Photon echoes from (In,Ga)As quantum dots embedded in a Tamm-plasmon microcavity

0.54 MB·

by  M. Salewski

APS/123-QED
Photon echoes from (In,Ga)As quantum dots embedded in a Tamm-plasmon
microcavity
M. Salewski,1,∗ S. V. Poltavtsev,1,2 Yu. V. Kapitonov,3 J. Vondran,1 D. R. Yakovlev,1,4 C. Schneider,5
M. Kamp,5 S. Ho¨fling,5 R. Oulton,6 I. A. Akimov,1,4 A. V. Kavokin,2,7,8 and M. Bayer1,4
1Experimentelle Physik 2, Technische Universita¨t Dortmund, 44221 Dortmund, Germany
2Spin Optics Laboratory, St. Petersburg State University, St. Petersburg 198504, Russia
3St. Petersburg State University, St. Petersburg 199034, Russia
7 4Ioffe Physical-Technical Institute, Russian Academy of Sciences, 194021 St. Petersburg, Russia
1 5Technische Physik, Universita¨t Wu¨rzburg, 97074 Wu¨rzburg, Germany
0 6Quantum Engineering Technology Labs, H. H. Wills Physics Laboratory and
2
Department of Electrical & Electronic Engineering, University of Bristol, BS8 1FD, UK
n 7School of Physics and Astronomy, University of Southampton, SO17 1 BJ, Southampton, United Kingdom
a 8CNR-SPIN, Viale del Politecnico 1, I-00133 Rome, Italy
J (Dated: January 13, 2017)

2
We report on thecoherent optical response from an ensemble of (In,Ga)As quantumdots (QDs)
1
embeddedin aplanarTamm-plasmon microcavity with aqualityfactor ofapprox. 100. Significant

enhancementofthelight-matterinteractionisdemonstratedunderselectivelaserexcitationofthose
]
l quantumdotswhichareinresonancewiththecavitymode. Theenhancementismanifestedthrough
l
a Rabi oscillations of the photon echo, demonstrating coherent control of excitons with picosecond
h pulsesat intensitylevelsmorethanan orderof magnitudesmaller ascompared withbarequantum
- dots. Thedecayofthephotonechotransientsisweaklychangedbytheresonatorindicatingasmall
s
e decrease of the coherence time T2 which we attribute to the interaction with the electron plasma
m in the metal layer located close (40 nm) to the QD layer. Simultaneously we see a reduction of
thepopulationlifetimeT1,inferredfromthestimulatedphotonecho,duetoanenhancementofthe
.
t spontaneousemission bya factor of 2, which is attributed tothePurcell effect, while non-radiative
a
processes are negligible as confirmed from time-resolved photoluminescence.
m
- PACSnumbers: 42.50.Ex,42.50.Md,42.25.Kb,42.70.Nq
d
n
o
Light-matter interaction in photonic nanostructures used. So far, efforts have mainly been focused on time-
c
attracts strong attention in all areas of optics. Efficient integrated and time-resolved studies of the emission un-
[
couplingatthenanoscaleplaysadecisiveroleforrealiza- der non-resonant excitation. Thereby the Purcell effect

1 tionofsinglephotonemittersandothernonclassicallight of single QD excitons coupled to a localized TP mode,12
v sources of importance in quantum information technolo- enhancementofthespontaneousemission13andcoherent
5
gies.1,2 Various structural concepts based on photonic laser emission9,14 were demonstrated.
5
crystals,patternedmicrocavitiesorplasmonicstructures
3 Coherent spectroscopy such as transient four-wave
3 havebeenintenselystudiedinthatrespect.3 Anotherin-
mixing (TFWM) is a powerful tool for investigating
0 teresting system is a Tamm-plasmon (TP) resonator in
the nonlinear optical phenomena and coherent dynam-
. whichconfinementoftheopticalfieldisobtainedbetween
1 ics of optical excitations confined in semiconductor
0 a distributed Bragg reflector (DBR) and a thin metal nanostructures.16 It allows one to study ensembles of
7 layer,leadingtotheappearanceofaTPphotonicmode.4
light emitters and to perform direct measurements of
1 In addition, TP structures support surface plasmon po-
their times for dephasing, decoherence, and popula-
: laritons(SPPs),evanescentelectromagneticwavesatthe
v tion decay. Moreover, in the strong field regime Rabi
Xi metal-semiconductorinterfacewhichcanpropagatealong oscillations can be used for direct evaluation of the
this interface. Especially interesting in such a system
light-matter interaction strength. TFWM and two-
r is the coupling between SPPs and TP cavity modes.5–7
a dimensional Fourier spectroscopy were used to investi-
Therefore these structures are appealing for generation
gate the coherent optical response of exciton-polaritons
of SPPs via optical or electrical pumping of a close ac-
inQWandQDbasedmicrocavitieswithhighqualityfac-
tive layer containing, e.g., semiconductor quantum dots
tors(strongcouplingregime).2,17,18QDswerealsoimple-
(QDs). Ingeneral,theintegrationofsemiconductorsinto
mentedinlowqualityDBR-basedcavitiesinordertoin-
plasmoniccircuitsisappealingforcompensatinglossesor
creasethestrengthofTFWMsignalandtostudythecor-
switching in these circuits. Furthermore, the metal mir-
responding coherent optical phenomena.19–21 However,
rormaybeusedasanelectrodetoapplyabiasvoltagefor
to the best of our knowledge there are no studies of the
controllingthechargingstateoftheQDsorpumpingthe
nonlinear optical phenomena under resonant excitation
optically active layer electrically.10 As active material in
in TP resonators.
