Biomedical Engineering Reference
In-Depth Information
due to cQDs. The labels s, p, and d denote the ground, first, and second excited
states, respectively. The nature of excited states has been confirmed by excitation-
dependent experiments. For 1.8/15/1.2 QDMs in Fig.
3.5
a, the spacing between the
ground and first excited states, or the s-p spacing, is 45 meV. For 1.8/25/1.5 QDMs
in Fig.
3.5
b, the s-p and p-d spacings are approximately 37 meV. These values
are within the expected range for QDs of similar size [
28
]. The equi-distance of
the s-p and p-d spacings results from the harmonic oscillator-type potentials in the
growth plane and is also expected [
29
]. The overall spectra can be fitted with multi-
Gaussian functions which are shown as examples in the 20-K spectra in both figures.
The fits allow the FWHM of all the spectral peaks to be accurately determined
which together with the associated peak intensity yield the integrated intensity
(II) vs. temperature plots of the 1.8/15/1.2 and 1.8/25/1.5 QDMs in Fig.
3.5
c, d,
respectively. Non-monotonous variations of integrated intensity with temperature
are evident in both cases.
For ensembles of isolated QDs, a simple, monotonous decrease of II with increas-
ing temperature is expected due to the presence of NRR channels [
30
] or reduced
PL yields [
31
]. For laterally-coupled QDMs, the temperature dependency of the PL
spectra is complicated by tunnel coupling which has many possible routes and is also
temperature dependent. The overarching monotonous II reduction with increasing
temperature for 1.8/15/1.2 QDMs in Fig.
3.5
c results from carriers gaining sufficient
energy, escaping into the WL and/or GaAs matrix before recombining with NRR
centers/channels as is typical for InAs/GaAs QD systems [
32
]. Close examination
of the constituent IIs shows that the monotonous reduction of the total II results
from a rapid, monotonous decrease of sQDs II and a non-monotonous decrease of
cQDs II. In fact, a slight increase of the cQDs GS is registered at around 75 K. This
increase can be interpreted as typical carrier redistribution between QD ensembles
of different nominal size, or bimodal QDs where carriers in sQDs may escape,
diffuse towards the cQD, and be captured. Or it can be interpreted as resulting from
direct tunnel coupling. The underlying mechanism(s) that governs the temperature
dependency of the II for this particular QDMs is not clear.
The 1.8/25/1.5 QDMs exhibit a qualitatively different II temperature dependency
as shown in Fig.
3.5
d. This is due to the bigger cQDs and sQDs as compared to the
1.8/15/1.2 QDMs above. The total II in this case is virtually constant from 20 to
75 K, above which it decreases monotonously. The overall change results from the
complex behaviors of the s and p peaks of the cQDs while the sQDs exhibit a simple,
monotonous decrease in intensity. As the temperature increases from 20 to 50 K,
the II of the p peak significantly increases. The intensity gain cannot originate from
carrier redistribution as the temperature is still too low for excitons to appreciably
escape the potential barriers. The most likely mechanism is thus tunnel coupling,
from sQDs to cQDs. AFM images show that the cQDs and sQDs in 1.8/25/1.5
QDMs almost merge which supports the conclusion. The II of the p peaks keeps
on increasing with temperature up to 75 K before carrier loss from sQDs to NRR
channels begin to aversely impact the availability of carriers tunneling to cQDs, and
thus the intensities of the s, p, and d peaks. The II of the s peak is approximately
constant between 20 and 50 K, indicating saturation. It, however, increases as the
