The need for low density materials with superior mechanical properties has been the subject of research and development for many years. Mg-base alloys with the hcp structure have been studied extensively and will be described in papers in this symposium. This talk will focus on several approaches to optimize the mechanical properties of hcp Mg alloys and describe a new low density high entropy hcp structure alloy. Two methods to improve the strength of Mg-base hcp alloys will be presented. One strategy is to produce the Mg alloy with a nanocrystalline structure. We present results of this method using mechanical alloying to synthesize nanostructured Mg- 2 at. % Y - 1 at. % Zn which had an average grain size of 21 nm. Very high hardness values were obtained for the as-milled alloy, and even higher hardness values after annealing at 573K. Another approach to enhance the mechanical behavior of hcp Mg was to use conventional hot rolling on a Mg-8.5 Gd - 2.3Y - 1.8 Ag - 0.4 Zr ( wt. %) alloy which had a low stacking fault energy. A high density of nanoscale stacking faults was introduced and resulted in high strength (~ 575 MPa yield strength) and reasonable ductility ( ~ 5.2% uniform elongation).
We have also recently used mechanical alloying to produce nanocrystalline Al₂₀Li₂₀Mg₁₀Sc₂₀Ti₃₀ (subscripts are atomic %). This high entropy alloy was single phase fcc in the as-milled condition. After annealing at 500^oC the structure transformed to single phase hcp. The very high hardness , 5.8 GPa, of the as-milled 12 nm grain size fcc alloy only dropped to 4.9 GPa in the hcp phase which had a grain size of 26 nm. Density functional theory calculations confirmed that the equilibrium solid solution phase is hcp.

Mg is an archetypal hexagonal close-packed simple metal. Due to a renaissance in the use of and research on Mg alloys over the past decade, the scientific community has learned a great deal about the mechanical behavior, including the effects of deformation twinning. Some of these things about twinning were known qualitatively in prior decades, but we have now developed a proficiency in characterization and computational modeling which permits a quantitative description of these twinning-induced effects over a wide range of strain rates, temperatures, loading conditions, and in a variety of alloy families. These capabilities could only be dreamed of by prior generations. This lecture will review the effects of the main twinning mode, {10.2} extension twinning (e.g., tension-compression yield asymmetry, yield plateau, anisotropy, rapid strain hardening, detwinning, etc.); characterization, primarily by diffraction-based techniques (electron, X-ray, and neutron); and modeling by crystal plasticity-based methods. Some recent results obtained on two precipitation strengthened commercial Mg alloys, ZK60A which contains zinc and zirconium and WE43 which contains yttrium, neodymium and zirconium, will be used as illustrations. The ZK60A is strengthened by c-axis rod-shaped precipitates while WE43 is strengthened by prismatic plane oriented plates. Strategies to control the mechanical behavior of through microstructure, texture, and alloy design will be proposed. Finally, outstanding questions which merit further research will be highlighted.

Mg alloys have attracted attention for recent years because of their low density and high specific strength. They are expected to be the next structural material. However, due to poor formability at room temperature, currently industrial applications of these alloys are limited. As a result of limitation in slip systems, twinning is very common in Mg alloys. Mg alloys show poor ductility during biaxial tension. Contraction twins are considered to be one of the important causes of early fracture. Therefore, understanding the twinning behavior is important for increasing the formability of these alloys.
Contraction twins are formed rapidly around necking region right before fracture. This makes it difficult to study them. Many researches have previously studied contraction twins by observing fractured samples. Here, we have successfully observed the increase of contraction twins by making steel/Mg multilayer composites. In such multilayer composites, Mg alloy has deformed 30 to 60% (maximum elongation of same monolithic samples are 10 to 20%). In both of composites and monolithic Mg sheets, volume fraction of contraction twins increased notably at the same strain level around 10 to 15% and exhibits the same value, approximately 20% at fracture. Thus, contraction twins were formed slowly with the increasing strain. Electron backscatter diffraction analysis was performed on the strained multilayer composites to verify the grain orientation and identify the twins. In-situ observation was done by observing the same region at different strain levels. This helped to understand the deformation mechanism from beginning to fracture. Based on the presented results, deformation mechanisms and the slip system which correlates with twin nucleation could be described.

Interest on Mg alloys has significantly increased in recent years for weight-critical applications. However, poor formability and low strength especially at low temperatures, which limits their extensive use as structural materials. Severe plastic deformation and detailed knowledge of multiple deformation mechanisms can be utilized to design microstructure, grain size distribution, crystallographic texture, and flow anisotropy of magnesium alloys. Low temperature shear formability of Mg-3Al-1Zn (AZ31) alloy samples with three different starting textures has been investigated via the equal-channel angular processing (ECAP) technique. Through careful texture modifications, the low temperature limit for ECAP of AZ31 has been reduced below 150°C by suppressing the compression twinning activity and instead, promoting prismatic slipA crystal plasticity model was utilized to gain a deeper understanding of operating deformation modes and dynamic recrystallization during ECAP of different starting textures at 150°C, the latter of which was found to play a significant role in shear localization during ECAP. Detailed electron backscatter diffraction (EBSD) analyses, transmission electron microscopy, and the simulation results clearly indicated that the formation of compression twins causes deformation localization, followed by dynamic recrystallization within the compression twins at 150°C. This non-uniform DRX causes local softening and macro shear banding, and eventual failure during ECAP. However, it was shown that with proper modification of the texture, the shear localization and macro shear bands can be suppressed. These studies led to achieve grain refinement down to few hundred nanometers and ultrahigh strength levels in magnesium alloys without rare earth alloying, and significantly enhanced their low temperature formability, which are otherwise impossible to achieve at low temperatures due to their inherent poor formability.

Al_xGa_1-xN alloys, particularly of high Al content (x>0.60), are actively pursued for deep UV optoelectronics, high-power and high-frequency devices. While significant challenges remain in the growth of AlGaN films, it has been demonstrated that high crystalline and optical quality can be achieved by using AlN single crystals as substrates. However, AlGaN/AlN heteroepitaxy results in compressively strained AlGaN layers due to the lattice mismatch, with the strain varying with alloy composition. Knowledge of relaxation mechanisms becomes important in order to control dislocation behavior to achieve stress management at different alloy compositions and thicknesses. Possible relaxation mechanisms include the introduction of a misfit component into pre-existing substrate dislocations and/or nucleation of misfit dislocations. Since AlN single crystals have very low dislocation densities (~10⁴ cm^-2), misfit dislocation nucleation becomes an essential mechanism in the relaxation of AlGaN films. Composition and strain relaxation of high Al content AlGaN layers grown by metalorganic chemical vapor deposition (MOCVD) on different miscut AlN substrates will be discussed. High-resolution X-ray diffraction (HRXRD) measurements were conducted to determine wafer miscut, alloy composition and relaxation. Transmission electron microscopy (TEM) was used to investigate the presence, type of dislocations and dislocation dynamics in AlGaN films. It was found that the nucleation of misfit dislocations in AlGaN layers depends on the presence of either macrosteps or bilayers steps on the growth surface. It is expected that stepped surfaces act as strain concentrators and nucleation sites for misfit dislocations, emphasizing the importance of substrate miscut and consequent alloy critical thickness. This follows from the fact that the hexagonal structure has a low number of primary slip systems and, for nitrides, relatively high critical shear stresses. Following these results, dislocation densities below 10⁵ cm^-2 have been obtained in films for Al content higher than 60%. Achieving such low dislocation densities allowed for the identification and control of compensating defects that affect n-type conductivity. First principles calculations based on point defect data were employed to predict defect type, defect concentrations and corresponding optical transitions. This approach is currently guiding the development of Fermi level control of point defect schemes that have allowed for the achievement of high carrier concentrations at high Al content AlGaN films.

Magnesium (Mg) hydride is a promising hydrogen storage material, yet its application has been limited by the slow hydrogen sorption kinetics. Recently, Mg nanoparticles have been shown to significantly improve hydrogen storage properties in terms of dimensional stability upon cycling with the trend that the smaller the particle the better the sorption kinetics. Since the volume change during sorption generates stress, leading to plastic deformation, the fundamentals of the mechanical deformation of the Mg nanoparticles is a significant issue. By using in situ compression tests inside a transmission electron microscope (TEM) coupled with molecular dynamics simulation of Mg nanoparticles, it is observed that deformation of a larger Mg particle is dominated by the nucleation of defects from stress concentrations at the surface, while the smaller particles deform more homogeneously with a larger distribution of dislocation sources. Consequently, the smaller nanoparticles demonstrate better plastic stability which can potentially help to accommodate volume expansion upon cycling and distribute pathways for hydrogen diffusion more homogeneously, contributing to the more promising hydrogen storage properties of Mg nanoparticles.

AlGaN-based semiconductor devices are particularly promising for novel optoelectronic applications, such as UV LEDs and deep-UV laser diodes. For many years, the use of foreign substrates for AlGaN epitaxy resulted in highly-mismatched heterostructures with high dislocation densities that hindered recombination rates and limited device performance. With the recent availability of AlN single crystals, close matching to AlGaN lattice parameters and thermal expansion coefficients can be achieved, leading to quantum-well structures with excellent crystallinity and higher internal quantum efficiencies. However, the lattice mismatch between AlGaN and AlN results in compressively strained alloy films, varying in strain magnitude with composition. In order to further guide device design, it is necessary to understand strain relaxation mechanisms. In this study, we report on the observation of crystallographic tilting of AlGaN layers as an in-plane stress relief mechanism in differently miscut AlN substrates. By means of high-resolution X-ray diffraction (HRXRD) characterization, epilayer tilt, alloy composition and relaxation were measured on Al_xGa_1-xN (x>0.6) films grown by metallorganic chemical vapor deposition (MOCVD). AlN substrates were cut and processed from [000-1]-oriented single crystal AlN boules grown by physical vapor transport (PVT), with ranging miscut from 0.2° to 4.0° off c-axis. Results show that for layers grown on low miscut substrates (0.2°-0.5°), the measurable tilt and partial film relaxation values were very small, insignificant for drawing conclusions on tilt-related relaxation. Growth on higher miscut substrates (0.5°-0.1°) resulted in AlGaN films that varied in tilt magnitude (0.005°-0.02°) and relaxation degree, independent of alloy composition. For significantly higher miscut substrates (>2°), AlGaN tilt behavior was in agreement with the Nagai tilt model, increasing in value with increasing miscut angle according to Nagai equation (>0.1°) and preserving miscut direction. While the misfit dislocation (MD) density in these heterostructures is expected to be low, epilayer tilt-related shear stresses are relieved by dislocation-mediated structural relaxation that is uniform and does not result in large mosaicity. The variation in tilt magnitude with miscut degree is dependent on step terrace width and its role on MD nucleation. Further knowledge on mismatch strain relief mechanisms will allow for stress management to be achieved, broadening the practical range of AlGaN/AlN systems.

The improvement of structure quality of AlGaN-based heterostructures with high Al content (x>0.4) is a critical issue for fabrication of high efficiency UV optoelectronic devices with a working wavelength below 300 nm. Here we review several approaches to defect engineering in AlGaN heterostructures which allow reduction of threading dislocation (TD) density, control of residual stress, carrier transport and localization in the optically-pumped UV laser structures grown on c-sapphire substrates by plasma-assisted MBE.
The samples were analyzed by transmission electron microscopy (TEM) using FEI Titan 80-300 electron microscope operated in both high resolution TEM (HRTEM) and scanning TEM (STEM) modes. Based on TEM data and X-ray diffraction analysis, it has been demonstrated that migration-enhanced epitaxy (MEE) nucleation of AlN buffer layer and N-rich 3D growth of ultra-thin GaN interlayers in AlN layers are efficient tools for TD density reduction in AlGaN-based heterostructures. Both AFM and high resolution TEM studies of MEE-nucleated AlN buffer layers indicate the increase of a lateral size of AlN islands and subsequent reduction of TD generation at epi-substrate interface. Atomic mechanisms behind that may include the initial formation of separated large-area AlN islands on the substrate, followed by lateral overgrowth of the AlN film.
Reflection high energy electron diffraction (RHEED) monitoring of the 3D GaN-interlayer growth in AlN layers demonstrates the abrupt stress variation that may force TDs to bend into the GaN/AlN interface with their following annihilation (“blocking” effect), which results in reduction of TD density to 10⁸-10⁹cm^-2 in the top part of (2-3)-mu;m-thick buffer layers.
High angle annual dark field (HAADF) STEM imaging evidences the presence of aperiodic superlattices (alternating Al- and Ga-rich AlGaN monolayers) formed in AlGaN cladding and waveguide layers grown under the strong metal-rich conditions. Development of the substrate-temperature modulated epitaxy is shown to improve homogeneity along the c-direction of AlGaN layers grown at Me(Ga)-rich conditions at relatively low temperatures (690-740°C).
Analysis of HAADF STEM images of AlGaN single QW (SQW) structures grown by sub-monolayer digital alloying technique reveals presence of 1-ML-thick Ga-enriched AlGaN disks within the nominally 2-nm-thick Al_xGa_1-xN/Al_yGa_1-yN(x=0.5-0.7, y-x=0.1) SQW. Photoluminescence studies of these structures evidences for strong localization of the carriers in such SQWs which can be regarded as a main reason of achieving in them stimulated and laser emissions within the spectral range of 258-290 nm at relatively low threshold power density (150-400 kW/cm²). The influence of AlGaN waveguide homogeneity along the growth direction on the carriers transport in the waveguide layers towards SQW is also considered.

