preview all symposia



Start atSubject View AllNum.Add
Plenary : Bernard Gil
Authors : Chris G. Van de Walle
Affiliations : Materials Department, University of California, Santa Barbara, California, USA

Resume : The adoption of nitride semiconductors has progressed startlingly fast, from initial discoveries to widespread applications. This is not because the materials are simple and trouble-free; to the contrary, throughout their history they have confronted us with physics problems that were known to be nuisances in other materials but could have been showstoppers for the nitrides. The issues include lattice-mismatched heteroepitaxy, dopant passivation, shallow-versus-deep character of acceptors, the role of polarization fields, and nonradiative recombination processes such as Auger and defect-assisted Shockley-Read-Hall. These problems have forced us to develop new understanding, resulting in insights into physical mechanisms that will now benefit other semiconductors and encourage us to explore other materials, such as the II–IV–V2 analogues of nitrides. Now we are entering the era of ultrawide-band-gap semiconductors, with band gaps significantly wider than the 3.4 eV of GaN, which will enable deep-UV optoelectronics and have compelling advantages in high-power and RF electronics, quantum information science, and extreme-environment applications. Building on what we have learnt, I will highlight areas in which further progress could have high payoffs. Work performed in collaboration with A. Alkauskas, C. Dreyer, A. Janotti, E. Kioupakis, J. Lyons, L. Gordon, J. Shen, J. Varley, D. Wickramaratne, and Q. Yan, and supported by DOE and NSF.

Localization, LED physics : Aurélien David
Authors : Aldo Di Carlo(a), Matthias Auf der Maur (a), Alessandro Pecchia (b)
Affiliations : (a) Dept. Electronics Engineering, University of Rome “Tor Vergata”, Rome (Italy) (b) CNR-ISMN, Rome (Italy)

Resume : Although current phosphor-converted white GaN-based LEDs have high electricity-to-light conversion efficiencies, it has been pointed out that the full potential of solid state lighting could be exploited only by color mixing approaches without employing phosphor-based wavelength conversion. Such an approach requires direct emitting LEDs of different colors, including, in particular, the green-yellow range of the visible spectrum. This range, however, suffers from a systematic drop in efficiency, known as the “green gap,” whose physical origin has not been understood completely so far. In this talk, we present advanced Multiscale/Multiphysics simulation approaches to design and analyze GaN LED devices. In particular, we will show by atomistic simulations that a consistent part of the green gap in c-plane InGaN/GaN-based LEDs may be attributed to a decrease in the radiative recombination coefficient with increasing indium content due to random fluctuations of the indium concentration naturally present in any InGaN alloy.[1] Moreover, we show that the unexpected temperature behavior measured recently [2] on high quality MQW samples can be quantitatively described by alloy fluctuations, supporting our assumption that alloy fluctuations are playing an important role in III-nitride LEDs. References: 1) M. Auf der Maur et al. Phys. Rev. Lett. 116, 027401 (2016) 2) F. Nippert et al., Applied Physics Letters 109, 161103 (2016)

Authors : Shuan Wang (1), Chi-Kang Li(1), Marcel Filoche(2), Claude Weisbuch (2,3), James S. Speck(3), and Yuh-Renn Wu(1)
Affiliations : (1) Graduate Institute of Photonics and Optoelectronics and Department of Electrical Engineering, National Taiwan University, Taipei 10617, Taiwan (2) Laboratoire de Physique de la Matière Condensée, CNRS-Ecole Polytechnique, 91128 Palaiseau Cedex, France (3) Materials Departments, University of California, Santa Barbara, California 93106, USA