the resonator single or multiple quantum well (QW)8–11
or QD12,13 layers as well as organic materials14,15 were The unique feature of TP structures is the possibility
to use an arbitrary in-plane design of a metallic layer
2
(a) k2 =A ku3 40 nkm1 kFWM = 2kx2 - kz1 yInGaAs QDs(c) ensity (arb. units)112205050000 b TaPre QQDDss | (arb. units)11M0500 1 12 2 P1E223 3SP1E2 (e)
DBR PL Int 50 |EFW50
0 0
GaAs substrate 1.30 1.32 1.34 1.36 1.38 1.40 -10 0 10 20 30 40 50 60
(b) Re(n) 24 (d) units)1.0 Photon energy (eV) s) 400 12 = 13ps Time t (ps) (f)
2 |E| Re(n)(a.u.) 002424 PQoDs iltaioyne rof ectivity (arb. 000...468 e cxaplceurilmateionnt | (arb. unitWM230000 27ps40ps53pTs =6 72pKs 80ps
2 |E| (a.u.)100 Refl0.2 |EF100
0 0.0 0
-0.5 0.0 0.5 1.0 1.5 1.30 1.32 1.34 1.36 1.38 1.40 0 20 40 60 80 100120140160180
Structure depth ( m) Photon energy (eV) Time t (ps)
FIG. 1. (Color online) (a) Schematic presentation of the investigated TP structure and the TFWM geometry. (b) Refractive
index and electric field distribution under normal incidence of light (along the z-axis) for the uncovered part (top) and the
metal-coveredpart(bottom)ofthesample. IntheTPmicrocavitytheintensityoflightattheQDlayerpositionisincreasedby
afactorof34comparedtothebarewaver. (c)PhotoluminescencespectraoftheinhomogeneouslybroadenedQDensemblewith
and without TP microcavity. Photon excitation energy ~ωexc = 2.33 eV, temperature T = 2 K. (d) Measured and calculated
reflectivity spectra of the TP microcavity. Quality factor Q = 130, temperature T = 8 K. The calculation used a transfer
matrix method with dw = 118 nm.25 (e) and (f) Photon echoes from the TP cavity with pulse energies P1 = 0.08 nJ and
P2 ≈P3≈0.3 nJ, temperatureT =2 K.(e) TFWM amplitudes measured at fixed delay times of τ12 =13psand τ23 =20ps.
The photon echo (PE) and stimulated photon echo (SPE) signals appear at t=2τ12 and t=2τ12+τ23, respectively. (f) PEs
for different time delaysτ12 indicated at thearrows.
on top of the semiconductor. Covering only a part of axis x (see Fig. 1(a)). Half of the sample is coveredwith
the sample with a gold film allows comparison of the a40nmthick goldlayer,whichleadsto formationofthe
strength of light-matter interaction within the same QD TP photon mode.
ensemble so that size distribution and QD density re- TheQDlayerislocatedatthemaximumoftheelectric
main the same. Here, we show that the coherent optical field distribution, 40nm away from the gold layer (see
responseofplanarQD-TPstructuresto picosecondlaser Fig. 1(b)). It is close enough to the gold to ensure that
pulses occurs in form of photon echoes (PEs): (i) The the coupling to the TP mode is strong, but far enough
magnitude of the PE signal and its dependence on exci- so that quenching of the photoluminescence (PL) due
tation intensity differ drastically from bare QDs. From to surface states, in particular at the uncovered surface,
Rabioscillationsweestimate the enhancementfactorfor does not occur. The QD density is about 2 109cm−2,
the driving optical field in the TP structure. (ii) From theirheightandlateralsizesbeforeovergrowt×hareabout
two-pulse and three-pulse PE transients we evaluate the 2.3 nm and 25 nm, respectively.22 In order to prevent
decoherence and population decay times of excitons in tunneling of photoexcitedcarriersinto the metal layer,a
the TP microcavity and compare them with bare QDs. 10 nm Al0.2Ga0.8As barrier was introduced between the
The investigated structures and experimental ap- QDs andthe surface (20nm below the surface). We also
proach are summarized in Fig. 1. The sample was studied structures with a single layer of (In,Ga)As QDs
grownbymolecular-beamepitaxyonaGaAs(001)semi- which is embedded in a GaAs λ microcavity of similar
insulating substrate. It comprises 20 pairs of λ/4n quality factor formed by two DBR mirrors.20
GaAs/AlAs layers forming the back DBR mirror with The low temperature (T = 2 K) photoluminescence
the stop-band located in the spectral region of interest spectrum from the bare QDs under non-resonantexcita-
( 1.2 1.4 eV). A single (In,Ga)As QD layer in the tion with photon energy ~ω = 2.33 eV shows a broad
exc
∼ −
GaAs cavity layer is located about 125 nm above the spectralbandcenteredatabout1.35eV(Fig.1(c)). The
