studied
quantum system. Therefore, the main task is to nd which type of the dynamical
group is for this quantum system. In this work, we take the quantum system for
the MPT potential as an example to study the controllability since this potential is
widely and popularly used in molecular physics.
S.-H. Dong et al. / Annals of Physics 315 (2005) 566581 571
3. The MPT potential
3.1. Brief introduction
We now review its importance in physics and chemistry since the Po schlTeller
(PT) potential representing one of the most studied anharmonic systems has been
studied by many authors [2630]. On the other hand, the Schro dinger equation with
this potential plays an important role in soliton mathematics, from which the
multi-soliton solutions of the nonlinear Kortewegde Vries (KdV) equation can be
explicitly constructed [31].
Before further proceeding to study the MPT potential following the approach ap-
plied in [17], it is necessary to review the another algebraic method [32] to compare
the dierence between these two methods and conrm the coincidence of the result.
For the xed j, the dynamic symmetry is characterized by the group chain
U(2) O(2). Following [33], we dene the following operators:
^
J
z
i
o
o/
; 4a
^
J

e
i/

o
oh
i cot h
o
o/
_ _
; 4b
^
J
2

1
sin h
o
oh
sin h
o
oh
_ _

1
sin
2
h
o
2
o/
2
_ _
: 4c
The simultaneous eigenfunctions of
^
J
2
and
^
J
z
satisfy
^
J
2
v
m
j
jj 1v
m
j
;
^
J
z
v
m
j
mv
m
j
5
with
v
m
j
h; / u
m
j
he
im/
; 6
where u
m
j
h satises the following dierential equation:

1
sin h
o
oh
sin h
o
oh
_ _

m
2
sin
2
h
_ _
u
m
j
h jj 1u
m
j
h: 7
The substitution of cos h = tanhy into Eq. (7) allows us to obtain

d
2
dy
2

jj 1
cosh
2
y
_ _
u
m
j
y m
2
u
m
j
y; m j; . . . ; 0; . . . ; j: 8
The corresponding eigenvalue is thus given by
E
m
bm
2
; 9
572 S.-H. Dong et al. / Annals of Physics 315 (2005) 566581
where the scale factor b is inserted to provide the appropriate dimensions [32]. It
is shown that
^
J
2
is the square of the angular momentum with the commutation
relations

^
J
i
;
^
J
j
i
ijk
^
J
k
; 10
where the operators
^
J
i
i 1; 2; 3 can be interpreted as the generators of the dynam-
ical group SU(2) for the PT dynamical system. Nevertheless, we attempt to obtain
the creation and annihilation operators only from the normalized eigenfunctions
without introducing any other auxiliary parameter (e.g., / used above) and then con-
stitute the dynamical group SU(2) naturally as will be shown in Section 3.3.
As shown by Tarn et al. [16], the purpose for constructing the creation and anni-
hilation operators is to constitute a suitable dynamical algebra for the studied quan-
tum system and then to apply these operators to nd the corresponding
combinations of the Hamiltonian. From the dynamical algebra as well as some
well-known and established theorems, we can determine the studied quantum system
is strongly completely controllable or not.
3.2. The eigenvalues and eigenfunctions
For completeness and clearness, it is necessary to study the construction of the
creation and annihilation operators and then study the controllability of this quan-
tum system. Generally, the MPT potential can be described as [26,29,30]
V x
D
cosh
2
ax
; 11
where D is the depth of the well and a is related with the range of the potential, while
x gives the relative distance from the equilibrium position. The corresponding Schro -
dinger equation with this potential can be expressed as
d
2
W
j
n
x
dx
2

2l
h
2
E
D
cosh
2
ax
_ _
W
j
n
x 0; 12
where l is the reduced mass of the molecule and j is related with the depth of the
potential well as dened below. Before further proceeding, we introduce the follow-
ing notations [27]:

