Bose 2.2 User Manual Page 63

  • Download
  • Add to my manuals
  • Print
  • Page
    / 155
  • Table of contents
  • BOOKMARKS
  • Rated. / 5. Based on customer reviews
Page view 62
with the parameter
A
def
= E/E
R
U
lat
/
(2
E
R
) and the lattice parameter
Q
def
= U
lat
/
(4
E
R
) =
s/
4. The solutions of Eq.
(4.5)
are the so-called Mathieu functions
M
(
A, Q, ˜z
), which can
be computed numerically for known parameters
A
and
Q
. Recalling the periodicity of
the lattice potential and according to Bloch’s theorem, any eigenfunction of the system
may be written in the form [170]
ψ
(
z
) =
e
iqz
u
(
z
). Here, the function
u
(
z
) is periodic in
z
with a period
d
lat
, and
q
is the quasi-momentum of the wavepacket. For the moment, we
consider negligible center of mass motion (
q
= 0) such that any eigenfunction
ψ
(
z
) of the
system exhibits the characteristic lattice periodicity.
As a consequence of Floquet’s theorem, any Mathieu function can be written in the form
M
(
A, Q, ˜z
) =
e
ir˜z
f
(
A, Q, ˜z
), with
f
(
A, Q, ˜z
) being a 2
π
-periodic function in
˜z
. While in
general complex and non-periodic, the Mathieu functions become real and periodic
62
for
integer values of
r
, as requested by Bloch’s theorem. Then, for a given
Q
, the parameter
A
=
A
r
can only take a finite number of values and is called the characteristic parameter,
with
r
being the characteristic Mathieu exponent. Note that, at given
Q
and
r
, there
exists only one value
A
r
solving the Mathieu equation: for even values of
r
the Mathieu
functions are symmetric, while they are anti-symmetric if r is odd.
4.2.2 Calibration of the Lattice Depth by BEC Diffraction
Since the lattice spacing
d
lat
is given via the laser wavelength and the geometry of the
experimental setup (see section 4.1.1), the lattice depth
U
lat
is the only parameter that
we need to determine from the experiment to have a full characterization of our system
63
.
Among the existing methods [170], we choose the BEC diffraction method for a daily
calibration of the lattice depth as it provides a precise result, is applicable for deep lattices,
and consumes only little measurement time.
The underlying principle of this calibration scheme is closely related to the diffraction
of laser light from a phase grating (see e.g. [175, Ch.12.6]). The difference is mainly the
inverted role of matter and light: We first prepare a coherent matter wave (BEC) in
its trap. Then, we suddenly turn on the periodic lattice potential and after a variable
evolution time
t
pulse
we switch off all optical trapping potentials to observe the diffracted
BEC in the “far field”, i.e. after a sufficiently long time-of-flight
64
.
At the time
t
= 0, i.e. directly before switching on the lattice potential, we can assume
the wave function of the BEC
ψ
(
z
) to be constant over the extent of two lattice sites.
When we suddenly turn on the lattice, the system does not adiabatically evolve from the
62
The periodic Mathieu functions are normalized via the relation
π
R
π
M(A
r
, Q, ˜z) M
(A
r
, Q, ˜z)d˜z = π.
63
In our system, we do not control the absolute position of the lattice sites, i.e. there is no active phase
stabilization of the lattice laser. This does not play a role in most of our experiments, as the lattice
spacing is typically much smaller than the overall size of the BEC. We will comment on the influence of
the random lattice phase shift when it is relevant for the measurements.
64
Not only the role of matter and light are exchanged, but also the role of space and time: while in optics
the components are located at different positions in space, we apply the elements (lattice potential,
imaging) at different instants in time.
63
Page view 62
1 2 ... 58 59 60 61 62 63 64 65 66 67 68 ... 154 155

Comments to this Manuals

No comments