DBR and 40 nm below the surface. The thickness of the large inhomogeneous broadening of 100 meV origi-
∼
GaAs layer dw beneath the QDs is slightly varied by a nates from fluctuations of QD size and composition in
wedge design which allows tuning the cavity resonance the ensemble. The emission from the same part of the
by the position of the excitation spot along the gradient sample,butcoveredwithgoldshowsaresonantenhance-
3
ment of the PL signal around the photon energy of the
) 0.08
cavity mode. The PL has a significantly narrower band- s
width with a full width at half maximum (FWHM) of unit 0.06 (a) bare QDs
only 12 meV. The maximum of the PL peak and its b. 0.04
r
width correspond to the dip in the cavity reflectivity a
spectrum (Fig. 1(d)). The latter is attributed to a TP (PE 0.02
photonic mode with energy ~ω = 1.36 eV and width P 0.00
TP 0 0.2 0.4 0.6
∆ = 10.5 meV, corresponding to a quality factor of
Q=130. Calculations using the transfer matrix method Pulse energy P2 (nJ)
are in good agreement with the measured reflectivity
0 0.1 0.2 0.3 0.4
spectrum for dw = 118 nm. The expected enhancement
1.0
factor of the light intensity at the position of the QD
(b) TP QDs
layer is 34 (see Fig. 1(b)).25 )
s 0.8
Thecoherentopticalresponseismeasuredusingdegen- nit
u
erate TFWM with a sequence of three spectrally narrow b. 0.6
ps-pulses in non-collinear reflection geometry as shown ar
(
in Fig. 1(a). A mode-locked tunable Ti:Sa laser with a E 0.4
P
repetition frequency of 75.75 MHz was used as source P
of the optical pulses. Pulse 1 with wavevector k1 hits 0.2
◦
the sample at an incidence angle of 6 while the follow-
ing pulses 2 and 3 hit the same spot at 7◦ (k2 = k3). 0.0
The beams are focused into a spot with a diameter of 0 2 4 6 8
approx. 200 µm. The sample is kept at T = 2 K. The
Pulse area 2 d (rad)
transients are measured by taking the cross-correlation
of the TFWM signal E (t) with the reference pulse FIG.2. (Color online) Dependenceof thePE amplitudePPE
using heterodyne detectFioWnM.23 The spectral width of the on the energy of the second pulse P2 for the bare QDs (a)
and the TP microcavity (b). P1 = 0.026 nJ, τ12 = 67 ps,
opticalpulsesof0.8meVissignificantlysmallerthanthe
T =2K.Thescalingofthex-axes(bottomin(a)andtopin
width of the resonator mode, i.e. the lifetime of the TP
(b))ischosensuchthatitwouldscalelinearlywiththesquare
mode in the microcavity is shorter than the pulse du- root ofP2. Thebottomx-axisin (b)corresponds tothearea
ration of τd = 2 ps. At zero magnetic field resonantly of pulse 2. The dashed curve is a fit using Eq. (1). The
excited QDs can be considered as isolated two-level sys- solidcurvefollowsfromEq.(1)whenincludingthestatistical
tems. Notethatexcitationoffew-particlecomplexessuch distribution of dipole moments of the two-level systems with
as biexcitons is excluded because the spectral width of a standard deviation of 30%.
the laser line is below the biexciton binding energy of
about2meV. Fromthese parameterswe expectthatthe
crocavity, respectively, for identical experimental condi-
TFWM signal is determined by photon echoes due to
tions. Stronginhomogeneousbroadeningofopticaltran-
the inhomogeneous broadening of the optical transitions
sitions can lead to a significant modification of the pho-
in the QD ensemble. In the case of the TP microcav-
ton echo transients when the energy of the first pulse is
ity the broadening is given by the spectral width of the
photonic mode.20 scanned.20 First, inthe TPmicrocavityoneachievessig-
nificantly stronger PE amplitudes for low pulse energies
An example of the TFWM signalmeasuredunder res-
0.1nJ.Second,inthe TPmicrocavityoneobserves
onant excitation in the TP mode is shown in Fig. 1(e). P2 ≤
a non-monotonous behavior which resembles the one ex-
Here, the delay time between pulses 1 and 2 is set to
pected for Rabi oscillations. Both features indicate a
τ = 13 ps, while the delay time between pulses 2 and
12 considerableenhancementofthelight-matterinteraction
3 is τ = 20 ps. In full accordance with our expecta-
23 intheTPcavitywhichallowsustoworkinthenonlinear
tion,weobservespontaneous(PE)andstimulated(SPE)
regime already at moderate intensities.