2lE
a
2
h
2
_
; jj 1
2lD
a
2
h
2
; j
1
2
1 2k 13
with
k

1
4

2lD
a
2
h
2
_
; m 2k 2j 1; 14
where m has been introduced because of its relevance for the identication of
the creation and annihilation operators with an su(2) algebra. It is shown that
S.-H. Dong et al. / Annals of Physics 315 (2005) 566581 573
Eqs. (8) and (12) describe the same dynamical system since jj 1
2lD=a
2
h
2
.
From these denitions given above and the change of variable u = tanh(ax), Eq.
(12) can be expressed as
d
du
1 u
2

dW
j
n
u
du
_ _
jj 1

2
1 u
2
_ _
W
j
n
u 0: 15
If we introduce the variable n
1
2
1 u and make the substitution
W
j
n
u 1 u
2

=2
wu, we obtain
n1 n
d
2
wu
dn
2
1 1 2n
dwu
dn
j1 j wu 0: 16
The nite solution for u = 1 is given by
W
j
n
u 1 u
2

=2
F j; j 1; 1;
1
2
1 u: 17
For W
j
n
u to remain nite for u = 1, we nd the condition j = n, where
n = 0,1,2, . . . ; then F j; j 1; 1;
1
2
1 u is a hypergeometric polynomial
of degree n. In this case, the eigenvalue can be determined by the constraint condi-
tion j = n and expressed as
E
n

a
2
h
2
2l
j n
2

a
2
h
2
2l
m 1
2
n
_ _
2
: 18
From the relation between the Gegenbauer polynomials and the hypergeometric
functions [34]
C
k
n
x
C2k n
n!C2k
F n; 2k n;
1
2
k;
1 x
2
_ _
; 19
substitution of this into Eq. (17) leads to
W
j
n
u N
j
n
1 u
2

jn=2
C
j1=2n
n
u; 20
with the normalization constant N
j
n
to be determined below. After complicated alge-
braic manipulation and from the normalization condition
_
1
1
W
j
n
x
2
dx
N
j
n

2
a
_
1
1
1 u
2

jn1
C
jn1=2
n
u
2
du 1; 21
we can obtain
N
j
n

aCn 1Cj n 1=2C2j 2n 1

p
1
2
Cj nC2j n 1

: 22
574 S.-H. Dong et al. / Annals of Physics 315 (2005) 566581
We note that for j integer the state associated with null energy is not normalizable.
Thus the last bounded state corresponds to jn = 1, which implies that
n
max
= j1 = (m3)/2.
3.3. The creation and annihilation operators
Following our previous work [17,35], we rst dene the diagonal operator ^n as
^nW
m
n
nW
m
n
: 23
We now address the problem of nding the creation and annihilation operators
with the factorization method, i.e., we intend to nd the dierential operators
^
L

with the property

^
L

W
j
n
u

W
q
n1
u: 24
Specically, we look for operators with the form
^
L

u
d
du
B

u; 25
which can be obtained by acting the dierential operator
d
du
on the MPT wave
functions. Such a method, which is dierent from the traditional one where an aux-
iliary variable was introduced [19], has been successfully applied to other quantum
systems [35].
We now construct the annihilation operator
^
L

. To this end, using the formula

[36]
dC
k
n
t
dt
2kC
k1
n1
t; 26
together with Eq. (20), we obtain
dW
j
n
u
du

uj n
1 u
2
W
j
n
u
2j 2n 1

1 u
2
p
N
j
n
N
j
n1
W
j
n1
u: 27
This can be further rearranged as

1 u
2
p
d
du

um 2n 1
21 u
2

_ _

m 2n 1
m 2n 1
_
W
m
n
u

nm n
_
W
m
n1
u; 28
where we have used the relation m = 2j + 1.
We thus dene the annihilation operator
^
L

as
^
L

1 u
2
p
d
du

um 2^n 1
21 u
2

_ _

m 2^n 1
m 2^n 1
_
29
with the property
^
L

W
m
n
u

W
m
n1
u

nm n
_
W
m
n1
u: 30
As we see, this operator annihilates the ground state W
j
0
u, as expected from an
annihilation operator.
S.-H. Dong et al. / Annals of Physics 315 (2005) 566581 575
We now proceed to nd the corresponding creation operator
^
L

