(Invited) Growth and Characterization of Bulk Hvpe-GaN. Pathway to Highly Conductive and Semi-Insulating GaN Substrates

Wednesday, 4 October 2017: 09:30
Chesapeake A (Gaylord National Resort and Convention Center)
M. Bockowski (Institute of High Pressure Physics PAS)
Development of GaN-based optoelectronic and electronic devices is closely linked to ongoing work devoted to crystallizing bulk GaN. Hydride vapor phase epitaxy (HVPE) is the most popular method for obtaining commercial-grade substrates. Significant advantages of HVPE are high growth rate and high purity of new-grown GaN. Using ammonothermal GaN substrates of high crystallographic and structural quality as seeds enables growth of HVPE-GaN of the same high quality [1]. Without intentional doping the material has a very low concentration of impurities and n-type conductivity comes from silicon with concentration close to 1×1017 cm-3. Free carrier concentration is at the level of 3-5×1016 cm-3. Controllable doping of HVPE-GaN to prepare substrates of specific parameters is still a challenge. In this work influence of different dopants on optical and electrical properties of GaN is presented.

Results showing highly conductive n-type HVPE-GaN doped with silicon and germanium and semi-insulating material doped with carbon and iron with manganese will be presented. Characterization of the crystals includes their structural (X-ray diffraction, defect selective etching), optical (Raman spectroscopy, photoluminescence) and electrical (Hall measurements, Capacitance–voltage profiling) properties. Concentrations of dopants will be examined by Secondary Ion Mass Spectrometry (SIMS).

Detailed characterization of the grown GaN:Si and GaN:Ge will be presented. The way of introducing precursors to the HVPE reactor and details of crystallization process will be demonstrated. A difference between concentration of Si, which is always higher, and free electron concentration is observed in GaN:Si. This suggests that part of Si donors is compensated by an acceptor state. The strong yellow luminescence (YL) peaks suggest that this acceptor is gallium vacancy (VGa) or its complexes. The assumption is in good agreement with theoretical calculations of energy of VGa formation decreasing for highly n-type material [2]. However, when we incorporate Ge into our HVPE-GaN we observe no YL in the PL spectra and there is no significant difference between concentration of Ge and the free carrier concentration. This is probably connected to lack of VGa in GaN:Ge. Uniformity, in terms of free carrier concentration, of GaN:Si and GaN:Ge on c-plane and m-plane of crystals will be studied. It will be shown that highly conductive n-type material can be used as substrates for building laser diodes on them.

When C is introduced to HVPE-GaN, using CH4 as a precursor, a strong YL peak is observed in PL spectra of crystallized material. The crystals are highly resistive (>108 Ωcm) at room temperature. Hall measurements performed up to 1000 K showed p-type conductivity with hole concentration 4×1015 cm-3. Activation energy of ~1 eV was calculated from Arrhenius plot. This is an experimental confirmation of DFT calculations performed for C substituted for N [3]. It is also an explanation for strong yellow luminescence (YL) present in PL spectra for GaN:C in this case not related to VGa but to C.

Gallium nitride crystals can also be grown using solid iron as a source of dopants. Then, no yellow luminescence and only weak near band edge luminescence are visible in the crystals. A sharp peak is observed at 1.3 eV. This was shown before as an intrinsic transition of Fe impurity in GaN [4]. The grown crystals are highly resistive at room temperature. High-temperature Hall effect measurements revealed n-type conductivity with activation energy equal to 1.8 eV. Secondary ion mass spectrometry indicated the presence of manganese in the samples. The concentration of manganese was always higher than concentration of iron in the doped GaN.

[1] T. Sochacki et al., Appl. Phys. Express 6, 075504 (2013).

[2] J. Neugebauer et al., Appl. Phys. Lett. 69 (4), 503 (1996).

[3] J. L. Lyons et al., Appl. Phys. Lett. 97, 152108, (2010).

[4] R. P. Vaudo, et al., Phys. Stat. Sol. (a) 200, No. 1, 18– 21 (2003)