Titanium and zirconium have a similar plastic behaviour arising from their hexagonal close-packed crystallography and from their similar electronic structure. In particular, plasticity in these two transition metals is controlled by screw dislocations gliding in the prism planes. Cross-slip of these screw dislocations in the first-order pyramidal planes or in the basal planes is also observed. We use ab initio calculations to study core properties of the screw dislocations and their mobility in both titanium and zirconium.
In zirconium, the screw dislocation spontaneously dissociates in a prism plane in two partial dislocations, each with a pure screw character. This leads to an easy glide of the screw dislocation in the prism planes, with a Peierls stress lower than 21 MPa [1], in agreement with experimental data. Using the NEB method, we also determine the Peierls barrier for the same screw dislocation, initially spread in the prism plane, but gliding in the pyramidal plane. The minimum energy path goes through an intermediate metastable configuration with a complex core structure that can be rationalized in terms of partial dislocations linked by a metastable stacking fault in the pyramidal plane. Interestingly, we show that the same metastable core is involved in glide in basal planes [2].
Ab initio calculations show that the stable structure of the screw dislocation in titanium is dissociated in a first order pyramidal plane, with the same stacking fault involved as in zirconium. This configuration can glide in its pyramidal habit plane, but with a high energy barrier. A configuration dissociated in a prism plane also exists, but is less stable. Like in zirconium, this metastable configuration can easily glide in its prism habit plane. NEB calculations show that the screw dislocation can go from one configuration to the other. As a consequence, prismatic glide in titanium operates by a locking-unlocking mechanism, in agreement with in situ TEM experiments [3].
[1] E. Clouet, Phys. Rev. B 86, 144104 (2012).
[2] N. Chaari, E. Clouet, and D. Rodney, Phys. Rev. Lett. 112, 075504 (2014).
[3] S. Farenc, D. Caillard, and A. Couret, Acta Metall. Mater. 41, 2701 (1993).

Dislocation behaviour and strain relaxation in lattice-mismatched epitaxial heterostructures are predicted successfully using a 3D dislocation dynamics model incorporating the effects of anisotropic elasticity, epilayer growth rates, epilayer growth temperatures, pre-existing threading dislocation densities and dislocation climb and glide. The model accurately reproduces experimental dislocation microstructures in Al_xGa_1-xN/AlN semiconductor heterostructures and reveals that dislocation climb can dominate strain relaxation behaviour in device heterostructures, in agreement with experiment and in contrast to classical equilibrium model predictions. In practice, for a given composition, strain relaxation may be controlled by selecting an appropriate substrate dislocation density and miscut, by varying growth temperatures and by controlling factors that influence dislocation climb, including doping and surface morphologies.

The results of molecular dynamics (MD) simulations of dislocation glide in GaN using a Tersoff potential are presented. The simulation methodology involves applying a constant shear stress to a single crystal system containing an individual dislocation, with multiple slip systems and dislocation types considered. Upon reaching a steady state, dislocation velocities are determined as a function of applied stress and temperature. Edge dislocations with a-type Burgers vectors in the basal, prismatic and pyramidal planes have been analyzed over the temperature range of 300-1300K. Results from systems with screw and mixed dislocations will be presented, as well as challenges related to performing simulations with these dislocation types. The mechanisms of dislocation glide will be discussed, as will the development of dislocation dynamics (DD) models from the present results.

Epitaxial AlN films with hexagonal structure have been widely studied to realize high-performance deep-ultraviolet optoelectronic devices. For epitaxial growth of high quality AlN films, templates with trench-patterned structure on substrates such as sapphire have been frequently used [1]. On the other hand, the use of the templates gives rise to a variation in the microscopic crystalline morphology (MCM), such as lattice tilting, residual strain and domain texturing at local areas inside the AlN film, depending on the patterning pitch and shape of the templates. To clarify the variation in the MCM in the AlN film, the microstructure analysis has been demanded [2]. In this paper, we clarify the variation in the MCM in the AlN film by position-dependent high-resolution tilt-2theta; maps of X-ray microdiffraction (XRMD).
A 5-µm-thick AlN film was first grown on sapphire substrate by metalorganic vapor phase epitaxy. Reactive ion etching was then used to prepare a trench-patterned template with 5-µm-wide terraces with the spacing of 3 µm. The depth of the trench was 5 µm. The trench direction was [1-100] periodically arranged to the [11-20] direction. Then, a 15-µm-thick AlN film was grown on the trench-patterned template by low pressure hydride vapor phase epitaxy (HVPE). The sample structure was confirmed by scanning electron microscopy. Two types of voids were formed at the trench regions, both forming tunnels running the [1-100] direction. The voids were periodically arranged along the [11-20] direction. XRMD was performed at the hard X-ray undulator beam line (BL13XU) in the SPring-8 facility. Typical size of X-ray microbeam was about 0.3 µm × 0.3 µm. The X-ray microbeam size was sufficiently smaller than the patterning pitch of the template. Position-dependent tilt-2theta; maps within a two-dimensional area on the film surface at 0.5 and 1.0 µm intervals along the [11-20] and [1-100] directions, respectively, by moving the sample stage were taken with the incident beam of the [11-20] direction for AlN 11-24 Bragg reflection and the [1-100] direction for AlN 0004 one.
The position-dependent tilt-2theta; maps revealed that the MCM in the AlN film is highly anisotropic and periodic depending on the morphology of the trench-patterned AlN/sapphire template. The variations of lattice tilting and residual strain become larger in void-containing trench regions than in terrace regions. These variations also become larger in the [11-20] direction than in the [1-100] one. These variations are possibly attributed to the influence of the growth evolution with epitaxial lateral overgrowth of the AlN crystal during HVPE and the distribution of dislocations in the AlN film.
[1] Y. Katagiri, et al., J. Crystal Growth 311, 2831 (2009).
[2] D. T. Khan, et al., J. Crystal Growth 381, 37 (2013).

Widespread application of many Mg-alloys has partly been hindered by the poor mechanical strength properties as compared to other lightweight structural materials. In this lecture, we will provide insight on a recently discovered mechanism for engineering Mg-alloys with ultrahigh strength (~500 MPa) and moderate ductility (~5%). This combination of properties can be enabled through the introduction of nano-spaced stacking faults, which serve to assist in dislocation accumulation and impede dislocation slip. Insights will be provide on appropriate alloy selection and processing, and phenomenological models for strengthing will be presented.

The crystal structure of hexagonal close-packed (HCP) crystals dictates complex deformation mechanisms, including dislocation-slip and twinning. As a result, the mechanical behavior of HCP metals displays strong anisotropy and strong orientation dependence, and identifying the fundamental aspects of plastic deformation in this class of metals is necessary to help improve their performance through alloying and microstructure design. In this talk we present a new implementation of a twin boundaries (TBs) into the framework of discrete dislocation dynamics (DDD) to simulate the collective evolution of dislocations and their interactions with TBs in magnesium (Mg) single-crystals. We report on the effect of crystal size and orientation on the deformation of Mg microcrystals. In addition, the new model is also used to investigate the mechanisms by which TBs and their size influence the yielding stress and hardening behavior of Mg. Finally, we report on the influence of dislocation interactions with TB and the glide of TB dislocations on the evolution and propagation of twins.

Laminated metallic composites consisting of an AZ31 magnesium alloy layer and two SS304 austenitic stainless steel layers were fabricated by using a reactive transient liquid phase bonding method, in order to achieve a metallic composite exhibiting an exceptional specific strength and ductility combination. Texture after deformation shows alignment of planes perpendicular to the tensile axis. In-situ tensile test has been performed and IPF maps were plotted at 0, 20 and 30% deformation to track the orientation change in various grains. Deformation and crystal rotation has been carefully examined in several grains. Results indicate prismatic slip is activated within grains in a specific range of orientations. Activation of this slip system is due to the buildup of extra shear stress on the prismatic planes by transverse tension caused by the multilayer structure.

Magnesium alloys attract attentions as high-specific strength materials. However, they show anisotropy of CRSS (Critical Resolved Shear Stress) for plastic deformation mechanisms. This anisotropy is one of the reasons why Mg alloys show poor cold workability. In contrast, it is reported that Mg-Y alloys show better ductility than other Mg alloys^[1]. To understand this good property in terms of Von-Mises criterion, it is necessary to evaluate CRSSes for various plastic deformation mechanisms. Nevertheless evaluation of CRSSes for non-basal plastic deformation mechanisms is difficult by only compression test and tensile test due to the anisotropy. To evaluate CRSSes for prismatic slip, we established a pure-shear test^[2]. This test method provides higher shear factor which convert applied stress to RSS for a certain plastic deformation mechanism than schmid factor by uni-axial compression test and tensile test. CRSSes for various plastic deformation mechanisms of Pure-Mg and Mg-Y alloy single crystals were also evaluated by pure-shear test and compression test at room temperature. As the results, it was revealed that CRSS ratio between basal slip and non-basal plastic deformation mechanisms decreases with increase of Y composition^[3]. From these experimental results, applied stresses which were required to operate 1th, 2nd, 3rd, 4th and 5th independent plastic deformation mechanism, respectively, with various stress axes were estimated by a method established by Cotton, et al. As the results, in Mg-Y alloy, several plastic deformation mechanisms can simultaneously operate more easily than in pure-Mg. This result is consistent with good ductility of Mg-Y alloys.
[1] S. Sandlöbes, et al., Acta Mater., 59 (2011) 429-439
[2] T. Mineta, et al., J.Japan Inst.Met.Mater, 77 (2013) 466-472
[3] T. Mineta, et al., Abstracts of the 2014 Spring Meeting of JIM (2014) 482

When magnesium alloys are deformed at elevated temperatures, deformation is accommodated by slip and/or twinning until a critical value is reached. After this the material often exhibits dynamic recrystallization (DRX), which results in softening instead of hardening of the flow curve and can be detrimental or beneficial to Mg processing depending on the application. In this paper, the crystal plasticity based finite element (CPFEM) method is coupled with a probabilistic model to simulate DRX. The proposed framework accounts for both nucleation of a new grain and growth of the nucleus. In the model, the local gradient in the dislocation density tensor in the deformed material is employed to determine the nucleation of recrystallized grains. Furthermore, morphological aspects (length, width, etc.) of deformation twinning are introduced into the numerical analyses for both compression and tension twins. Twins are assumed to initiate at stress hot spots on the grain boundaries. These locations are then used as potential nucleation sites for DRX by neglecting any twin growth. Simulations of DRX are performed for AM30 and AZ31 magnesium alloys deformed at high temperatures and the predicted textures as well as the stress-strain curves are compared with experimental data to identify what further refinement of the model would improve accuracy.

Despite the fact that the classical line tension model for dislocations lacks dislocation-dislocation interactions, we observe that the model underestimates the yield strengths of binary junctions in hcp crystals, in comparison to estimates from discrete dislocation dynamics in which the interactions are present, consistently by a factor of two. This is observed for all of the combinations of stable junction orientations and configurations we have examined. The present focus is primarily on junctions involving a (0 1 -1 0) prismatic plane and any of the adjacent planes: basal (0 0 0 1) , prismatic (1 -1 0 0), primary pyramidal (1 -1 0 1), or secondary pyramidal (-2 1 1 2) planes. The major basal, prismatic, and pyramidal Burgers vectors have been considered. Based on these observations, a polynomial expression was deduced to represent the junction yield surface in stress space of any general hcp crystal. The formulation is based on the idea that the interaction effects can be accounted for by scaling the result from the orientation-dependent line tension model by the observed factor. The line tension model is a function of the Burgers vectors and isotropic elastic constants which makes the resulting analytic expressions applicable to any hcp material. The aspect ratio and tilting angle of the yield surfaces were subsequently calculated using the coefficients of the simple equation to explain the apparent similarity of all yield surfaces regardless of the differences in the junction and dislocation configurations.

In this talk, we present measurements of slip-band grain boundary interactions with high resolution electron backscatter diffraction and AFM to reveal systematic variations in stress, slip transfer and dislocation storage depending on the nature of the grain boundary. This study focusses on understanding local stress variations, grain boundary character and surface morphology to probe the behaviour of a very important hexagonal material at the micro-scale. Our findings reveal the fundamental nature of slip transfer and how this is governed by the nature of the grain boundary geometry.
At a low strain level, three types of interactions were observed: blocked slip band with stress concentration; slip transfer; and blocked slip band with no stress concentration. The stress concentration induced by the blocked slip band was fitted with Eshelby&’s theoretical model, from which a Hall-Petch coefficient was deduced. It was found that the Hall-Petch coefficient varies with the individual grain boundary. We investigated the geometric alignment between the slip band and various slip systems to the neighbouring grain. Stress concentration can be induced by the blocked slip band if the slip system is poorly aligned with < a > prismatic, pyramidal or basal slip systems in the neighbouring grain. Transfer of slip across the boundary occurs when there is good alignment on < a > prismatic or pyramidal slip systems. Other stress-relieving mechanisms are possible when the best alignment is not with the slip system that has the lower critical resolved shear stress.

