.

Letter
.
SCIENCE CHINA
Physics, Mechanics & Astronomy
September 2013 Vol. 56 No. 9: 18101814
doi: 10.1007/s11433-013-5152-z
c _ Science China Press and Springer-Verlag Berlin Heidelberg 2013 phys.scichina.com www.springerlink.com
Schr odinger cat states prepared by Bloch oscillation
in a spin-dependent optical lattice
WU BaoJun & ZHANG JiangMin
*
Beijing National Laboratory for Condensed Matter Physics, Institute of Physics, Chinese Academy of Sciences, Beijing 100080, China
Received March 27, 2012; accepted July 2, 2012; published online July 22, 2013
We propose to use Bloch oscillation of ultra-cold atoms in a spin-dependent optical lattice to prepare Schr odinger cat states.
Depending on its internal state, an atom feels dierent periodic potentials and thus has dierent energy band structures for its
center-of-mass motion. Consequently, under the same gravity force, the wave packets associated with dierent internal states
perform Bloch oscillation of dierent amplitudes in space and in particular they can be macroscopically displaced with respect to
each other. In this manner, a cat state can be prepared.
Bloch oscillation, Schr odinger cat, spin-dependent optical lattice
PACS number(s): 03.75.Lm, 03.65.Ud, 03.65.Sq
Citation: Wu B J, Zhang J M. Schr odinger cat states prepared by Bloch oscillation in a spin-dependent optical lattice. Sci China-Phys Mech Astron, 2013, 56:
18101814, doi: 10.1007/s11433-013-5152-z
Bloch oscillation is a peculiar response of a particle in a peri-
odic potential to a weak external force [1,2]. Under the drag
of a constant force F, the wave packet of the particle oscil-
lates back and forth periodically in space without accelerat-
ing catastrophically in the direction of the force. The strange
phenomena is that a DC bias generates an alternating current.
The reason underlying is because as long as the force is weak
enough, interband transitions are prohibited by the gaps be-
tween the energy bands. Conned in a specic energy band
by such a mechanism, the motion of the particle is to a good
extent captured by the semi-classical equations [3]
dr
dt
=
E
n
(q)
q
,
dq
dt
= F, (1)
where r and q are the center-of-mass and wave vector of the
particle, respectively, and E
n
(q) is the dispersion relation in
the nth band. In the k-space, the particle traverses the Bril-
*Corresponding author (email: wdlang06@gmail.com)
louin zone and follows the topology of a circle, repeatedly at
a constant rate. From these equations, one solves readily the
displacement of the wave packet as
r(t) =
1
F

E
n

q
0
+
Ft

E
n
(q
0
)

, (2)
where q
0
is the initial value of q. We see that r is a pe-
riodic function of time with the period T = 2/aFthis
states the Bloch oscillation in the semi-classical theory. Here
a is the period of the periodic potential and 2/a is the size
of the Brillouin zone. Note that T is independent of the de-
tailed structure of the periodic potential but depends only on
its period.
Suppose initially the particle is at the bottom of the low-
est energy band (n = 0) with q
0
= 0. The maximum dis-
placement r
m
is reached at t = T/2, when the wave vector
q = /a arrives at the edge of the rst Brillouin zone. After-
wards, the velocity of the particle dr/dt reverses. The value
Wu B J, et al. Sci China-Phys Mech Astron September (2013) Vol. 56 No. 9 1811
of r
m
is simply
r
m
=
B
F
, (3)
where B E
0
(/a) E
0
(0) is the band width of the lowest
band. Simple as it is, this equation is used as a reference point
in our work below.
So far, Bloch oscillation has been observed in a vari-
ety of systems, such as semiconductor superlattices [4,5],
cold atoms in optical lattices [611], and photonic lattices
[1214]. For applications, it has found use in microwave gen-
eration [4,5,15], precision force measurements [811], and
coherent transport of matter waves [1618].
In this paper, we propose that Bloch oscillation can also
be used to prepare Schr odinger cat states [1923] in a spin-
dependent optical lattice. The idea is actually very simple.
In a spin-dependent optical lattice, atoms in dierent inter-
nal states see dierent potentials (For example, of dierent
strengths). This non-trivial fact means that they also have
dierent energy band structures for their center-of-mass mo-
tion, which in turn means they will have dierent Bloch os-
cillation modes under the same force (eq. (2)). In particular,
their maximum displacements will be dierent according to
eq. (3) since the Bs may dier, and this implies that the wave
packets corresponding to dierent internal states will be dis-
placed with respect to each other. That is, a cat state can be
prepared.
First of all we need a spin-dependent optical lattice. As
proposed in refs. [24,25] and realized in ref. [26], such an
optical lattice can be constructed by interfering two counter-
propagating laser beams linearly polarized but with an an-
gle between the polarization vectors. The resulting stand-
ing light eld can be decomposed into a
+
and a

