Appl. Phys. Lett. 117, 224001 (2020); https://doi.org/10.1063/5.0030991 117, 224001
© 2020 Author(s).
Deterministically fabricated strain-tunable
quantum dot single-photon sources emitting
in the telecom O-band
Cite as: Appl. Phys. Lett. 117, 224001 (2020); https://doi.org/10.1063/5.0030991
Submitted: 26 September 2020 • Accepted: 10 November 2020 • Published Online: 01 December 2020
N. Srocka, P. Mrowiński, J. Große, et al.
COLLECTIONS
Paper published as part of the special topic on Non-Classical Light Emitters and Single-Photon Detectors
This paper was selected as Featured
ARTICLES YOU MAY BE INTERESTED IN
Deterministically fabricated quantum dot single-photon source emitting indistinguishable
photons in the telecom O-band
Applied Physics Letters 116, 231104 (2020); https://doi.org/10.1063/5.0010436
Purcell-enhanced single-photon emission from a strain-tunable quantum dot in a cavity-
waveguide device
Applied Physics Letters 117, 254002 (2020); https://doi.org/10.1063/5.0033213
Progress in quantum-dot single photon sources for quantum information technologies: A
broad spectrum overview
Applied Physics Reviews 7, 021309 (2020); https://doi.org/10.1063/5.0010193
Deterministically fabricated strain-tunable
quantum dot single-photon sources emitting
in the telecom O-band
Cite as: Appl. Phys. Lett. 117, 224001 (2020); doi: 10.1063/5.0030991
Submitted: 26 September 2020 .Accepted: 10 November 2020 .
Published Online: 1 December 2020
N. Srocka,
1
P. Mrowi
nski,
1,2
J. Große,
1
M. Schmidt,
1,3
S. Rodt,
1
and S. Reitzenstein
1,a)
AFFILIATIONS
1
Institut f€
ur Festk€
orperphysik, Technische Universit€
at Berlin, Hardenbergstraße 36, D-10623 Berlin, Germany
2
Laboratory for Optical Spectroscopy of Nanostructures, Department of Experimental Physics, Wrocław University of Technology,
Wybrze_
ze Wyspia
nskiego 27, Wrocław, Poland
3
Physikalisch-Technische Bundesanstalt, Abbestraße 2-12, 10587 Berlin, Germany
Note: This paper is part of the APL Special Collection on Non-Classical Light Emitters and Single-Photon Detectors.
a)
Author to whom correspondence should be addressed: stephan.reitzenstein@physik.tu-berlin.de
ABSTRACT
Most quantum communication schemes aim at the long-distance transmission of quantum information. In the quantum repeater concept,
the transmission line is subdivided into shorter links interconnected by entanglement distribution via Bell-state measurements to overcome
inherent channel losses. This concept requires on-demand single-photon sources with a high degree of multi-photon suppression and high
indistinguishability within each repeater node. For a successful operation of the repeater, a spectral matching of remote quantum light sour-
ces is essential. We present a spectrally tunable single-photon source emitting in the telecom O-band with the potential to function as a
building block of a quantum communication network based on optical fibers. A thin membrane of GaAs embedding InGaAs quantum dots
(QDs) is attached onto a piezoelectric actuator via gold thermocompression bonding. Here, the thin gold layer acts simultaneously as an
electrical contact, strain transmission medium, and broadband backside mirror for the QD-micromesa. The nanofabrication of the QD-
micromesa is based on in situ electron-beam lithography, which makes it possible to integrate pre-selected single QDs deterministically into
the center of monolithic micromesa structures. The QD pre-selection is based on distinct single-QD properties, signal intensity, and emission
energy. In combination with strain-induced fine tuning, this offers a robust method to achieve spectral resonance in the emission of remote
QDs. We show that the spectral tuning has no detectable influence on the multi-photon suppression with g
(2)
(0) as low as 2%–4% and that
the emission can be stabilized to an accuracy of 4 leV using a closed-loop optical feedback.