photon echoes which appear exactly at time delays of
For isolated two-level systems in absence of decoher-
t = 2τ and t = 2τ +τ , respectively. In addition,
12 12 23 ence processes and when optical pulses with rectangular
Fig.1(f)demonstratesthatthePEpulseappearsattwice
intensity profile are assumed the amplitude of the PE
the delay between pulses 1 and 2. In what follows, we
follows the simple relation26
consider the maximum value of the PE amplitude at the
peaks of the PE, PPE, or the SPE, PSPE, signals. P sin2 Ω2τd , (1)
The dependence of the PE amplitude on the inten- PE ∝ (cid:18) 2 (cid:19)
sity of the excitation pulses can be used to compare
the strength of light-matter interaction in both systems. where Ω = 2E d/~ is the Rabi frequency which is de-
2 2
| |
For simplicity we consider the dependence of the two- termined by the electric field amplitude of the second
pulse PE on the energy of second pulse, , shown in pulse E and the dipole matrix element of the two-level
2 2
P
Figs. 2(a) and 2(b) for the bare QDs and the TP mi- systemd. InourcaseτdisfixedwhileE2 √ 2isvaried.
∝ P
4
We use this expression to evaluate the electromagnetic
fieldamplitudeinsidethemicrocavityE2TP−QD. Thisap- nits) 1 (a)
proximationshould be valid also in the strong field limit u
because the lifetime of the cavity mode is the shortest rb.
a
time scale compared to the duration of the excitation (E0.1
P
pulses and the radiative lifetime in the QDs. For P
2
P ≤
0.15nJ the PE oscillatorybehaviorin the TP-QDstruc-
ture is reasonably well reproduced by Eq. (1) with the
0 100 200 300 400 500 600
π-rotation occuring at 0.06 nJ. However, for larger
P2 ≈ Delay time 12 (ps)
pulse areas significant damping of the Rabi oscillations
s) 1
takes place. This is mainly due to a statistical distribu- nit (b)
tion of the dipole moments d in the QD ensemble which u
resultsinavariationofthepulseareaΘ=Ω2τdandblur- arb.
ringofthe Rabioscillations.24 Assuming a Gaussiandis- (E
tribution with a standard deviation σΘ centered around SP
Θ0 we obtain PPE ∝ 0∞sin2[Θ/2]exp[−(Θ2σ0−2ΘΘ2)2]dΘ. P 0.1
The best fit to the expRerimental data is obtained for 0 100 200 300 400
σ = 0.3 as shown by the solid curve in Fig. 2(b). This
value is in good agreementwith previous results on sim- s) Dealy time 23 (ps)
ilar QD ensembles.24 In addition other mechanisms such unit 1 (c)
as the interaction with acoustic phonons can lead to b.
r
a
dtoambepisntgroonfgtehstefRorabpiuolssecsillwaittihondsu.rTathieonlastotferseivseerxaplepcsteads en. (
used in our experiment.28,29 nt
L I0.1
In contrast to the TP microcavity structure, the P P
PE
forthebarewaferfollowsaquadraticdependenceon√ 2 0 500 1000 1500
P
which indicates that the pulse area is significantly below
Delay time t (ps)
π even for pulse energies as large as 0.5 nJ. Thus a
2
P ≈ FIG. 3. (Color online) Normalized time-resolved measure-
significantenhancementofthelight-matterinteractionis
ments on the bare QDs (circles) and the QDs in the TP mi-
clearlypresentintheTPcavity. Directcomparisonofthe
crocavity (triangles). T = 2 K. (a) PE amplitude as func-
intensities allows us to estimate the enhancement factor
tion of the pulse delay τ12. (b) SPE amplitude as function
ofthe electromagneticfieldforcoherentexcitationofthe of the pulse delay τ23 in order to measure the lifetime T1.
QDs. TakingintoaccountthatthePEsignaldependson τ12 =26.7 ps. The pulse areas in (a) and (b) did not exceed
both 1and 2,weestimatethatthepulseareainEq.(1) π. (c) Time-resolved PL for excitation with photon energy
is enhPancedPby a factor of 6, i.e. the amplitude of the ~ωexc =1.494 eV.Thedashed curvesare fitswith double(a)
electromagnetic field ETP 6Eb. This is in accordance and single (b,c) exponential decays, respectively. The result-
with the calculated inc2rea≈se of2the light intensity by a ing time constants are summarized in Table I.
factor of34 in the TP microcavityas comparedwith the
bare QD system (see Fig. 1(b)).
Let us now consider the transient decay of the photon due to the strong inhomogeneous broadening of the op-
echo signal which gives insightinto the coherentdynam- tical transition energies in the studied QDs. Apparently
ics and the relaxationprocesses in the QD systems. The the long component decays faster in the TP microcavity
dependences ofthePEandSPEamplitudes onthepulse with T2TP =170 ps, while in the bare QDs T2b =350 ps,
delays τ and τ are presented in Figs. 3(a) and 3(b), i.e the coherent decay is twice slower (see Table I).
