Ultimate fast optical switching of a planar microcavity in the telecom
Georgios Ctistis,1, a)Emre Yuce,1,2Alex Hartsuiker,1,2Julien Claudon,3Maela Bazin,3Jean-Michel G´ erard,3and
Willem L. Vos1, b)
1)Complex Photonic Systems (COPS), MESA+ Institute for Nanotechnology, University of Twente,
7500 AE Enschede, The Netherlands
2)Center for Nanophotonics, FOM Institute for Atomic and Molecular Physics (AMOLF), 1098 XG Amsterdam,
3)CEA-CNRS-UJF ”Nanophysics and Semiconductors” joint laboratory, CEA/INAC/SP2M,
38054 Grenoble Cedex 9 France
We have studied a GaAs–AlAs planar microcavity with a resonance near 1300 nm in the telecom range by
ultrafast pump-probe reflectivity. By the judicious choice of pump frequency, we observe a ultimate fast
and reversible decrease of the resonance frequency by more than half a linewidth due to the instantaneous
electronic Kerr effect. The switch-on and switch-off of the cavity is only limited by the cavity storage time
of τcav= 0.3ps and not by intrinsic material parameters. Our results pave the way to supra-THz switching
rates for on-chip data modulation and real-time cavity quantum electrodynamics.
Switches are widely applied and necessary ingredients
in modulation and computing schemes1.
progress on photonic integrated circuits2,3promises to
overtake boundaries set by conventional switching tech-
nology. To do so, ultrafast switching of photonic cavities
is crucial as it allows the capture or release on demand
of photons4–6, which is relevant to on-chip communica-
tion with light as information carrier7, and to high-speed
miniature lasers8. Ultrafast switching would also permit
the quantum electrodynamical manipulation of coupled
cavity-emitter systems9in real-time. Switching the opti-
cal properties of photonic nanostructures is achieved by
changing the refractive index of the constituent mate-
rials. To date, however, the switching speed has been
limited by material properties11–14, but not by optical
considerations. To achieve ultimate fast switching of a
cavity two challenges arise. Firstly, both the switch-on
and switch-off times τonand τoffmust be shorter than all
other relevant time scales for the system, i.e., the cavity
storage time in photon capture/release experiments τcav,
or the vacuum Rabi period τRabifor a strongly coupled
emitter-cavity system10. Secondly, the refractive index
change must be large enough to switch the cavity reso-
nance by at least half a linewidth.
Here, we demonstrate the ultimate fast switching of
the resonance of a planar cavity in the well-known
GaAs/AlAs system in the telecom wavelength range. We
exploit the instantaneously fast electronic Kerr effect by
the judicious tuning of the pump and probe frequencies
relative to the semiconductor bandgap. We observe that
the speed of the switching is then only limited by the
dynamics of the light in our cavity (τcav= 0.3 ps), but
not by the intrinsic material parameters.
Instantaneous on- and off-switching with vanishing τon
and τoff is feasible with the well-known nonlinear re-
a)Electronic mail: email@example.com
b)Electronic mail: W.L.Vos@tnw.utwente.nl
fractive index from nonlinear optics15.
electronic Kerr effect is the fastest Kerr phenomenon on
account of the small electron mass. In many practical
situations, however, non-degenerate two-photon absorp-
tion overwhelms any instantaneous effect and therefore
also the dispersive electronic Kerr effect13,15. In order
to avoid two-photon absorption and to access the elec-
tronic Kerr switching regime, we designed our exper-
iment to operate with low energy pump photons, see
Fig. 1 (a). First of all, the energy of the pump pho-
tons is chosen below half the semiconductor band gap
pump and probe photons are chosen so that their sum
does not exceed the bandgap energy of the semiconduc-
tor (Eprobe+ Epump≤ Egap)16. These conditions serve
to suppress instantaneous two-photon generation of free
carriers and to perform electronic Kerr switching at el-
evated frequencies (Eprobe>1
com band, with a broad range of semiconductors.
We have performed our experiments on a planar micro-
cavity grown by means of molecular-beam epitaxy. The
sample consists of a λ-thick GaAs layer (d = 376 nm)
sandwiched between two Bragg stacks consisting of 7
and 19 pairs of λ/4-thick layers of nominally pure GaAs
(dGaAs= 94 nm) or AlAs (dAlAs= 110 nm). The cav-
ity was designed such that the resonance occurs at λ0=
1280 nm in the Original (O) telecom band. Measuring
the linewidth of the cavity resonance we obtained a qual-
ity factor Q = 320±30 corresponding to a cavity storage
time of τcav= 0.3 ps.