Single crystal Aluminum Nitride (AlN) possesses excellent opto-electronic properties, but remains difficult to produce in bulk. The two leading approaches to making bulk AlN are epitaxial growth processes, such as metalorganic vapor-phase epitaxy (MOVPE), and sublimation growth. These are high temperature processes leading to thermal stresses generated within the crystal, which cause stress-induced cracks and defects, primarily in the form of dislocations, which limit the performance and lifetime of the devices. Understanding and modeling the thermal-mechanical behavior of single crystal AlN under growth conditions can be used to accelerate improvements in process design to allow for more efficient production of high quality crystals. In this work, thermal-mechanical constitutive relations that are suitable for modeling crystal growth processes are developed for bulk, single crystal, wurtzite AlN. A temperature dependent, elastic-viscoplastic crystal model that considers slip and the evolution of mobile and immobile dislocation densities on the prismatic and basal slip systems is developed. The model is implemented into a finite element framework, and growth processes are modeled. A parametric study is carried out for the sublimation growth process to study the effect of various growth conditions on the stresses developed in the crystal as well as on the level of plastic deformation (slip) and dislocation density. Conclusions are drawn on what are the most important growth parameters and recommendations are made.

Titanium is an attractive material for high-end applications where advanced properties such as strength-to-weight ratio and corrosion resistance are vital. It is well known that the crystal structure of titanium is predominantly HCP at ambient temperatures and BCC at elevated temperatures. Here we show that titanium can undergo a thermally induced HCP to FCC phase transition in freestanding thin foils that has not been anticipated on the equilibrium phase diagram. This phase was found to be stable at ambient temperatures as well. In situ heating in the transmission electron microscope coupled with high-resolution strain mapping of titanium foils revealed that this phase transformation is accompanied with dislocation nucleation rather than a martensitic transformation mechanism. The mechanical properties of this new phase is explored through nanomechanical in situ TEM testing which allows us to isolate the individual phases as well to explore the mechanical properties of the phase boundary. It was found that the FCC/HCP phase boundary presents a strong barrier to dislocation propagation, significantly contributing to the work hardening capabilities of the dual-phase materials. Our findings shed light on both fundamental understanding on structure transformation and engineering design of advanced Ti alloys.

There has been renewed interest recently in the Zn-IV-N₂ subclass of heterovalent ternary compounds. The heterovalent ternaries have already proved to be important functional materials in areas such as thin film photovoltaics and nonlinear optics. The Zn-IV-N₂ semiconductors are of special interest because of their close relation to the technologically important III-nitrides and because of their many earth abundant constituent elements and alloy band gap range that spans from 1.7 to 5.5 eV.
Here we report on the photoluminescence, Raman and x-ray diffraction (XRD) spectroscopy of polycrystalline ZnSnN₂. These measurements provide important insight into the issue of cation disorder on the crystal lattice.
ZnSnN₂ has an orthorhombic crystal structure, with atomic positions similar to the wurtzite III-N structure and a cation sublattice consisting of equal parts Zn and Sn atoms. There are two predicted phases for ZnSnN₂, space groups Pna2₁ and Pmc2₁, which correspond to two arrangements of cations on the lattice that fulfill the octet rule. Our calculations show that the phases are nearly identical in energies of formation and band gaps. To date, all XRD of ZnSnN₂ show only wurtzite diffraction peaks, indicating a high degree of disorder on the cation sublattice. Further evidence of disorder is seen in the Raman spectra of ZnSnN₂ grown between 450 and 550 °C. These spectra consist only of density-of-states features, typical of amorphous material and phonon glasses, over the entire range of growth temperature.
Photoluminescence spectroscopy at 77 K of our ZnSnN₂ reveals a near-band-edge luminescence peak at 1.7 eV, in agreement with our band gap calculations for the pure phases. This agreement is surprising because the current model of disorder in the heterovalent ternaries is a random arrangement of cations, which results in many violations of the octet rule. These violations are energetically unfavorable and drastically reduce the band gap; calculations of the band gap of disordered ZnSnN₂ predict a decrease of over 46% from the band gap of the pure phases.¹ Our photoluminescence data demonstrate that, contrary to this model, the band gap of ZnSnN₂ is insensitive to disorder.
We propose that the lattice disorder is due to the random stacking of planar layers of the Pna2₁ or Pmc2₁ phases, in such a way that the octet rule is preserved, rather than to the random arrangement of cations. We describe this system using a two-pseudospin Ising model. Using this model, we calculate the XRD spectra as a function of the proportions of the two phases, reproducing the signatures of disorder in the experimental spectra. This model applies to all wurtzite-based heterovalent ternary compounds.
¹ Feldberg, N. et al. Appl. Phys. Letts. 103, 042109 (2013).
The authors acknowledge support from the NSF grants DMR-1006132, -0420765, -0907477, and -1004595, and the U.S. Department of Education grants P200A090276 and P200A120018.

The dislocation mechanisms responsible for the high-temperature mechanical properties of Zr alloys are still poorly understood. Indeed, whereas the role of dynamic strain ageing close to 350°C seems well established, various mechanisms have been proposed to account for a more or less constant stress exponent n ~ 6 above this temperature.
In situ straining experiments have been carried out in a M5 alloy between 250°C and 450°C. At 250°C and 300°C, the results show a steady and homogeneous dislocation motion in prismatic planes, with little cross-slip in the pyramidal planes. At 350°C, the kinetics of mobile dislocations becomes very jerky and inhomogeneous, in agreement with a dynamic strain ageing mechanism.
Above this temperature, the motion is again steady and homogeneous. Extensive cross-slip between prismatic and pyramidal planes forms super-jogs which are efficient pinning points against the glide motion. These super-jogs can eliminate by glide along the Burgers vector direction, never by climb. The glide velocity between super-jogs is linear as a function of the total driving stress (applied stress minus line-tension stress due to dislocation curvature), in agreement with the solute dragging mechanism. The origin of the stress-strain rate dependence with an exponent larger than unity is then discussed.

GaN is the material of choice in various applications such as HEMTs, UV LEDs, chemical sensors, UV photo detectors, and power amplifiers. Currently, there are mainly two deposition techniques most widely used for the utilization of epitaxial GaN films, namely MOCVD and MBE. Both of these techniques offer single crystalline films; however, both necessitate high deposition temperatures [6-7]. In order to utilize GaN in settings with limited thermal budget, such as back end of line (BEOL) and flexible substrates, utilization of alternative deposition techniques carries vital importance. With this aim low temperature deposition of GaN has been achieved by the utilization of several methods. One of this methods is atomic layer deposition. Differing from other methods ALD offers the most uniform and conformal deposition even at sub-nanometer thickness levels. In this study, we employ the ALD grown GaN thin films as the channel material of the bottom gate thin film transistors (TFT). Proposed device of this work is the GaN based transistor with the lowest thermal process budget reported to the date.
Fabrication of the bottom gate TFT starts with the RCA cleaning of the highly doped (1-5 mOmega;-cm) p-type Si wafer. Plasma-enhanced chemical vapor deposition of a 200-nm-thick SiO₂ is performed at 250 °C. The SiO₂ film is patterned to define the active device areas. An HF-last clean is immediately followed by the growth of 77-nm-thick Al₂O₃ and 11-nm-thick GaN subsequently deposited at a single ALD process in a modified Fiji F200-LL ALD Reactor (Ultratech/Cambridge NanoTech Inc.), where the process temperature is kept at 200 °C. Active device areas are isolated by Ar-based dry etching of the GaN layer. Source and drain contacts are formed by sputtering a multilayer metal stack consisting of Ti/Al/Ti/Al/Ti/Au.
Output electrical characteristics of the HCPA-ALD-based GaN TFTs show that fabricated devices have clear pinch-off and saturation characteristics, and they exhibit n-type field effect transistor behavior. Transfer characteristics of the devices reveal the I_ON/I_OFF ratio of 2x10³. The advantage of using a thick gate insulator is that the gate leakage current was kept below 0.5 pA for all the bias conditions. The threshold voltage of the device is extracted from the transfer characteristics and it is found to be 11.8 V. Sub-threshold swing (SS) of the device is extracted to be 3.3 V/decade. Charge mobility in the channel is extracted in the linear region of the device operation and calculated to be 0.025 cm²/V-sec. Finally the effect of the positive gate bias stress on threshold voltage of the devices is studied and reasonable threshold voltage shifts for a device with a considerably thick gate insulator are obtained. This study demonstrates the possibility of using low-temperature ALD-grown GaN layers for alternative and stable flexible/transparent TFT devices upon further materials and process optimization.

ZnO is a remarkable multi-functional material with a distinctive property set and a huge range of applications. Recently, some research has been directed towards development of ZnO based materials with smaller band gap, which allow light emission/absorption over a broad spectrum from the UV to the visible region. In this context, we have developed a new pseudo-binary alloy, ZnO-InN (ZnInON). ZnInON has a tunability of the band gap over the entire visible spectrum as well as high optical absorption coefficient of 10⁵ cm^-1 [1], which makes this compound a promising candidate for light absorbing layers in solar cells. Furthermore, we found from the simulation of electron/hole wave functions that solar cells with ZnInON multi quantum wells (MQWs) possess low recombination rate of photo-generated carriers, being three orders of magnitude lower than that in conventional GaAs-based MQWs [2]. Here, we first demonstrate the sputtering growth of single crystalline ZnInON films by using ZnO templates, aiming to obtain high quality films for solar cell applications. The photo-electric properties of MQWs consisting of the single crystalline ZnInON films are evaluated through measurement of I-V characteristics under photo-irradiation.
30-nm-thick ZnInON films were epitaxially grown on ZnO templates by RF magnetron sputtering at room temperature in N₂-O₂-Ar atmosphere. MQWs with 7 periods consisted of 6-nm-thick ZnInON well layers (3.0 eV) and 10-nm-thick ZnO barrier layers that were fabricated on ZnO templates by RF magnetron sputtering at room temperature. I-V characteristics were measured under AM1.5 solar simulator light (100 mW/cm²) together with a diode pumped solid state (DPSS) green laser light (532 nm, 2.15 W/cm²).
Single crystalline ZnInON films with atomically flat surface have been successfully fabricated on ZnO templates by RF magnetron sputtering. X-ray diffraction measurements reveal high quality of ZnInON films, where the full width at half maximum (FWHM) of rocking curves from both (002) and (101) planes is significantly small of around 320 arcsec. From Hall measurements, the electron mobility was found to be high of 90 cm²/Vsec at the band gap of 3.0 eV. ZnInON based MQWs show interesting photo-response. The photo-conductivity is drastically increased by 1.8 times when the irradiation of green laser light was superimposed in addition to the solar simulator light, indicating long lifetime (>1 mu;sec) of photo-generated carriers in the ZnInON quantum wells. From these results, we conclude that ZnInON is a promising material for solar cells, especially for MQW solar cells, where ultra-high conversion efficiencies are expected.
This work was partially supported by JSPS and JST-PRESTO.
[1] N. Itagaki, et al., U.S. Patent No. 8274078 (2008). [2] N. Itagaki, et al., PCT/JP2013/055973 (2013). [3] N. Itagaki, et al., Appl. Phys. Express 4 (2011) 011101. [4] K. Kuwahara, et al., Thin Solid Films 520 (2012) 4674.

Emerging markets in electronics industry drive the development of flexible and transparent electronic devices. In the next decade, it is projected that new generation electronic and optoelectronic devices will be widespread from household needs to military applications. GaN is a promising semiconductor due to its superior properties in optoelectronic applications. However, conventional techniques to grow GaN films require very high temperatures. On the other hand, low cost and flexible substrates can withstand low temperatures; therefore, film growth on such substrates requires low temperature processing. To overcome this barrier, we demonstrate GaN film growth at 200C with decent crystal quality. We also demonstrate operational metal-semiconductor-metal (MSM) ultraviolet (UV) photodetectors on such layers, for proof-of-concept. GaN MSM photodetector devices exhibit dark current value as low as 3 pA when biased at 30V reverse bias, and 15X UV/VIS rejection ratio. Photocurrent measurements indicated existence of persistent photoconductivity effect due to deep level acceptor states.

Nanomaterials which have strong absorption in the near infrared (NIR) region are of great importance [1]. Among nanomaterials, Indium nitride nanocrystals (InN-NC) exhibit absorption onset of 0.7-0.9 eV and have localized surface plasmon resonance around 3000 nm. Moreover, recent studies revealed that InN has heavy effective hole mass and slow relaxation which can be a promising candidate for next generation photovoltaic technologies [2]. Our group synthesized 3.24-36nm sized InN-NCs using laser ablation of a high pressure chemical vapor deposition (HPCVD) grown InN thin film and reported the optical characteristics of InN-NCs [3]. Despite various experimental efforts to synthesize InN-NCs, the use of these nanocrystals in NIR optoelectronic devices is lacking. Typically, InN nanostructures are synthesized in vacuum environments that limit the throughput and scalability of InN-NCs applications to large area and flexible optoelectronic applications. In this work, we present a proof-of-concept NIR range photodetector based on InN-NCs obtained through laser ablation of a high pressure chemical vapor deposition (HPCVD) grown indium nitride thin film and are used as optically active absorption region.