12 23
respectively. ThePEdecayreflectsanirreversiblelossof TheSPEdecaygivesinsightintothepopulationdecay
opticalcoherence(i.e.,ofthemedium’spolarization)due dynamics and allows the evaluation of the exciton life-
to interaction of the two-level systems with the environ- time T . Here, we obtain for the bare QDs Tb = 390 ps
1 1
mentand/orradiativedamping. ItfollowsfromFig.3(a) and for the TP microcavity TTP = 170 ps. Using
1
that the signalcan be describedby a double exponential T2−1 = 1/2T1−1 +τc−1 we calculate the pure dephasing
′
dcleucdaiyngPaPEsho∝rtAcoehxepre(n−c2eτt1i2m/Te2T)2′+=B30expps(a−nd2τa1n2/aTp2p)roinx-- tfiinmeedτinc.thτcetQuDrnssinoutthetoTbPemshicorrotcearvfiotyr,tτhcTePex=ci3t4o0nspsc,oans-
imately 10 times longer coherence time T . The short compared with the bare QDs, τb =635 ps. As this time
2 c
dynamics are attributed to fast energy relaxation of QD exceeds the exciton lifetime, the pure dephaing is nev-
excitons excited in higher energy states, e.g. in p-shell erheless weak. In a studied fully dielectric DBR struc-
states,30 into the ground state. On the other hand, the ture with Q 200 the pure dephasing is negligible and
≈
long-lived signal with T decay time corresponds to the thegroundstateexcitoncoherenceisradiativelylimited:
2
coherent response of excitons in the ground s-shell. Ex- TDBR = 2TDBR with TDBR 300 400 ps. The lat-
2 1 1 ≈ −
citation of excitons in different energy shells is possible ter value has approximately the same magnitude as for
5
bare QDs. From the comparison we conclude that the but has to be attributed to the Purcell effect with an
exciton coherence in the bare QDs is somewhat reduced enhancement factor of about 2.
by charges at the surface 40 nm separated from the dot Acceleration of the spontaneous emission rate T−1 by
1
layer. The gold layer of the TP cavity induces further a factor of 2 represents a significant change of the radia-
decoherence which can be attributed to the interaction tive emission dynamics. In λ/2 microcavities with ideal
between excitons and plasmons. planar metal mirrors the Purcell factor is limited to 3,
The lifetime measurements deduced from the SPE in while in structures with DBR mirrors the modification
Fig. 3(b) show that for the TP microcavity T1 is ap- ofthespontaneousemissionrateistypicallysmallerthan
proximately halved compared to the bare QDs. This 20%.31,32 For planar resonators the spontaneous emis-
shorteningmostlikelyisduetothePurcelleffect.12 How- ±sion becomes mostly redistributed spatially, while the
ever, we have to consider also non-radiative processes vacuum field becomes only weakly squeezed leading to
due to tunneling of photoexcited carriers from the QDs a moderate enhancement of the local density of photon
into the nearby metal as potential origin. Further in- modes and the associated emission rate at best. In our
sightcanbeobtainedfromPLtransientsmeasuredusing case,inadditionSPPsmaybecomerelevantfortheshort-
a streak camera for below-barrier pulsed excitation with ening of T . On the other hand, the resonator-induced
1
photon energy ~ωexc = 1.494 eV. These measurements enhancementoflight-matterinteractionismuchstronger
are shown in Fig. 3(c). The PL signals decay with life- for resonantexcitation of coherentprocesses which is re-
times τ0b = 1350 ps and τ0TP = 590 ps for the bare QDs latedtotheselectiveexcitationofthecontributingquan-
andthe TP structure,respectively. Interestingly,the PL tum dots.
decay times τ are significantly larger as compared to
0 In conclusion, we have demonstrated that the coher-
the T -values from the SPE decay. Indeed, the two tech-
1 ent optical response from self-assembled (In,Ga)As QDs
niques measure different population dynamics. The PL
embeddedinaTPplanarmicrocavityisgivenbyphoton
from the ground state is an incoherent process after re-
echoes. Despite the low quality factor of about 100 we
laxation of the involved carriers. While the rise time of
demonstrate a substantial enhancement of the selective
this signal is in the few ten ps-range, the PL decay time
optical excitation of QDs whose optical transitions are
is significantly extended by several factors. For the cho-
in resonance with the TP cavity mode. The intensity of
sen conditions a carrier reservoir is excited also in the
thedrivingopticalfieldisamplifiedbymorethanoneor-
wetting layer,from where carriershave to be transferred
der of magnitude. Such enhancement allows to observe
totheQDs. Afterbeingcapturedbythedots,relaxation
Rabi oscillations in the photon echo and to perform co-
canoccureitherbyphononemissionorbycarrier-carrier
herent controlof excitons with picosecond optical pulses
scattering. In the latter case a carrier, for example an
of moderate intensities, while the statistical distribution
electron, relaxes at the expense of the other carrier, the
of dipole moments still represents a significant problem.
