Epitaxial growth and characterization of dilute nitride based “W”-quantum well heterostructures for laser applications
In this present thesis, the growth by metal organic vapor phase epitaxy (MOVPE) of GaAs-based materials is investigated to shift the emission wavelength to longer wavelengths. This was addressed by incorporating dilute amounts of nitrogen into “W” type-II heterostructures (WQWH). As materials, studi...
|Online Access:||PDF Full Text|
No Tags, Be the first to tag this record!
|Summary:||In this present thesis, the growth by metal organic vapor phase epitaxy (MOVPE) of GaAs-based materials is investigated to shift the emission wavelength to longer wavelengths. This was addressed by incorporating dilute amounts of nitrogen into “W” type-II heterostructures (WQWH). As materials, studies regarding Ga(N,As)/Ga(As,Sb)/Ga(N,As), Ga(N,As)/(Ga,In)As/Ga(N,As), and (Ga,In)(N,As)/Ga(As,Sb)/(Ga,In)(N,As) WQWH were detailed, ultimately leading to the very first dilute nitride based WQWH laser device exhibiting room temperature lasing.
At first, the respective single quantum wells were analyzed with a particular focus on the possible interactions between the epitaxial growth of the entire “W” heterostructures, especially the influence of surface segregated antimony on the nitrogen incorporation in the subsequent dilute nitride containing layer. For that, the consideration was divided into gas phase and surface effects. Investigations of gas phase effects were carried out using an in-situ mass spectrometer integrated into the MOVPE reactor and examining the decomposition temperatures of the respective nitrogen precursors under the application of different amounts of antimony precursor supply. In detail, the nozzle of the mass spectrometer was overgrown with polycrystalline GaAs to comprise catalytic effects. The decomposition as a function of the ambient temperature while supplying varying amounts of antimony precursor was analyzed. For the conventional most established nitrogen precursor 1,1-dimethylhydrazine (UDMHy), an elevated decomposition temperature was observed for increased antimony precursor supply, indicating that UDMHy decomposition is highly dependent on the underlying surface due to significant changes of surface reconstructions occurring because of surface segregation of antimony atoms. In contrast to that, the novel combined nitrogen and arsenic precursor di-tert-butyl-amino-arsane (DTBAA) showed a much weaker sensitivity on the surface and no influence on the antimony supply. Besides that, surface effects were investigated by growing relatively thick Ga(N,As) layers under a constant supply of antimony. As a result, no influence of this antimony on the incorporation efficiency was found when DTBAA was used, while a strong reduction to almost no nitrogen incorporation occurred when UDMHy was utilized. However, the effect of reduced nitrogen incorporation of UDMHy combined with antimony was reported to be perfectly compensable with an application of higher UDMHy supplies. All in all, the effect of antimony could not be exhaustively clarified, but first hints point towards a decomposition effect occurring at the growth surface. Future experiments will be conducted investigating the precise changes of precursor decomposition paths under antimony influence.
These insights were exploited for growing the full Ga(N,As)/Ga(As,Sb)/Ga(N,As) WQWH in which the first Ga(N,As) quantum well can be deposited undisturbed while the growth of the second Ga(N,As) quantum well suffers from the influence of surface segregated antimony from the previous Ga(As,Sb) layer deposition causing the nitrogen incorporation to drop by 60%. An optimization process was developed to overcome this issue, which included a compensation of the reduced nitrogen incorporation by adjusting the UDMHy supply. Besides that, optimization of the internal interfaces between Ga(N,As) quantum wells to the Ga(As,Sb) quantum well was carried out. For that, an antimony predeposition before growing Ga(As,Sb) layers helped to achieve an abrupt chemical composition profile. The second internal interface was prepared by applying TBAs stabilized growth interruptions, smoothing the surface while desorbing excess antimony. However, the duration of such growth interruptions must be carefully chosen since too long durations induce the effect of antimony desorption from the already grown crystal, hence worsening the interface morphology. As optimized durations, 20 s or 30 s were determined for growth at 550°C or 525°C, respectively. Unfortunately, the full antimony surface coverage cannot be abolished within 20 s but only after a growth interruption of 120 s. Nevertheless, the antimony surface coverage is already significantly reduced after 20 s long growth interruptions, and the compensation of the reduced nitrogen incorporation can easily be achieved. Alternatively, DTBAA can be used as nitrogen precursor removing the need for further optimization of the nitrogen incorporation equality because the incorporation efficiency was not influenced by antimony.