Resume : In this paper, we presented a three-dimensional simulation of GaN-based green LEDs by considering alloy fluctuations. In GaN-based multiple quantum well LEDs, the alloy fluctuations of indium in InGaN layers have significant effects to the properties of LEDs due to the strong localization effects[1]. We applied 3D model considering alloy fluctuations to solve the potential, current and emission properties of LEDs. To consider random alloy fluctuations, we solve local strain, polarization and potential. The local strain and polarization are solved using 3D finite element method strain solver. A self-consistent 3D FEM Poisson, drift-diffusion and Schrodinger localized landscape solver[2] is used to solve the effective quantum potential and carrier distribution. The Schrodinger localized landscape model is solving Schrodinger equations with Anderson localization theory. To further understand the emission properties, we calculate the effective ground eigenstates in quantum well regions with the result of localized landscape model. The quantum efficiency and emission spectrum of the carrier recombination are calculated. With this model, we also study influence of band engineering to the LED performance. AlGaN is used to partially replace GaN as a part of the quantum barrier. Quaternary AlInGaN appears at the edge of InGaN and AlGaN due to diffusion. The calculations of strain, polarization and potential of the AlInGaN are introduced to the simulation model. The design of gradual indium composition in quantum well is also simulated. The transport physics in these random alloy fluctuation especially the quaternary alloy barrier will be presented in detail in this paper. [1] Wu, Yuh-Renn, et al. "Analyzing the physical properties of InGaN multiple quantum well light emitting diodes from nano scale structure." Applied Physics Letters 101.8 (2012): 083505. [2] Arnold, Douglas N., et al. "Effective confining potential of quantum states in disordered media." Physical review letters 116.5 (2016): 056602.

Authors : Piotr A. Dróżdż(1,2), Krzysztof P. Korona(1), Marcin Sarzyński(2), Tadeusz Suski(2), Czesław Skierbiszewski(2), Henryk Turski(2)
Affiliations : (1) Faculty of Physics, University of Warsaw, Pasteura 5, 02-093 Warsaw, Poland; (2) Institute of High Pressure Physics “Unipress”, Polish Academy of Sciences, Sokołowska 29/37, 01-142 Warsaw, Poland;

Resume : The subject of In segregation in InGaN quantum wells (QW) is important for understanding the optical properties of nitride optoelectronic devices. In this work, the new model of radiative recombination in semiconductor layer with potential fluctuations is proposed. Our model assumes Gaussian shape of density of states (DOS) in fluctuating potential and continuous delocalization of carriers with increasing temperature. The distribution of exciton gas is characterized by effective temperature. Based on these assumptions, we have obtained equations describing energy vs. temperature S–shaped dependence and exciton lifetime dependence on temperature. Two samples were studied in order to verify our model with experiment. First has been grown using MOCVD technique and was composed of many regions with different indium content in its quantum wells. The second sample set, with identical QW structure, has been grown by MBE. It has regions of different indium content obtained due to different growth rates. Time–resolved micro–photoluminescence measurements were performed on these two sets of samples and the results were compared with the proposed model. We determined such parameters as width of DOS distribution, s, and delocalization energy, Ea. It has been confirmed that our formulas describe well the emission energy and PL lifetime data between 4 K and 300 K, for samples grown by different methods and containing various amount of indium in the InGaN quantum wells.

Authors : Liverios Lymperakis 1, Christoph Freysoldt 1, Tobias Schulz 2, Martin Albrecht 2, Jörg Neugebauer 1
Affiliations : 1 Computational Materials Design department, Max-Planck-Institut für Eisenforschung GmbH, 40237 Düsseldorf, Germany; 2 Leibniz-Institut für Kristallzüchtung, Max-Born-Straße 2, 12489 Berlin, Germany

Resume : The growth of high quality InGaN alloys with high In content to achieve green emission, is challenging: Its constituents, GaN and InN, exhibit a high lattice mismatch and a large difference in the formation enthalpies. Furthermore, in epitaxial growth, surface reconstructions may strongly influence the alloy composition and distribution. Employing DFT calculations we investigated the growth of single ML thick InGaN layers. For growth under N rich conditions the (0001) surfaces reconstruct in order to obey the electron counting rule. The later results in a 2×2 N adatom reconstruction which leaves ¾ of the surface metal atoms ×4 coordinated in a sp3 configuration and ¼ of them ×3 coordinated in a planar sp2 configuration. Although, the stronger Ga-N bonds are expected to result in preferential incorporation of In at the low coordinated surface sites, our calculations reveal a novel reconstruction mechanism that contradicts this conventional picture: Due to their large atomic radius, In atoms at the triply coordinated sites are hindered from re-hybridizing to the sp2 configuration. Instead, they preferentially incorporate to ×4 coordinated surface sites where re-hybridization of the triply-coordinated Ga is allowed and results in ?1 eV lower energy. Furthermore, efficient in-plane strain accommodation favors spatial distribution of In atoms as 2nd nearest neighbors aligned along the < 1-100> direction. Both mechanisms act simultaneously and result in a 2?3×2?3 reconstruction for the In0.25Ga0.75N layers. Based on these insights we will elaborate the influence of these mechanisms on ordering, growth temperatures and stoichiometric limits and we will discuss in details recent TEM analysis findings on MBE grown short period InGaN/GaN SLs.