Published under license by AIP Publishing. https://doi.org/10.1063/5.0030991
The emission of single photons with controllable wavelength and
high indistinguishability is a key parameter in quantum light sources
for quantum nanophotonics. It is, for example, required to implement
boson sampling experiments
1
and to realize entanglement swapping
via Bell-state measurements in large-distance quantum communica-
tion networks based on the quantum repeater concept.
2–4
In fact, in
the case of the quantum repeater, it is crucial to fabricate a chain of
wavelength-tunable quantum light sources that provide identical
photons on demand. Even more, it is important to operate such light
sources at telecommunication wavelengths, i.e., within the telecom O-
band (1.3 lm) or C-band (1.55 lm) with enhanced brightness
5
and a deterministic fabrication scheme,
6
to pave the way toward the
real-world implementation of long-distance quantum communication
networks via optical fibers, as recently reported using electrical Stark
tuning of a quantum dot (QD) device.
7
In this work, we demonstrate technological advances and experi-
mental findings to realize wavelength-tunable quantum emitters of
high single-photon purity in the telecom O-band. The strain-tunable
emitters are based on self-assembled InGaAs quantum dots that are
deterministically integrated into photonic nanostructures attached to
piezoelements by means of a flip-chip process.
8,9
Here, the emission
range of the InGaAs QDs is redshifted to the telecom O-band in the
epitaxial growth process by using a strain-reducing layer (SRL),
10,11
while the piezoelectric actuator allows us to apply an external strain
fieldtofinetuneandstabilizetheQDemissionwavelengthduring
operation of the quantum light sources. To demonstrate their
Appl. Phys. Lett. 117, 224001 (2020); doi: 10.1063/5.0030991 117, 224001-1
Published under license by AIP Publishing
Applied Physics Letters ARTICLE scitation.org/journal/apl
application potential, we examined the optical properties of two
strain-tunable O-band single-QD mesa structures. The emission of
their charged excitonic (CX) states, which are relevant for efficient
single-photon generation,
12
is first studied regarding multi-photon
suppression. Second, we test the wavelength tuning capabilities via
strain-tuning and we introduce active spectral stabilization based on a
closed-loop proportional–integral–derivative (PID) controller. Finally,
we demonstrate strain-tuning and active stabilization of two remote
QDs, which is a crucial step over existing results toward the real-world
application of QD single-photon source (SPS) in advanced quantum
communication schemes where it can pave the way toward two-
photon interference (TPI) of remote sources in the telecom O-band
required for the implementation of fiber-based quantum repeater
networks.
4
We manufactured our device in three main technological steps: (i)
growing a semiconductor heterostructure by metal organic chemical
vapor deposition, (ii) gold bonding an about 900 nm thick membrane
of this structure onto a piezoelectric actuator, and (iii) nanostructuring
the QD membrane deterministically by low-temperature cathodolumi-
nescence (CL) scanning and in situ electron beam lithography (EBL).
First, a 200 nm GaAs buffer is grown on an n-doped GaAs (100)
substrate followed by a 1 lmthickAl
0.90
Ga
0.10
As etch stop layer, a
2lm thick GaAs layer, and another 100 nm thick Al
0.90
Ga
0.10
As layer.
These three layers and the substrate are sacrificial layers to be removed
during post-growth processing. The final 879 nm thick region of GaAs
forms the active device membrane and includes a single InGaAs QD
layer combined with the SRL, with the QD layer located 637 nm above
the second etch stop layer [see Fig. 1(a)]. The QD layer is formed by
1.5 monolayers of In
0.7
Ga
0.3
As followed by a 0.5 monolayer GaAs
flush. The subsequent 5.5nm thick InGaAs SRL has a gradual decrease
in the In content from 30% to 10% over the first 3.5nm.
