The influence of the growth conditions on the optical properties of InGaN QWs
Homogeneous InGaN QWs with a low concentration of dislocations and point defects have to be grown for the achievement of long-wavelength lasing. As we demonstrated, the increase of the In content in InGaN layers can be realized by i) reduction of the growth temperature or ii) increase of the N flux. The first approach is effective in increasing the In content, but at low growth temperatures the optical quality of the InGaN is dramatically deteriorated, as illustrated by a strong drop in LED efficiency (Siekacz et al., 2009). Here we examined whether an increase of N flux provides a better alternative to decreasing growth temperature for the production of InGaN structures with an optical quality sufficient to sustain laser action at longer wavelengths. For LD development we need to minimize localization potentials and non-radiative recombination processes in QWs. Time-resolved photoluminescence (TRPL) spectroscopy is an ideal tool for monitoring both of these parasitic effects.
The TRPL experiments provide information on the carrier dynamics during and after their pulse laser excitation. The observed PL decay time, tpl, depends on the radiative (tr) and non-radiative (tn) recombination times (t—L = t—1 + t—1). For a perfect QW (i.e., a sample without point defects, etc.) the tpL is determined by the characteristics of the QW confining potential and only weakly changes with temperature (for this situation, tpl = tr). For practically attainable QWs the factor t- cannot be neglected because of the presence of non-radiative recombination centers, such as point defects or dislocations. The tn is known to decrease with increasing temperature, and as a result, at room temperature, tpl is strongly reduced for poor-quality QWs. This means that the role of non-radiative recombination centers in InGaN QWs can be compared effectively by measuring the tpl for QWs at room temperature. In order to perform such a comparison, TRPL in InGaN QWs deposited using different growth parameters was measured at nominally identical conditions. The samples were excited by a 350-nm laser line with a 76 MHz repetition rate and a pulse duration of 150 fs from a mode-locked Ti:Sapphire laser. In order to observe effects associated with point defects, measurements were performed at low excitation conditions (the average excitation power was a few mW). The PL signal was dispersed by a 0.3-m focal length monochromator and detected by a Si streak camera.
In Fig. 2.20 we plot as an example the decay of PL intensity at 425 nm for one of our samples, together with the laser pulse temporal profile. The tpl has been obtained by fitting the experimental data with a one-exponential decay of PL signal. The tpl determined in such a way are plotted in Fig. 2.21, together with the PL spectrum taken just after the excitation.
As we pointed out, by decreasing the growth temperature we can shift the PL from 420 nm (see Fig. 2.21(a), sample A) to 520 nm (Fig. 2.21(b), sample B) for the constant nitrogen flux of 4.1 nm/min. For sample A we observe no spectral dispersion of tpl = 0.8 ns. Such a behaviour of tpl is typical of homogeneous QWs, i.e., QWs where the carrier localization induced by indium content fluctuations can be neglected. For sample B, the PL decay time is qualitatively different, and tpl increases from 3 ns at A = 485 nm to 13 ns at A = 570 nm. Such a spectral dispersion of tpl implies carrier localizations in this QW, and can be attributed to indium content fluctuations. In general, it is expected that
Fig. 2.20. The photoluminescence transient from an MQW structure at 425 nm.
such a localization can effectively reduce the influence of non-radiative recombination channels, but on the other hand it is much harder to obtain the population inversion necessary for lasing in such a system. Therefore, the spectral dispersion of tpl can be used as a fingerprint of carrier localization, giving valuable feedback for optimizing InGaN QWs growth procedure. As we have already shown, when we increased the N flux we were able to grow InGaN QWs with the same In content (around 26%) at higher growth temperature, as indicated by PL (see Fig. 2.21(c), sample C). For this sample the TRPL measurements show no spectral dispersion of tpl. This indicates that samples A and C are comparable from the viewpoint of carrier localization. Note that sample A has considerably lower In content in QW: 12%. In addition to the homogeneity improvement, the increased nitrogen flux probably contributes to the reduction in the number of non-radiative recombination centers in the layers by reducing the number of
Fig. 2.21. The photoluminescence intensity and its decay time for blue (a) and green (b) and (c) emission, respectively.
nitrogen vacancies. Such a possibility is supported by the observed increase in the absolute value of tpl for the sample shown in Fig. 2.21(c) in comparison to that in Fig. 2.21(a). Since at room temperature tpl is expected to be dominated by a non-radiative recombination component tn , this increase is not likely to be caused by the larger built-in electric field for the high-In-content QWs of sample C and the related decrease in the oscillator strength for this transition.