po-
larized one with intensities I
+
= I
m
cos
2
(kx + /2) and I

=
I
m
cos
2
(kx /2), respectively. Here k = 2/ is the wave
vector of the laser beams. Such a decomposition is helpful
since the dipole potential for an atom in a state F, m
F
) (the
quantization axis of the atom is along the optical lattice) is
simply the sum of the contributions of the two components.
Below we will use the same system as in ref. [26]. That is,
we choose
87
Rb as the atom, and 1) F = 2, m
F
= 2)
and 0) F = 1, m
F
= 1) as the two atomic internal states.
The spin dependence of the dipole potential is realized by
choosing the laser frequency to resolve the ne structure of
the Rubidium D line. Specically, as in ref. [26], by tuning
the wave length of the optical lattice laser to = 785 nm,
the dipole potentials for an atom in the 1) and 0) states, are
respectively,
V
1
(x; ) = V
m
cos
2

kx

2

,
V
0
(x; ) =
3
4
V
m
cos
2

kx +

+
1
4
V
m
cos
2

kx

2

,
where V
m
I
m
. The point is that if 0 or , V
1
(x) and
V
0
(x) are shifted relative to each other, and more importantly,
have dierent amplitudes. The latter eect results in dierent
band structures as we see in Figure 1. There the potentials
V
1,0
and the corresponding energy bands for the two internal
states are depicted. The parameters chosen are V
m
/E
r
= 5
and = /2, where E
r
=
2
k
2
/2m = 2 3.72 kHz is the
recoil energy of the atom. Note that the width of the lowest
band for the 1) state is 2 0.983 kHz, while that for the
0) state is 21.925 kHz. The two dier almost by a factor
of 2.
Our scheme to generate a Schr odinger cat is according to
the following. Suppose initially the angle = 0 (for this
value of , V
1
= V
0
) and the atom is in the 1) state. As for its
external state, it is assumed to be

i
=

+/a
/a
dq f (q)
0
(q). (4)
Here
0
(q) is the Bloch state in the lowest band with wave
vector q. The weight function f (q) is localized around q = 0
but otherwise unspecied. This condition is easily satised
as long as the spatial size of the wave packet
i
is much
larger than the lattice constant a = /2. Actually, the con-
densate wave function should satisfy this condition if the
condensate is loaded adiabatically from a magnetic trap into
the lattice as is usually done in cold atom experiments. It
can be conrmed that the results presented in the following
are barely aected with dierent choices of f (q), as long as
the localization condition is satised. This fact is consis-
tent with the semi-classical theory in which the details of the
wave packets are irrelevant. Specically, in the simulations
to be presented, f (q) is of the form f (q) exp(q
2
/w
2
) with
wa/ = 0.1 < 1. This value of w corresponds to a wave
packet with a size on the order of 10a.
Then at some moment, by using a microwave pulse we can
prepare the internal state of the atom into an arbitrary super-
position of the 0) and 1) states. Let it be 0) + 1) with

2
+
2
= 1. Subsequently, is adjusted to /2 (suddenly
or smoothly, it does not matter; but in our simulation we sup-
pose the sudden scenario), and the lattice is tilted by with
respect to the horizontal plane. The Bloch oscillation then
begins. At a later time t, the wave function of the atom is of
the form (t) = 0)
0
(x, t) + 1)
1
(x, t). The evolution of
the external wave function
j
(x, t) is given by ( j = 0, 1)
i