Next, the as-grown sample and a PIN–PMN–PT [Pb(In
1/2
Nb
1/2
)O
3
–
Pb(Mg
1/3
Nb
2/3
)O
3
–PbTiO
3
] piezoelectric crystal are sputtered with
250 nm of gold. Their surfaces are then activated under argon
plasma and, immediately after being removed from the deposition
tool, bonded face-to-face under ambient air and thermocompres-
sion. In this step, a pressure of 6 MPa and a temperature of about
600 K are applied. This process allows for a stable bond that is not
affected by the inherent surface roughness of the PIN–PMN–PT
crystal. For more details on this process, see Ref. 13.
The flip-chip process is finalized by a layer-by-layer wet-etching
process. First, the 400 lm thick GaAs substrate is removed by a fast and
aggressive etchant (H
2
O
2
/NH
3
, 10:1) and stopped at the first
Al
0.9
Ga
0.1
As layer due to the high selectivity of the wet-chemical etching
process.
14
HCl acid is then used to solely remove the exposed etch stop
layer. In our process, we use HCl instead of the typically applied HF
mainly because of the lower etch rate that leads to better process control
and higher quality of the etched surfaces. The now-exposed second and
thinnerGaAslayer(2lm) is lifted off by a slower etchant (citric acid/
H
2
O
2
, 4:1), which leads to a significantly improved surface roughness
compared to a single-step etch process. Again, the second etch stop layer
is removed by HCL acid. At this point, an 885 nm thick membrane,
containing the single QD layer, is gold bonded to a PIN–PMN–PT crys-
tal and completely freed from sacrificial layers.
In preparation for the following main processing step, the sample
is spin coated with a nominal 300 nm thick layer of AR-P 6200 (CSAR
62) electron-beam resist. The sample is then loaded in a special scan-
ning electron microscope (SEM) that enables CL measurements at
cryogenic temperatures. The sample is cooled to 10K, and the CL sig-
nal is mapped over a 20 lm20 lm large area with a voxel size of
0.5 lm to identify the position of bright and isolated QDs that emit at
the desired wavelength in the telecom O-band. A corresponding map
showing spatially resolved CL in a spectral range of (1300–1305) nm is
shown in Fig. 2(a). By applying a 2D Gaussian fit, we can select the
two circled QDs in Fig. 2(a) with a lateral accuracy of about 40 nm.
In the following electron beam lithography (EBL) step, deterministic
circular mesa structures are written into the resist at low temperature
(10 K) in the same SEM system. When the resist is developed, the
image areas are cleared and only the EBL-structured mesa structures
remain and form an etching mask for the final reactive ion etching in
an inductively coupled plasma. For the circular QD micromesas with a
gold mirror on the back, we expect a moderately high photon
extraction efficiency gin the range of 5%–10%.
13,15
This parameter
can possibly be increased in the future by applying the present
manufacturing method to sources based on circular Bragg resonators
(CBRs) that promise broadband photon extraction efficiency up to
g¼95% in the telecom O-band.
16
In fact, our developed device proc-
essing is fully compatible with the realization of CBRs, and using a
soon available commercial in situ EBL system, we plan to implement
1.3 lm QD-CBRs. In this context, it will be interesting to explore the
effect of piezostrain tuning on both the QD and cavity properties,
which could probably be used to induce a piezocontrolled slight ellip-
ticity in the CBRs to realize high-performance QD-SPS with linearly
polarized emission under resonant excitation.
17
Noteworthily, and in contrast to previous work on piezoelectric
tuning of QD-micromesas emitting in the 930 nm range,
18
the etching
depth used in the present work corresponds to the thickness of the
gold-bonded membrane in our design. Thus, the base diameter of the
mesa structure is the only cross section to transfer the strain-field effect
at the piezoelectric actuator to the QD position. As we discuss below,
the limited contact area between the semiconductor and the piezoelec-
tric actuator does not have a strong impact on the achievable tuning
range, which is rather limited by non-ideal wafer-bonding.