Authors : S. Yu. Karpov
Affiliations : STR Group - Soft-Impact, Ltd., P.O.Box 83, 27 Engels ave., St.Petersburg, 194156 Russia

Resume : More than for decade, the “green gap” problem remains one of the principal obstacles for development of efficient long-wavelength, i.e. green, yellow, and orange, InGaN-based light-emitting diodes (LEDs). “Green gap” means substantial reduction in the emission efficiency of the LEDs from blue to green/orange spectral range of ~500-600 nm. Two natural mechanisms were regarding as potential origins of the “green gap” for a long time. One attributed the LED efficiency decline in the “green gap” to increasing defect density in InGaN quantum wells (QWs) at higher indium contents necessary for long-wavelength light emission. Another mechanism invoked quantum-confined Stark effect (QCSE) due to polarization field embedded in InGaN QWs: this field separated electron and hole wave functions, leading to lowering the emission efficiency at high indium contents in the wells. Recent experimental estimations of the wavelength-dependent Shockley-Read-Hall recombination coefficients in industrial-grade LED structures enabled one, however, to withdraw the former mechanism from the list of potential origins of the “green gap” [1,2]. In turn, the latter mechanism was found to disagree with observations of pronounced “green gap” in the efficiency of non-polar and semi-polar LEDs with QCSE either completely eliminated or substantially reduced. Recently, implication of carrier localization by composition fluctuations in InGaN alloys into the “green gap” problem has been demonstrated by atomistic simulations [3] and those made by a simple semi-empirical model [2,4]. The atomistic approach considered the radiative recombination coefficient (RRC), attributing ~70% of the LED efficiency reduction in the “green gap” to QCSE and ~30% to the effect of uncorrelated electron and hole localization by compositional fluctuations in InGaN [3]. The semi-empirical model developed for bulk InGaN alloys [2,4] reproduced qualitatively well the experimental wavelength dependence of both radiative and Auger recombination coefficient (ARC) from [1], explained anomalous (ascending) temperature dependence of RRC reported in [2] but failed in prediction of the absolute values of the recombination coefficients. The latter was attributed to neglecting the QCSE in the polar LED structures. In this study, the above semi-empirical model has been extended to the case of polar InGaN QWs. The model is found to be capable of providing quantitative predictions for both RRC and ARC, retaining agreement with the experimental trends of their dependence on the emission wavelength. Here, holes are assumed to be strongly localized by composition fluctuations due to their heavy effective masses, whereas electrons are regarded to be completely delocalized. As a result, the overlap between the electron and hole wave functions acquires a three-dimensional rather than one-dimensional character specific for QCSE. The additional in-plane localization of holes reduces considerably the overlap integral making it to depend on the hole localization radius. In turn, the hole localization radius, decreasing with the emission wavelength because of disorder in InGaN alloys enhanced with the indium content, is found to affect both RRC and ARC. So, the dependence of the localization radius on the indium content/emission wavelength becomes the major factor determining the LED efficiency reduction in the “green gap”. Using the predicted RRC and ARC and the Shockley-Read-Hall recombination coefficients borrowed from [1,2], the wavelength dependence of the maximum internal quantum efficiency (IQE) has been simulated. The obtained results are found to be in reasonable agreement with the data estimated independently from the maximum values of external quantum efficiency reported for various LEDs in a wide spectral range. The agreement is the evidence for the fact that the carrier localization is indeed the important factor controlling the efficiency of InGaN-based LEDs in the “green gap”, at least, at the currents close to those corresponding to the efficiency maximum. Finally, the suggested model has been applied to predict the efficiency of non-polar/semi-polar LEDs with vanishing built-in polarization field. Assuming the same materials quality (i.e. the same Shockley-Read-Hall recombination coefficient), the maximum IQE of such LEDs has been calculated as a function of emission wavelength. In the absence of QCSE, the calculations predict nonetheless a pronounced “green gap” for the non-polar/semi-polar LEDs too. Comparison of the results with those obtained for the polar LEDs has shown that more than 50% of the efficiency reduction in the “green gap” can be attributed to the impact of carrier localization. The reported results refine current understanding of the role of carrier localization by composition fluctuations in InGaN-based LEDs. Nevertheless, there still exists a big room for further studies. One of the important issues for this is experimental evaluation of the localization energy, which is directly related to the localization radius of holes, and its dependence on the indium content/emission wavelength. Another important question to be addressed in future is distinguishing between the localization caused by compositional fluctuations in InGaN alloys and local fluctuations of the QW widths. [1] D. Schiavon, et al., Phys. Stat. Solidi B, 250, 283 (2012). [2] F. Nippert, et al., Appl. Phys. Lett. 109, 161103 (2016). [3] M. Auf der Maur et al., Physical Review Letters 116, 027401 (2016). [4] S. Yu. Karpov, accepted in OSA Photonics Research (2017).

