Date of Award

5-1-2024

Language

English

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

College/School/Department

Department of Nanoscale Science and Engineering

Dissertation/Thesis Chair

F. Shadi Shahedipour-Sandvik

Committee Members

Michael A. Reshchikov, Denis O. Demchenko, Mengbing Huang, Vincent LaBella

Keywords

AlGaN, Beryllium, Doping, GaN, III-Nitride, MOCVD

Subject Categories

Nanoscience and Nanotechnology

Abstract

The III-nitride material system is well-suited to a broad range of applications in light emitting diodes and laser diodes, photodetectors, power management, and telecommunications. Much of III-nitrides’ utility derives from their direct, wide, and tunable bandgap, but it is their ability to be doped n- and p-type that makes them especially useful and attractive. However, despite the decades of academic interest and commercial success, p-type doping in III-nitrides is infamously difficult and inefficient. Magnesium, the only currently viable p-type dopant in III-nitrides, is not a shallow acceptor, and as a result, the activation efficiency of Mg in GaN is only ~1%. Because of this, III-nitrides have not been utilized to their theoretical limits. It is therefore necessary to employ a novel doping approach to realize the next generation of III-nitride technologies. Beryllium, a common p-type dopant in other III-V materials such as GaAs, is a natural choice as an alternative to Mg. Furthermore, early density functional theory (DFT) calculations predicted a shallower acceptor level in GaN than Mg. Photoluminescence (PL) measurements of GaN:Be indeed indicate an acceptor ~113 meV above the valence band maximum. This thesis represents the culmination of several years of effort toward the first systematic study and demonstration of (Al)GaN:Be grown by metal-organic chemical vapor deposition (MOCVD), as well as the development of a deeper understanding of in situ doping of the Be acceptor. Under optimal MOCVD growth conditions and using a beryllium acetylacetonate (Be(acac)2) metal-organic precursor for Be, high quality GaN:Be was grown using a Veeco D75 vertical cold wall MOCVD. Atomic force microscopy measurements reveal a surface with <1 nm root mean square (RMS) roughness, and X-ray rocking curve (XRC) measurements indicate dislocation density ~5×108 cm-2. PL measurements show strong luminescence peaks at 3.38 eV and 2.2 eV, characteristic of GaN:Be. Electrical testing of the films indicated their highly resistive nature. From secondary ion mass spectrometry (SIMS) measurements, the concentration of Be was found to be <1019 cm-3. This relatively low [Be], coupled with the omnipresence of the deep BeGa state, giving rise to the yellow luminescence at 2.2 eV, is a likely reason for the semi-insulating nature of the GaN:Be. Since simply increasing Be(acac)2 flow primarily results in material degradation without significantly increasing [Be], other methods to increase Be incorporation and activation efficiency were pursued. By using a pulsed growth method, the [Be] was found to increase by nearly an order of magnitude, but early results show that this caused significant surface degradation, however this may be a potentially strong method to increase [Be] in the film. The small atomic size of Be relative to Ga is a likely reason for the poor solubility of Be in GaN as well as the deep acceptor level. As such, methods of strain mitigation, including co-doping with (larger) Mg and the use of indium surfactant during growth were investigated. Despite its larger atomic size and resulting local compressive strain, the presence of Mg did not increase Be incorporation efficiency. Instead, Mg seemed to impede Be incorporation, likely in part due to competitive incorporation. With the growth conditions tested, the use of In surfactant had very little impact on [Be], though PL measurements show a suppression of YL relative to UVL. This implies that the In facilitated the formation of the shallow Be-related defect over the deeper isolated BeGa. Despite this, the GaN:Be was still highly resistive, underscoring the need for future work. Finally, MOCVD AlGaN:Be was demonstrated for the first time. Because of the AlGaN composition, PL using a HeCd laser could not be completed on AlGaN:Be with Al>8%. Measurements of 8% AlGaN:Be sample showed strong YLBe, indicating some Be incorporation. The shallow acceptor-related UVL was not observed. These results lay the foundation for further development of ultra-wide bandgap (UWBG) AlGaN:Be, which is arguably the most promising application for Be-doping in III-nitrides. Be-doping in AlGaN, particularly Al-rich AlGaN, requires significantly more investigation and will undoubtedly be the subject of many future studies.

Download

Share

COinS