The properties of the optical device are investigated by means
of microphotoluminescence (lPL), lPL-excitation (lPLE), and
FIG. 1. (a) Sample structure as grown. (b) Sample mounted in the customized
spring holder of the cryostat. (c) Optical microscope image of the sample after
resist development. (d) SEM image of the processed sample showing the mapping
area with the two mesas hosting QD1 and QD2. (e) SEM image of the investigated
mesa hosting QD1.
Applied Physics Letters ARTICLE scitation.org/journal/apl
Appl. Phys. Lett. 117, 224001 (2020); doi: 10.1063/5.0030991 117, 224001-2
Published under license by AIP Publishing
photon-correlation spectroscopy at 10 K. The used helium-flow cryo-
stat has an extra-high-voltage feedthrough to electrically connect the
piezoelectric actuator to a voltage supply and a customized spring
holder to facilitate strain-tuning of the QD-micromesas. The deter-
ministic QD mesa structures are optically excited by a continuous-
wave (cw) diode laser (785 nm) or a tunable pulsed laser providing
ps-pulses at a repetition rate of 80MHz. The photoluminescence
signal is collected with an objective [numerical aperture (NA) ¼0.4]
and spectrally resolved in a grating spectrometer (spectral resolution
20 leV). The photon stream can either be detected by a liquid nitro-
gen cooled InGaAs-array detector or fiber-based Hanbury Brown and
Twiss (HBT) and Hong–Ou–Mandel (HOM) configurations attached
to the output slit of the monochromator. The quantum optical config-
urations include two superconducting nanowire single-photon detec-
tors (SNSPDs) (a temporal resolution of approximately 50 ps and a
detection efficiency of approximately 80% at 1310 nm).
Device characterization starts by basic lPL measurements under
non-resonant cw excitation (785 nm) at 10 K. We identified two pho-
tonic structures with bright CX emission lines at 1303.98nm (QD1)
and 1303.78 nm (QD2). The spectral fingerprints of these two QDs, as
shown in Fig. 2(b), have three characteristic and representative emis-
sion lines, which are identified as a neutral exciton (X), singly charged
exciton (CX), and biexciton (XX) by excitation-power and polari-
zation dependent measurements (see supplementary material
Figs. S1 and S2).
To demonstrate strain-tuning and to quantify the corresponding
spectral shift, the QD emission is studied depending on the applied
voltage to the piezoelectric actuator. The applied voltage was varied
from 400 V to þ400 V, which corresponds to an electric field Fof
13.4 kV cm
1
to þ13.4 kV cm
1
and results in a spectral tuning
range of 0.5 nm for the present sample, as shown in Fig. 3.
Noteworthily, other samples showed a tuning range up to 4nm when
the device allows for applying higher voltages of up to 633 kV cm
1
(cf. Fig. S3), however, at the risk of sample damage. Comparison
between these results and values reported in Ref. 18 for a non-
completely etched QD-mesa shows that the complete removal of the
semiconductor material around the QD-mesa does not significantly
reduce the available tuning range since in the electric field range
of 620 kV/cm, the achieved relative spectral shift of (0.24 60.02)%
for device B presented in the supplementary material, which is consis-
tent with (0.20 60.02)% that can be extracted for the device discussed
in Ref. 18. In fact, we observe strong device-to-device variations of the
available tuning range, which indicates non-ideal wafer-bonding using
our home-made bonding tool, and we expect higher tuning ranges
and better reproducibility by using a commercial wafer bonder in the
future.
In addition to the desired wavelength shift of our piezocontrolled
QD-micromesas, we observe spectral fluctuation over time due to
piezoelectric creep, which is most pronounced directly after a voltage
change. To illustrate this point, Fig. 4(a) shows a systematic wave-
length change of about 0.05 nm over the first 5 min after setting the
piezovoltage. The creep effect decreases significantly after 30 min but
does not vanish completely over time. This behavior is similar to
FIG. 2. (a) CL map of the pre-selection step at 10 K with QD1 and QD2 encircled. The spectral range of 1300 nm–1305 nm is depicted. (b) Corresponding lPL spectra of QD1
and QD2 after mesa processing was completed.