. To this end, we
consider the formula [36]
2k12k1xC
k
n
x4kk11x
2
C
k1
n1
x2kn1n1C
k1
n1
x: 31
Substitution of this into Eq. (26) leads to
dW
j
n
u
du

uj n
1 u
2
W
j
n
u
n 12j n

1 u
2
p
2j 2n 1
N
j
n
N
j
n1
W
j
n1
u; 32
from which we have

1 u
2
p
d
du

um 2n 1
21 u
2

_ _

m 2n 3
m 2n 1

W
m
n
u

n 1m n 1
_
W
m
n1
u: 33
Likewise, we can dene the creation operator as
^
L

1 u
2
p
d
du

um 2n 1
21 u
2

_ _

m 2n 3
m 2n 1

34
with the eect on the wave functions
^
L

W
m
n
u

W
m
n1
u

n 1m n 1
_
W
m
n1
u: 35
We now establish the algebra associated to the operators
^
L

. Based on Eqs. (30)

and (35), we calculate the commutator
^
L

;
^
L

^
L

;
^
L

W
m
n
u 2
0
W
m
n
u; 36
where we have introduced the eigenvalue

0
n
m 1
2
: 37
We thus dene the operator
^
L
0
^n
m 1
2
: 38
It should be pointed out that the operator
^
L
0
plays the role of the angular
momentum operator
^
J
z
, one of the generators
^
J
;z
of the SU(2) group in the
usual angular momentum commutation relations, as studied in [33], namely,
the operator
^
L
0
and the ladder operators
^
L

satisfy the following commutation

relations:

^
L

;
^
L

2
^
L
0
;
^
L
0
;
^
L

^
L

;
^
L
0
;
^
L

^
L

; 39
which correspond to the commutation relations of the generators of the SU(2)
group. This result coincides with the description of a nite discrete spectrum for
the MPT potential [32].
576 S.-H. Dong et al. / Annals of Physics 315 (2005) 566581
Additionally, the Casimir operator is given by
^
CW
m
n
u
^
L
2
0

1
2

^
L

^
L

^
L

^
L

_ _
W
m
n
u jj 1W
m
n
u; 40
where j, the label of the irreducible representations of the su(2) algebra, is given by
j
m 1
2

N
2
: 41
which has been introduced. Thus the operator
^
J
2
is related with the number operator
^
N through
^
J
2

1
4
^
N
^
N 2: 42
If j takes an integer or half integer value, it determines the irreducible representation
D
j
of the su(2) algebra, which is the dynamical algebra of the MPT potential for the
discrete bound states.
It is known from the commutation relations (39) that
^
L
0
is the projection of the
angular momentum m, consequently we have
n
m 1
2
m: 43
Therefore the ground state (n = 0) corresponds to m = j, while the maximum
number of states n
max
= (m3)/2 and consequently m
max jn
max
1, in accordance
with the constraint condition = jn = 1 for the last bound state. The MPT wave
functions are then associated to one branch (in this case to m 6 1) of the su(2)
representations.
Moreover, it is shown from Eqs. (13), (18), (30), and (35) that, for the su(2) alge-
bra, the Hamiltonian acquires the simple form
^
H
0

hx
m
^
L
2
0

hx
m
m 1
2
^n
_ _
2

hx
2m
^
L

^
L

^
L

^
L

1
2

m
2
2
_ _
; 44
where
x
hb
2
m
2l
: 45
On the other hand, the wave functions can be expressed as
W
m
n
u N
m
n
^
L
n

W
m
0
u 46
with
N
m
n

m n 1!
n!m 1!

: 47
To study the harmonic limit of the MPT potential, we rst introduce the
renormalization
S.-H. Dong et al. / Annals of Physics 315 (2005) 566581 577
b
y

^
L

m
p ; b
^
L

m
p ; b
0

2
^
L
0
m
; 48
which implies that the operators b

have the properties

b
y
W
m
n

n 1 1
n 1
m
_ _

W
m
n1
; bW
m
n

n 1
n
m
_ _
_
W
m
n1
: 49
Thus, we can reexpress the Hamiltonian (44) as
^
H
0

hx
2
^
b
y
b bb
y

1
2m

m
2
_ _
: 50
Actually, by studying the harmonic limit of the MPT potential we can further
prove whether the dynamical group SU(2) for the bound states of the MPT poten-
tial is correct or not. This can be obtained when a 0 and D 1, but keeping
the product k = 2a
2
D nite, so that the expansion of the exponential functions in
(11), leads to
lim
D!1
V x
1
2
kx
2
: 51
For D 1, i.e., m 1, together with the approximation cosh(ax) . 1 and
sinh(ax) . ax in the harmonic limit a 0, from Eqs. (29), (34), and (38), we
have
lim
m!1
^
b
y