Authors : Kazunobu Kojima1, Hirotaka Ikeda2, Kenji Fujito2, Shigefusa F. Chichibu1
Affiliations : Tohoku University1, Mitsubishi Chemical Corporation2

Resume : For realizing GaN-based power devices with a high breakdown voltage, recent development of bulk GaN crystals is remarkable. Point defects become important to characterize the high quality GaN, because structural defects like threading dislocation (TD) and stacking fault are nearly extinguished. The authors' group has been investigating the correlation between the near-band-edge (NBE) photoluminescence (PL) lifetime at room temperature and the concentration of intrinsic nonradiative recombination centers (NRCs) in n-GaN, of which origin is point defect complexes containing VGa, by combining positron annihilation spectroscopy (PAS) and time-resolved PL methods [Chichibu, et al., APL86, 021914 (2005)]. However, PAS becomes less sensitive below the concentration of VGa ([VGa]) of 10^15 cm^-3. Thus the development of an alternative way to detect dilute point defects (<10^15 cm^-3) in high quality GaN crystals is essential. In this presentation, we will present the quantification of absolute quantum efficiency of radiation (AQE) by employing the omnidirectional PL (ODPL) methodology to determine internal quantum efficiency (IQE) of the emission in GaN crystals with different excitation conditions. A high AQE of 8.22% corresponding to IQE of 70.9% was obtained at room temperature for the NBE emission of a freestanding-GaN crystal grown by hydride vapor phase epitaxy on a GaN seed crystal manufactured with the acidic ammonothermal method, when cw photo pumping density was 66 W/cm^2.

Authors : Dan Ritter (1), Chris G. Van de Walle (2)
Affiliations : (1) Department of Electrical Engineering, Technion- Israel Institute of Technology, Haifa, Israel (2) Materials Department, University of California, Santa Barbara, California 93106-5050, USA

Resume : The bound surface polarization charge in bulk polar crystals is compensated by an opposite charge to avoid excessive potential buildup. Since surfaces of polar crystals such as GaN are not conductive, it is believed that the compensating charge resides in localized surface states. However, UV assisted CV measurements on GaN metal-insulator-semiconductor capacitors reported in the literature detect a density of surface states lower by a factor of at least four than that required to compensate accepted values (2x10^13 q/cm^2 ) of polarization charge. Moreover, recent theory by Dreyer et al. indicates that the value of the bound polarization charge in GaN is much higher, about 8x10^14 q/cm^2. This implies that the concentration of compensating charge, if manifested in the form of electrons in surface states, would be far beyond the Mott transition and should be conductive. We conclude that polarization must somehow induce fixed charge at the surface, not residing in surface states; we will discuss some options for its origin. This fixed surface charge may also compensate the 2DEG charge in the AlGaN/GaN system, resolving modeling difficulties of MIS HFET structures when the surface donor approach is used. It is worth noting that the presence of the 2DEG at epitaxial interfaces indicates that no fixed charge is induced at epitaxial interfaces.

