Optical characterization of InGaAsN / GaAs quantum wells: Effects of annealing and determination of the band offsets
In the last decade great attention has been given to the characteristics of dilute nitrides. Both their peculiar physical properties and their wide range of possible applications have attracted the interest of many experimental and theoretical groups. In this thesis work some open questions about th...
|Online Access:||PDF Full Text|
No Tags, Be the first to tag this record!
|Summary:||In the last decade great attention has been given to the characteristics of dilute nitrides. Both their peculiar physical properties and their wide range of possible applications have attracted the interest of many experimental and theoretical groups. In this thesis work some open questions about the fundamental properties of dilute nitrides have been answered. Two important topics have been investigated: the correlation between the optical and morphological properties of InGaAsN / GaAs single quantum well samples (SQWs) and a quantitative, model-independent determination of the band offsets for the same types of structures.
In chapter 3, a combined study of photoluminescence (PL) measurements and transmission electron microscopy (TEM) analysis has allowed to find a direct correlation between the degree of carrier localization in the sample and the homogeneity of the material. In particular, the degree of localization increases with increasing the inhomogeneity of the QW layer. On the basis of that, it has been found that the growth temperature (Tg) and the indium content strongly influence the morphology of the InGaAsN QW samples. With increasing Tg or with increasing [In], the inhomogeneity of the sample increases.
The growth temperature affects also the optical properties of InGaAsN SQWs. By raising Tg, the PL intensity degrades and the peak emission energy red shifts. On the other hand, the indium content does not remarkably influence the PL efficiency of the QW. The only exception is for very high indium contents ([In] > 34%). In this case, dislocations due to strain relaxation and / or other types of non-radiative recombination centres are created causing a drastic decrease of the PL intensity.
After annealing both the morphological and the optical properties are modified. Most notably, by employing samples grown in the range of temperatures between 360 °C and 480 °C annealed in different environments, two important conclusions have been found. First of all, morphology and PL efficiency are not always correlated and secondly, the PL efficiency of a QW directly depends on the density of non-radiative centres. Annealing samples in different atmospheres is a novelty in the literature and it has been the key-point to reach these findings.
The first conclusion has been obtained by performing photoluminescence measurements on samples annealed in hydrogen and argon environment, and comparing the results with those of as-grown samples. It has been shown that while the PL intensity of H2-annealed samples is maximum for low values of Tg (400 °C) and minimum for high Tg (450 °C), the PL intensity of the Ar-annealed samples is maximum for high values of Tg (450 °C) and minimum for low Tg (400 °C). In contrast, the degree of localization and the TEM images have shown the same Tg-behaviour, independently of the annealing environment.
The second conclusion has been reached by performing time resolved photoluminescence measurements on the same series of samples. It has been shown that whilst the radiative decay time varies with Tg in the same manner for the two annealing atmospheres, i.e. it increases with increasing Tg, the non-radiative decay time varies with the growth temperature in a different way for different annealing environments. In particular, the non-radiative decay time decreases with increasing Tg for H2-annealed samples and increases with increasing Tg for Ar-annealed samples. This behaviour correlates in both cases with the dependence of the PL intensity on the growth temperature.
In addition to that, by performing power dependent PL measurements, it has been verified that changes of degree of localization after annealing are only due to morphological modification of the sample.
By comparing the results obtained performing PL measurements on GaAsN / GaAs, InGaAs / GaAs, and InGaAsN / GaAs SQW samples, it has been shown that at least two different type of defects are created during the growth of InGaAsN SQWs. One type of defect is related to the presence of nitrogen. The density of these defects increases with Tg and decreases by annealing. Defects of another type are related to the simultaneous presence of indium and nitrogen. They are created at low Tg and tend to agglomerate under annealing. These two types of defects have been employed in a simple model in order to justify the main results obtained in this chapter.
In chapter 4, a much debated topic has been analysed: the evolution of the band offsets of InGaAsN / GaAs structures with varying QW parameters. The chapter has been initially focussed on the refinement of the information which can be obtained employing an experimental method developed at Infineon Technologies based on surface photovoltage (SPV) measurements. With this method it is possible to identify optical transitions involving bound states and extended states in a QW sample. In particular, in addition to the bound-to-bound transitions, also the indirect transition from the extended state of the valence band to the first confined state of the conduction band can be identified. This allows the easy determination of the practical band offsets of the QW. These quantities represent the energy values of the conduction (valence) band offset of the heterostructures without the value of the first quantized state of the electrons (holes). For the design of a device, the practical band offsets are fundamental quantities because they quantify the real confinement of the carriers in the well.
SPV measurements have been performed on several dedicated series of samples. The results have been compared with those obtained employing other optical techniques and performing theoretical simulations. It has been shown that by using this method, it is possible to gather comprehensive information about a single quantum well which otherwise could be obtained only by combining different experimental techniques and theoretical calculations. With this method transitions related to the ground states of the QW involving both the heavy and light holes states can be detected. Also, the excited states can be identified. As a main condition, it has been shown that only bound-to-bound transitions having the same parity can generate a step in the spectra.
This method has been employed to investigate the band states of dilute nitrides SQWs. In particular, the effect of varying nitrogen and indium content on the practical band offsets of InxGa1-xAs1-yNy /GaAs SQW samples has been analysed. As a main result, it has been found that with increasing nitrogen content, the conduction band offset strongly increases (with a rate of about 100 meV / [N]), while the valence band offset is almost unchanged. Moreover, with increasing indium concentration both the conduction and the valence band offsets are modified. In particular, the conduction band offset varies with indium content as in the case of N-free samples. These results represent the first quantitative analysis which directly, i.e. independently of any model, determines the band offsets in dilute nitrides quantum wells. More importantly, it allows to analyse the effect of nitrogen and indium on the conduction and valence band states separately.
The practical band offsets are highly important parameters in the design of many devices. Therefore, in the end of this thesis, it has been shown that the SPV method can be employed to determine the practical band offsets of real device structures. In particular, the practical conduction and valence band offsets of lasers emitting at 1.3 µm and 1.5 µm have been determined from the SPV spectra.|