AVS2012 Session EM+TF+AS-ThA: Growth and Characterization of Group III-Nitride Materials

The effect of low temperature on the surface photovoltage (SPV) in semiconductors is rarely studied and not well understood. We studied the SPV behavior for Mg-doped, p-type GaN using a Kelvin probe at temperatures from 80 to 300 K. Under band-to-band UV illumination at room temperature, the measured SPV signal in p-type GaN becomes negative as electrons are swept to the surface. However, we observed that at low temperatures, the SPV signal becomes positive under UV illumination, contrary to the SPV behavior of p-type GaN at room temperature. This positive SPV resembles the behavior of an n-type semiconductor. We assume that under UV illumination and at low temperatures, the conductivity of Mg-doped GaN does indeed convert from p- to n-type. This conversion was predicted from photoluminescence studies on Zn-doped GaN.^[1] At low temperatures, photo-generated electrons may accumulate in the conduction band which causes an upward shift in the bulk Fermi level towards the conduction band. This results in a positive SPV signal, since the Kelvin probe uses the bulk Fermi level as a reference for the measured SPV signal. Interestingly, the characteristic temperature at which we observe this transition from p- to n-type behavior depends on illumination intensity. As the excitation intensity increases from 10¹⁵to 10¹⁷ cm^-2 s^-1, the characteristic temperature increases from 130 to 170 K. This result also agrees with previously reported photoluminescence data and further authenticates the above assumption.^[1]

[1] M. A. Reshchikov, A. Kvasov, T. McMullen, M. F. Bishop, A. Usikov, V. Soukhoveev, and V. A. Dmitriev, Phys. Rev. B 84, 075212 (2011).

Gallium nitride is a high quality semiconductor widely used in both optical and electronic devices. The polarity of GaN (+/- c-direction) influences many properties of the resultant material, including chemical reactivity and electric field in these ‘piezoelectric’ materials. Control over the polarity of GaN grown on sapphire and SiC substrates has been previously demonstrated by controlling the growth conditions, doping levels, and buffer or nucleation layer properties. Further, in the case of heavily doped p-type layers, spontaneous polarity inversion has been demonstrated in GaN homoepilayers, switching the doped layer from Ga-polar to N-polar. However, this approach leads to uncontrolled inversion domain boundaries and often results in dopant clustering within the film, impacting film quality and resultant device performance.

In this work we investigate the fabrication of Mg-free inversion layers (ILs) to control the polarity of MOCVD-grown GaN on GaN substrates. By changing the IL material, we demonstrate conversion of GaN polarity in both directions (N-polar to Ga-polar and Ga-polar to N-polar). By employing a patented selective growth method to deposit the IL, the lateral polarity of the GaN can also be alternated, allowing control of the polarity in both vertical and lateral directions. A one-dimensional grating of periodically oriented (PO) GaN stripes was achieved over square-centimeter (or large) areas. The boundaries between polarities are found to be both sharp and vertical, and the growth conditions have been adjusted to result in equal growth rates of both polarities. Chemical etching of the material verifies the polarity of the material. Transmission electron microscopy (TEM) rules out the presence of alternating polar inclusions in the inverted material while showing a strong inversion domain boundary at the vertical interfaces. Dislocation density and grain size are determined through the use of electron channeling contrast imaging. The MOCVD-grown PO GaN structures have been extended in thickness by further HVPE growth. TEM and photoluminescence imaging confirms that the PO GaN structure is maintained throughout the extended growth (up to 80 μm in thickness). This method of GaN polarity inversion offers the promise of engineering both lateral and vertical polarity heterostructures and the potential of novel engineered polarity-based devices.

Solid state lighting by means of GaInN/GaN light emitting diodes (LEDs) is rapidly progressing to a major factor in energy savings technology. By convergence of lighting and lighting control, however, smart lighting is an opportunity to elevate lighting to a holistic experience of human wellbeing beyond the obvious economic benefits. Full epitaxial control of the GaInN/GaN active region is prime to fulfill the promise of an optical bandgap tunable across the entire visible spectrum. As such it will serve both, as tunable absorption layer for multijunction solar cells and emitter for direct emitting LEDs. The later aspect is of particular promise to outperform the traditional phosphor conversion approach known from historic fluorescence lamps and current white light LEDs.