Start atSubject View AllNum.Add
Tuesday poster : Bruno Daudin
Authors : Amina Benzina1, 2, Abdelhadi Lachebi1, Ahmad Shuhaimi2, Saadah Abdul Rahman2, Hamza Abid1.
Affiliations : 1Applied Materials Laboratory, Research Center, Djillali Liabes University of Sidi Bel Abbes, 22000, Sidi Bel Abbes, Algeria. 2Low Dimensional Materials Research Centre (LDMRC), Department of Physics, Faculty of Science, University of Malaya, 50603 Kuala Lumpur, Malaysia.

Resume : Numerical simulation based on FP-LAPW calculations is applied to study the structural and optoelectronic properties of wurtzite InxGa1-xN matched on GaN substrate using 16-atom supercell. The generalized gradient approximation of Wu and Cohen (WC-GGA) and the modified Becke?Johnson potential were used for better reproduction of the band structure and optical properties for binary and ternary. Whenever possible, results were compared with experiments and calculations completed with other computational schemes. Keywords: FP-LAPW, InxGa1-xN, TB-mBJ

Authors : Tomonori Ito, Toru Akiyama, Kohji Nakamura
Affiliations : Department of Physics Engineering, Mie University

Resume : Using thin film growth technique, AlN with metastable 4H and graphite-like hexagonal (G) structures can be fabricated by appropriately selecting the substrates and their orientations. On the 4H-SiC(11-20) substrate, 4H-AlN(11-20) were found depending on the III/V ratio. G-AlN nanosheets with smaller energy band gap to open a new field of nanoelectronic device applications were successfully grown on Ag(111). Despite the importance of the polytypism in AlN, there have been very few systematic studies for the structural stability of AlN even in bulk state. In this study, we investigate the polytypes of AlN using ab initio calculations and simple energy formula to interpret the origin of these polytypes. The calculated results imply that AlN strongly favors the 2H. However, the calculated energy differences among 3C, 6H, 4H, and 2H structured AlN with Al-vacancy (VAl) and N-vacancy (VN) suggests that the 4H becomes competitive with the 2H in AlN with VN. Moreover, it is found that Al-rich condition producing large amount of VN tends to favour the 4H-AlN, while the 2H-AlN is preferable under N-rich condition suppressing VN formation. This is consistent with experimental results. The stability of G-AlN is also discussed in terms of bond order and bond length comparing the other III-N such as BN, GaN, and InN.

Authors : Toru Akiyama, Harunobu Nakane, Motoshi Uchino, Kohji Nakamura, Tomonori Ito
Affiliations : Department of Physics Engineering, Mie University

Resume : The structures and polarity of III-nitrides are of importance for the epitaxial growth of III-nitrides on various substrates such as sapphire and SiC, since the polarity causes different surface morphologies, chemical properties, and growth kinetics. The structures and polarity of III-nitrides on SiC substrates have been intuitively understood in terms of the dipole contribution and charge neutrality between III-nitrides and SiC interface. However, quantitative analysis for the polarity of III-nitrides on substrate incorporating the epitaxial growth conditions such as temperature and pressure is still lacking. Here, we theoretically analyze the polarity of III-nitride using absolute surface and interface energies. Surface and interface phase diagram calculations as functions of temperature and pressure are demonstrated for polar AlN on 6H-SiC(0001) substrates as well as polarity inversion of AlN and GaN on N-polar AlN(000-1) substrate. The calculations reveal that there are four types of structures stabilized depending on temperature and Al pressure. The H-terminated Al-polar surfaces with substitutional Al atoms at AlN/SiC interface are found to be stabilized below 1380 K, while the Al-polar surface with N adatom on ideal AlN/SiC interface are favorable for temperature beyond 1380 K.These results suggest that the stability of interface between III-nitride and substrate is crucial for epitaxial growth of III-nitrides.

