Effect of InAs quantum dots capped with GaAs on atomic-scale ordering in Ga0.5In0.5P

Abstract

The CuPt ordering of the group III elements in GaxIn1-xP (x ≃ 0.5) has been observed to vary during growth by metalorganic vapor-phase epitaxy of InAs quantum dots capped with GaAs in a GaInP matrix. While ordering is not affected by the insertion of a GaAs layer, the growth of InAs quantum dots capped with GaAs results in ordered, partially ordered, or fully disordered GaInP. We show that the degree of ordering depends on the deposition time of the InAs quantum dots and on the thickness of the GaAs capping layer. Our results indicate that disordered GaInP occurs in the presence of excess indium at the growth surface, which results from the growth of strained InAs quantum dots. Ordering resumes when the excess indium is consumed. Cathodoluminescence, scanning transmission electron microscopy, and atomic force microscopy are used in this study to correlate the electronic properties and the microstructure of the thin films.

Panning ideal In(Ga)As(P)/InGaP quantum dot structures for intermediate band solar cells

“…The observed PL peak for the QDs capped with a 3 nm thick GaAs layer at around 1.32 eV in figure 8(a) is attributed to a QD distribution limited by the capping layer. In figure 8(b) the PL spectra for the QDs capped by a 4 nm thick GaAs layer presents two peaks attributed to emission from two QDs' families: one at 1.30 eV, due to small QDs fully buried, and the other at 1.17 eV, due to larger QDs, which include QDs limited by the capping layer and others that surpass the capping layer due to a partial In-flush, as previously described [12]. In fact, a small contribution of the emission of such large QDs is also observed in figure 8(a) for the sample with a 3 nm capping layer, but it is negligible.…”

“…First, the growth of disordered lattice matched InGaP as a barrier material shall be addressed. In a previous publication [12] it has been shown, by HAADF imaging and diffraction patterns in a transmission electron microscope, that lattice matched InGaP on GaAs grows in its CuPt ordered phase under the MOVPE standard growth conditions used in this investigation. The ordered phase however is highly affected by the presence of QDs.…”

“…If the QDs' deposition time increases by 25% (larger QDs), the InGaP spacer layer is then fully ordered. These results are thoroughly explained, based on the availability of In atoms in the gas phase during growth [12]. Figure 2 shows the room temperature PL spectra of these samples, where one clearly observes the emission peak at 1.83 eV associated with the ordered InGaP layer for the structure with 9 nm capping layer and QDs deposited for 3 s and another one at 1.91 eV, corresponding to the disordered phase for the other two samples, QDs deposited for 2.4 s and capping layer of 6 nm and 9 nm, in full agreement with the previously published data [12].…”

“…These results are thoroughly explained, based on the availability of In atoms in the gas phase during growth [12]. Figure 2 shows the room temperature PL spectra of these samples, where one clearly observes the emission peak at 1.83 eV associated with the ordered InGaP layer for the structure with 9 nm capping layer and QDs deposited for 3 s and another one at 1.91 eV, corresponding to the disordered phase for the other two samples, QDs deposited for 2.4 s and capping layer of 6 nm and 9 nm, in full agreement with the previously published data [12]. These results indicate that for QD IBSCs, which require a matrix material with a bandgap energy of around 1.95 eV, the GaAs capping layer should not be thicker than 6 nm and the QDs should be nucleated during a time not longer than 2.4 s, corresponding to 1.32 MLs.…”

“…Therefore, ideally InGaP should be disordered to be adequate for IBSCs. In a previous publication it has been shown that, under standard metalorganic vapor phase epitaxy (MOVPE) growth conditions, InGaP is ordered, but if the nucleated InAs QDs are small enough and thin capping layers (<∼6 nm) are deposited, the subsequent InGaP layers become disordered, leading to a bandgap close to the ideal [12].…”