1

m
p
a
d
dx

m
p
a
2
x

lx
2h
_
x

h
2lx

d
dx
^a
y
; 52a
lim
m!1
^
b
1

m
p
a
d
dx

m
p
a
2
x

lx
2h
_
x

h
2lx

d
dx
^a; 52b
lim
m!1
^
b
0
1 52c
with
x
a
2
h
2l
m

2Da
2
l

;
where we have used the relation m=2 k

2lD
a
2
h
2
_
in the harmonic limit D 1. The
operators ^a
y
and ^a thus satisfy the bosonic commutation relations
^a; ^a
y
1; ^a; ^a ^a
y
; ^a
y
0, as expected. That is to say, the su(2) algebra is con-
tracted to the Weyl algebra in the harmonic limit lim
m!1
G
su2
lim
m!1
f
^
b
y
;
^
b;
^
b
0
g
f^a
y
; ^a; 1g:
The advantage of the present method is that we can construct the creation
and annihilation operators only from the normalized eigenfunctions (20) without
introducing any other auxiliary parameter, which is dierent from the traditional
operator method [19,32]. Also, the creation and annihilation operators can be
expressed explicitly.
578 S.-H. Dong et al. / Annals of Physics 315 (2005) 566581
4. The controllability
As we know the Hamiltonian
^
H
0
studied above dominates this quantum system,
but the others
^
H
i
i 1; 2; :::; m can be interpreted as the eect of the external per-
turbation. We now analyze the controllability of this quantum system following [16].
It is shown from Eq. (13) that, for a given quantum number j which implies that the
potential well is xed, the energy level will change with the quantum number
m(m 6 |j|). Unlike the work by Tarn et al. [16], where this potential was studied
by the noncompact Lie group SU(1,1) following the potential group approach,
i.e., the energy level is xed but the potential will change with the quantum number
m. The quantum system studied by them corresponds to the MPT potential with the
continuous spectra. However, in the present work, the quantum system corresponds
to the MPT potential with the discrete bound states only. There exists the essential
dierence between them. On the other hand, it is shown from Eq. (40) that the Casi-
mir operator C keeps invariant for the given quantum number j. As addressed in [16],
if the free evolution is driven by a Casimir invariant (a dynamical symmetry), the
control Hamiltonian can be physically realized by the group generators (
^
L
;0
used
in this work), the quantum system can be guaranteed to be strongly completely con-
trollable. In the physical language, this quantum system can be solved from the com-
bination of the creation and annihilation operators as shown in Eqs. (44) and (50).
More importantly, it is found from Eq. (46) that the eigenfunctions of this system
can be obtained directly from the action of the creation operator
^
L

on the ground
state W
m
0
u, which can be guaranteed by the fact that the transition from the initial
state to the selected target state can be realized theoretically and experimentally [3,4].
Therefore the controllability of this quantum system will conrm powerfully that
such a combination exist. We now construct the quantum system
dWt
dt

^
H
0
u
1
t
^
L

^
L

iu
2
t
^
L

^
L

Wt; W0 W
0
; 53
where the quantities W(t) and W
0
should be interpreted as abstract state vectors
rather than wave functions since the eigenfunctions of the quantum system can form
a nite-dimensional Hilbert space, which constructs the compact Lie group SU(2) as
shown by Eq. (39), i.e., the quantum system evolves on the nite-dimensional state
spaces with the dimensions 2j + 1. On the other hand, it is shown from Eq. (53) that
we can take
^
H
1

