Multi-normal-mode splitting of a cavity in the presence of atoms - towards the superstrong
coupling regime
OS
O
o
(N
in
:3 .
>
<N
o
o
Xudong Yu^, Dezhi Xiong^, Haixia Chen^, Pengjun Wang^, Min Xiao^'^, Jing Zhang
^ The State Key Laboratory of Quantum Optics and Quantum Optics Devices,
Institute of Opto-Electronics, Shanxi University, Taiyuan 030006, RR. China and
^ Department of Rhysics, University of Arkansas, Fayetteville, Arkansas 72701, USA
Multi-normal-mode splitting peaks are experimentally observed in a system with Doppler-broadened two-
level atoms inside a relatively long optical cavity In this system, the atoms-cavity interaction can reach the
"superstrong coupling" condition with atoms-cavity coupling strength to be near or larger than the cavity
free-spectral range Afsr- In such case, normal-mode splitting can occur in many cavity longitudinal modes to
generate the multi-normal-mode splitting peaks, which can be well explained by the linear dispersion enhance-
ment due to the largely increased atomic density in the cavity Many new interesting phenomena might come
out of this superstrong atoms-cavity coupling regime.
Studies of strong coupling between atoms and cavity field
have been very active in the past 30 years 1 1] due to its impor-
tance in the fundamental understanding of quantum electro-
dynamics and potential applications in quantum computation
and quantum information processing [2]. In the traditional
cavity-quantum electrodynamics (C-QED), high finesse mi-
crocavities are normally used to enhance the single-photon
coupling strength ^ (=
2hVM
where cOr is the resonant fre-
quency of the cavity, /i is the atomic dipole matrix element,
Vm is the cavity mode volume), so the strong-coupling con-
dition of ^ > 7 can be satisfied even with a single atom
in H 0] (where k is the cavity decay rate and 7 is the atomic
decay rate). Two normal-mode splitting peaks (i.e. Rabi side-
bands) appear in the cavity transmission spectrum due to such
strong atom-cavity interaction, with the frequency space be-
tween the two side peaks given by 2^. This atom-cavity cou-
pling strength g can be enhanced by using an assemble of
atoms to be gy/N, where N is the number of atoms in the
cavity mode volume JH Normal-mode splitting in as-
semble of two-level atoms has been demonstrated in atomic
beams ItIIsI], cold atomic cloud and Bose-Einstein
condensate 1I12I1 . Recently, such normal-mode splitting was
even observed in Doppler-broadened two-level atoms in a hot
atomic vapor cell inside a low finesse ring cavity lO .
A different regime of atoms-cavity interaction has been
investigated in a system with cold atoms in an optical lat-
tice formed inside a high-finesse, macroscopic optical cavity
iS 13 EH]. Interesting effects, such as cavity-mediated col-
lective light scattering due to the self-organized atoms in the
intracavity optical lattice and collective atomic motion, were
observed in such system Hsll . Recently, normal-mode split-
ting and collective mechanical effects have also been studied
in such atoms-cavity system (9^. The coherent backscatter-
ing between the two propagating directions of a longitudinal
mode has enhanced the coupling between the atoms in the op-
tical lattice and the cavity fields, even when the fields are de-
tuned far from the atomic resonance Ill5i1 .
In this Letter, we present our experimental demonstration
of multi-normal-mode splitting peaks due to strong coupling
between high-density, inhomogeneously-broadened two-level
atoms and multiple longitudinal cavity modes in a long Fabre-
Perot optical cavity. Here, we have reached a new regime of
strong atoms-cavity interaction, i.e. with g\^ > Afsr,
where Afsr is the free-spectral range (FSR) of the optical
cavity. Meiser and Meystre theoretically studied a system
under such condition with a microscopic number of atoms
lfl6ll and named it as the "superstrong coupling" regime in the
C-QED. Such superstrong coupling condition has not been
reached in previous experiments due to either short optical
cavity (therefore very large Afsr) used to increase g or rel-
atively low atomic density. In the current experimental sys-
tem with inhomogeneously-broadened two-level atoms in a
long optical cavity, the atomic density can be easily increased
to satisfy this superstrong coupling condition. With normal-
mode splittings for multiple longitudinal cavity modes, dif-
ferent cavity modes and their atom-cavity polaritons can in-
teract with each other, which can provide the atom-mediated
coupling between different cavity modes via x(3) nonlinearity
II17I1 . Such coupling between different cavity field modes can
be used for many applications in quantum information pro-
cessing, such as phase gates and multimode entangled states.
