We present the recent progress in the intermixing of InGaAs/GaAs quantum dot (QD) material. Quantum dot intermixing (QDI) allows the tuning of the energy bandgap in selected areas of the wafer or optoelectronic device, thus modifying its emission or absorption properties, in much the same way as in quantum well (QW) material. QDI has recently received increasing interest, as it combines bandgap engineering with the predicted advantages that quantum dots offer, such as low temperature-sensitivity of threshold current, high modulation frequency and low chirp.

We have applied the dielectric-cap-based techniques that were originally developed for QW structures, to the intermixing of InGaAs/GaAs/AlGaAs QD material with an emission wavelength of 1280 nm. Intermixing was achieved by sputtering a QDI-enhancing cap in some areas, and a QDI-suppressing cap in other areas, followed by a high-temperature anneal cycle. Extremely large bandgap blue-shifts of up to 280 nm have been obtained with an anneal temperature of 800 °C. The shifts were inferred from the photoluminescence (PL) spectra measured at 77 K under red-laser excitation.

To be of use in many applications, QDI must be able to provide a multiplicity of bandgaps on a single substrate. Multiple bandgaps can be created by varying the thickness of the QDI-enhancing cap, repeating the anneal cycle several times, or varying the coverage density of QDI-enhancing features over that of QDI-suppressing ones. The latter approach, termed selective intermixing in selected areas (SISA), involves the deposition of QDI-enhancing patterns of various area fill factors, which, upon annealing, will cause different degrees of intermixing in the underlying regions.

To demonstrate the SISA process in the QD material, we defined patterns containing lines and squares of various sizes (3 - 100 μm) and area fill factors (5% - 95%). The wafer was then annealed at 725 °C for 1 minute. As expected, the observed bandgap shifts were commensurate with the fill factor, with a 5% coverage providing a minimum shift (0 - 10 nm) and ∼40% a maximum one (∼ 120 nm). At fill factors above 40%, the shifts appeared to saturate and even decrease slightly. The effect of the feature size and shape was not very pronounced, with smaller features generating somewhat larger shifts. This may be due to the fabrication-related size bias that will have the strongest effect on the fill factor of smaller features. The PL spectra measured from patterns of large-size features (20 μm or more) often had a lopsided shape and broader peak width, which may be attributed to the limited spatial resolution of the measurement probe.