FIG. 3. Energy tuning of the CX emission line of QD1 (black to blue). An energy shift
of 380 leV (0.5nm) was observed when changing the electric field applied to the pie-
zoelectric actuator from 13.4 kV/cm to þ13.4 kV/cm. The CX emission line of QD2
without the applied electric field (red) is within the energetic tuning limits of QD1.
Applied Physics Letters ARTICLE scitation.org/journal/apl
Appl. Phys. Lett. 117, 224001 (2020); doi: 10.1063/5.0030991 117, 224001-3
Published under license by AIP Publishing
earlier observations in piezocontrolled QD devices.
8,18,19
In addition,
we examined the relative change of the emission energies of X, XX,
and CX and we found that the binding energy is only slightly influ-
enced, on a scale of 0.3 meV, by the applied strain (see supplementary
material Fig. S3). This change in binding energy is significantly smaller
than values on the order of meV observed for planar QD samples with
obviously better strain transfer.
20
For some QDs, the excitonic fine
structure splitting (FSS) could be reduced from 40 leV to 20 leV [see
supplementary material Fig. S3(d)]. Further FSS reduction is not possi-
ble in our case. In fact, a second degree of freedom in the piezoelectric
actuator would be required to fully symmetrize the distorted confine-
ment potential, being responsible for the FSS.
21
Long-term spectral stability is one of the most important aspects
of a practical single-photon source for quantum communication appli-
cations based, e.g., on entanglement swapping via Bell-state measure-
ments. Therefore, the effect of creep on the wavelength stability is an
important issue to address. One can efficiently stabilize the emission
wavelength, as it was reported for a quantum light emitting diode
8
and
later also realized for a deterministically fabricated InAs/GaAs QD
microlens,
18
by using an active optical feedback (closed-loop)
algorithm. In the implemented concept, the emission wavelength is
monitored by Lorentzian fitting of lPL spectra taken regularly within
an integration time of 0.2–1.0 s, depending on the signal intensity. The
center position of the fit is compared with the target wavelength, and
the determined deviation initiates a re-adjustment procedure of the
inputvoltagetoshifttheemissionlinebacktothesettargetvalue.
This re-adjustment is based on a PID algorithm.
The implemented wavelength stabilization is demonstrated for
the CX emission of QD1 for several PID settings. In this test, we set
the proportional gain (PG) to 100, and the integral time (IT) parameter
to 0.01 s, 0.05 s, and 0.1 s to study its impact on the QD wavelength
control after a voltage change applied to the piezoactuator. A graphical
representation of the associated results is shown in Fig. 4(b),wherethe
following settings are characterized within a time window of 5 min for
each setting. The corresponding parameter set and achieved wavelength
stability are shown in Table I. Evaluation of the recorded lPL time
traces shows that the best spectral stabilization with a standard devia-
tion of the emission wavelength (energy) of 5.6 pm (4.1 leV) could be
achieved for PG ¼100 and IT ¼0.01 s. Moreover, we found that by
increasing PG, the stabilization time decreases. However, in order to
maintain the stabilization quality, the IT value must be increased pro-
portionally. We did not observe any clear dependence of the QD spec-
tral stability on the change in the derivative term of the PID controller.
Next, we demonstrate resonance tuning and stabilization of CX
emission lines of the two different QD-micromesas (QD1 and QD2)
by applying the optimized PID settings determined before. The time
evolution of the piezoelectrically stabilized CX emission wavelength is
presented in Fig. 5(a), showing standard deviations of 12 leV for QD1
and 6 leV for QD2, respectively. In Fig. 5(b), we present lPL spectra
of both QDs tuned into resonance and being stabilized at 1304.05 nm
with respect to the CX emission line. The good spectral overlap of the
inhomogeneously (by spectral diffusion) broadened emission lines of
QD1 and QD2 under active PID control reflects the efficient fine tun-
ing and spectral control of the QDs via the PID controlled piezoactua-
tor. Ultimately, it has to be demonstrated that a high spectral overlap
can also be achieved on a scale of the homogenous linewidth, which
can be probed by HOM experiments on remote sources.