To perform Kerr-switching on a semiconductor micro-
cavity, we have built a set-up illustrated in Fig 1 (b), con-
sisting of two independently tunable optical parametric
amplifiers (OPA,Topas) that are the sources of the pump
and probe beams. The pump beam can be tuned down
to 4200 cm−1(2400 nm) and the probe beam is tuned to
cavity resonance at 7810 cm−1(1280 nm). Under these
conditions, we solely pump GaAs (the refractive index
change of AlAs is much smaller and can be neglected
here). The pump beam has a larger Gaussian focus (75
2Egap). Secondly, the energies of the
2Egap), including the tele-
arXiv:1102.3351v2 [physics.optics] 30 Mar 2011
frequency of the probe beam is resonant with the cavity. The
pump frequency is tuned such that the sum of pump and
probe energies are less than the energy gap of GaAs to avoid
two-photon absorption. (b) Schematic of the set-up.
probe-beam path is shown in blue, the pump-beam path in
(Color online) (a) Schematic energy diagram. The
µm full width at half maximum) than the probe beam
(28 µm), to ensure that only the central flat part of the
pump focus is probed and that the probed region is ho-
mogeneously pumped. The OPAs have pulse durations
τP= 0.12 ± 0.01 ps. The delay ∆t between pump and
probe pulse is set by a delay stage with a resolution of
0.01 ps. A versatile measurement scheme was developed
to compensate for possible pulse-to-pulse variations in
the output of our laser17.
Figure 2 shows transient reflectivity versus frequency
spectra for three different pump-probe delay times. One
can see that the cavity resonance red-shifts when ap-
proaching pump-probe overlap. Afterwards, at positive
delay times, the cavity resonance blue-shifts back to its
original frequency. By modeling the transient reflectivity
trough with a Lorentzian, both the resonance frequency
ω0 and the width of the resonance Γ are obtained for
every time delay ∆t. The unswitched cavity resonance
frequency (ω0= 7810 cm−1) and width (Γ = 12 cm−1)
are obtained at a delay of ∆t = −5 ps, confirming a
cavity quality factor of Q = 320.
Figure 3 (a) shows the switching of the cavity res-
onance due to the electronic Kerr effect.
nance frequency as a function of delay near pump and
probe coincidence is obtained from spectra similar to
those shown in Fig 2. The dashed horizontal line de-
notes the unswitched resonance frequency ω0. The mea-
sured resonance frequency shifts to a lower frequency
starting at -0.8 ps and reaching its maximum shift at
-0.2 ps.Subsequently, the resonance frequency shifts
back to its original frequency ω0.
quency shift ∆ω = 7 cm−1clearly exceeds one half of
the cold cavity linewidth. We conclude from the shift to
a lower frequency that the refractive index is increased by
0.1 %, which corresponds to a positive Kerr coefficient for
GaAs16. To confirm our interpretation we performed cal-
culations with a dynamic Fabry-P´ erot model18,19, taking
solely into account the refractive index change of GaAs
The dynamic fre-
Transient Reflectivity (%)
FIG. 2. (Color online) Transient reflectivity versus frequency
spectra for three different delay times, ∆t = −2.9 ps (trian-
gles), ∆t = −0.2 ps (circles), and ∆t = +2.9 ps (squares).
We observe an ultrafast shift of the cavity resonance to lower
frequencies and back to its original frequency.
due to the electronic Kerr effect. We observe an excellent
quantitative agreement between the measured and calcu-
lated shifts for the amplitude and the temporal evolution.
Our results furthermore demonstrate that our method
serves to truly inhibit excitation of free carriers since
there is no blue-shift of the resonance at ∆t > 0 ps. For
comparison, the inset shows the result when the pump
frequency is increased to 5000 cm−1in the free carrier
regime. There is no instantaneous shift of the resonance
in the given range as expected; the free-carrier excitation
leads to a blue-shift after a time of 3 ps13. The very good
agreement between the calculation and our experimental
results firmly confirm the ultimate fast electronic Kerr
switching of our photonic microcavity.
Figure 3(b) shows the transient reflectivity of the
unswitched resonance frequency as a function of pump-
probe delay. The relatively high reflectivity of the trough
(Rtrough= 90 %) is a result of the asymmetric cavity de-
sign. One can see that the transient reflectivity quickly
changes from high to low (at pump-probe overlap) and
back to high, within 1.5 ps. This decrease is a result
of the change of the refractive index of GaAs not only in
the cavity but also in the mirrors, which leads to a higher
contrast in the Bragg stack.