We examine chemically-exfoliated 2D layered dielectric hexagonal boron nitride (h-BN) as the surface passivation material to modify the surface characteristics of defective semiconductors. Cadmium selenide (CdSe) quantum dot solar cell is demonstrated in which h-BN passivated TiO₂ is served as the electron acceptor. We observe nearly 46% improvement in photoelectric efficiency in quantum dot solar cell employing h-BN for surface passivation, as compared with the solar cell without passivation. Dominant mechanism of interfacial carrier recombination, including electron capturing by TiO₂ surface states and recombination at the valence band of CdSe, are efficiently suppressed by h-BN based surface passivation, leading to significantly improved photovoltaic performance. While most of the commonly used passivation materials (such as HfO₂ and Al₂O₃) are amorphous in nature, we demonstrate that highly crystalline, dangling bond-free 2D layered dielectric could be effectively used for defect passivation.

Sapphire (α-Al₂O₃) is important for a variety of technological applications due to its hardness, transparency, chemical stability, and proven usefulness as a substrate for nitride semiconductors. Currently, the fabrication of sapphire substrates involves the crystallization of alumina from the molten state. Given the very high melting temperature of alumina (~2072^oC), the process is high cost and requires significant capital investment. Thin, amorphous alumina films can be deposited by atomic layer deposition (ALD) which uses self-limiting surface reactions to create coatings of precise thicknesses. The formation of large-grained polycrystalline sapphire films via ALD remains an outstanding challenge.
In this study, we examined the seeded lateral crystallization of amorphous ALD alumina films. ALD was used to deposit amorphous Al₂O₃ films on SiO₂-coated (300nm) silicon substrates. Nanoparticle sapphire seeds were then deposited to initiate crystal growth through seeded lateral solid phase epitaxy. The buffer layer of SiO₂ was used to separate the silicon wafer and alumina film to prevent crystallization induced by the substrate. Seeded samples were annealed at 900-10500C and analyzed using scanning electron microscopy to measure crystal growth rates. Growth rates ranged from 6.2 nm/hr at 9000C to 92 nm/hr at 10500C. Growth rates were observed to plateau after long crystallization times, possibly due to solid state reactions between the amorphous Al₂O₃ and SiO₂ substrate. Electron back scatter diffraction maps indicated randomly oriented single crystalline domains below each seed crystal. The results support the potential for large-grain sapphire thin film growth on a variety of substrates at low temperatures (9000C). Future efforts will involve controlling crystal orientation as well as investigation of other variables that may influence growth rate, including seed density, addition of impurities, and alternative substrates.

The minimum Mg {10-12} <10-1-1> twin thickness is estimated by density functional theory (DFT) calculation. We found the minimum stable twin thickness is 1.14 nm, which corresponding to thickness of four twin deformation units [1][2]. Estimate twin boundary energy is 8.012\times 10^2 eV/nm^2. We also estimate an activation energy of twin boundary migration from a potential energy profile along twin boundary migration path. Estimated activation energy is quite small, which is less than 1.0meV/atom. Thus, twin can quickly thicken once stable embryo is nucleated as shown in the recent experimental observation [3]
[1] B. Li and E. Ma, Physical Review Letters 103, 035503 (2009).
[2] J. Wang, S. K. Yadav, J. P. Hirth, C. N. Tome, and I. J. Beyerlein, Materials Research Letters 1, 126 (2013).
[3] B.-Y. Liu, J. Wang, B. Li, L. Lu, X.-Y. Zhang, Z.-W. Shan, J. Li, C.-L. Jia, J. Sun, and E. Ma, Nature communications 5, 3297 (2014).

Nitrogen doped carbon has received a lot of attention because nitrogen doping can improve the electrical, mechanical and structural characteristics of carbon. Therefore, it has been applied in various fields which including fuel cells, energy storage, solar cells and water oxidation. We introduce a simple synthesis method, solution plasma process(SPP), to fabricate conductive nitrogen-doped ordered carbon with a graphitic wall structure by the simple adjustment of the condition of power without the addition of any external heat treatment. The most important merits of the SPP for the N-doped carbon synthesis, compared with the conventional plasma in vacuum and liquid, does not require the use of high temperature heat treatment, graphite electrodes, long process time and addition of poisonous gas feed like ammonia for nitrogen source. By simply controlling the condition of power, the structural order and intrinsic properties such as surface area, conductivity, and pore volume, and the nitrogen content of ordered graphitic carbon can be controlled. N-doped carbon was synthesized after discharge by bipolar pulsed power supply in Aniline solvent for carbon and nitrogen sources with varied solution plasma parameters. The obtained suspension was filtered and dried in a vacuum desiccator at room temperature. X-ray diffraction (XRD) pattern, X-ray Photoelectron Spectroscopy (XPS) and Brunauer-Emmett-Teller (BET) were conducted to examine the crystal structure, chemical bonding structure and specific surface area, respectively. The electrical resistivity of N-doped carbon in the bulk state was measured. In addition, the preliminary results of electrocatalysis will be further discussed.

Nonpolar-based devices for optoelectronic applications are actively pursued as part of highly efficient next-generation nitride technology. The use of nonpolar surfaces can eliminate piezoelectric and spontaneous polarization fields in nitride quantum-well structures that are known to reduce the recombination efficiency of polar oriented c-plane devices. However, growth of nonpolar nitride systems remains challenging, mostly due to the lack of native substrates that can significantly improve surface morphology and structural quality of nonpolar epilayers. M-plane AlN single crystals with dislocation densities around 10³ cm^-3 are excellent candidates for native nonpolar nitride substrates. In this study, we report on high-quality nonpolar AlN films homoepitaxially grown on m-AlN substrates by metalorganic chemical vapor deposition (MOCVD). Substrates were cut and processed from [000-1] AlN single crystalline boules by physical vapor transport (PVT). Growth temperatures were varied among the films, while all other growth parameters were kept identical. Atomic force microscopy (AFM) and high-resolution X-ray diffraction (HRXRD) measurements were utilized to monitor the evolution of surface features and crystalline quality as a function of growth temperature. For all m-AlN layers, the full width at half maximum (FWHM) of symmetric (10-10) rocking curves ranged from 15 to 78 arcsecs, indicating high crystallinity and no mosaicity. As growth temperature was increased, a reduction in (h0-h0) rocking curve broadening was observed, along with a decrease in anisotropic variation as samples were aligned either parallel or perpendicular to the [0001]; this was more evident in the film grown at the lowest temperature (1150°C) and disappeared for the film grown at the highest temperature (1500°C). Symmetric (10-10) 2theta;-omega; line scans exhibited interference fringes on films grown at temperatures lower than 1350°C, which could be attributed to the different incorporation of impurities in each film. For higher temperature growths, no strain features or interference fringes were observed. Surface roughness effects were observed in the (10.0) RSMs and confirmed by AFM imaging. A reduction in the pole diffuse scatter and elongation of crystal truncation rod (CTR) was observed as the growth temperature was increased. Additional (20-20) and (30-30) RSMs were recorded to image basal or prismatic stacking faults. However, no stacking fault related diffuse scatter was observed in any of the m-AlN homoepitaxial films.

When dislocations dissociate, generalized stacking fault energy landscape determines the dissociation plane and the dissociation width. For the case of pyramidal dislocations in Mg, two inequivalent slip planes are associated with one Burgers vector, and the determination of the morphology of the split dislocation core is highly non-trivial problem. Based on the first-principles calculations of stacking fault energy of pyramidal planes and structural relaxation of dislocation core, we present unique properties of pyramidal dislocations in Mg.

The recently discovered, graphene has attracted intensive attention from the scientific community in regard to their extraordinary physical and chemical properties. The excellent electrical conductivity and mechanical properties of graphene offers great potential for applications in electronic devices or high-performance composites. These exiting breakthroughs have triggered interest in new 2D ordered crystals constructed by elements other than carbon.
Boron nitride nanosheets (BNNSs) are two dimensional crystals composed of boron and nitrogen atoms covalently bonded in a hexagonal manner with exceptional physiochemical properties. Similar to graphene, BNNSs have remarkable mechanical properties and thermal conductivity.
In the present, the researches on BNNS are limited due to the lack of an efficient method to prepare BNNS. So we demonstrate a simple method to produce few layers boron nitride nanosheets (BNNSs) based on a solvothermal method by functionalization of the BNNSs with oleyl amines. By this simple method, BNNS of 4~5 layers with lateral size ca. several hundreds of nanometers were produced. The BNNSs were characterized by transmission electron microscopy, atomic force microscopy and infrared spectroscopy, X-ray photoelectron spectroscopy measurement. The concentration was verified by filtering the BNNSs dispersion as 40mg/ml and We also confirmed that the dispersion stability of the functionalized BNNSs by oleyl amines is better than the BNNSs in some organic solvent. Furthermore, we have researched a variety of applications for using mechanical properties of the functionalized BNNSs.

Multiple quantum well (MQW) solar cells have received much attention as high conversion efficiency and low cost “third generation photovoltaic cells”. In order to realize high efficiency (>50%), reduction of recombination rate of photo carriers is important. We have recently proposed ZnInON based MQW solar cell structure with strain-induced piezoelectric field [1, 2]. ZnInON has a wurtzite structure, and a bandgap energy from 1.7 eV to 3.4 eV depending on the chemical compositional ratio [3]. Furthermore, we have succeeded in fabricating single crystalline ZnInON films by sputtering method. From simulation of carrier recombination rate in the ZnO/ZnInON/ZnO QW with piezoelectric field, we found that the recombination rate is 3 orders of magnitude lower than that in GaAs based QW. Here we have fabricated MQW solar cells utilizing ZnInON MQW for i-layer, and have studied effects of the piezoelectric filed in the MQW on photo current density-voltage (J-V) characteristics of MQW solar cells.
We fabricated two kinds of ZnInON MQW solar cells: cells with and without piezoelectric filed in the MQW. Both kinds of cells were composed of n-type ZnO:Al (AZO), 7-period ZnInON/ZnO MQW, and p-type GaN template. We fabricated all films at room temperature by RF magnetron sputtering. The well layers were fabricated in Ar-N₂-O₂ atmosphere of 0.27 Pa and the barrier layers were fabricated in Ar-N₂ atmosphere of 0.70 Pa. The bandgap energies of ZnInON well layers and ZnO barrier layers were 3.1 eV and 3.4 eV, respectively. The film thickness of the well layers, the barrier layers and AZO films were 6 nm, 10 nm and 50 nm, respectively. J-V characteristics were measured under AM1.5 solar simulator light (100 mW/cm²) and under solar simulator light together with a diode pumped solid state (DPSS) green laser light (532 nm, 2.15 W/cm²).
Cells without piezoelectric field have nearly the same J-V characteristics under solar simulator light and under solar simulator light together with laser light, whereas cells with piezoelectric field show high open circuit voltage and high short circuit current density under solar simulator light together with laser light compared with those under solar simulator light. These results indicate that for cells with piezoelectric field (i) high density photo carriers exist in the wells probably because of a low recombination rate of carriers and (ii) the photo carriers are excited to the barrier layers by superposing laser light. Therefore, ZnInON QW with piezoelectric field is effective for reduction of recombination rate of photo carriers.
This work was partially supported by Grant-in-Aid for JSPS Fellows, JSPS and PRESTO.
[1] N. Itagaki, et al., “Metal oxynitride semiconductor containing zinc”, U.S. Patent No. 8274078 (2008).
[2] N. Itagaki, et al., “Multi-quantum well solar cell and method of manufacturing multi-quantum well solar cell”, PCT/JP2013/055973 (2013).
[3] K. Matsushima, et al., Jpn. J. Appl. Phys. 52, 11NM06 (2013).

Hexagonal close packed (HCP) metals, including magnesium and titanium alloys are currently a developing area of research for a number of applications due to their attractive properties, in particular their excellent strength to weight ratio. One such application is as a candidate material for the next generation of ultra light armor. The design and optimization of these materials for armor applications can be significantly accelerated by improvements in the understanding of the deformation and failure mechanisms associated with the target material undergoing ballistic impact. Such deformation conditions include both high strain rate deformation and shock loading.
In this study, large-scale molecular dynamics (MD) simulations are used to investigate the response of magnesium and titanium to extreme environments, including extreme strain rate (10⁹ s^-1) uni-axial stress and strain deformation, as well as shock loading. The capabilities and limitations of interatomic potentials to predict these mechanisms is first investigated and will be discussed. The deformation response has been investigated for single crystal systems as well as polycrystalline systems with varying nanocrystalline grain size. The first simulations suggest the failure of nanocrystalline magnesium occurs predominantly along grain boundaries under both uni-axial stress and strain deformation at extreme strain rates (10⁹ s^-1), and that it is highly resistant to the formation of stacking faults. Additionally preliminary results show a transformation to the high pressure body centered cubic magnesium phase along grain boundaries during deformation. The relationships between the microstructure and the strength of magnesium and titanium at high strain rates and the underlying micromechanisms related to deformation and failure will be discussed.