To verify the quantum nature of emission, we measured the
photon-autocorrelation of the spectrally tunable CX lines of QD1 and
QD2. Here, the CX transition is driven at saturation under pulsed
non-resonant excitation with a repetition rate of 80 MHz. In Fig. 6,we
present histograms of the correlation events recorded in the HBT con-
figuration representing the second-order autocorrelation function
g
(2)
(s), which is used to evaluate the multi-photon suppression at zero
time delay. First, the autocorrelation functions of QD1 and QD2 were
measured without the influence of the piezoinduced strain-field emit-
ting at different wavelengths, as shown in the upper panels of Fig. 6.
The measurements show pronounced antibunching for the central
FIG. 4. (a) Impact of the piezoelectric actuator’s creep on the wavelength of the CX emission line of QD1 over time. (b) PID stabilization tests on QD1 (see Table I for the
summary).
TABLE I. Quantified results of the emission energy stability for various PID integral
times evaluated in terms of the standard deviation of emission energy and stabiliza-
tion time required for the adjustment.
PID settings (prop. gain ¼100) PID
1
PID
2
PID
3
PID integral time (s) 0.01 0.05 0.1
Standard deviation (leV) 4.12 4.45 5.26
Stabilization time (s) 30 50 80
Applied Physics Letters ARTICLE scitation.org/journal/apl
Appl. Phys. Lett. 117, 224001 (2020); doi: 10.1063/5.0030991 117, 224001-4
Published under license by AIP Publishing
peak at s¼0 ns, and the evaluation of the experimental data with two-
sided mono-exponential fit functions convoluted with the timing
response of 50 ps of the SNSPDs used in this experiment yields
g2
ðÞ0
ðÞ
fit ¼0:020þ0:030
0:020 for QD1 and g2
ðÞ 0
ðÞ
fit ¼0:052 60:030 for
QD2. When the electric field is applied to the piezoactuator and both
QDs’ CX emissions are set to the resonant wavelength and stabilized
by PID for several minutes at 1304.05nm (the measurements
were taken independently), the antibunching behavior is preserved
and the corresponding fits yield g2
ðÞ0
ðÞ
fit ¼0:040 60:030 for QD1
and g2
ðÞ 0
ðÞ
fit ¼0:053 60:030 for QD2, confirming that the spectral
tuning has no significant impact on the single-photon purity of the
QD-micromesas. The achieved results are very promising with regard
to the target applications requiring the emission of indistinguishable
single photons from remote quantum light sources. However, we
would like to point out that the photon indistinguishability of these
QD-micromesas is not yet high enough for this purpose. In fact, we
examined the TPI of tunable and stabilized CX emission of QD1 under
pulsed p-shell excitation in the Hong–Ou–Mandel configuration. As
we discuss in the supplementary material, this experiment yields a
HOM visibility of ð16 68Þ%, a post-selected visibility of 79 615
ðÞ
%;
with a corresponding coherence time of ð470 685Þps, and a lifetime
of s1¼ð1:54 60:05Þns (see the supplementary material,Figs.5and
6). These values are in agreement with results achieved recently for
non-tunable QD-micromesas,
13
which indicates that the optical qual-
ity and coherence of the 1.3 lm QD have to be improved in the future,
e.g., by optimizing the SRL, in order to fully exploit their potential in
advanced quantum communication applications. For instance, the
gradual increase in the In content in the SRL and the overall thickness
of the SRL could be varied as an optimization parameter to achieve a
smoother transition from the strongly lattice mismatched In
0.7
Ga
0.3
As
QD layer to the GaAs matrix, thereby reducing the density of (charge
trapping) defect states. Alternatively, one could also consider slowing
down the GaAs growth on top of the SRL and try to better “heal” out
the roughness introduced by the In-containing layers. HOM measure-
ments can be a sensitive tool to evaluate the impact of design modifica-
tions on the optical and quantum optical properties of the 1.3lm
QDs.