Authors : A. Redondo-Cubero1, E. David Bosne2, U. Wahl2, P. Miranda3, M.R. da Silva4, J.G. Correia2, K. Lorenz5
Affiliations : 1Department of Applied Physics, Universidad Autónoma de Madrid, 28049 Madrid, Spain; 2Centro de Ciências e Tecnologias Nucleares, Instituto Superior Técnico, Universidade de Lisboa, 2695-066, Bobadela LRS, Portugal; 3Universidad Tecnológica Metropolitana, Las Palmeras, Santiago de Chile, Chile; 4Centro de Física Nuclear, Universidade de Lisboa, 1649-003 Lisboa, Portugal; 5IPFN, Instituto Superior Técnico, Universidade de Lisboa, Estrada Nacional 10, 2695-066, Bobadela LRS, Portugal

Resume : Due to lattice mismatch between the different components, nitride heterostructures are usually subject to strain. While strain is frequently analyzed by means of X-ray diffraction (XRD), an excellent complement to XRD is Rutherford backscattering spectrometry of MeV 4He particles under channeling conditions (RBS/C). An added advantage is that RBS also directly provides the atomic composition of the layers independently from the validity of Vegard?s law [1,2]. The availability of 2-dimensional position-sensitive detectors for RBS/C has meanwhile opened a new and more detailed way to extract this structural information. In this work we analyze the accuracy of RBS/C for the determination of strain in AlInN/GaN heterostructures with different strain states (?, from -1% to 1%) and thicknesses (from 50 to 250 nm). The blocking patterns of 2 MeV 4He particles from In in AlInN and Ga in GaN in the vicinity of the < -2113> axis were recorded with a position sensitive Timepix pixel detector developed at CERN [3]. The 4He trajectories were modeled with Monte Carlo simulations, which takes into account steering phenomena at the heterostructure interface [1,2]. The strain in the AlInN layer can then be derived from a shift of the In blocking pattern in AlInN with respect to the Ga blocking pattern in GaN. We demonstrate the good agreement between RBS/C experiment, Monte Carlo simulations and XRD strain data, showing the general applicability of this method. [1] K. Lorenz et al., Phys. Rev. Lett. 97 (2006) 85501 [2] A. Redondo-Cubero et al., Appl. Phys. Lett. 95 (2009) 051921 [3] J. Jakubek et al., Nucl. Instr. Meth. A 591, 155 (2008)

Authors : Nasir Alfaraj, Somak Mitra, Feng Wu, Idris A. Ajia, Bilal Janjua, Aditya Prabaswara, Renad A. Aljefri, Haiding Sun, Tien Khee Ng, Boon S. Ooi, Iman S. Roqan, Xiaohang Li
Affiliations : King Abdullah University of Science and Technology, Thuwal 23955-6900, Kingdom of Saudi Arabia

Resume : The photoinduced entropy of InGaN/GaN p-i-n nanowires was investigated using temperature-dependent (6–290 K) photoluminescence. We also analyzed the photocarrier dynamics in the InGaN active regions using time-resolved photoluminescence. An increasing trend in the amount of photoinduced entropy of the system above 250 K was observed, while we observed an oscillatory trend in the generated entropy of the system below 250 K that stabilizes between 200 and 250 K. Strong exciton localization in indium-rich clusters, carrier trapping by surface defect states, and thermodynamic entropy effects were examined and related to the photocarrier dynamics. We conjecture that the amount of photoinduced entropy of the system increases as more non-radiative channels become activated and more shallowly localized carriers settle into deeply localized states; thereby, additional degrees of uncertainty related to the energy of states involved in thermionic transitions are attained. The thermal energizing of shallowly localized states, where the lifetimes are affected by thermal activation, results in carrier relaxation into broadly distributed deeply localized states via hopping, thus increasing the amount of generated randomness in the InGaN layers as the temperature rises.