^
L

^
L

and
^
H
2
i
^
L

^
L

. Accordingly, it is shown from The-

orems 1, 2,3 that this quantum system is strongly completely controllable since SU(2)
group generated by
^
H
0
;
^
H
1
; and
^
H
2
is connected, compact and nite-dimensional
Hilbert space with the dimensions 2j + 1 [37,38]. Therefore, we draw a conclusion
that this quantum system is strongly completely controllable.
5. Concluding remarks
The controllability of a quantum system for the MPT potential with a compact
Lie group SU(2) are investigated. The creation and annihilation operators of this po-
S.-H. Dong et al. / Annals of Physics 315 (2005) 566581 579
tential are constructed and associated to the su(2) algebra. It is shown that this quan-
tum system with the nondegenerate discrete bound states can, in principle, be
strongly completely controlled, i.e., the system eigenfunction can be guided by the
external eld to approach arbitrarily close to a selected target state at any chosen
time, which can be theoretically realized by the actions of the creation and annihila-
tion operators on the ground state. The harmonic limit of this quantum system has
been briey analyzed and it is shown that the su(2) algebra is contracted to the Weyl
algebra. This implies that the expressions of the ladder operators
^
L

and the
^
L
0
intro-
duced in this work are correct. Otherwise, it is impossible to get the operators a