With such demonstrated new regime in atoms-cavity interac-
tion, many more interesting applications can be envisioned,
including ones in solid-state systems and photonic crystal cav-
ities.
In the model of the coupled atoms-cavity system, the cavity
is a standing- wave cavity with length Lc, and the intracavity
medium is an assemble of two-level atoms, whose resonance
frequency uOa is near a cavity resonance at frequency uOc- The
two-level atomic medium has a length La < Lc with N atoms
in the cavity volume. The intensity transmission function of a
probe laser (00 l) for this coupled atoms-cavity system can be
easily found and is given by Eq.l in Ref. The frequency
-dependent intensity-absorption coefficient and the refractive
index of the atomic medium are given by
a
4A2 + 72
c 2A7a
I - ao
'a 4A2
(1)
(2)
2
respectively, where = uJaN^l/j^l'^ /eohcja is the Hne-center
absorption coefficient. Notice that Eqs.l and 2 are vaUd only
when 27rao/Xa <C 1 IH}]. Here, the angle sustained by
the cavity mode is small, and the transverse decay rate can
be very closely approximated by the atomic free-space de-
cay rate = jJ^'^ujl/Snheoc^ . Thus, the line-center absorp-
o 2
tion coefficient becomes ao = ^^Nd- A = col —
and Aac = — are the laser- atom and atom-cavity fre-
quency detunings, respectively. The cavity linewidth is given
by = Afsr/F, where F is the cavity finesse. This linear-
dispersion theory was used to explain the experimental ob-
servation in the system with a two-level atomic beam pass-
ing through an one-centimeter long standing- wave cavity JtI],
where only one cavity longitudinal mode (m=0) was consid-
ered.
Here, we can also apply this linear-dispersion theory to the
case with superstrong coupling, in which other cavity longi-
tudinal modes (m = ±1 and m = ±2, et al) have to be in-
cluded. We will only consider the case of Aac = 0. For an
empty cavity, the cavity transmission peaks are Lorenzian in
shape and occur at (I){ujl) = ±m27r, where (j) is the round-trip
phase shift experienced by the intracavity field going through
the cavity and m = 0,1,2,..., with equal mode spaces given
by Afsr, can be easily seen from the cavity transmission
function (Eq. (1) in Ref. [Tfl). With an intracavity (two-
level) atomic medium, the cavity transmission structure is sig-
nificantly modified. First, when the atomic density is high
enough, so g^N 7a, but much smaller than Afsr,
= will have two real solutions due to the dispersion
introduced by the atoms as given in Eq. (2). This indicates that
the center peak (m=0) in the cavity transmission is split into
two side peaks located at ±^a/]V, respectively, which is the
standard normal-mode splitting, as shown in Fig. 1(a). Under
this condition, the other cavity modes (m = ±1 and m = ±2,
et al) are not affected by the atoms and still have equal spaces
between them given by A^^i^. However, as the condition that
gVN is near or larger than Afsr is satisfied, then not only
the center cavity mode (m = 0) has normal-mode splitting,
other cavity modes (such as m = ±1) will interact with the
atoms and have their own normal-mode splitting peaks (i.e.
^(^l) = 27r or = — 27r will also have two real so-
lutions), as shown in Fig. 1(b). The positions of their two
normal-mode splitting peaks for |m| > 1 locate in the two
sides of the atomic resonant frequency and present asymmet-
ric structure as shown in Fig. 1(b). To see this, let's consider
only the m = 1 cavity mode, which can be viewed as a mode
with an effective cavity detuning of Afsr- As it interacts
with atoms under the "superstrong coupling" condition, the
Rabi sidebands become very asymmetric (due to large effec-
tive detuning), with one large peak on the right (near the orig-
inal empty cavity peak position) of the atomic resonance and
a smaller peak on the left of the atomic resonance (all labeled
as "1" in Figs. 1(b)- 1(d)). When the atomic density gets even
higher, more cavity modes (such as m = ±2 and m = ±3, et
al) will participate in the mode- splitting process, which form
the multi-normal-mode splitting structure, as shown in Figs.
1(c) and 1(d).