In summary, we demonstrated a deterministically fabricated
and tunable QD single-photon source emitting in the telecommu-
nication O-band at 1.3 lm. Our device consisting of a QD micro-
mesa and attached to a piezoelectric actuator allows us to fine tune
the QD’s emission wavelength by up to 0.5 nm via a voltage-
controlled strain field. The tuning range and an implemented
closed-loop PID control system enabled us to stabilize the emission
energy with an accuracy (standard deviation) of up to 4leV and
to hold emission lines of two remote QD-micromesas with multi-
photon-suppression better than 5% in resonance with excellent
spectral overlap. These results are very promising for the future
implementation of long-distance quantum communication
networks.
See the supplementary material for information on the identifica-
tion of excitonic complexes in QD1 and QD2, Hong–Ou–Mandel
studies on the photon indistinguishability of QD1, and on optical
properties of a second device with a larger strain-tuning range.
FIG. 5. (a) Emission wavelength traces of QD1 and QD2 over time to quantify the stabilization and (b) lPL spectra when both QDs are tuned in resonance via the applied pie-
zoelectric strain field.
FIG. 6. (a) Autocorrelation experiments performed on the charged excitons of
QD1 and (b) QD2 in the case of “on” and “off” PID controllers, which are used for
stabilizing the piezocontrolled wavelength of the emitters.
Applied Physics Letters ARTICLE scitation.org/journal/apl
Appl. Phys. Lett. 117, 224001 (2020); doi: 10.1063/5.0030991 117, 224001-5
Published under license by AIP Publishing
This work was funded by the FI-SEQUR project jointly financed
by the European Regional Development Fund (EFRE) of the European
Union in the framework of the programme to promote research,
innovation, and technologies (Pro FIT) in Germany within the 2nd
Poland-Berlin Photonics Programme, Grant No. 2/POLBER-2/2016.
Support from the German Research Foundation through CRC 787
“Semiconductor Nanophotonics: Materials, Models, Devices” and the
Volkswagen Foundation via project “NeuroQNet” is also
acknowledged. P.M. gratefully acknowledges the financial support
from the Polish Ministry of Science and Higher Education within the
“Mobilnosc Plus–Vedycja” program and from the Polish National
Agency for Academic Exchange (NAWA) via project PPI/APM/2018/
1/00031/U/001.
We thank T. Heindel for technical support.
DATA AVAILABILITY
The data that support the findings of this study are available
from the corresponding author upon reasonable request.
REFERENCES
1
H. Wang, J. Qin, X. Ding, M. C. Chen, S. Chen, X. You, Y. M. He, X. Jiang, L.
You, Z. Wang, C. Schneider, J. J. Renema, S. H€
ofling, C. Y. Lu, and J. W. Pan,
Phys. Rev. Lett. 123, 250503 (2019).
2
P. Kok, C. P. Williams, and J. P. Dowling, Phys. Rev. A 68, 022301 (2003).
3
H. J. Kimble, Nature 453, 1023 (2008).
4
K. Azuma, K. Tamaki, and H.-K. Lo, Nat. Commun. 6, 6787 (2015).
5
J. Yang, C. Nawrath, R. Keil, R. Joos, X. Zhang, B. H€
ofer, Y. Chen, M. Zopf, M.
Jetter, S. L. Portalupi, F. Ding, P. Michler, and O. G. Schmidt, Opt. Express 28,
19457 (2020).
6
M. Sartison, L. Engel, S. Kolatschek, F. Olbrich, C. Nawrath, S. Hepp, M. Jetter,
P. Michler, and S. L. Portalupi, Appl. Phys. Lett. 113, 032103 (2018).
7
Z. H. Xiang, J. Huwer, J. Skiba-Szymanska, R. M. Stevenson, D. J. P. Ellis, I. Farrer,
M.B.Ward,D.A.Ritchie,andA.J.Shields,Commun. Phys. 3, 121 (2020).