t

j
=

2
2m

2
x
2
+ V
j
(x) Fx

j
, (5)
with the force F = mg sin and the initial condition

j
(t = 0) =
i
.
Though the semi-classical theory above gives us an overall
idea of the motion of the wave packets
j
(x, t), here we shall
solve eq. (5) numerically. Snapshots of
j
(x, t) are shown in
1812 Wu B J, et al. Sci China-Phys Mech Astron September (2013) Vol. 56 No. 9
2 1 0 1 2
0
1
2
3
4
5
2x/
V
1
,
0
(
x
)
/
E
r
(a)
0.5 0 0.5
2
3
4
5
6
7
8
q/4
E
n
(
q
)
/
E
r
(b)
Figure 1 (Color online) (a) Dipole potentials V
1
(x) (solid line) and V
0
(x) (dashed line) for an atom in the internal states 1) and 0), respectively. The
parameters are V
m
/E
r
= 5 and = /2. Here E
r
=
2
k
2
/2m is the recoil energy of the atom. (b) The three lowest energy bands for an atom in the 1) (solid
lines) and 0) (dashed lines) internal states.
0 50 100 150
0
0.5
1.0
1.5
2.0
2.5
3.0
2x/
|

1
,
0
(
x
)
|
2

(
1
/

)
(a)
0 50 100 150
0
0.5
1.0
1.5
2.0
2.5
3.0
2x/
|

1
,
0
(
x
)
|
2

(
1
/

)
(b)
0 50 100 150
0
0.5
1.0
1.5
2.0
2.5
3.0
2x/
|

1
,
0
(
x
)
|
2

(
1
/

)
(c)
0 50 100 150
0
0.5
1.0
1.5
2.0
2.5
3.0
2x/
|

1
,
0
(
x
)
|
2

(
1
/

)
(d)
0 50 100 150
0
0.5
1.0
1.5
2.0
2.5
3.0
2x/
|

1
,
0
(
x
)
|
2

(
1
/

)
(e)
0 50 100 150
0
0.5
1.0
1.5
2.0
2.5
3.0
2x/
|

1
,
0
(
x
)
|
2

(
1
/

)
(f)
Figure 2 (Color online) Time evolution of the wave packets
1
(x, t) and
0
(x, t) in a Bloch oscillation cycle. (a) t/T = 0; (b) t/T = 1/6; (c) t/T = 1/3; (d)
t/T = 1/2; (e) t/T = 3/4; (f) t/T = 1. The parameters are the same as in Figure 1 and Fa/E
r
= 0.005. Initially,
1,0
=
i
as dened in (4). Note that the wave
packets are not normalized to unity.
in Figure 2. As expected, in the interval 0 t T/2, the
two wave packets move rightward and gradually depart. At
the turning point t = T/2 (Figure 2(d)), the distance between
the two reaches the maximum. Remarkably, at this point,
the two wave packets are well separatedthe separation be-
tween themis about 50a, which is much larger than their sizes
(= 10a). Note that the wave packets themselves are large
enough to deserve the name macroscopic. Therefore we ob-
tain a desired Schr odinger cat state
(t = T/2) = 1)
1
(t = T/2) + 0)
0
(t = T/2), (6)
for which the internal and external states of the atom are
entangled. The critical condition is that the latter is in two
macroscopically displaced macroscopic states, which corre-
spond to a live and a dead cat, respectively. The coher-
ence between the two packets can be checked by applying a
Wu B J, et al. Sci China-Phys Mech Astron September (2013) Vol. 56 No. 9 1813
microwave pulse to achieve a rotation in the 1), 0) space,
and then turning o the lattice and observing the momen-
tum distribution of the atom in the 1) state by the absorption
imaging method [26]. Note that
1,0
(T/2) are both peaked
around q = /a in momentum space and thus the interfer-
ence pattern will primarily consist of two peaks whose am-
plitudes depend on the rotation as well as and .
Afterwards, the two wave packets move backwards and
return approximately
1)
to their original states at t = T (Fig-
ure 2(f)). That is,
1,0
(T) =
i
up to some global phases [27].
The global wave function is then (T) = (0) + e
i
1))
i
,
where is the dierence of the phases
1,0
accumulated in a
cycle. Thus the atom completes a Bloch oscillation cycle by
getting its internal state rotated somehow. From the point of
view of the cat state, a cycle of Bloch oscillation is a cycle of
birth-growth-death. The cycle can be interrupted by putting
F = 0 at an appropriate time, e.g., at t = T/2, when the cat is
in its largest size.
We have followed the center-of-mass motion of the wave
packets in time. The results are shown in Figure 3. We see
that the semi-classical theory is consistent quantitatively. The
numerically exact results deviate signicantly from the semi-
classical predictions only in the vicinities of t = 0 and t = T.
The reason is that due to the sudden change of , both wave
packets are partially excited and some packets (not visible in
the snapshots) belonging to higher bands are emitted around
t = 0, which may perform Bloch oscillation also and return
around t = T.
Now let us turn to the problem of the feasibility of the
scheme in experiment. In our simulation, Fa/E
r
= 0.005.
It is essential to make sure that this ratio is much smaller
than unity. There are two reasons. First, the potential drop
Fa between two neighboring sites should be much smaller
than the gap between the zeroth (lowest) and the rst bands
(Figure 1(b)) so as to suppress Zener tunneling [2]. Or equiv-
alently, the Brillouin zone, especially its boundary, should
be traversed slowly so that transition into excited bands can
be neglected. Second, according to eq. (3), the distance be-
tween the two wave packets is inversely proportional to F,
thus smaller F means larger separation or larger cat. Of
course, there should be an optimal value of F since the pe-
riod T is also inversely proportional to F. For
87
Rb and a
lattice constant a = /2 = 392.5 nm, the ratio above corre-
sponds to a tilt angle = 4