Rigorous defect reduction approaches have enabled us to continuously improve he emission efficiency in ever longer wavelength emission reaching beyond green, deep green to yellow and orange (590 nm). In contrast to conventional phosphor or AlGaInP-based LED, such emitters show a superior temperature stability of their light output performance. A further leap in defect reduction has been demonstrated by the implementation of heteroepitaxy on nanotextured templates. Unlike widely explored lateral epitaxial overgrowth, growth zones primarily coalesce without the generation of threading dislocations. Implemented at the sapphire substrate level in green LEDs, the texturing substantially boosts both, internal quantum efficiency and light extraction. Furthermore, by control of the crystallographic orientation of growth we achieve a modulation of the piezoelectric polarization within the active region. This for once results in the emission of highly linear polarized light but on the other hand holds the promise to move the actual sweet spot of LED performance from the blue into the green and yellow spectral region. We discuss our approaches in light of our latest achievements.

This work was supported by a DOE/NETL Solid-State Lighting Contract of Directed Research under DE-EE0000627. This work was supported in part by the Engineering Research Centers Program of the National Science Foundation under NSF Cooperative Agreement No. EEC-0812056.

The influence of the substrate growth temperature on the structural and optoelectronic properties of group III-nitride layers grown by various growth techniques has been extensively studied and reported on, due to the close relationship of substrate temperature with crystalline quality and the point defect chemistry of the alloy. Most thin film growth systems only control the substrate temperature and have limited control to adjust the gas phase decomposition dynamic independent to influence to growth surface chemistry.

In this contribution, we present results on the growth of InN epilayers grown the high-pressure chemical vapor deposition (HPCVD), studying in influence of and independent from the substrate temperature controlled gas phase temperature above the substrate reactor zone. The HPCVD reactor system has two heater elements: one that controls the substrate temperature and a second radiative heat source above, which allows the control of the gas phase temperature. While the substrate temperature dominantly controls the growth process and the crystalline layer properties, the heater above the substrate surface influences strongly the precursor decomposition processes and the diffusion and concentrations of the precursor fragments in the boundary layer and at the growth surface. InN epilayers grown with different gas phase heating settings where grown and analyzed with the respect to their short- and long-range crystalline ordering and their optoelectronic properties as function of the gas phase temperature. The long-range and the short-range crystalline order of the layers have been analyzed by x-ray diffraction 2Θ-ω scans FWHM and the Raman E₂ (high) FWHM, respectively. The optoelectronic properties have been studied by reflectance spectroscopy and are related to the structural properties and the additional gas phase heating.

The figure depicts the FWHM values of Raman-E₂ (high) peak of the InN epilayers as a function of reactor pressure for higher (red line) and lower (blue line) gas phase temperature. The results indicate that there is an improvement of the short-range crystalline order of the layers with lower gas phase temperature. However, the FWHM values of XRD 2Θ-ω scans, which are not shown here, are indicating that there is an improvement of long-range crystalline order of the layers with increasing gas phase heating.

InN in principle opens up the possibility of using only one ternary III-V semiconductor alloy (InGaN) in optoelectronic devices to cover the whole visible spectral range. Despite this, key material properties of InN are still under debate. The intrinsic energetic position of the Fermi level is unclear, i.e., whether the Fermi level is located within the fundamental band gap or shifted slightly into the conduction band. The latter case induces electron accumulation at the surfaces of the crystal. Such an electron accumulation is typically observed at InN surfaces upon air contact, raising the question whether it is an intrinsic material property or not?

In order to probe intrinsic bulk properties by STM and not only contamination or surface effects, a clean and stoichiometric surface is necessary. This can be achieved by cleaving InN along non-polar planes. To analyze the origin of the different electronic states in detail, we investigated the clean non-polar (11-20) cleavage surface using cross-sectional scanning tunneling microscopy (XSTM) and spectroscopy (XSTS).