and
a from the ladder operators
^
L

in the harmonic limit. It should be pointed out that

the study in the present work is concerned with the case of driving the system to a
particular target state. How to generate a particular transformation rather than a
state is the main task of the quantum information community.
Acknowledgments
The authors thank the kind referee for the positive and invaluable suggestions
from which the manuscript has been improved greatly. This work was partly sup-
ported by IMP of Mexico.
References
[1] G.M. Huang, T.J. Tarn, J.W. Clark, J. Math. Phys. 24 (1983) 2608.
[2] D.P. DiVincenzo, Science 270 (1995) 13.
[3] H. Rabitz, R. de Vivie-Riedle, M. Motzkus, K. Kompa, Science 288 (2000) 824.
[4] V. Ramakrishna, H. Rabitz, Phys. Rev. A 61 (2000) 032106.
[5] P. Kral, Z. Amitay, M. Shapiro, Phys. Rev. Lett. 89 (2002) 063002.
[6] N. Khaneja, S.J. Glaser, Chem. Phys. 267 (2001) 11;
Phys. Rev. A 63 (2001) 032308.
[7] A.I. Solomon, S.G. Schirmer, Int. J. Mod. Phys. B 16 (2002) 2107.
[8] D. DAlessandro, in: Proceedings of the 39th IEEE Conference on Decision and Control, Sydney,
Australia, December, 2000.
[9] M. Zhao, S. Rice, J. Chem. Phys. 95 (1991) 2465.
[10] J.L. Herek, W. Wohlleben, R.J. Cogdell, D. Zeidler, M. Motzkus, Nature 417 (2002) 533.
[11] D. DAlessandro, M. Dahleh, IEEE Trans. Automat. Contr. 46 (2001) 866.
[12] S. Chu, Nature 416 (2002) 206;
T. Brixner, N.H. Damrauer, P. Niklaus, G. Gerber, Nature 414 (2001) 57;
S.M. Hurley, A.W. Castleman, Science 292 (2001) 648;
W.P. Schleich, Nature 397 (1999) 207;
D. Meshulach, Y. Silberberg, Nature 396 (1998) 239;
de Araujo LEE, I.A. Walmsley, C.R. Stroud, Phys. Rev. Lett. 81 (1998) 955;
D. Deutsch, Proc. R. Soc. Lond. A 400 (1985) 97.
[13] P.W. Shor, in: Sha Godwasser (Ed.), Proceeding of 35th Annual Symposium on Foundations of
Computer Science, Sante Fe, NM, 1994, IEEEComputer Society Press, Los Alomitos, CA, 1994, p. 124.
[14] L. Grover, in: Proceeding of the 28th Annual ACM Symposium on the Theory of Computation,
ACM Press, New York, 1996, p. 212.
580 S.-H. Dong et al. / Annals of Physics 315 (2005) 566581
[15] P.A. Benio, J. Stat. Phys. 22 (1980) 563;
Phys. Rev. Lett. 48 (1982) 1581;
J. Stat. Phys. 29 (1982) 515;
Ann. N.Y. Acad. Sci. 480 (1986) 475;
R.P. Feynman, Opt. News 11 (1985) 11;
Found. Phys. 16 (1986) 507;
Int. J. Theor. Phys. 21 (1982) 467.
[16] T.J. Tarn, J.W. Clark, D.G. Lucarelli, in: Proceeding of the 39th IEEE Conference on Decision and
Control, Sydney, Australia, December, 2000.
[17] S.H. Dong, Y. Tang, G.H. Sun, Phys. Lett. A 320 (2003) 145.
[18] S.G. Schirmer, A.D. Greentree, V. Ramakrishna, H. Rabitz, J. Phys. A: Math. Gen. 35 (2002) 8315;
S.G. Schirmer, A.I. Solomon, J.V. Leahy, J. Phys. A: Math. Gen. 35 (2002) 8551;
S.G. Schirmer, A.I. Solomon, J.V. Leahy, J. Phys. A: Math. Gen. 35 (2002) 4125;
S.G. Schirmer, I.C.H. Pullen, A.I. Solomon, J. Phys. A: Math. Gen. 35 (2002) 2327.
[19] M. Berrondo, A. Palma, J. Phys. A: Math. Gen. 13 (1980) 773.
[20] A. Frank, P. Van Isacker, Algebraic Methods in Molecular and Nuclear Structure Physics, Wiley,
New York, 1994.
[21] F. Iachello, R.D. Levine, Algebraic Theory of Molecules, , Oxford Univ. Press, 1995.
[22] J.W. Clark, D.G. Lucarelli, T.J. Tarn, nucl-th/0205005.
[23] V. Jurdjevic, H. Sussmann, J. Dier. Equa. 12 (1972) 313.
[24] H. Kunita, Appl. Math. Optim. 5 (1976) 89.
[25] E. Schro dinger, Proc. R. Irish Acad. A 46 (9) (1940) 183;
L. Infeld, T.E. Hull, Rev. Mod. Phys. 23 (1951) 21;
G. Haidinger, N. Bessis, G. Bessis, J. Math. Phys. 15 (1974) 716.
[26] G. Po schl, E. Teller, Z. Phys. 83 (1933) 143.
[27] S.H. Dong, R. Lemus, Int. J. Quantum Chem. 86 (2002) 265;
S.S. Dong, S.H. Dong, Czech. J. Phys. 52 (2002) 753.
[28] T. Carrington, J. Mol. Phys. 70 (1990) 757;
F. Iachello, S. Oss, Chem. Phys. Lett. 205 (1993) 285.
[29] L.D. Landau, E.M. Lifshitz, Quantum Mechanics, Non-Relativistic Theory, third ed., Pergamon,
1977.
[30] S. Flu gge, Practical Quantum Mechanics, Springer-Verlag, Berlin, 1971.
[31] P. Lax, Commun. Pure Appl. Math. 27 (1968) 97.
[32] Y. Alhassid, F. Gu rsey, F. Iachello, Ann. Phys. 148 (1983) 346.
[33] A. Messiah Quantum Mechanics, vol. II, North-Holland, Amsterdam, 1962.
[34] Z.X. Wang, D.R. Guo, Special Functions, World Scientic, Singapore, 1989.
[35] S.H. Dong, R. Lemus, A. Frank, Int. J. Quantum Chem. 86 (2002) 433;
S.H. Dong, Can. J. Phys. 80 (2002) 129;
S.H. Dong, Z.Q. Ma, Am. J. Phys. 70 (2002) 520;
S.H. Dong, Z.Q. Ma, Int. J. Mod. Phys. E 11 (2) (2002) 155;
S.S. Dong, S.H. Dong, Int. J. Mod. Phys. E. 11 (2002) 265;
S.H. Dong, App. Math. Lett. 16 (2003) 199;
S.H. Dong, G.H. Sun, Y. Tang, Int. J. Mod. Phys. E 12 (2003) 809;
S.H. Dong, Comput. Math. Appl. 47 (2004) 1035;
S.H. Dong, G.H. Sun, M. Lozada-Cassou, Phys. Lett. A 328 (2004) 299.
[36] J.D. Talman, Special Functions: A Group Theoretic Approach, Mathematical Physics Monograph
Series, W.A. Benjamin, New York, 1968.
[37] A. Perelomov, Generalized Coherent States and their Applications, Springer, New York, 1985.
[38] R. Gilmore, Lie Groups, Lie Algebras and Some of Their Applications, Wiley, New York, 1974.
S.-H. Dong et al. / Annals of Physics 315 (2005) 566581 581