and a period T = 53 ms. On the

contrary, under the chosen detuning and strength of the opti-
cal lattice, the spontaneous radiation rate
e
of the atom is
about 0.2 s
1
. Thus the atom is long lived enough to oscillate
several circles before incoherent processes set in.
0 0.2 0.4 0.6 0.8 1.0
0
20
40
60
80
100
120
140
t/T

2
x
/

|1 exact
|0 exact
|1 semi-classical
|0 semi-classical
Figure 3 (Color online) Evolution of the center-of-mass of the wave pack-
ets
1,0
(x, t) in a Bloch oscillation cycle. The parameters are the same as in
Figures 1 and 2. The results obtained by solving eq. (5) numerically/exactly
agree very well with those by the semi-classical theory (eq. (2)).
In conclusion, we have proposed that the spin-dependent
optical lattice may oer an opportunity to create Schr odinger
cat states by using Bloch oscillation. Our scheme has several
interesting advantages. First, the cat state experiences birth-
growth-death cycles repeatedly. It would be worthy to study
experimentally how this process is damped in a real optical
lattice, which is believed to be well isolated from the envi-
ronment. Second, if we start from a Bose-Einstein conden-
sate and minimize the atom-atom interaction which is delete-
rious to the Bloch oscillation, it might be possible to create
a collection of atoms condensed in a cat state. We note that
some generalizations are also possible. For example, though
here we focused on the one dimensional case, the scheme
can be directly extended to higher dimensions [18] since
two-dimensional spin-dependent optical lattices have already
been demonstrated experimentally [28]. Furthermore, in con-
trast to the static force considered here, periodically modula-
tions [16,17] are worth consideration also since they may help
to increase the size of Schr odingers cat.
We are grateful to DUAN L M, PANY, and ZHOUD L for stimulating discus-
sions and valuable suggestions. ZHANG J M was supported by the National
Natural Science Foundation of China (Grant No. 11091240226).
1 Bloch F.

