Photocarrier transport dynamics in InAs/GaAs quantum dot superlattice solar cells using time-of-flight spectroscopy

We studied time-resolved photocarrier transport through InAs/GaAs quantum dot superlattice (QDSL) solar cells (SCs) using time-of-flight spectroscopy with an optical probe QD structure beneath the QDSL. Carriers optically pumped in the top $p$-GaAs layer were transported through the intrinsic layer, including the QDSLs, before arriving at the probe QDs. The photoexcited carrier density significantly influenced the time-resolved photoluminescence (PL) of the QDSLs and probe QDs. The time-resolved PL profile of the probe QDs indicated that excitation densities in excess of $25\mathrm{nJ}/\mathrm{c}{\mathrm{m}}^2$ drastically decreased the rise time, suggesting rapid carrier transport through the QDSLs. This was also confirmed by QDSL carrier transport dynamics, for which the PL intensity of the excited states decayed rapidly above this excitation power density, $25\mathrm{nJ}/\mathrm{c}{\mathrm{m}}^2$, while the ground state remained constant. These results demonstrate that filling the ground states of QDSLs and starting to populate the excited state miniband accelerates carrier transport in QDSL SCs. Furthermore, according to two-step photon absorption measurements taken with a 1.3-μm infrared laser light source, electrons play a key role in the generation of extra photocurrent by sub-band-gap photon irradiation.

Figure 1

Schematic diagram of the sample structure used in our time-of-flight measurement, GaAs QDSL-SC with probe QDs. Detailed structure in the undoped GaAs is shown on the right-hand side. The nine-layer-stacked InAs/GaAs QDs and InAs QDs capped by a strain reducing InGaAs layer were inserted into the intrinsic region.

Figure 2

Carrier transport dynamics in QDSL SCs that have a probe structure. The main three steps of our time-of-flight measurement are (i) carriers are selectively excited only in the top $p$-GaAs layer according to both the thickness of the $p$-GaAs layer (150 nm) and the penetration depth of the excitation wavelength (approximately 30 nm). (ii) After photoexcited carriers reach the intrinsic region by diffusion, electrons are driven by the internal electric field and then arrive at the QDSLs, where they are repeatedly trapped and detrapped. Holes also penetrate towards the $n$ side by diffusion. A portion of the carriers recombine in the QDSLs. (iii) Other carriers can escape from the QDSL region and reach the probe QDs, where they recombine and emit photons.

Figure 3

PL spectrum of a QDSL SC with probe QDs at 17 K. QDSLs and probe QDs have peak emission wavelengths of 1050 and 1190 nm, respectively.

Figure 4

Typical streak-camera images of (a) QDSLs and (b) probe QDs with excitation power density of $40\mathrm{nJ}/\mathrm{c}{\mathrm{m}}^2$ at 4 K.

Figure 5

Time-resolved PL profiles of (a) QDSLs and (b) probe QDs at various excitation power densities at 4 K. The zero delay time corresponds to the threshold of the increase in the QDSLs’ PL intensity. The time-resolved PL profiles were recorded to within ${\lambda }_0\pm 3\mathrm{nm}$ of the time-resolved PL spectrum, where ${\lambda }_0$ is the peak PL wavelength of the signal, and the time scale for magnification ranges from 0 to 2.0 ns. The arrows in both figures at each excitation power density indicate the time at which PL begins to decay. Solid lines in Fig. 5 indicate simulated results based on a model taking into account diffusion current and drift current for electrons and holes influenced by multiple trap and detrap processes in QDSLs.

Figure 6

Excitation power density dependences of (a) delay time indicating the maximum of a hump developed in the decay curves, (b) PL decay lifetime, and (c) PL peak intensity for probe QDs at 4 K. The dotted lines represent the critical point. The scales of both axes in (c) are logarithmic. The low- and high-excitation regions have linear and superlinear slopes, respectively. The dashed line is a linear plot for a guide to the eye. The inset figure shows the linear relationship of the excitation power density dependence.

Figure 7

(a) The excitation-power-dependent PL spectrum of QDSLs. The GS signal of the QDSLs is indicated by the arrow labeled GS (QDSLs) at 1030 nm. The signal appearing at 1060 nm is attributed to the GS of the base QDs, which is labeled GS (base QDs). Carrier population increases with increasing excitation power, and then the ES signal of the QDSLs begins to dominate, as indicated by the arrow labeled “ES (QDSLs)” at 980 nm. (b) Time-resolved PL spectra of QDSLs, measured at excitation power densities of $10\mathrm{nJ}/\mathrm{c}{\mathrm{m}}^2$ (red line) and $40\mathrm{nJ}/\mathrm{c}{\mathrm{m}}^2$ (purple line) at 4 K. The PL spectra were obtained by time-integrating the streak-camera image by 0.5 ns, from 0 to 2.0 ns.

Figure 8

PL decay times that were analyzed for (a) GS (QDSLs) at 1030 nm, (b) ES (QDSLs) at 980 nm, and (c) GS (base QDs) at 1060 nm.

Figure 9

(a) Time-resolved PL profiles detected at the GS (base QDs) transitions at 10 and $40\mathrm{nJ}/\mathrm{c}{\mathrm{m}}^2$. The time-resolved PL profile was integrated between 1060 and 1070 nm. (b) The difference between the two profiles in Fig. 9.

Figure 10

(a) PL spectra of QDSLs at various IR powers. Time-resolved PL of QDSLs (b) without and (c) with IR irradiation, 5 mW, focusing on the ES (QDSLs) at 980 nm and the GS (base QDs) at 1060 nm. Solid lines indicate fitted single-exponential decay curves.