hole. Thereby populations in higher states and poten-
The decoherenceandpopulationdynamicsofexcitonsin
tially evenagainin the wetting layerare createdslowing
TPstructuresalsoexperiencemodifications. We observe
downthecarrierrecombination. Incontrast,SPEisaco-
a decrease of the radiative recombination time from 350
herent phenomenon following resonant excitation. If for
to 170 ps due to the Purcell effect. The presence of the
some reason exciton relaxation occurs, it will not con-
metal layer gives rise to pure dephasing of the QD exci-
tribute to the echo signal.
tonswithcharacteristictimesofabout200psto400psso
that pure dephasing remains quite weak. We note that
TABLE I. Decay constants evaluated from Fig. 3. T2 follows the metal layer at the top of the TP microcavity can
fromthePE,T1 fromtheSPE,andτ0 fromthetimeresolved be used to control the charge state of QDs electrically.13
PL measurements. Therefore,suchstructuresareappealingforinvestigation
T1 (ps) T2 (ps) τc (ps) τ0 (ps) oflong-livedphotonechoes fromchargedQDs where the
bare QDs 390 350 635 1350 decay rate is governedby the spin relaxation of the resi-
TP-QDs 170 170 340 590 dent electrons.33
We are grateful to M. Glazov and B. Glavin for use-
Non-radiative processes lead to shortening of the life- fuldiscussions. Weacknowledgethe financialsupportby
time. PLtransientswithextendeddynamicalrangeallow the Deutsche Forschungsgemeinschaft through the Col-
us to give an upper estimate for the non-radiative rate, laborative Research Centre TRR 142 and the Interna-
τ−1, in the TP microcavity which should contribute to tional Collaborative Research Centre 160. S.V.P. and
NR
the SPEdecayrate. The excitondecayratecanbe writ- Yu.V.K.thanktheRussianFoundationofBasicResearch
tenasτ−1 =τ˜−1+τ−1,whereτ˜−1 istheradiativedecay. for partial financial support (contracts no. ofi m 16-29-
0 0 NR 0
Neglecting non-radiative processes in the bare QDs, we 03115andno. 15-52-12016NNIO a). M.B.acknowledges
setτ˜b τb =1350ps. Thenfromτ =590ps,thelower partial financial support from the Russian Ministry of
limit0fo≈r th0e non-radiative decay tim0e is τTP 1 ns in Science and Education (contract no. 14.Z50.31.0021).
NR ≥
the TPmicrocavity. ThereforetheshorteningofT from Yu.V.K. acknowledges Saint Petersburg State Univer-
1
the 350 ps in the bare QDs to the 170 ps in the TP mi- sity for a research grant 11.42.993.2016. The project
crocavitycannot be initiated by non-radiativeprocesses, SPANGL4Q acknowledges financial support from the
6
TABLEII.Tamm-plasmonstructurecompositionwithlayerthicknessesandmaterialparametersusedinnumericalcalculations.
Material Thickness (nm) Refractive index n Function
Au 40 0.17+5.6i34 Gold mirror (Top)
GaAs 20 3.6 Cap
Al0.2Ga0.8As 10 3.5 Electron barrier
GaAs 10 3.6 Spacer
InAsQDs 2.3 3.6 QDslayer
GaAs 10 3.6 Spacer
Si ∼0 - Si-delta doping
GaAs 108 (variable) 3.6 Mode matching layer
20× AlAs / GaAs 78.2 / 66.7 2.9 / 3.6 DBR mirror
GaAs thick 3.6 Substrate(bottom)
1 within the Seventh Framework Programme for Research
oftheEuropeanCommission,underFET-Opengrantno.
0.8
FP7-284743.
y
vit 0.6
cti Q ≈170
efe 0.4
R
APPENDIX - SAMPLE OF TAMM-PLASMON
0.2
STRUCTURE AND SIMULATION METHOD
0
1 1.1 1.2 1.3 1.4 1.5 1.6 1.7
Energy (eV) The composition of the TP structure is presented in
Table II. During the growth process substrate was ro-
FIG. 4. (Color online) Reflectivity spectrum simulated for tated in order to produce homogeneous thicknesses of
theTamm-plasmon structure. the layers. The substrate rotation was interrupted for
the mode matching layer which resulted in a gradient of
4 thicknessofthislayer. Owingtothisgradientitwaspos-
15 with Au
w/o Au sible to tune the spectral position of the photon cavity
y
nsit 3 mode on that part of the sample which was coated with
nte 10 a 40 nm golden layer.
ve feld i 5 2nRe() TPInstorrudcetrurteotshimeutlraatnesfoeprtmicaaltrpixrompeertthieosdowfatsheussetdu.dieIdn
elati 1 calculation, normal light incidence was considered. The
R
thicknessofthemodematchinglayerwasadjustedtoget
0 0
−0.2 −0.1 0 0.1 0.2 0.3 0.4 0.5 anaccordanceofthesimulatedreflectionspectrumtothe
Depth (µm) measuredone. Resultant spectrumshownin Fig.4 gives
quality factor Q 170, which is somewhat greater than
FIG.5. (Coloronline)Relativefieldintensitydistributionin- measured value o≈f 130.
sidethestructurewith topgolden layer(blueline) andwith-
Exactly at the center of the photon mode (E =
outit(redline);fallinglightintensityis1. Realpartofrefrac-
1.3607 eV) the distribution of field intensity inside the
tiveindexdistributedinsidethestructureisshownwithgray
line. Thesampletopislocated atzero; layerabovesampleis structurewascalculatedforthecaseofTPstructureand
vacuum. Dashed line indicates location of QDs layer. for the structure without top golden layer. This calcula-
tionis shownin Fig.5 together withthe refractiveindex
distribution inside the structure. Enhancement of field
Future and Emerging Technologies (FET) programme intensity at the location of QDs layer is 34.
