An InAsBi photodiode has been grown, fabricated and characterized to evaluate its performance in the MWIR
region of the spectrum. Spectral response from the diode has been obtained up to a diode temperature of 225 K.
At this temperature the diode has a cut off wavelength of 3.95 μm, compared to 3.41 μm in a reference InAs
diode, indicating that Bismuth has been successfully incorporated to reduce the band gap of InAs by 75 meV.
Similar band gap reduction was deduced from the cut off wavelength comparison at 77 K. From the dark current
data, R0A values of 590 MΩcm2 and 70 MΩcm2 at temperatures of 77 and 290 K respectively, were obtained in
our InAsBi photodiode.
The performance of lasers with self assembled quantum dot active regions is significantly affected
by the presence of the two dimensional wetting layer and the other states necessary for carrier
injection due to the manner in which carriers are distributed amongst the various states. In this work
we describe three approaches to overcome the low value of maximum saturated gain, which has
been observed by many groups worldwide, and explain the approaches in terms of the impact on the distribution of carriers within the available states. We present results of direct measurements of the modal gain and measurements that indicate the form of the carrier distribution within the samples to justify our argument. The structures examined include the use of a high growth temperature to smooth the matrix layer, the use of p-type modulation doping and the use of InAlAs capping layers and all have been grown by solid source molecular beam epitaxy. We demonstrate CW operation at 1.3&mgr;m for 1mm long devices with uncoated facets and very low threshold current density (< 40Acm-2) in longer devices. We also demonstrate that the negative T0 (reducing threshold current density with increasing temperature) obtained around room temperature in our p-doped devices is due to the temperature dependence of the gain.
We have calculated recombination rates of an inhomogeneous ensemble of 106 dots by summing localized
recombination rates at individual dots, with occupation of dot states in the inhomogeneous distribution specified by
Fermi Dirac statistics. We assign the same single dot recombination lifetime (1 ns) to all recombination processes to
reveal the effect of localization on the overall rates. For the simplest system of the ground states alone deep state,
radiative and Auger recombination processes depend in a similar manner upon the population of electrons in the ground
states Consequently the light-current curves for the ground state are approximately linear and are not sensitive to the
dominant non-radiative process. When excited states are included Auger recombination becomes dominant at high
ensemble populations due to the higher degeneracy assigned to the excited states. While the form of the light-current
curves of the total dot system do depend upon the dominant recombination process, an analysis based on power law
relations with respect to the ensemble electron population are not appropriate.
Self-assembled In(Ga)As quantum dot (QD) lasers incorporating p-type modulation doping have generated much interest recently due to reports of a temperature insensitive threshold current and increased modulation bandwidth. The mechanism by which p-type doping improves the performance of QD lasers is thought to be similar to that envisaged for quantum well lasers, where increased gain is expected for a given quasi-Fermi level separation due to a shift in both quasi-Fermi levels towards the valence states. However, the benefits may be much more pronounced in quantum dot structures since the population of the smaller number of dot states can be dramatically affected using relatively low doping levels, which may incur less penalty with regard to increased non-radiative recombination and internal optical mode loss. We present results of direct measurements of the modal gain measured as a function of the quasi-Fermi level separation for samples with different degrees of doping, which demonstrate unambiguously the increased gain that can be obtained at a fixed quasi-Fermi level separation. In addition, we have measured the internal optical mode loss and radiative and non-radiative recombination currents for samples containing 0, 15 and 50 dopant atoms per dot and show that, although the internal optical mode loss is similar for all three samples, the non-radiative recombination current increases for samples containing p-doping. We show that our experimental results are consistent with a simple computer simulation of the operation of our structures.
Quantum dots have demonstrated improved performance relative to quantum wells in lasers and amplifiers for structures where the total optical loss, and hence the gain required from the dot active material, has been kept low. In many applications higher gain and/or high differential gain are required and high gain structures must be routinely produced if quantum dots are to replace quantum wells in more than a few niche applications. The obvious approach is to use multiple layers of quantum dots in the active region of the laser or amplifier. However, stacking multiple quantum dot layers modifies the growth of subsequent layers and in the extreme case leads to defect formation.
In this work we study an approach where the negative effects caused by the introduction of multiple layers of quantum dots are minimised using a high growth temperature spacer layer (HGTSL) to planarize the surface before deposition of the subsequent layer of dots. We show that this has a dramatic affect on the threshold current of our 1.3μm emitting lasers and by use of detailed characterisation show that this is due to 4 physical effects. Samples containing the HGTSL exhibit less inhomogenous broadening, have an increased dot density, a lower internal optical mode loss and contain fewer defects than samples containing a conventional spacer layer. Our results demonstrate the importance of going beyond an approach based on defect reduction alone.