For our experimental situation, we consider a system with
two-level rubidium atoms (in a vapor cell) inside an optical
standing-wave cavity of 17.7 cm long. Here, 7^ = 27r x
6MHz. The original normal-mode splitting gVN can reach
and even be larger than A^^^i^ by increasing the temperature
of the atomic cell (the corresponding atomic density is in-
creased quickly with temperature). Due to Doppler effect, the
absorption (and therefore also the dispersion) profile is much
broader with a width of Sujd = ^ simplicity
in discussion, we can replace the homogeneous absorption
linewidth 7^ by the Doppler width 5uj d in Eqs. 1 and 2, which
is approximately valid. At the same time, the line-center ab-
sorption coefficient is changed into ao = ^^^Nd. In or-
der to accurately calculate the absorption and dispersion prop-
erties in the Doppler-broadened atomic system, we need to
replace Eqs.l and 2 by Eqs. 5 and 6 of Ref. 1 13]. Notice that
the second-term in the index of refraction of the intracavity
medium (Eq. 2) depends on the atomic density in the cav-
ity (ao cx Nd)- So, as the temperature of the atomic cell
increases, more cavity modes will participate in the mode-
splitting interactions to generate the multi-normal-mode split-
ting peaks.
Parameters used in plotting Fig. 1 are basically the same as
in our experiment. For very low absorption coefficient ao (cor-
responding to low temperature or atomic density), the Doppler
absorption and dispersion width (Sud = 27r x 34:3MHz) is
smaller than the cavity free-spectral range (Afsr = 27r x
SbO MHz). Therefore, except for the two middle normal-
mode peaks, all other peaks are the cavity transmission peaks
without mode-spitting (corresponding to Fig. 1(a)), since they
do not interact strongly with the atoms. The shorter transmis-
sion peak height of the two middle normal-mode peaks is due
to the atomic absorption. As the absorption coefficient ao gets
larger (higher temperature), the dispersion changes and more
FSR cavity modes join the mode- splitting interactions, as
shown in Figs. 1(b)- 1(d). Figures 1(e) and 1(f) (which are the
re-plots of Figs. 1(b) and (d), respectively, with the FSR cavity
mode number m as the horizontal axis) give a more clear in-
sight into the positions and heights of the multi-normal-mode
splitting peaks. Such plots are the typical avoided-crossing
plots commonly used in cavity-QED.
The experiment was done using a standing-wave cavity of
17.7 cm long, as shown in Fig. 2. The cavity composes of two
mirrors with same radius of curvature of 100 mm. The re-
flectivity is 90% at 780 nm for the input coupler Ml, which
is mounted on a PZT to adjust the cavity length. The out-
put coupler M2 has a reflectivity of 99.5% at 780 nm. The
finesse of the cavity including the losses of two faces of the
atomic cell is about F = 20. The length of the vapor cell is
5 cm. The temperature of the vapor cell can be controlled by
a heater. A grating- stabilized diode laser, which first passes a
standard polarization maintaining single-mode fiber, was used
as the cavity input beam with an input power of 4 mW. The
3
0.10
0.00
^ 0.050
1.0 0.0250
0.5 0.0125
0.0 0.0000
c
5 -
2 -
1 A
1
I 3 4
1
3
1
1
1
2
1
i 2
-1
Oi 1 1
til
!aUUU
u
1.0
-10 -5 5 10
4 -8 -
I -
M i i
I
t
b
i
j 1
i i
-2
2
1-1 C
'i
u
y
ii
-10 -5 5 10
-7 -6 -5 -4 -3 -2 -
-.01 234567
-5 -4 -3 -2-1 1 2 3 4 5 6 7
FIG. 1: (Color online). Theoretical calculations of the transmis-
sion spectra of the coupled atoms-cavity system with the Doppler-
broadened two-level atoms with different absorption coefficients.
For comparison, the cavity transmission spectrum for the empty cav-
ity (blue dashed) is plotted in (a), (b), (c) and (d). The vertical scale
is normalized to the light intensity transmitted through the empty
cavity. This is for an optical cavity with Afsr = 27r x 850
MHz. m = 0,±1,±2,... label the multi-normal-mode splitting
peaks coming from the FSR cavity modes. The finesse of the empty
cavity F = 20 and A = 780 nm. (a) aoLa = 12 (correspond-
ing to NoLa = 9.4 X lO^V^^); (b) aoLa = 70 (NoLa =
5.5 X lO^V^^); (c) aoLa = 130 (NoLa = 1.0 x 10^7^^); W
aoLa = 170 (NoLa = 1.3 x 10^'^/m^); (e) and (f) are the re-plots
of (b) and (d), respectively.
laser beam with the spatial mode filter by an optical fiber is
easer to mode-match to the TEMqq mode of the optical cav-
ity. The diode laser has also been divided into several parts
to be used in the saturated absorption spectroscopy (using an-
other rubidium atomic cell) and in the F-P cavity for monitor-
ing the frequency and mode of the laser. The optical cavity
length was kept unchanged and the input laser frequency was
scanned to measure the transmission spectra. The transition
between the 551/2, F = 1 and 5P3/2, F' in ^^Rb was used
for the two-level system.