∗ [email protected] controlledcouplingofexcitonicqubits,NatureCommun.4,
1 P.Lodahl,S.Mahmoodian,andS.Stobbe,Interfacingsin- 1747 (2013).
gle photons and single quantum dots with photonic nanos- 3 O.Benson,Assemblyofhybridphotonicarchitectures from
tructures, Rev.Mod. Phys.87 , 346 (2015). nanophotonic constituents, Nature 480, 193 (2011).
2 F. Albert, K. Sivalertporn, J. Kasprzak, M. Strau, 4 M. Kaliteevski, I. Iorsh, S. Brand, R.A. Abram,
C. Schneider, S. H¨ofling, M. Kamp, A. Forchel, S. Re- J.M.Chamberlain,A.V.Kavokin,andI.A.Shelykh,Tamm
itzenstein, E.A. Muljarov, and W. Langbein, Microcavity plasmon-polaritons: Possible electromagnetic states at the
7
interface of a metal and a dielectric Bragg mirror, Phys. a semiconductor microcavity, Phys. Rev. B 91, 201304
Rev.B 76, 165415 (2007). (2015).
5 B.I. Afinogenov, V.O. Bessonov, A.A. Nikulin, and 19 F. Fras, Q. Mermillod, G. Nogues, C. Hoarau, C. Schnei-
A.A. Fedyanin, Observation of hybrid state of Tamm der, M. Kamp, S. H¨ofling, W. Langbein and J. Kasprzak,
and surface plasmon-polaritons in onedimensional pho- Multi-wave coherent control of a solid-state single emitter,
tonic crystals, Appl.Phys. Lett. 103, 061112 (2013). NaturePhoton. 10, 155 (2016).
6 M.Lopez-Garcia,Y.-L.D.Ho,M.P.C.Taverne,L.-F.Chen, 20 S.V. Poltavtsev, M. Salewski, Yu.V. Kapitonov, I.A. Yu-
M.M. Murshidy, A.P. Edwards, M.Y. Serry, A.M. Adawi, gova, I.A. Akimov, C. Schneider, M. Kamp, S. H¨ofling,
J.G. Rarity, and R. Oulton Efficient out-coupling and D.R.Yakovlev,A.V.Kavokin,andM.Bayer,Photon echo
beaming of Tamm optical states via surface plasmon po- transients from an inhomogeneous ensemble of semicon-
lariton excitation, Appl.Phys.Lett. 104, 231116 (2014). ductor quantum dots, Phys.Rev. B 93, 121304 (2016).
7 S.Azzini, G. Lheureux, C. Symonds, J.-M. Benoit, P. 21 Q. Mermillod, D. Wigger, V. Delmonte, D.E. Reiter,
Senellart, A. Lemaitre, J.-J. Greffet, C. Blanchard, C. C.Schneider,M.Kamp,S.H¨ofling,W.Langbein,T.Kuhn,
Sauvan, and J. Bell, Generation and spatial control of hy- G. Nogues, and J. Kasprzak, Dynamics of excitons in in-
bridTammplasmon/surfaceplasmonmodes,ACSPhoton- dividual InAs quantum dots revealed in four-wave mixing
ics 3, 1776 (2016) spectroscopy, Optica3, 377 (2016).
8 C. Grossmann, C. Coulson, G. Christmann, I. Farrer, 22 T. Braun, S. Betzold, N. Lundt, M. Kamp, S. H¨ofling,
H.E.Beere,D.A.Ritchie,andJ.J.Baumberg,Tuneablepo- and C. Schneider, Impact of ex situ rapid thermal anneal-
laritonicsatroomtemperaturewithstronglycoupledTamm ingonmagneto-opticalpropertiesandoscillatorstrengthof
plasmon polaritons in metal/air-gap microcavities, Appl. In(Ga)As quantum dots, Phys.Rev.B 93, 155307 (2016).
