Biomedical Engineering Reference
In-Depth Information
temperature reaches 75 K which contradicts the saturation interpretation. Closer
inspection of the PL spectra shows that the II increases not by increasing peak
intensity but by broadening. The saturation conclusion thus remains valid.
The observation of intensity transfer and the implied underlying mechanism
via tunnel coupling in our lateral QDMs are at first surprising, considering the
significant dot-to-dot separation in the order of 50 nm. Tunnel coupling decreases
exponentially with distance and Szafran and Peeters predicted that coupling is
negligible for a dot-to-dot distance of 20 nm [
33
]. This, however, is strictly true for
isolated QDs with a thin WL acting as the main coupling path. Our lateral QDMs,
especially the sQDs, form on the nanomound template which in itself can act as a
low barrier region connecting the constituent QDs in a QDM. A similar “basin” has
recently been demonstrated to be acting as a coupling channel for double InGaAs
QDs spaced as far as 40 nm apart [
34
].
3.4.3
Bimodal Optical Characteristics
Depending on growth procedures, lateral QDMs can take shape in many different
geometries, with varying degrees of uniformity and size distribution. The latter can
be categorized into a mono-, bi-, or multi-modal size distribution. This section
describes a novel bimodal
optical
characteristics which results from the bimodal
size
distribution of QDMs. The converse is not necessarily true. The unique bimodal
optical characteristics stems from specific spatial arrangement of constituent QDs
resulting from the partial-cap and regrowth process.
To demonstrate the unique bimodal optical properties, temperature-dependent
GS PL spectra of the 1.8/25/1.2 and 1.8/25/1.5 QDMs are shown in Figs.
3.6
aand
3.7
a, respectively [
35
]. The two QDM ensembles mainly differ in the nominal size
of sQDs, and consequently the degree of cQD-sQD coupling. The excitation power
density at 2 W/cm
2
is sufficiently low to avoid filling up the GS and complicating
peak analyses with ES. The spectra can be fitted with double Gaussian functions;
examples are shown as the dashed lines in the 20-K spectra of both figures. The
fittings allow temperature variations of peak position, intensity, FWHM, and hence
II to be accurately determined.
The cQD and sQD peak energy variations with temperature of the 1.8/25/1.2
and 1.8/25/1.5 QDMs are shown in Figs.
3.6
band
3.7
b, respectively. In both cases,
the cQDs and sQDs exhibit fundamentally different behaviors: the cQDs-related
peak energies exhibit a slow red-shift with increasing temperature throughout the
experimental temperatures, while the sQDs-related peak energies exhibit a slow red-
shift only up to a certain temperature (
75 K for 1.8/25/1.2 QDMs and 100 K
for 1.8/25/1.5 QDMs) before they rapidly decrease at a rate of 1 meV/K. Dashed
(dotted) lines in the figures approximate the slow (fast) red-shift of peak energies.
The cQDs's energy slow red-shift can be readily explained in terms of bandgap
variations with temperature. The dashed lines in the figures are obtained from
Varshni's equation [
36
] using bulk InAs bandgap parameters, shifted up the energy
∼