8
R. Trotta, P. Atkinson, J. D. Plumhof, E. Zallo, R. O. Rezaev, S. Kumar, S.
Baunack, J. R. Schr€
oter, A. Rastelli, and O. G. Schmidt, Adv. Mater. 24, 2668
(2012).
9
B. H€
ofer, F. Olbrich, J. Kettler, M. Paul, J. H€
oschele, M. Jetter, S. L. Portalupi,
F. Ding, P. Michler, and O. G. Schmidt, AIP Adv. 9, 085112 (2019).
10
V. M. Ustinov, N. A. Maleev, A. E. Zhukov, A. R. Kovsh, A. Y. Egorov, A. V.
Lunev, B. V. Volovik, I. L. Krestnikov, Y. G. Musikhin, N. A. Bert, P. S. Kop’ev,
Z. I. Alferov, N. N. Ledentsov, and D. Bimberg, Appl. Phys. Lett. 74, 2815
(1999).
11
M. Paul, F. Olbrich, J. H€
oschele, S. Schreier, J. Kettler, S. L. Portalupi, M. Jetter,
and P. Michler, Appl. Phys. Lett. 111, 033102 (2017).
12
S. Strauf, N. G. Stoltz, M. T. Rakher, L. a Coldren, P. M. Petroff, and D.
Bouwmeester, Nat. Photonics 1, 704 (2007).
13
N. Srocka, P. Mrowi
nski, J. Große, M. von Helversen, T. Heindel, S. Rodt, and
S. Reitzenstein, Appl. Phys. Lett. 116, 231104 (2020).
14
A. R. Clawson, Mater. Sci. Eng. R: Rep. 31, 1 (2001).
15
N. Srocka, A. Musiał, P.-I. Schneider, P. Mrowi
nski, P. Holewa, S. Burger, D.
Quandt, A. Strittmatter, S. Rodt, S. Reitzenstein, and G. SeRk, AIP Adv. 8,
085205 (2018).
16
L. Rickert, T. Kupko, S. Rodt, S. Reitzenstein, and T. Heindel, Opt. Express 27,
36824 (2019).
17
H. Wang, Y. M. He, T. H. Chung, H. Hu, Y. Yu, S. Chen, X. Ding, M. C. Chen,
J. Qin, X. Yang, R. Z. Liu, Z. C. Duan, J. P. Li, S. Gerhardt, K. Winkler, J. Jurkat,
L. J. Wang, N. Gregersen, Y. H. Huo, Q. Dai, S. Yu, S. H€
ofling, C. Y. Lu, and J.
W. Pan, Nat. Photonics 13, 770 (2019).
18
M. Schmidt, M. V. Helversen, S. Fischbach, A. Kaganskiy, R. Schmidt, A.
Schliwa, T. Heindel, S. Rodt, and S. Reitzenstein, Opt. Mater. Express 10,76
(2020).
19
T. Zander, A. Herklotz, S. Kiravittaya, M. Benyoucef, F. Ding, P. Atkinson, S.
Kumar, J. D. Plumhof, K. D€
orr, A. Rastelli, and O. G. Schmidt, Opt. Express
17, 22452 (2009).
20
F. Ding, R. Singh, J. D. Plumhof, T. Zander, V. K
r
apek, Y. H. Chen, M.
Benyoucef, V. Zwiller, K. D€
orr, G. Bester, A. Rastelli, and O. G. Schmidt, Phys.
Rev. Lett. 104, 067405 (2010).
21
R. Trotta, J. Mart
ın-S
anchez, J. S. Wildmann, G. Piredda, M. Reindl, C.
Schimpf, E. Zallo, S. Stroj, J. Edlinger, and A. Rastelli, Nat. Commun. 7, 10375
(2016).
Applied Physics Letters ARTICLE scitation.org/journal/apl
Appl. Phys. Lett. 117, 224001 (2020); doi: 10.1063/5.0030991 117, 224001-6
Published under license by AIP Publishing