The epitaxial growth of Ga(N,As)/(Ga,In)As/Ga(N,As) WQWH raises a similar challenge, namely a reduction of nitrogen incorporation due to segregated indium atoms. An application of TBAs stabilized growth interruption after the (Ga,In)As quantum well growth of up to 120 s resulted in improved optical properties but showed no effect on the nitrogen incorporation into the subsequent Ga(N,As) quantum well due to the lower volatility of indium atoms, compared to antimony atoms.
So far, the nitrogen incorporation issue was not solved for (Ga,In)(N,As)/Ga(As,Sb)/(Ga,In)(N,As) WQWH if UDMHy is used due to even more severe effects if indium and antimony atoms affect the nitrogen incorporation simultaneously. Thus, DTBAA was successfully employed for growing these structures resulting in symmetric nitrogen incorporation.
For all the above-mentioned material systems, laser devices were fabricated according to the insights gained by all epitaxial growth experiments. All materials have in common that they feature diluted amounts of nitrogen as an essential element for reducing the conduction band edge. Experimental determination of the optical efficiency of multiple laser devices reveals a strong dependence of the optical efficiency on the nitrogen content which was attributed to the formation of nitrogen clusters and carbon incorporation that act as non-radiative recombination centers. The investigation of Ga(N,As)/(Ga,In)As/Ga(N,As) based laser structures exhibited the least suitability for laser applications due to too low band offsets while requiring relatively large nitrogen contents. Better suitability was found for Ga(N,As)/Ga(As,Sb)/Ga(N,As) WQWH because only 2% nitrogen is necessary to reach emission at 1.3 µm. Even less nitrogen (less than 1%) is needed if (Ga,In)(N,As) replaces Ga(N,As) as electron quantum well material. With this, even 1.5 µm are reachable with reasonable nitrogen contents.
Valuable insights were gained when examining the device properties of Ga(N,As)/Ga(As,Sb)/Ga(N,As) WQWH based devices. For relatively low nitrogen contents, a significant saturation of the optical output power was observed, while for higher nitrogen contents above 3%, the optical output power was found to be weak. These effects were attributed to two different major loss channels that come into play at different nitrogen contents. Under low nitrogen conditions non-radiative recombination paths increase the threshold current density to comparatively high values. Before sufficient photon densities could be reached, electrons occupy energetically higher states facilitating electron leakage to conduction band barrier states. In the case of large nitrogen contents, the non-radiative recombination processes become the major limiting factor. Carrier leakage was additionally proven by realization of Ga(As,Sb)/Ga(N,As)/Ga(As,Sb) “M” type-II structures in which one instead of two Ga(N,As) quantum wells provide confined electron states.
These issues can be partly overcome by replacing Ga(N,As) by (Ga,In)(N,As). It was found that less nitrogen content is required to achieve a sufficiently high conduction band barrier due to the band edge reduction introduced already by indium incorporation alone. This reduced requirement of nitrogen incorporation enables a reduced non-radiative recombination path. Finally, these structures led to the outstanding result of first room temperature lasing for this material class. A device containing 1.1% nitrogen showed lasing above a current density of 9.5 kA/cm² with a differential efficiency of 1% and an optical efficiency of 4.6 mW/A. A pump current limited maximum pulse output power of 27.5 mW was reached. Spectral characterization revealed a lasing wavelength of 1.28 µm, which could be shifted to longer wavelengths by further optimizing this structure.
When considering Ga(N,As)/(Ga,In)As/Ga(N,As) WQWH devices, the barriers especially in the valence band are too low and thus electronic states spanning over the full WQWH are forming, as indicated by the weak blueshift at high excitation densities.
First attempts to find suitable annealing conditions were made by applying in-situ reactor annealing or additional rapid thermal annealing with different temperatures and ambients. However, all investigated WQWHs seem to deteriorate under all examined conditions in structural quality due to the applied annealing steps. This was attributed to an over-annealing since all structures endured already a long annealing step in the form of cladding growth.
All in all, the development of a proof of principle laser device was successful, with the first ever demonstrated room temperature laser emission of dilute nitride-based type-II WQWH.|
|Physical Description:||212 Pages|