Uber die quantenmechanik der elektronen in kristallgittern. Z
Phys, 1929, 52: 555600
2 Zener C. A theory of the electrical breakdown of solid dielectrics. Proc
R Soc Lond A, 1934, 145: 523529
3 Kittel C. Introduction to Solid State Physics. 7th ed. New York: John
Wiley & Sons, 1996
4 Esaki L, Tsu R. Superlattice and negative dierential conductivity in
semiconductors. IBM J Res Dev, 1970, 14: 6165
1) The recurrence of
i
is imperfect since on the one hand both
1,0
are partially excited into higher bands at the beginning due to the sudden change of
, and on the other hand the subsequent evolution is not strictly adiabatic and higher band components still mix in ref. [2]. This feature is not captured by the
semi-classical theory
1814 Wu B J, et al. Sci China-Phys Mech Astron September (2013) Vol. 56 No. 9
5 Waschke C, Roskos H G, Schwedler R, et al. Coherent submillimeter-
wave emission from Bloch oscillations in a semiconductor superlattice.
Phys Rev Lett, 1993, 70: 33193322
6 Dahan M B, Peik E, Reichel J, et al. Bloch Oscillations of atoms in an
optical potential. Phys Rev Lett, 1996, 76: 45084511
7 Salger T, Ritt G, Geckeler C, et al. Bloch oscillations of a Bose-
Einstein condensate in a subwavelength optical lattice. Phys Rev A,
2009, 79: 011605
8 Ferrari G, Poli N, Sorrentino F, et al. Long-lived Bloch oscillations
with Bosonic Sr atoms and application to gravity measurement at the
micrometer scale. Phys Rev Lett, 2006, 97: 060402
9 Poli N, Wang F Y, Tarallo M G, et al. Precision measurement of grav-
ity with cold atoms in an optical lattice and comparison with a classical
gravimeter. Phys Rev Lett, 2011, 106: 038501
10 Gustavsson M, Haller E, Mark M J, et al. Control of interaction-
induced dephasing of Bloch oscillations. Phys Rev Lett, 2008, 100:
080404
11 Fattori M, DErrico C, Roati G, et al. Atom interferometry with a
weakly interacting Bose-Einstein condensate. Phys Rev Lett, 2008,
100: 080405
12 Pertsch T, Dannberg P, Elein W, et al. Optical Bloch oscillations in
temperature tuned waveguide arrays. Phys Rev Lett, 1999, 83: 4752
4755; Morandotti R, Peschel U, Aitchison J S. Experimental observa-
tion of linear and nonlinear optical Bloch oscillations. Phys Rev Lett,
1999, 83: 47564759; Sapienza R, Costantino P, Wiersma D. Optical
analogue of electronic Bloch oscillations. Phys Rev Lett, 2003, 91:
263902
13 Trompeter H, Krolikowski W, Neshev D N, et al. Bloch oscillations
and Zener tunneling in two-dimensional photonic lattices. Phys Rev
Lett, 2006, 96: 053903
14 Longhi S. Quantum-optical analogies using photonic structures. Laser
Photon Rev, 2009, 3: 243261
15 Grenzer J, Ignatov A A, Schomburg E, et al. Microwave oscillator
based on Bloch oscillations of electrons in a semiconductor superlat-
tice. Ann Phys-Berlin, 1995, 3: 184190
16 Alberti A, Ivanov V V, Tino G M, et al. Engineering the quantum trans-
port of atomic wavefunctions over macroscopic distances. Nat Phys,
2009, 5: 547550
17 Haller E, Hart R, Mark M J, et al. Inducing transport in a dissipation-
free lattice with super Bloch oscillations. Phys Rev Lett, 2010, 104:
200403
18 Zhang J M, Liu W M. Directed coherent transport due to the Bloch
oscillation in two dimensions. Phys Rev A, 2010, 82: 025602
19 Wheeler J A, Zurek W H. Quantum Theory of Measurement. Prince-
ton: Princeton University Press, 1983
20 Schr odinger E. Die gegenwartige situation in der quantenme chanik.
Naturwissenschaften, 1935, 23: 807812, 823828, 844849. Re-
printed in English in ref. [19]
21 Monroe C, Meekhof D M, King B E, et al. A Schr odinger cat super-
position state of an atom. Science, 1996, 272: 11311136
22 Brune M, Hagley E, Dreyer J, et al. Observing the progressive deco-
herence of the meter in a quantum measurement. Phys Rev Lett, 1996,
77: 48874890
23 Friedman J R, Patel V, Chen W, et al. Quantum superposition of distinct
macroscopic states. Nature, 2000, 406: 4346
24 Brennen G K, Caves C M, Jessen P S, et al. Quantum logic gates in
optical lattices. Phys Rev Lett, 1999, 82: 10601063
25 Jaksch D, Briegel H J, Cirac J I, et al. Entanglement of atoms via cold
controlled collisions. Phys Rev Lett, 1999, 82: 19751978
26 Mandel O, Greiner M, Widera A, et al. Coherent transport of neutral
atoms in spin-dependent optical lattice potentials. Phys Rev Lett, 2003,
91: 010407
27 Zak J. Berrys phase for energy bands in solids. Phys Rev Lett, 1989,
62: 27472750
28 Soltan-Panahi P, Struck J, Hauke P, et al. Multi-component quantum
gases in spin-dependent hexagonal lattices. Nat Phys, 2011, 7: 434
440