The shape of the resonance shift in Fig. 3 (a) is strik-
ingly asymmetric. The asymmetry is a direct result of
3 Download full-text
the light dynamics in the cavity. In our experiment we
are probing the change in the dielectric function ∆ε of
the material induced by P(3)15. The magnitude of the
observed frequency shift ∆ω is given by the overlap inte-
gral of the pump and probe, the latter has been stored in
the cavity and subsequently escaped, at a certain delay
Transient Reflectivity (%)
Pump-Probe Delay Δt (ps)
FIG. 3. (Color online) (upper panel) Measured (open sym-
bols) and modeled (line) resonance frequency vs. pump-probe
delay. The pump frequency is at 4166 cm−1. The resonance
frequency reaches its maximum shift at -0.2 ps and quickly
returns to its unswitched value. The model matches the ex-
periment well. The inset shows no switch when the pump
frequency is at 5000 cm−1for comparison.
Transient reflectivity vs. pump-probe delay at the cold cavity
resonance frequency. The transient reflectivity shows similar
reversible behavior as the resonance.
Hence, we tomographically sample the probe with the
shorter pump. Thus, while scanning ∆t, we obtain a shift
that follows the cavity envelope. The shift is maximal at
−0.2 ps when the cavity field is also maximum.
We have demonstrated for the first time the switching
of a semiconductor microcavity at telecom wavelengths
using the electronic Kerr effect. Our system also serves
as a model system, since the nature of the switch process
can be employed in any realization of a semiconductor
cavity. We observed that the switching speed is limited
by the cavity storage time and not by material proper-
ties. The 0.3 ps storage time in our work paves the way to
sub-ps real-time data modulation. Since switching using
the electronic Kerr effect is repeatable, on-chip supra-
THz switching rates are feasible. Our results also open
an avenue towards ultrafast control of all solid-state cav-
ity quantum electrodynamical systems that exploit the
strong coupling regime of quantum wells or quantum dots
in optical microcavities10.
We thank Allard Mosk, Harm Dorren, Huib Bakker,
and Pepijn Pinkse for stimulating discussions. This re-
search was supported by Smartmix Memphis, NWO-Vici
(to WLV), and the QSWITCH ANR project (to JMG).
This work is also part of the research program of FOM,
which is financially supported by NWO.
1S. M. Sze,
(John Wiley&Sons, New York, 1985).
2S. Noda and T. Baba, Roadmap on Photonic Crystals, (Kluwer,
3S. Assefa, F. Xia, and Y.A. Vlasov, Nature (London) 464, 80
4P. M. Johnson, A. F. Koenderink, and W. L. Vos, Phys. Rev. B
66, 081102 (2002).
5V. R. Almeida, C. A. Barrios, R. R. Panepucci, and M. Lipson,
Nature (London) 431, 1081 (2004).
6T. Tanabe, M. Notomi, E. Kuramochi, A. Shinya, and H.
Taniyama, Nature Photon. 1, 49 (2007)
7M. F. Yanik and S. Fan, Phys. Rev. Lett. 92, 083901 (2004).
8S. G. Hense and M. Wegener, Phys. Rev. B 55, 9255 (1997).
9J.-M. G´ erard, Top. Appl. Phys. 90, 269 (2003).
10G. Khitrova, H. M. Gibbs, M. Kira, S. W. Koch and A. Scherer,
Nature Phys. 2, 81 (2006).
11D. A. Mazurenko, R. Kerst, J. I. Dijkhuis, D. A. Kurdyukov, and
A. B. Pevtsov, Appl. Phys. Lett. 86, 041114 (2005).
12S. W. Leonard, H. M. van Driel, J. Schilling, and R. B.
Wehrspohn, Phys. Rev. B 66, 161102 (2002).
13P. J. Harding, T. G. Euser, Y. R. Nowicki-Bringuier, J.-M.
G´ erard, and W. L. Vos, Appl. Phys. Lett. 91, 111103 (2007).
14X. Hu, P. Jiang, C. Ding, H. Yang, and Q. Gong, Nature Photon.
2, 185 (2008).
15R. W. Boyd, Nonlinear Optics, (Academic Press, San Diego,
16A. Hartsuiker, P. J. Harding, Y. Nowicki-Bringuier, J. M. G´ erard,
and W. L. Vos, J. Appl. Phys. 104, 083105 (2008); P. J. Harding,
T. G. Euser, and W. L. Vos, J. Opt. Soc. Am. B 26, 610 (2009).
17T. G. Euser, P. J. Harding, and W. L. Vos, Rev. Sci. Instrum.
80, 073104 (2009).
18S. Marzenell, R. Beigang, and R. Wallenstein, Appl. Phys. B 71,
19M. Born and E. Wolf, Principles of Optics, Cambridge University
Press, 7th edition, Cambridge, 2002.
Semiconductor Devices: Physics and Technology,