Phys. Lett. 98, 231105 (2011). 23 W. Langbein and P. Borri in Semiconductor Qubits (eds
9 C. Symonds, G. Lheureux, J.P. Hugonin, J.J. Greffet, F. Henneberger and O. Benson), Ch. 12, 269 (Pan Stan-
J. Laverdant, G. Brucoli, A. Lemaˆıtre, P. Senellart, and ford, 2008).
J. Bellessa, Confined Tamm Plasmon Lasers, Nano Lett. 24 P. Borri, W. Langbein, S. Schneider, U. Woggon,
13, 3179 (2013). R.L.Sellin, D.Ouyang,andD.Bimberg, Rabi oscillations
10 J. Gessler, V. Baumann, M. Emmerling, M. Amthor, in the excitonic ground-state transition of InGaAs quan-
K.Winkler,S.H¨ofling,C.Schneider,andM.Kamp,Elec- tum dots, Phys. Rev.B 66, 081306 (2002).
tro optical tuning of Tamm-plasmon exciton-polaritons, 25 Seesupplementingmaterialfordetailsofsamplestructure
Appl.Phys. Lett. 105, 181107 (2014). and calculations using tranfer matrix method.
11 K. Sebald, SK.S. Rahman, M. Cornelius, J. Gutowski, 26 P.R. Berman and V.S. Malinovsky, in Principles of Laser
T. Klein, S.Klembt, C. Kruse,and D.Hommel, Tailoring Spectroscopy andQuantum Optics,Ch.9(PrincetonUniv.
the optical properties of wide-bandgap based microcavities Press, 2011).
via metal films, Appl.Phys. Lett. 107, 062101 (2015). 27 P.Borri,W.Langbein,S.Schneider,andU.WoggonUltra-
12 O. Gazzano, S. Michaelis de Vasconcellos, K. Gauthron, long dephasing time in InGaAs quantum dots, Phys. Rev.
C. Symonds,J. Bloch, P. Voisin, J. Bellessa, A. Lemaˆıtre, Lett. 87, 157401 (2001).
and P. Senellart, Evidence for confined Tamm plasmon 28 A. Kru¨gel, V.M. Axt, T. Kuhn, P. Machnikowski, and
modes under metallic microdisks and application to the A. Vagov, The role of acoustic phonons for Rabi oscilla-
control of spontaneous optical emission, Phys. Rev. Lett. tions in semiconductor quantum dots, Appl. Phys. B 81,
107, 247402 (2011). 897 (2005).
13 T. Braun, V. Baumann, O. Iff, S. H¨ofling, C. Schnei- 29 D.E. Reiter, T. Kuhn, M. Gla¨ssl, and V.M. Axt, The role
der,andM.Kamp,Enhanced single photon emissionfrom of phonons for exciton and biexciton generation in an op-
positioned InP/GaInP quantum dots coupled to a con- tically driven quantum dot, J. Phys.: Condens. Matter 26
finedTamm-plasmonmode,Appl.Phys.Lett.106,041113 423203 (2014).
(2015). 30 H. Kurtze, J. Seebeck, P. Gartner, D.R. Yakovlev,
14 R. Bru¨ckner, A.A. Zakhidov, R. Scholz, M. Sudzius, D. Reuter, A.D. Wieck, M. Bayer, and F. Jahnke, Car-
S.I. Hintschich, H. Fr¨ob, V.G. Lyssenko, and K. Leo, rier relaxation dynamics in self-assembled semiconductor
Phase-locked coherent modes in a patterned metalorganic quantum dots, Phys.Rev.B 80, 235319 (2009).
microcavity, NaturePhoton. 6, 322 (2012). 31 S.D. Brorson, H. Yokoyama, E.P. Ippen, Spontaneous
15 Y. Chen, D. Zhang, L. Zhu, R. Wang, P. Wang, H. Ming, emission rate alteration in optical waveguide structures,
R. Badugu, and J.R. Lakowicz, Tamm plasmon- and sur- IEEE Journal of Quantum Electronics 26, 1492 (1990).
face plasmon-coupled emission from hybrid plasmonicpho- 32 M.Bayer,T.L.Reinecke,F.Weidner,A.Larionov,A.Mc-
tonic structures, Optica1, 407 (2014). Donald, and A. Forchel, Inhibition and Enhancement of
16 J. Shah, Ultrafast Spectroscopy of Semiconductors and the Spontaneous Emission of Quantum Dots in Structured
Semiconductor Nanostructures, Vol. 115 of Springer Se- Microresonators, Phys. Rev.Lett. 86, 3168 (2001).
ries in Solid-State Sciences (Springer Science & Business 33 L. Langer, S.V. Poltavtsev, I.A. Yugova, M. Salewski,
Media, New York,1999). D.R. Yakovlev, G. Karczewski, T. Wojtowicz, I.A. Aki-
17 V. Kohnle, Y. Leger, M. Wouters, M. Richard, mov, and M. Bayer, Access to long-term optical memories
M.T. Portella-Oberli, and B. Deveaud, Four-wave mixing using photon echoes retrieved from semiconductor spins,
excitations in a dissipative polariton quantum fluid, Phys. NaturePhoton. 8, 851–857 (2014).
Rev.B 86, 064508 (2012). 34 P.B. Johnson and R.W. Christy, Optical Constants of the
18 B.L. Wilmer, F. Passmann, M. Gehl, G. Khitrova, and Noble Metals, Phys. Rev.B 6, 4370 (1972).
A.D. Bristow, Multidimensional coherent spectroscopy of