Using combined XSTM and XSTS we were able to locate an InN layer grown on an AlN buffer layer on top of a Si(111) substrate [1]. XSTS spectroscopy on the InN(11-20) cleavage surface yield normalized conductivity spectra, where three contributions to the tunneling current can be observed: (i) the contribution from the conduction band density of states for biases above the conduction band minimum at +0.3 V, (ii) a defect induced current, dominating the spectra between biases of 0 and -0.4 V, and (iii) a valence band related tunneling current rising at a bias of about -0.4 V and dominating the spectrum for biases below. The defect induced current arises from semi-filled defect states being present at the surface steps, and probably also from other (point) defects at the surface. Within the bulk band gap of E_G = 0.7 eV no intrinsic surface states could be observed. Furthermore, the Fermi level pinning at about 0.3 eV below the conduction band minimum indicates the absence of an electron accumulation layer.

The results illustrate that electron accumulation at InN surfaces is not a universal property on InN. For clean stoichiometric cleavage surfaces no electron accumulation is observed. Thus, electron accumulation results primarily from the details of the surface structure and is hence not an intrinsic property of the bulk InN material.

[1] Ph. Ebert, S. Schaffhausen, A. Lenz, A. Sabitova, L. Ivanova, M. Dähne, Y.-L Hong. S. Gwo, and H. Eisele, Appl. Phys. Lett. 98, in press (2011).

With the aim to find the optimum S₂ separation for high quality indium-rich InGaN epilayers, a set of In_xGa_1-xN samples with nominal x=0.9 has been grown with different S₂ timings. It will be shown that the S₂ separation is critical for the incorporation of gallium into the epilayers. In order to maintain single-phase epilayers, the S₂ separation has to be increased from S₂=400 ms for InN to over 1200 ms for In_xGa_1-xN. Raman spectroscopy and X-ray diffraction (XRD) spectroscopy are used to study the structural properties while the Fourier Transform Infra-red (FTIR) and transmission spectroscopy are utilized to investigate the electrical and optical properties of the epilayers.

We have recently proposed “SMART” III-N tandem solar cells in which all sub-cells could be coherent-structure high-quality pn junctions with low leakage current, resulting in high performance solar cells. SMART means “Superstructure Magic Alloys fabricated at Raised Temperature”. The most important feature in the proposed SMART solar cell is a novel idea for realizing ordered and/or quasi InGaN-ternary alloys with InN/GaN Short-Period Superlattices (SPS) enabling coherent-structure band engineering for the (InN)_n/(GaN)_m SPSs with simple integer pairs of (n, m)≤4. In this symposium, detailed epitaxy processes, structural and physical properties of SPSs, and also the idea and features of proposed “SMART” III-N tandem solar cells are reported.

We have ever reported successful growth of fine and coherent-structure 1-ML InN/GaN matrix QWs, and they can be fabricated so under self-limiting and self-ordering growth processes at remarkably higher and/or “raised” temperatures (~650 °C) than the critical one (~500 °C) for growing thick InN layer under +c growth regime in MBE. We are now underway to extend this understanding and the corresponding epitaxy technology to realize the proposed (InN)_n/(GaN)_m SPSs, and we have started to achieve (InN)₁/(GaN)_m (m=1-20) SPSs. When fabricating high structural quality those SPSs, very careful surface stoichiometry control such as (In+Ga)/N and In/Ga composition in adlayers, and also periodical complete surface dry-up of In and Ga for each one-cycle growth of SPSs are necessary and quite important.

In brief, 50-100 periods of (InN)₁/(GaN)_m SPSs were grown on MOCVD-grown +c-GaN template at 650 °C by a conventional plasma-assisted MBE. Surface stoichiometry and surface dry up were quite carefully monitored and controlled by in-situ Spectroscopic-Ellipsometry. First, structural properties of 50 periods of (InN)₁/(GaN)_m SPSs were characterized with XRD diffraction patterns taking the m as a parameter. It was found that coherent structure SPSs could be fairly easily fabricated even when the m was decreased down to 4. Generally, much more careful surface stoichiometry control was necessary with decreasing the m, though it was confirmed coherent structure (InN)₁/(GaN)₄ SPSs could be grown finally after quite careful control, such as selective re-evaporation between In and Ga consuming a long time. This leads to complete In re-evaporation leaving only some Ga metals on the surface. Of course those Ga metals must be completely dried up with irradiating plasma-excited nitrogen just before the following deposition of 1ML InN on it. It is still difficult at present, however, to grow fine structure InN/GaN SPSs with the m below 3.