ZnO has a potential to replace rare-metal-based materials such as GaN in light emitting diodes (LEDs) and laser diodes (LDs). For ZnO applications in such devices, it is required to fabricate single crystalline ZnO films with low defect density on cost-effective substrates. Recently, we have developed a new fabrication method based on RF magnetron sputtering, “Nitrogen Mediated Crystallization (NMC)”, where nitrogen atoms inhibit the crystal growth and thus bring a high surface concentration of crystal grain [1, 2]. By utilizing the ZnO films fabricated via NMC (NMC ZnO) as buffer layers, we have succeeded in fabricating single crystalline ZnO films even on c-plane sapphire substrates with a large lattice mismatch of 18%. This is because the buffer layers with small crystal grains reduce the interfacial and strain energies coming from the lattice mismatch. Here, aiming to clarify the crystal growth mechanism when the NMC buffers are utilized, we perform detailed investigation on the effects of the surface morphology of buffer layers on the crystal growth of ZnO films, especially, in terms of the defect formation.
First, NMC-ZnO buffer layers were deposited on c-plane sapphire substrates by RF magnetron sputtering. The deposition temperature of the buffer layers (T_b) was 700-780°C. Ar and N₂ were used and the total gas pressure was 0.35 Pa. The film thickness of NMC-ZnO buffer layer was 10 nm. Then, ZnO films were fabricated on NMC-ZnO buffer layers at 700°C in Ar-O₂ atmosphere. The total gas pressure was 0.70 Pa. The surface morphology was evaluated by atomic force microscopy (AFM).
The surface morphology of NMC-ZnO buffer layers were significantly changed with changing T_b, associated with the variation in nitrogen desorption behavior at the growth surface. We found a strong correlation between the height distribution profiles, which is derived from AFM images, of NMC-ZnO buffer layers and the crystal quality of ZnO films. On the buffer layer with a sharp peak in height distribution, a pit-free single-crystalline ZnO film with atomically-flat surface was grown. Our results indicate that homogeneous and high-density nucleation at the initial growth stages is critical in heteroepitaxy of ZnO on lattice mismatched substrates.
This work was supported in part by JSPS No.25630127 and JST-PRESTO.
[1] N. Itagaki, et al., Appl. Phys. Express 4 (2011) 011101.
[2] K. Kuwahara, et al., Thin Solid Films 520 (2012) 4674.

Due to its inherent optical and electrical properties, InN is a very attractive material for the fabrication of high efficient optoelectronic devices. Particularly by alloying with GaN, InGaN ternary materials can span the band gap from 0.7 eV (InN) to 3.4 eV (GaN) that covers the most of the solar spectrum. However, unavailability of lattice matched substrates and the low dissociation temperature of InN are significant factors that make the growth of high quality InN materials an extremely challenging process. The commonly used substrates for InN are sapphire and GaN, with lattice mismatch ~26% and 10%, respectively. This lattice mismatch between substrate material and overgrown InN layers result in formation of point defects and extended defects which critically affect the physical properties of the overgrown layers of InN.
In this contribution, we report on the influence of substrate growth templates (e.g. sapphire substrates, metal-polar and N-polar GaN and AlN templates, AlN based lateral polarity structures (LPS), and GaN grown on LPS) on the properties of bulk InN epilayers, grown at elevated reactor pressures (8bar - 15bar) as well as the III/V precursor ratio. The growth temperature has been optimized in the range of 800°C to 900°C based on Raman E₂(high) mode evolution. The various templates introduce different lattice strain and polarization fields during the initial nucleation process, affecting the extended defect generation and their propagation and relaxation processes. To assess this effect on the bulk properties of thick InN epilayers, Raman spectroscopy has been used to characterize the localized crystalline structure (e.g. E₂(high) mode analysis) and the coupling with the free carriers (e.g. A₁(LO) mode analysis). Results are confirmed by XRD rocking and omega;-2theta; scans. Using atomic force microscopy and scanning electron microscopy we demonstrate that the used substrate and specifically the polarity (III- or N-polar) of the substrate have a significant influence on the surface morphology. In order to assess the optical quality of the layers and incorporation of point defects, photoluminescence (PL) and transmission measurements were performed. It is found that in general, N-polar material seems to be more favorable for the incorporation of point defects. However, it is expected that this can be controlled by using adjusted growth conditions. Finally, Fourier transform infra-red reflectance (FTIR) spectra are analyzed to obtain the carrier concentration, mobility and thickness of the epilayers. Results from FTIR are compared to those obtained from optical and structural analysis and a strong correlation is found.

Many hexagonal chalcogenides have emerged as extremely important functional materials. With their remarkable mechanical and electronic properties, they have been regarded as promising candidates for energy storage devices, non-volatile memory devices, topological insulators, thermo-electrics and optoelectronics. Many of the properties vary significantly with the material morphology, thickness, orientation and size. Here we show that using chemical vapor deposition (CVD) with custom-synthesized single source precursors, we are able to control these properties to a significant extent through the use of patterned substrates on which the TiSe₂, SnSe₂, Bi₂Te₃, and Sb₂Te₃ chalcogenides grow with great selectivity. The TiN/SiO₂ micro- and nano-patterns allow precise control of the position of individual crystals. Scanning Electron Microscopy (SEM) and Atomic Force Microscopy (AFM) imaging show hexagonal plate crystallites of SnSe₂ and TiSe₂ which grow perpendicular to the substrate surface in the thicker films, but align mostly parallel to the surface when the quantity of reagent is reduced to limit the film thickness [1, 2]. We have demonstrated that regular arrays of high quality single Bi₂Te₃ nanocrystals can be deposited by LPCVD from a single source bismuth chloride telluroether complex with a very high degree of positional control, and that the preferred orientation of the nanocrystals is strongly governed by the dimensions of the underlying micro- or nano-patterned substrate [3]. Here we report the growth, materials, and electronic properties of the above compounds and extend the work to include data on Sb₂Te₃ growth and characteristics. The latter is a complementary material to Bi₂Te₃ for thermo-electric applications. SEM, AFM, X-ray diffraction, Raman spectroscopy, Hall and Seebeck measurements will be reported. The origin of the selective growth will be discussed.
[1] Highly Selective Chemical Vapor Deposition of Tin Diselenide Thin Films onto Patterned Substrates via Single Source Diselenoether Precursors C.H. De Groot, C. Gurnani, A.L. Hector, R. Huang, M. Jura, W. Levason, and G. Reid; Chemistry of Materials 24, 4442-4449 (2012)
[2] Area Selective Growth of Titanium Diselenide Thin Films into Micropatterned Substrates by Low-Pressure Chemical Vapor Deposition S. L. Benjamin, C. H. de Groot, C. Gurnani, A. L. Hector, R. Huang, K. Ignatyev, W. Levason, S. J. Pearce, F. Thomas, and G. Reid; Chem. Mater. 25, 4719minus;4724 (2013)
[3] Controlling the nanostructure of bismuth telluride by selective chemical vapour deposition from a single source precursor S.L. Benjamin, C.H. de Groot, C. Gurnani, A.L. Hector, R. Huang, E. Koukharenko, W. Levason and G. Reid; Journal of Mat. Chemistry A, 2, 4865-4869 (2014)

Dislocation pile-ups occur in crystals when a number of similar dislocations are driven by an applied stress towards a low stress region (soft pile-up) or obstacle (e.g. a grain boundary or pinned dislocation) that they cannot overcome. Giving rise to long-range stress fields, their properties strongly influence plastic behaviour of the crystal as a whole, including grain size strengthening and slip transfer in polycrystals. Dislocation walls, on the other hand, form by the alignment of geometrically necessary dislocations (GNDs) in parallel slip planes, acting to minimize the long range stress field, and provide the finite lattice rotations required by deformation These walls form the boundaries between subgrains after the polygonization stage of post-deformation annealing.
If a hexagonal crystal is loaded such that only the three prismatic slip systems are active, with the dislocation lines parallel to the axis, the deformation can be modelled as plane strain. This allows the use of flexible, accurate and efficient quasi-two-dimensional discrete dislocation (DD) plasticity simulations. Periodic boundary conditions can be applied to predict the relaxation of random dislocation populations, and complex geometries such as micropillars and cantilevers can be treated by coupling the DD simulations to a finite element scheme for handling the free surfaces. This allows a direct comparison between simulations and micromechanical experiments. Much of the previous work assumes elastic isotropy, though most hexagonal materials are anisotropic to some extent. We generalize the simulation approach to anisotropic elasticity, assess the validity of the isotropic assumption, and investigate its effects on the mechanical response.
At this lengthscale, we evaluate dislocation structures with high (angular) resolution electron backscatter diffraction (HR-EBSD). This technique measures lattice rotations in a map with very high precision (1E-4 rads), which can be interrogated to calculate lattice curvatures, and hence the dislocation content via the Nye tensor. In practice, a maximum of only nine curvature components can be measured (e.g. three mapped directions with 3D EBSD techniques and three rotation axes), yet in hexagonal systems more than nine slip systems are present. This means the GND densities are not uniquely determined from the curvature maps, and require assumptions on the expected dislocation types and minimization pathways to be made. To identify the suitability of different strategies, we compare GND density distributions extracted from experimentally obtained curvature maps (and certain postulated dislocation populations and minimization pathways) with those predicted using anisotropic DD simulations.

The transition metal carbides (TMC) are a class of refractory materials known for having very high melting temperatures, high electrical and thermal conductivity and exceptional hardness. The 4B and 5B transition metal carbides form the cubic rocksalt structure near equal parts metal and carbon atoms, but some carbides can precipitate hexagonal and trigonal based carbides with carbon loss. Though these phases have a lower number of available close-packed slip options, these phases tend to be more ductile than their cubic counterparts. These differences are associated with the mixed bond type, metallic and covalent, present in the unit cell. In this talk we will link first principle computational simulations of slip energies with experimental identification of the slip systems to elucidate the differences in mechanical responses across these different carbide symmetries. In particular, we will discuss why HfC, a group 4B TMC, prefers {110} slip at room temperature whereas its equivalent cubic TaC counterpart, a group 5B TMC, slips on {111}. These differences are related to the formation of a stable, intrinsic stacking fault found to exist in group 5B TMCs and not in group 4B TMCs. The trigonal based carbides have been reported to have both basal and pyramidal slip. In Ta2C, it was found that the Ta-Ta out-of-plane basal bonding that dictated an asymmetric slip behavior on basal plane. This demonstrates that ductility in these materials is controlled more by bonding type than crystal structure, a result quite different than on observes in metals. The talk will also address the application of a Pierels-Nabarro model to the DFT calculations, with the associated difficulties of its application to these types of mixed-bonded systems. Collectively, these carbides offer unique microstructure engineering opportunities to tailor mechanical properties by controlling carbon content.

Due to the layered structure of graphite, it is enough to comprise only two or more graphene layers in order to realize large anisotropic material properties, both electronic and mechanical. Properties like superlubricty and self-retracting motion were experimentally observed in highly oriented pyrolytic graphite (HOPG) demonstrating that “smart”, low-energy dissipation nano-electromechanical (NEM) elements can be realized using 2D layered materials. While these phenomenons were observed in such systems, this behavior is still not fully quantified nor explained. Furthermore, while superlubricity has been experimentally studied using frictional force microscopy (FFM) for constant contact sizes, we study here the case for variable contact sizes i.e. the frictional force is measured for different overlaps between two adjacent graphitic layers. This setup allows us to directly measure the exfoliation energy and the required forces for displacing one graphitic layer on top of the other. Thus, we obtain values for both the interlayer cohesion energy and for the dynamic friction (dynamic shear strength) of 2 MPa (0.31±0.1 J/m²) and 0.06 MPa, respectively. We then demonstrate how these displacement forces can be designed to realize variable, constant and zero applied force mechanical elements. We also discuss the combined electromechanical properties between two adjacent layers of graphene both in the pristine (AB-stacking) and in a turbostratic state.

Bimodal grain size distributions in titanium have recently received much attention as a method for enhancing the strength and ductility simultaneously. By contrast, the contribution of ultrafine grains to the plasticity has not been clarified. Titanium having a hexagonal close-packed crystal structure exhibits anisotropy in the elastic and plastic properties. Therefore, it is a key issue in the deformation process of titanium that the activation of slip and twin systems are strongly dependent on the crystallographic orientation. To clarify the deformation behavior of the ultrafine-grained pure titanium, the current study examined the uniaxial tensile straining process using the micrometer-sized specimens with different integration of the crystal orientation extracted from the disc processed by high-pressure torsion (HPT). The material used in this study was a grade 1 commercially pure titanium with an average grain size of sim;30 mu;m. Discs with a diameter of 19 mm and a thickness of sim;0.8 mm were processed by HPT under an applied pressure of 1.5 GPa for two turns at room temperature. An ultrafine-grained microstructure with an average grain size of 0.4 mu;m was obtained by annealing for 0.5 h at a temperature of 673 K after the HPT processing. Micro-tension specimens with a gauge section of 20 × 20 × 50 mm³ were extracted at 7 mm of spot from the center of the HPT-processed disc using focused ion beam machining. Two ultrafine-grained specimens were prepared so that their loading directions were parallel to the circumferential and radial directions of the HPT-processed disc, which are denoted as FGA and FGB specimens, respectively. For comparison, two coarse-grained specimens were prepared, where the gauge sections were covered by approximately one grain. Their loading directions were parallel to [11-20] and [0001], which are denoted as CGA and CGB specimens, respectively. Micro-tension testing was performed at a crosshead speed of 0.1 mu;m s^-1 at room temperature in atmospheric air. All the specimens exhibited a discontinuous yielding. The CGA and CGB specimens were deformed by the activation of a couple of prismatic slips and {1-102} twins, respectively. The critical resolved shear stress (CRSS) for the prismatic slip was determined to be 94 MPa, which was lower than 110 MPa for the {1-102} twin. The FGA and FGB specimens exhibited significantly high yield stress, i.e. 520#8210;700 MPa, when compared to the CGA and CGB specimens. For the FGA specimen, in which each grain is oriented in a direction favorable to prismatic slips, the moderate uniform elongation, i.e. sim;13%, was attained, whereas the FGB specimen exhibited no strain hardening after yielding. It is suggested that in the FGB specimen, the mixed microstructure composed of soft and hard grains to prismatic slips facilitates the strain localization.