At a low atomic cell temperature (T = 105^C, correspond-
ing to N]jLa = 2.5 x 10^^ /m'^ [19]), the cavity transmission
spectrum is given in Fig. 3 (a). At both sides of the resonant fre-
quency of atoms, the atoms-cavity normal modes can be ob-
served. Only two normal-mode splitting peaks appear in the
Laser Isolator PBSl A/ 2
Dl
A/2
D2
[[ ■ - II I- -
FP cavity ^1
HV-AMP
Rbcell A/4
PBS2
M2Fq
PZT
Fibre
L2
D3
-III
HV-AMP
^[2 PBS5
FIG. 2: (Color online). Schematic of the experimental setup of the
coupled atoms-cavity system. A/2: half-wave plate; Dl, L)2, D3:
detectors; HV-AMP: high voltage amplifier; PZT: piezoelectric
transducer; PBS: polarized beam splitter; LI, L2: optical lens.
transmission spectrum, and the other outside peaks are sim-
ply the cavity FSR peaks without mode- splitting, as predicted
in Fig. 1(a). Notice that the atomic density predicted theoret-
ically in Fig. 1(a) is about an order of magnitude smaller than
the density predicted by Ref. lEll for rubidium at T = 105°C,
which is mainly due to a large fraction of atoms populated at
55'i/2, F = 2 when without a pump light Il20|] . As the tem-
perature increases (T = llO^C), more FSR cavity modes par-
ticipate in the mode- splitting process and generate the multi-
normal-mode splitting peaks, as shown in Fig. 3(b), which also
agrees with the theoretical curve of Fig. 1(b). In order to la-
bel the two original normal-mode splitting peaks, we must
monitor the peaks carefully by slowly increasing the tempera-
ture. The asymmetry in the two transmission side peaks com-
pared with Fig.l is due to the influence of absorption from
bSi/2, F = 2 to 5P3/2, F' transition in ^^Rb, as shown in
Fig. 3(e). As the temperature is further increased (T = 115^C
and T = 120^C), more FSR cavity modes join in the mode-
splitting interactions as shown in Figs. 3 (c) and (d) (which
are also predicted in Figs. 1(c) and 1(d)) due to the largely
increased dispersion as indicated by Eq.2. Since the atomic
density (and therefore the absorption coefficient ao) depends
exponentially on the temperature, the dispersion change (and
therefore the number of FSR cavity modes generating multi-
normal-mode splitting) has a very sensitive dependence on the
temperature of the atomic cell, as shown in Fig. 3.
In summary, we have studied the cavity transmission spec-
tra in a system with Doppler-broadened two-level atoms in a
standingwave cavity under the "superstrong coupling" condi-
tion of larger or equal to Afsr- In this superstrong
atoms-cavity coupling region, mode- splitting occurs in many
FSR cavity modes due to the interactions with the intracav-
ity Doppler-broadened atoms, and multi-normal-mode split-
4
1.0 0.10
0.5 0.05
0.0 0.00
1.0 0.10
0.5 0.05
^ 0.0 0.00 ^
for Distinguished Young Scholars (Grant No. 10725416),
National Basic Research Program of China (Grant No.
2006CB921 101), NSFC Project for Excellent Research Team
(Grant No. 60821004), and NSFC (Grant No. 60678029).
-10 12 3
-10 12 3
-3-2-10123
Detuiiing(GHz)
FIG. 3: (Color online). Experimentally measured transmission spec-
tra of the coupled atoms-cavity system with the Doppler-broadened
two-level atoms at different temperatures of the atomic cell. The
curves (a), (b), (c) and (d) correspond to (e) with a narrow frequency
detuning range. For comparison, the cavity transmission spectrum
for the empty cavity (blue dashed) is added in (a), (b), (c) and (d).
The vertical scale is normalized to the light intensity transmitted
through the empty cavity, (a) T = 105° C; (b) T = 110°C7; (c)
T = 115°C; (d) T = 120°C; (e) The saturated absorption spec-
troscopy (curve 1) and the corresponding cavity transmission spectra
(curve 2) of the coupled atoms-cavity system. T — 120° C.
ting peaks were observed experimentally. Such multi-normal-
mode peaks depend sensitively on the atomic density. This
phenomenon can be qualitatively explained by using the linear
absorption and dispersion theory of the cavity transmission,
and by taking into account the sensitive dependence of the
index of refraction of the intracavity medium on the atomic
density. This work sheds new light on the investigations of
this novel "superstrong coupling" region in the exciting field
of collective coupling between atoms and cavity, and can lead
to interesting applications in quantum information processing.