HCP-type semiconductors are sp³ covalently bound compounds with the wurtzite, W, structure composed of two interpenetrating HCP structures shifted from each other by ~ #8540;c along the c-axis. The unit cell contains 4 atoms per lattice point, does not have a center of symmetry, and the bond parallel to the c-axis has a length different from the other 3. Thus, the dislocations have a well defined structure that minimizes the amount of bond bending, and the basal {0001} planes can have a large polarization charge that can be exploited to create unique electronic devices. The W structure differs from the cubic zinc blend, ZB, structure in that the stacking of its close packed (0001) planes is AαBβhellip; whereas it is AαBβCγhellip; for the ZB close packed {111} type planes, and all 4 of the sp³ hybrid bonds are the same length. As a result, the W structure has only 1 {0001} type plane while the ZB has 4 {111} type planes. W AlN and GaN can be combined to create W AlGaN with lattice a parameter that varies ~ linearly between that of AlN and GaN. The orientation of choice of the AlN (compression) or GaN (tension) substrate, particularly for electronic devices, is the (0001) plane where the mismatch creates a plane strain because the (0001) plane has isotropic properties. Because there is no shear stress on the basal or prismatic planes, dislocations will have to slip from the surface to the interface on pyramidal planes, stacking faults will have to be formed, or W will have to be converted to ZB to accommodate the mismatch. All of these options will be explored.

The dislocation dynamics (DD) method models dislocations behavior, in-
teractions and evolution in BCC and FCC materials. It is used to predict
the strength of a material that varies with pressure, strain rate, temperature
and evolving dislocation density by providing input parameters to continuum
based approaches. Continuum models based on constitutive equations built us-
ing dislocation dynamics and molecular dynamics data have been successfully
compared to high energy physics experiments in BCC tantalum and vanadium.
We will present two recent development in dislocation dynamics. First, large
scale dislocation dynamics simulations usually involve several millions of inter-
acting dislocation segments. The stress at a point and interaction force between
two segments need to be computed many times during simulations. Up to now,
DD simulations were restricted to isotropic elasticity calculations because using
anisotropic elasticity was perceived as too expensive. We evaluate the cost ver-
sus accuracy of using spherical harmonics series to approximate the anisotropic
elastic Green&’s function in calculating stresses and forces between segments. The
stress at a point is obtained by analytically integrating the spherical harmonics
series once and the forces by integrating it analytically twice. We analyze the
convergence and cost of using this approach and describe the elements of a fast
implementation.
Second, the challenge of extending dislocation dynamics to hexagonal close
packed will be explained in details. In particular, two topological operations
specific to hexagonal close packed crystals will be described. Large scale simu-
lations will be presented integrating the hexagonal close packed crystal lattice,
core energies, dislocation mobilities and these new topological operations.
Lawrence Livermore National Laboratory is operated by Lawrence Livermore
National Security, LLC, for the U.S. Department of Energy, National Nuclear
Security Administration under Contract DE-AC52-07NA27344.

The generation of non-basal c+a dislocations near the c-axis direction is the key to inducing the widespread plasticity in highly anisotropic hcp magnesium. Using the molecular dynamics method, we have studied the formation process of screw and edge dislocations from the perfect dislocation state, in particular, focusing on the temperature dependent properties of the splitting process. We also studied interaction processes of the a screw dislocation and the c dislocation for the formation of a c+a dislocation. Here, we employed various forms of c dislocations, such as straight and loop dislocations. In the case of the c edge dislocation loops, we confirmed that various forms of polyhedra appeared due to the splitting process. In this process, we applied the shear stress for the a screw dislocation to cross-slip in the c-direction to interact with the c dislocation. The effects of temperatures up to 500K and the strain rates have been assessed in relation to the final structure of the formed dislocation.

Crack free high reflective AlGaN/AlN-based distributed Bragg reflectors (DBRs) can be used to realize high performance vertical emitting laser structures like VCSEL or polariton lasers in the UV. Such microcavity-based lasers are sought after due to their low lasing thresholds and superior beam quality to conventional semiconductor lasers. Unfortunately, high quality AlGaN/AlN-based DBRs with a high reflectivity are difficult to realize. On one side, to assure a low absorption in the UV, a relative high Al content in the AlGaN layers of the DBR is crucial. On the other side the reflectivity of a DBR is dependent on the growth of alternating l/4 thick, high and low refractive index layers, leading to the need of a reduced Al composition. The trade-off is a composition of ~65% offering a sufficient large refractive index contrast of 6 % to the AlN; but a high number of layer pairs in the DBRs is needed. Unfortunately, with such low Al compositions, the lattice mismatched induced strain energy within the AlGaN layers increases, effectively limiting the growth of thick AlGaN/AlN DBRs.
However, we report on the growth of Al_0.65Ga_0.35N/AlN DBRs with more than 20 pairs on sapphire. These DBRS show high reflectance and are suitable to be used in the UV at 270 nm. We investigate the influence of the nucleation layer on the critical thickness of the DBR to be able to increase the number of pairs. Using a 300 nm thick high temperature (HT) AlN nucleation layer prior to the growth of the DBR structure leads to cracking of the films when exceeding the critical thickness of 1.4 µm. The strain energy becomes significant in the DBR leading to relaxation of the films by forming cracks. In contrast, it could be shown that the use of a 500 nm thick Al_0.85Ga_0.15N nucleation layer instead of the AlN nucleation layer can increase the critical thickness of the DBR by ~200 nm. Due to the fact of comparable dislocation densities in both templates, strain has been found to be the major mechanism leading to the cracking in DBRs. The Al_0.85Ga_0.15N template with an average composition between AlN and Al_0.65Ga_0.35N reduces the total strain energy of the DBR. The final DBR structures were analyzed using x-ray diffraction (XRD), scanning electron microscopy (SEM), x-ray reflectivity and optical reflectivity measurements. Reciprocal space maps of the structure on sapphire/AlN template indicate a pseudomorphic growth of all layers. Furthermore, XRD w/2q-scans and cross-section SEM investigations indicate a high periodic structure with a homogeneous thickness distribution across the DBR with abrupt interfaces. The reflectivity spectra exhibit a pronounced stop band region of 10 nm centered at a wavelength of 270 nm. The maximum reflectivity of the DBRs reaches 97 % at 25.5 pairs; becoming the highest reported reflectivity of a DBR in this wavelength region.

In this report, we present results of the structural and optoelectronic properties of GaN and InGaN epilayers on sapphire templates with an AlN buffer layer grown by migration enhanced plasma-assisted Metal Organic Chemical Vapor Deposition (MEPA-MOCVD).
Applied reactor control provisions provide a temporal and spatial injection of metal organic (MO) precursors (e.g. TMG, TMI) as well as remote plasma excited nitride precursors species via excited nitrogen hydrogen and ammonia. The plasma activated species are generated by a hollow cathode plasma source (13.56 MHz radio-frequency (rf) power source and scalable up to 600W). Real-time plasma emission spectroscopy (PES) and UV-absorption spectroscopy (UV-AS) have been utilized to characterize the type and quantity of active gas phase species in the remote plasma and the afterglow region, in order to assess and control the active nitride species in the growth surface. Raman spectroscopy, X-ray diffraction (XRD), atomic force microscopy (AFM), Fourier Transform Infrared spectroscopy (FTIR) and photoluminescence measurements (PL) have been used to assess the structural and optoelectronic properties of GaN and InGaN epilayers. This contribution will detail and correlate real-time growth monitoring with ex-situ structural and optoelectronic properties of the GaN and InGaN epilayers grown on sapphire templates with an AlN buffer layer in order to understand the growth process mechanism depending on the remote plasma activated species and their concentrations, substrate temperature and reactor pressure. We will demonstrate that the use of MEPA-MOCVD can lead to GaN and InGaN layer with increased crystal quality as indicated by the FWHM of X-ray rocking curves and the E₂(high) Raman line. This is further confirmed by the observation of a strong near band edge emission and decreased deep defect luminescence as observed by low temperature luminescence measurements.

The group III-nitrides (GaN, AlN, InN and related alloys) are an important class of III-V semiconductors that are widely used in high brightness light emitting diodes, laser diodes and power electronics. Due to difficulties associated with the bulk crystal growth of GaN, group-III nitride thin films are typically grown heteroepitaxially on substrates such as sapphire, silicon carbide and silicon. These substrates have significantly different lattice constants and coefficients of thermal expansion than GaN, which can result in thin film stress, dislocation formation and film cracking. In addition to the epitaxial and thermal mismatch stresses, which are well known, growth-related stresses due to evolving film morphology and impurities also play an important role, but are not as well understood for the group III-nitride materials system.
Our efforts have focused on using in-situ wafer curvature measurements to study the magnitude and evolution of intrinsic growth stress during the deposition of group III-nitride thin films by metalorganic chemical vapor deposition (MOCVD). The presentation will focus on three specific cases where the ability to monitor film stress during growth provided key insights into the mechanisms of stress generation and relaxation and film properties. This includes the growth of GaN on silicon substrates, the effects of dopant incorporation on film stress in GaN and AlGaN and the influence of film polarity on stress evolution in InGaN/GaN heterostructures. By combining the in-situ measurements with post-growth atomic force microscopy and cross-sectional transmission electron microscopy, we correlate the growth stress to microstructural changes in the film arising from island coalescence and dislocation inclination. Methods to mitigate stress and reduce film cracking will also be discussed.

The majority of III-Nitride based devices (e.g. light emitting diodes, laser diodes and high power transistors) are grown in form of heteroepitaxial structures on (0001)-oriented substrates. A predictive model for strain relaxation by misfit dislocations in these heterostructures is of crucial importance for design and improvement of devices in case on foreign (sapphire, SiC, Si) or native substrates (AlN, GaN). However, for III-Nitride wurtzite films as well as for other hexagonal crystals primary slip-systems are not active for (0001)-oriented biaxially strained heterostructures under plane stress conditions and plastic relaxation via pyramidal slip-systems has typically high kinetic barriers for nucleation and/or glide of dislocations. Though this is beneficial for compressively strained layers since they can be grown much thicker than equilibrium critical thickness theory (Matthews-Blakeslee model) predicts, it might lead easily to cracking in case of tensile strain. So far only little work exists in literature which addresses this problem quantitatively and also considers explicit the dislocation kinetics.
In this paper we present a detailed study on strain relaxation process of (0001)-oriented heterostructures in III-Nitride wurtzite films. For this purpose growth and relaxation mechanisms of Al_xGa_1-xN/GaN heterostructures were investigated by (scanning) transmission electron microscopy and atomic force microscopy. In the first part of the paper we will compare different possible relaxation mechanisms including also cracking and elastic relaxation due to three-dimensional growth while in the second part we will focus on the microscopic mechanism leading to plastic relaxation.
Our studies show that efficient strain relaxation is predominantly caused by formation of new a-type misfit dislocations via the 1/3<11-20>|{0001} slip system. This however requires a redistribution of strain through a three-dimensional film morphology. The presence of strain singularities at e.g. island edges or crack tips then produce a resolved shear stress on the {0001} slip planes and reduces the energetical barrier for homogeneous nucleation of dislocation half-loops and glide of the a-type misfit dislocations into the interface. With the help of our experimental results and finite element method calculations we will present a quantitative model for the growth and plastic relaxation mechanism including also the dislocation formation kinetics. Finally we will determine a critical thickness for the onset of plastic relaxation and compare our results with critical thickness values based on mechanical equilibrium models published in literature and with experimental data.

It has been established that doping of AlN, GaN, and their alloys to achieve high carrier concentrations is a challenge. For the case of Si doped AlGaN, an additional consequence of attempting high doping levels is the increase of tensile strain within the layer upon the increase of silicon concentration. Much work has been done to analyze the origin of strain in Si doped AlGaN, where the increase in tensile stress has been attributed to the inclination of dislocations. It has been proposed that dislocation inclination in Si doped GaN is due to surface-mediated climb facilitated by Ga-vacancies. Therefore, the strain at the onset of doping should be managed by controlling vacancy concentration through its Fermi level dependence. Charged point defects, such as vacancies, can be controlled during growth by external excitation in a steady state condition through the modification of the Fermi level position by using above bandgap UV illumination during growth.
In the present work, it is demonstrated that the strain in Si doped AlGaN films is controlled by manipulating the vacancy formation energy via control of the Fermi level during growth. By implementing above bandgap illumination, the cation vacancy formation energy is increased, thus reducing the concentration of vacancies at the growth surface. The first result of this reduction is the inability for further dislocation inclination by surface mediated climb. By reducing the dislocation inclination the corresponding tensile strain in the doped layer is reduced. Si doped AlGaN layers of 65% Al content were grown on sapphire substrates by metalorganic chemical vapor deposition (MOCVD) with and without UV illumination. Above bandgap illumination with a power density at the wafer of ~1 W/cm² was provided by a mercury arc lamp located at the top of the MOCVD reactor. Optical microscopy results showed that Si doped AlGaN layers grown without UV illumination exhibited severe cracking while the doped AlGaN of identical thickness and Si concentration grown with UV illumination showed no cracks. XRD measurements confirmed that the sample grown without UV contained more tensile strain. Transmission electron microscopy (TEM) results show the expected dislocation inclination upon the addition of mismatch strain by growth of AlGaN on AlN, however, Si doped AlGaN grown under illumination did not exhibit the additional dislocation inclination typically observed upon doping. These results show that dislocation inclination and the associated tensile strain was reduced in the growth of Si doped AlGaN under UV illumination. These results provide a novel technique for strain management through the manipulation of point defect concentration in Si doped AlGaN layers. Thus, point defect control extends beyond the control of compensation effects that reduce carrier concentration in doped films to the reduction of strain in doped AlGaN layers and overall improvement of film quality.

One issue for the reliability of heterostructure devices is the deteriorating impact of dislocations, originating from epitaxy, when a thin layer is grown on a substrate with different lattice parameters. Below a certain layer thickness, called the critical thickness, CT, the layer is grown pseudomorphically on a substrate. Consequently, the layer is strained. When CT is reached, a relaxation of the strain occurs via the introduction of the misfit dislocations (MDs) along the interface between the film and the substrate. Several models have been proposed to calculate CT. Assuming the film and the substrate with the same isotropic elastic properties, the film has a finite thickness (but neglecting the free surface effects), and the substrate is semi-infinite, Freund (F) compared the energy necessary to create a MD with the energy inside the fully strained thin film. The impact of the free surface of the film, and the difference between the elastic constants of the film and the substrate on the CT was discussed by Willis, Jain and Bullough (WJB). Nevertheless, they still worked within the framework of isotropic elasticity. Holec considered the anisotropy of the heterostructure in the CT calculation using the methodology developed by Steeds (S) for the energy of an infinitely long straight dislocation inside an anisotropic medium, ignoring, however, the free surfaces and differences in elastic response of the film and the substrate. We evaluated the impact of elastic anisotropy, difference between the elastic constants of the film and the substrate, and the free surface of the film on the CT calculation. A formula to calculate the energy of a misfit dislocation at the interface of a film with a finite thickness and a substrate with a semi-infinite thickness, when these are modeled anisotropic and with different elastic properties, is proposed. The results are compared with the treatments of S, WJB, F evaluating separately the influence of (i) free surface, (ii) different elastic constants in the film and substrate, and (iii) elastic anisotropy. Our results suggest that starting from homogeneous infinite isotropic medium, inclusion of free surface increases dislocation energy, the difference in elastic constants of the film and substrate has not any significant role (is more than an order of magnitude smaller than the impact of e.g. free surface), the elastic anisotropy decreases the dislocation energy. The novel treatment is used to calculate the CT for Al_xGa_1-xN/GaN, In_xGa_1-xN/GaN and Si_1-xGe_x/Si structures. Our model yields an excellent agreement with the available experimental data in that sense that no MDs are detected below the here predicted threshold. In order to include the dislocation core energy in the evaluation of CT, we compared the pre-logarithmic terms of the analytical models with the corresponding values obtained by atomistic simulations. A good agreement is found with the continuum predictions based on anisotropic elasticity.

Knowledge on dislocation characters and dynamical properties is essentially important as a basis for the control of dislocation generation and deformation during crystal growth and device processing. However, in contrast to diamond and sphalerite structure semiconductors such as Si, GaAs, and so forth, understanding of dynamic behavior of dislocations in wide band-gap semiconductors as typical GaN is far limited for the structure of hcp-based wurtzite.
Here, we summarize our recent results on mechanical strength as hardness obtained by micro and nano-indentation test and yield stress obtained by a conventional compressive deformation method at elevated temperatures for various wide bandgap single crystals of GaN, AlN, InN, ZnO etc., including II-VI sphalerite compounds as ZnSe. From the results, activation energies for dislocation motion are evaluated to be 2-2.7 and 0.7-1.2 eV in GaN and ZnO, respectively. Finally, we progress our comprehensive knowledge on dislocation mobilities and fundamental mechanism in various semiconductors.
Ref. 1. I. Yonenaga et al., Physica B 404 (2009) 4999.

A review will be presented of Synchrotron X-ray Topography and KOH etching studies carried out on n type 4H-SiC offcut substrates before and after n^- homo-epitaxial growth to study defect replication and strain relaxation processes and identify the nucleation sources of both interfacial dislocations (IDs) and half-loop arrays (HLAs) which are known to have a deleterious effect on device performance. We show that these types of defects can nucleate during epilayer growth from five different types of source: (1) from short segments of edge oriented basal plane dislocations (BPDs) in the substrate which are drawn into the epilayer; (2) from segments of half loops of BPD that are attached to the substrate surface prior to growth which glide into the epilayer; (3) from BPD half-loops created at 3C-SiC inclusions in the epilayer; (4) from BPD half- loops associated with micropipes; and (5) from BPD half-loop nucleation occurring at the growth surface. Comparison of measured and calculated critical thicknesses for epilayer relaxation will be discussed in light of the Matthews-Blakeslee [1] type of mechanism. The origin of the mismatch stress is shown to be associated with lattice parameter differences at the growth temperature, arising from the differences in doping concentration between substrate and epilayer. Observations of the nucleation sites for IDs and HLAs in (1) demonstrate that it is not necessary for a BPD to intersect the substrate surface in order for it to be replicated into the homo-epitaxial layer and take part in the nucleation of IDs and HLAs. Those in (2) demonstrate that the conversion of the surface intersections of a BPD half loop into threading edge dislocations (TEDs) does not necessarily prevent it from also becoming involved in the nucleation of IDs and HLAs. Those in (3)-(5) demonstrate that IDs and HLAs can also be nucleated from new BPDs generated in the epilayer during growth. In general, the significance of these observations is that while BPD to TED conversion can eliminate most of the BPD transfer into the epilayer, further mitigation may only be possible by continued efforts to reduce the BPD densities in the substrates by control of temperature gradient induced stresses during their PVT growth. Care also has to be taken to avoid nucleation of 3C inclusions during epilayer growth.
[1] J.W. Matthews, A.E. Blakeslee, J. Cryst. Growth 27, 118 (1974)

Semiconductors and piezoelectric heterostructures has always been in the substantial center of attention due to the growing demand for novel micro/nano-electronic devices appropriate for the applications in energy harvest, optoelectronics, and high-tech photovoltaics. The vast majority of semiconducting-based electronic devices are produced through heteroepitaxial growth and thin film technology thus have significant concentrations of extended crystal defects. Understanding the behavior of defects including threading and misfit dislocations in semiconductor film/substrate heterostructures is necessary in order to produce powerful high efficiency solar cells as well as multifunctional micro-electro-mechanical systems (MEMS). There is a growing experimental evidence that dislocations in piezoelectric semiconductors are electrically active and even highly charged. These charged dislocations act as lines of Coulomb scattering centers, affecting the electron mobility and the device efficiency. Although the role of dislocations on mechanical behavior has been well investigated, the impact of dislocation charge on the elastic and electric fields in piezoelectric and layered heterostructures is not presently understood and the lack of theoretical analyses on charged dislocations is perceptible. We present a continuum-based analytical approach to characterize line defects with charge in layered heterostructures. The methodology is based on the Green&’s functions in the piezoelectric medium and it is applied to analyze the fields induced by three-dimensional (3D) arbitrarily charged dislocation loops in bilayer heterostructures. The contributions from both uncharged-dislocation loop and charge-only sources are clearly separated so that one can easily address the relative importance of these two different types of sources. In addition, the force on the dislocation segments originating from the material stress field and the interfacial image stress has been proved as the key parameter in dislocation dynamics (DD) analysis and continuum plasticity (CP) since it is crucial relating to the dislocation motion, migration, and interactions. Therefore, we calculate the acting force on the dislocation loop since the magnitude and direction of the force on the dislocation core determines its tendency to glide or cross-slip. We conduct the analysis under various slip and twinning systems in hcp III-nitride heterostructures. The presented analytical method emphasizes on the contribution of the charge and its effect on the Peach-Koehler force as well as the importance of the moduli/lattice mismatch at the interface. We demonstrate through a series of calculations that the charge on the dislocation core contribute significantly on the dislocation mobility.

High AlN mole fraction AlGaN quantum wells have attracted attention for applications in DUV LEDs. To increase the external quantum efficiency (EQE) of the near-band-edge (NBE) emission, the use of low threading dislocation (TD) density bulk AlN substrates^1,2) is an essential approach. However, as EQE of DUV LEDs is still limited to 10%,^3,4) internal quantum efficiency (IQE), injection efficiency, and light extraction efficiency must be pursued. Among these, IQE is an intrinsic material talent and is a fraction of radiative rate over the sum of radiative and nonradiative rates; i.e. IQE=(1+tau;_R/tau;_NR)^-1, where tau;_R and tau;_NR are the radiative and nonradiative lifetimes, respectively. To understand what dominates and how we increase tau;_NR and decrease tau;_R, quantitative analysis of the lifetimes is mandatory, especially in view of the structural and point defects. However, due to limited availability of a desirable DUV femtosecond excitation source, only a few papers^5-7) have dealt with the emission dynamics of AlN and there has been no reported result on optical investigation of AlN using spectroscopic tools with sufficiently high spatial and temporal resolution.
In this presentation, we show the results of spatio-time-resolved cathodoluminescence (STRCL) spectroscopy⁸⁾ on various quality AlN epilayers grown on the bulk AlN substrates prepared by physical vapor transport.¹⁾ Reflecting the low TD density (lower than 10⁴ cm^-2), room-temperature CL intensity images mapped at the free A-exciton energy exhibited homogeneous contrasts. Then we show well-resolved low-temperature excitonic CL peaks to verify the exciton binding energy (51.3 meV). Low-temperature CL peaks at 6.0415 and 6.0287 eV, which were polarized parallel and perpendicular, respectively, to the c-axis, exhibited identical risetimes and short lifetimes, and the latter coincided with the temporal delay of neutral donor (Al) -bound exciton emissions. These results support the assumption that the two peaks originate from the recombination of free A-excitons of irreducible representations Γ₁ and Γ₅, respectively.
This work was supported in parts by NEDO by METI and The Asahi Glass Foundation, Japan, and AOARD budget monitored by G. Jessen.
[1] Rice et al., JAP 108, 043510 (2010). [2] Grandusky et al., Solid-State Electron. 78, 127 (2012). [3] See for example Pernot et al., APEX 3, 061004 (2010); Shatalov et al., ibid5, 082101 (2012) and references cited therein. [4] Kinoshita et al., APEX 5, 122101 (2012) [5] Nam et al., APL 82, 1694 (2003). [6] Onuma et al., APL 96, 061906 (2010). [7] Chichibu et al., APL 97, 201904 (2010). [8] Ishikawa et al., APL 101, 212106 (2012).

GaN is considered a model system for group III-N semiconductors. One of the controversial issues in the field of III-N semiconductors is the effect dislocations have on the electronic properties of the GaN epilayers. A-type screw dislocations are the most important threading dislocation in III-N heterostructures grown on nonpolar and/or semipolar substrates. Heterostructures with these orientations allow to reduce the quantum confined Stark effect and thereby to improve the efficiency of light emitting devices. Due to their importance we have therefore studied structure and electronic character of screw dislocations in GaN. For this defect the dislocation core is fully coordinated and electronic states are exclusively induced by its strain field. According to the conventional picture based on deformation potentials the shear strain field of a screw dislocation deforms the p-type VBM states but it does not affect the s-type CBM. Thus, excitons are formed by binding electrons to bound holes by Coulomb interactions.
In the present study we combine Density Functional Theory (DFT) calculations with cathodoluminescence (CL) and photoluminescence (PL) characterization experiments and we investigate the unusually strong excitonic emission at individual a-type screw dislocations in GaN. In order to investigate the electronic structure of both the core as well as the long range strain field of a screw dislocation we developed and applied a quasicontinuum approach that combines first principles calculations with elasticity theory. In a first step the effect of the dislocation core on the electronic structure has been investigated by calculating the electronic structure of a 400 atom supercell containing a pair of mutually compensating screw dislocations. In a second step and in order to address the effect of the long range strain field on the band edges we computed the strain field of an infinite array of screw dislocations in the quadrupol configuration. Then, we considered a series of bulk supercells which are placed at various positions in the strain field and deformed accordingly. The electronic structure of these supecells is then derived by DFT calculations and is used to sampled all relevant strain states in our geometry. Our results, clearly indicate that the shear stress applied by a screw dislocation bends both VBM (upwards) and CBM (downwards) states. A close analysis of our DFT calculations showed that the unexpected downward shift of the CBM is the result of rehybridization of the s-type lowest unoccupied state with the next highest p-like state. Temperature dependent PL and CL spectral map measurements on freshly induced a-type screw dislocation in HVPE grown GaN are in full agreement with this picture: Screw dislocations can be assumed as natural quantum wires that bound excitons and allow for radiative transitions.