^Corresponding author's email address:
jzhang74 @ sxu.edu.cn, jzhang74 @y ahoo.com
J. Zhang thanks K. Peng, C. Xie and T. Zhang for the help-
ful discussions. This research was supported in part by NSFC
[1] P. R. Herman, Cavity Quantum Electrodynamics (Advances in
Atomic, Molecular, and Optical Physics, Academic, New York,
1994).
[2] T. Pellizzari, S. A. Gardiner, J. 1. Cirac, P. Zoller, Phys. Rev.
Lett. 75, 3788 (1995); L. M. Duan, H. J. Kimble, Phys. Rev.
Lett. 92, 127902 (2004); J. 1. Cirac, P Zoller, H. J. Kimble, H.
Mabuchi, Phys. Rev. Lett. 78, 3221 (1997).
[3] A. Boca, R. Miller, K. M. Birnbaum, A. D. Boozer, J. McK-
eever, and H. J. Kimble, Phys. Rev. Lett. 93, 233603 (2004).
[4] P. Maunz, T. Puppe, L Schuster, N. Syassen, P W. H. Pinkse,
and G. Rempe, Phys. Rev. Lett. 94, 033002 (2005); T. Puppe,
1. Schuster, A. Grothe, A. Kubanek, K. Murr, P. W. H. Pinkse,
and G. Rempe, Phys. Rev. Lett. 99, 013002 (2007).
[5] M. Tavis, F W. Cummings, Phys. Rev. 170, 379 (1968).
[6] G. S. Agarwal, Phys. Rev. Lett. 53, 1732 (1984).
[7] Y. Zhu, D. J. Gauthier, S. E. Morin, Q. Wu, H. J. Carmichael,
and T. W. Mossberg, Phys. Rev. Lett. 64, 2499 (1990).
[8] R. J. Thompson, G. Rempe, and H. J. Kimble, Phys. Rev. Lett.
68, 1132 (1992).
[9] J. Klinner, M. Lindholdt, B. Nagorny, A. Hemmerich, Phys.
Rev. Lett. 96, 023002 (2006).
[10] A. K. Tuchman, R. Long, G. Vrijsen, J. Boudet, J. Lee, and M.
A. Kasevich, Phys. Rev. A 74, 053821 (2006).
[11] G. Hernandez, J. Zhang, and Y Zhu, Phys. Rev. A 76, 053814
(2007).
[12] S. Gupta, K. L. Moore, K. W. Murch, and D. M. Stamper-Kurn,
Phys. Rev. Lett. 99, 213601 (2007); Y Colombe, T. Steinmetz,
G. Dubois, F Linke, D. Hunger, J. Reichel, Nature 450, 272
(2007); F Brennecke, T. Donner, S. Ritter, T. Bourdel, M. Kohl,
T Esslinger, Nature 450, 268 (2007).
[13] J. Gea-Banacloche, H. Wu, and M. Xiao, Phys. Rev. A 78,
023828 (2008).
[14] P Domokos and H. Ritsch, Phys. Rev. Lett. 89, 253003 (2002);
J. K. Asboth, P. Domokos, H. Ritsch, and A. Vukics, Phys. Rev.
A 72, 053417 (2005).
[15] B. Nagorny, Th. Elsasser, A. Hemmerich, Phys. Rev. Lett. 91,
153003 (2003); D. Kruse, C. von Cube, C. Zimmermann, and
P W. Courteille, Phys. Rev. Lett. 91, 183601 (2003); A. T
Black, H. W. Chan, and V. Vuletic, Phys. Rev. Lett. 91, 203001
(2003); S. Slama, S. Bux, G. Krenz, C. Zimmermann, and Ph.
W. Courteille, Phys. Rev. Lett. 98, 053603 (2007).
[16] D. Meiser and P Meystre, Phys. Rev. A 74, 065801 (2006).
[17] L. M. Duan, J. L Cirac, P Zoller, E. S. Polzik, Phys. Rev. Lett.
85, 5643 (2000); L. M. Duan, G. Giedke, J. L Cirac, P Zoller,
Phys. Rev. Lett. 84, 4002 (2000); B. Julsgaard, A. Kozhekin, E.
S. Polzik, Nature 413, 400 (2001).
[18] R. W. Boyd, Nonlinear Optics, (Academic, San Diego, CA,
2003).
[19] http://steck.us/alkahdata
[20] H. Wu, J. Gea-Banacloche, and M. Xiao, Phys. Rev. Lett. 100,
173602 (2008).