Optical polarization properties of m-plane Al_1-xIn_xN epilayers grown by metalorganic vapor phase epitaxy were investigated both theoretically and experimentally. Approximately 500~600-nm-thick Al_1-xIn_xN epilayers subsequently grown on a 1-mm-thick GaN homoepitaxial underlayer on an m-plane freestanding GaN substrate prepared by hydride vapor phase epitaxy were examined.
At first, high-resolution x-ray reciprocal space mapping (X-RSM) measurement was carried out to quantify anisotropic strains of the Al_1-xIn_xN epilayers. Because of m-plane growth, a- and c- axes were anisotropically strained when a coherent growth was maintained. The X-RSM analysis confirmed that the Al_1-xIn_xN films of 0.15 < x le; 0.32 were fully strained, and the ones of x le; 0.15 were partially relaxed. We note that a- and c- axes of Al_1-xIn_xN alloys match those of GaN for x = 0.177 and x = 0.282, respectively.
Theoretical investigation was then carried out to correlate the anisotropic strains and optical polarization characteristics. For simplicity, the m-plane Al_1-xIn_xN layers were assumed to be fully strained. Valence band structures including the energies and polarization oscillator strengths were computed based on the k#8729;p perturbation technique. The results implied that valence bands were significantly mixed and showed anticrossing behavior at x = 0.17 and x = 0.20. Actually, there are two important x values for crystal symmetry in this regime. One is x = 0.177, where a-axis of Al_1-xIn_xN matches to that of GaN. Another particular x is 0.195, where the strains in a- and m-axes become the same. In this case, both the axes have slight compressive strains of -0.25% and crystal symmetry of Al_1-xIn_xN is C_6v, while it is C_2v for x ne; 0.195.
Consequently, the polarization oscillator strength of the optical transition between the conduction band and the topmost valence band was drastically changed at around x = 0.17. For x < 0.17, the oscillator strength had finite value only for the direction along m-axis and other components of the oscillator strength were nearly zero, where we represent this polarization as E//m in the usual manner. In this case, light extraction along surface normal might be difficult. For x > 0.17, on the other hand, most of the oscillator strength was focused along c-axis (E//c) and surface emission could be allowed. We will compare the experimental results with theoretically predicted optical characteristics described above in the conference.

High Al-content AlGaN alloys are attractive for fabrication of deep-UV optoelectronic devices such as light emitting diodes (LEDs) and laser diodes (LDs). Nitride-based LEDs and electrically pumped LDs have been demonstrated with wavelengths as short as 210 nm and 354.7 nm, respectively. Optically pumped laser cavities have been demonstrated with wavelengths as low as 242 nm but these had a high threshold pump power of 1.2 MW/cm². Lower threshold pump powers can be achieved by using low dislocation density (<10⁴ cm^-2) bulk AlN substrates. A low threshold pump power and high quantum efficiency are prerequisites for the realization of electrically pumped lasers.
We have carried out an extensive study on Al_xGa_1-xN/Al_yGa_1-yN multiple-quantum well (MQW) and Al_xGa_1-xN/AlN double heterojunction (DH) structures grown on AlN substrates that were targeted to emit between 235 nm and 280 nm. The structures were grown on 500 nm of pseudomorphic undoped Al_xGa_1-xN on AlN substrates using low-pressure metal-organic chemical vapor deposition (MOCVD). Well width, barrier width, and V/III ratio were varied to determine their dependence on lasing characteristics. Laser cavities were formed by cleaving wafers into 0.2 - 1.5 mm wide stripes. Narrow cavities were realized by first thinning the samples to 0.15 mm. They were characterized at room temperature (RT) and at 3 K (LT) using a 193 nm ArF excimer laser where the lowest RT pump power threshold was 50 kW/cm². All lasers had a peak FWHM ranging from 0.3 nm to 0.8 nm above threshold. There was a relatively small difference in RT and LT threshold indicating that the structures have high internal quantum efficiency. Longitudinal modes with line widths as narrow as 0.014 nm were observed where the mode separation was directly dependent on the cavity width. The stimulated emission was transverse electrically (TE) and transverse magnetically (TM) polarized for wavelengths above and below 245 nm respectively. It was found that a higher V/III ratio during growth resulted in brighter emitters with lower lasing thresholds. This is attributed to a reduction in point defect incorporation during growth. In conclusion, deep UV-C photo-pumped lasers have been realized on bulk AlN substrates with low threshold pump power. These promising results will further aid the fabrication of deep UV-C LDs in the near future.

Strain-induced piezoelectric fields within GaInN quantum wells grown on polar GaN are blamed as a potential reason for the fairly low efficiency of green LEDs. Therefore, non- and semipolar growth directions gained special attention over the recent years leading to reduced or even completely absent internal electric fields in such structures. Besides using semipolar GaN substrates cut from thick c-plane wafers, which are still very small in size and very expensive, foreign wafers like sapphire cut in non-c directions can be taken to achieve this goal. However, such less polar structures typically contain large dislocation and stacking fault densities. Therefore, we have concentrated on an approach, where metalorganic vapour phase epitaxial (MOVPE) growth is initiated by nucleating on inclined c-plane-like side-facets prepared by etching grooves into adequately oriented sapphire wafers. By this approach, we are still driving the growth mainly in the conventional c-direction. After coalescence of such initially striped nitride structures, they form large area planar semipolar surfaces on which GaInN quantum well structures can be grown. Following approach, several semipolar planes including (10-11), (11-22), and (20-21) can be produced on n-plane (11-23), r-plane (10#8209;12), and s-plane (22-43) sapphire wafers, respectively. By carefully optimizing the growth conditions and applying various defect reduction methods, we were able to decrease the dislocation and stacking fault densities substantially. Excellent structural properties have been obtained, being evident from very narrow X-ray diffraction peaks in the range of 200 arcsec and photoluminescence spectra dominated by the excitonic peaks. Only very weak stacking fault related signals were visible, probably due to the lateral overgrowth of defect-rich areas in our stripes from the neighbour stripes. Overgrowth by hydride vapour phase epitaxy typically leads to further improved layer qualities. GaInN quantum well structures grown on such semipolar (10-11) and (11-22) layers show strong luminescence at about 500 nm indicating that also on such planes large amounts of In can be incorporated. It should be noticed that even more In is needed for such semipolar quantum wells as compared to polar c-plane structures, because of the significantly reduced quantum confined Stark effect. First LED test structures show fair electroluminescence, although the doping profile of these structures is not yet optimized. In this contribution, we will mainly address the defect reduction mechanisms which we observed. Moreover, we have analysed the anisotropic strain and the resulting bow of our layers as a consequence of the growth conditions. Thermally induced strain may be helpful for the self-separation of thick HVPE-grown layers, whereas a too large bow would be a problem for later surface polishing.

Deep-UV (DUV) emitters have numerous applications. But most of the current DUV emitters like excimer lasers have large size. Recently, the III-N DUV emitters have drawn great attention due to suitable direct bandgap which can lead to compact size. Recently, low-threshold optically-pumped DUV lasers containing AlGaN multiple-quantum wells (MQWs) were demonstrated by using c-plane bulk AlN substrates. The AlN substrates have low-dislocation density and can reduce lattice mismatch, thus leading to active regions with low-dislocation density. But because of the high cost for the AlN substrates, it is more desirable to grow DUV lasers on inexpensive sapphire substrates.
In this abstract, we present stimulated emission from optically-pumped AlGaN MQW DUV lasers grown on (0001) sapphire substrates. The lowest threshold were obtained from the lasing at 249 nm (“the 249-nm laser”) and 256 nm (“the 256-nm laser”).
The AlGaN MQW laser structures were grown on c-plane sapphire substrates in a 3×2” MOCVD reactor. The laser structure firstly comprised an AlN template layer deposited on the sapphire substrate. The total dislocation density of the template layer was estimated to be 2.5×109/cm2 by TEM experiments, which represents one of the lowest dislocation densities reported for planar AlN/sapphire templates. The RMS surface roughness was 0.12 nm determined by 5×5 #61549;m2 AFM scans. Thus the template layer provided a relatively good crystal quality and smooth surface for growth of the AlGaN structure.
An AlGaN waveguide layer, AlGaN MQWs and a thin AlGaN cap layer were grown on the template layer sequentially. All the epitaxial layers were pseudomorphically grown confirmed by XRD (105) RSM. The 5×5 #61549;m2 AFM scans show RMS roughness of the finished structure was 0.55 nm, which was close to that of a comparable laser structure grown on an AlN substrate and thereby suggest relatively low dislocation density. The wafers were cleaved into laser cavities by mechanical scribing from back side of the substrate. The cavities were optically pumped at room temperature by a 193-nm excimer laser.
Power-dependent PL spectra of the 249-nm laser and 256-nm laser were obtained. The difference of the emission wavelengths between the two lasers was due to normal shift of sample condition across a two-inch diameter wafer grown by the MOCVD. The 249-nm laser and the 256-nm laser demonstrated thresholds of 90 kW/cm2 and 61 kW/cm2, respectively. Spectral linewidth of both the 249-nm laser and the 256-nm laser reduced with increasing pumping power density and reached 1.6 nm, indicating stimulated emission. The thresholds are more than an-order-of-magnitude lower than the previously-reported optically-pumped DUV laser grown on 4H-SiC substrates. In addition, the thresholds are comparable with the reported state-of-the-art optically-pumped AlGaN DUV lasers grown on bulk AlN substrates lasing at 266 nm, suggesting excellent candidacy of sapphire substrates for III-N DUV laser diodes.

Bulk AlN crystals with high structural perfection are considered as most promising substrate material for optoelectronic deep UV devices based on AlGaN layers with high Al content. From today&’s point of view, only bulk growth by Physical Vapour Transport (PVT) can provide crystals with highest crystalline quality (dislocation densities < 10⁴ cm^-2, no small-angle grain boundaries), which result in high quantum efficiencies of the UV-C device if this structural perfection can be kept in homoepitaxy.
In this work, we report on the PVT growth of high quality single crystalline AlN with continuous crystal enlargement and on the effect of the impurity/dopant concentrations of O, C, and Si on the optical and electrical properties of the bulk AlN crystals.
The developed growth technique comprises (i) the spontaneous nucleation of freestanding AlN crystals for the initial seed preparation and (ii) subsequent growth runs on AlN seeds cut from such crystals. The structural quality of the grown material is evaluated by rocking curves, X-ray Lang topography, and defect-selective wet chemical etching. Currently, up to Oslash;15 mm crystals with rocking FWHM values between 13 and 21 arcsec across the whole surface area are grown by subsequent growth experiments.
The impurity concentrations are investigated by secondary ion mass spectrometry (SIMS). In conjunction with UV-VIS spectrometry and Fourier-transform infrared (FTIR) absorption measurements a tri-carbon defect complex is correlated with a strong absorption band at 265 nm which impairs the application relevant deep UV transparency. The appearance and quenching of this absorption band will be discussed in dependence of the Fermi level which is correlated to the ratio of [C] to [O] + [Si]. Furthermore, the free carrier concentration in the samples is investigated in dependence of different growth facets by temperature dependent optical transmission imaging.
We will show that the impurity concentrations and thus the electical and optical properties can be controlled to some extent by adjusting the growth temperature or the thermal field-induced crystal faceting. Also, we have successfully grown Si-doped AlN bulk single crystals by adding silicon containing ceramics to the source material. These samples show a weakly n-type electrical behavior as evidenced by Hall effect and resistivity/admittance measurements.

Introduction
Acidic ammonothermal method is one of the most promising techniques which enables the mass production of large diameter bulk GaN crystal.High power light emitting diodes and laser diodes are now fabricated on GaN substrates made by hydride vapor phase epitaxy, but to realize electrical devices such as a vertical GaN-based power switching device, large-diameter GaN substrates are essential.Acidic ammonothermal method is expected for high purity, strain free, low dislocation, large scale and low cost bulk GaN manufacturing, similar to hydrothermal method for quartz crystal mass production.
We have been studying the characteristics of supercritical NH₃ using ammonium halides as mineralizers, and succeeded in growing the GaN crystals faster than 100mm/day along Ga-polar direction ^[1]. We have tried in-autoclave gas phase mineralizer synthesis ^[2], metal Ga nutrient ^[3] and also found that acidic ammonothermal crystals can be used for substrates for epitaxy ^[4].
To realize large scale ammonothermal GaN production, the pressure is one of the most important issues.The capacity of the superalloy autoclave depends on the operating pressure, because the size of the Ni-Cr material is limited.
Experimental procedure
The solubility was measured by weight loss method.The ammonothermal growth of GaN crystals was carried out using 4 kinds of ammonium halide mineralizers. To prevent the Ni-based superalloy autoclave from corroding, the inner wall was covered with Pt. A circular Pt baffle plate with holes was placed at the center of the autoclave to separate the growth region and nutrient region. After charging the precursor and the seed crystals, ammonium halide powder was added. Then, filter-purified NH₃ gas was fed into the autoclave. The charging amount was controlled to obtain the designated pressure for the temperatures during the growth period. The autoclave was heated using a two-zone vertical furnace.
Results and discussion
Because of the small inner diameter of the high pressure tubing, it is difficult to keep high purity environment only by vacuum deaeration.We have found that the deoxidizer is effective to grow high purity crystals.
The crystal quality and the growth rate strongly depend on mineralizer species.We have also studied the dependence on temperature and pressure, and found it possible to control growth direction by optimizing the growth condition.Finally, we have found that the growth rate faster than 1000mm/day can be achieved. Based on these studies and optimization, we have succeeded in growing high purity crystal with x-ray rocking curve FWHM less than 25arcsec at the pressure condition around 100MPa.
[1] D. Tomida et al., J.Cryst.Growth 353, 59 (2012).
[2] D. Tomida et al., J.Cryst.Growth 348, 80 (2012).
[3] Q. Bao et al., Cryst.Eng.Comm. 14, 3351 (2012).
[4] S. F. Chichibu et al., Appl. Phys. Express 4, 045501 (2011).