Authors : S. Sakai, T. Minami, K. Kojima, S. F. Chichibu, A. A. Yamaguchi
Affiliations : Kanazawa Institute of Technology; Institute of Multidisciplinary Research for Advanced Materials, Tohoku University

Resume : AlGaN quantum-well (QW) structures have been promising as active layers of deep-ultraviolet optical devices. Optical properties of the AlGaN QWs are largely affected by their valence-band structures, and comprehensive understanding of the valence-band structures is essential for the device structural design. Emission from an Al-rich AlGaN layer is strongly polarized in the c-axis direction, because its topmost valence band has |Z⟩-like characters. It is known, however, that the polarization direction changes into the perpendicular direction to the c-axis when strain or quantum confinement is introduced in the AlGaN layers, due to the replacement of the topmost valence band. In the polarization-crossover region, the energy spacing between the first and second valence bands is small, and the valence band structure will be strongly modified due to the valence-band mixing. Such modified valence band structures could be favorable for some optical devices. In this study, we have theoretically investigated the valence-band structures and optical properties of AlGaN QWs in the polarization-crossover composition region, using the kp perturbation method. It is found that the valence-band structures drastically change from parabolic dispersion to quartic-function-like one, in which the valence band maximum (VBM) does not coincide with the Gamma point, and that the density of states (DOS) diverges at the VBM like quantum wire (QWI). Such QWI-like DOS could be favorable for some new-functional optical devices.

Authors : K. Sekiguchi, H. Shirakawa, K. Chokawa, M. Araidai, Y. Kangawa, K. Kakimoto, and K. Shiraishi
Affiliations : Graduate School of Engineering, Nagoya University; Institute of Materials and Systems for Sustainability, Nagoya University; Research Institute for Applied Mechanics, Kyushu University

Resume : Modeling of GaN MOVPE has been performed but trimethylgallium (TMG) decomposition process is still unclear. We used thermodynamic analysis to investigate the TMG decomposition process. In our previous study, we concluded that the formation of the (CH3)2GaNH2 adduct cannot produce [K. Sekiguchi et al., JJAP 2017], in contrast with the hypothesis used in conventional MOVPE simulations [D. Sengupta et al., JCG 2005]. That is, the TMG decomposition proceeds by the reaction with H2 instead of NH3. In this study, we analyzed the TMG decomposition process in more detail, taking the Gibbs free energy and the activation energies into consideration. To clarify the TMG decomposition process, we considered the reaction through which TMG with H2 decomposes into GaH(CH3)2 with CH4 desorption. At 700 K, which is assumed to be the temperature near the nozzle, the final state (GaH(CH3)2+CH4) is energetically more stable than the initial state (TMG+H2) by 0.9 eV. Moreover, it was found that the activation energy is 1.21 eV at 0 K by NEB analyses, indicating that the energy barrier is relatively low. These results show that TMG reacts with H2 near the nozzle and decomposes into GaH(CH3)2. It is believed that GaH(CH3)2 decomposes into Ga(CH3), GaH and Ga gas after reacting with H2. These results are consistent with recent high resolution mass spectrometry experiments [K. Nagamatsu et al., PSS B in press]. In this presentation, we also discuss the subsequent decomposition reactions of GaH(CH3)2.

Authors : Pawe? Strak1, Konrad Sakowski1, Pawel Kempisty2, Agata Kaminska3, Stanislaw Krukowski1
Affiliations : 1Institute of High Pressure Physics, Polish Academy of Sciences, Sokolowska 29/37, 01-142 Warsaw, Poland ; 2Center for Integrated Research of Future Electronics (CIRFE), Institute of Materials and Systems for Sustainability (IMaSS), Nagoya University, Nagoya 464-8603, Japan; 3Institute of Physics, Polish Academy of Sciences, Al. Lotników 32/46, 01-142 Warsaw, Poland and Cardinal Stefan Wyszynski University, College of Science, Department of Mathematics and Natural Sciences, Dewajtis 5, 01-815 Warsaw, Poland

Resume : The up to date thermalization of the adatoms at surfaces of the solids is reviewed. It is argued that both scenarios, i.e. hot atom and electron-hole pair creation could not explain efficient dissipation of the excess energy of adsorbed atoms and molecules. Thus the melting of the surface should occur, the effect not observed in experiment. A new scenario of thermalization process of the adsorbate attached at solid surfaces is proposed. The scenario is based on existence of electric dipole layer created due to the electron wavefunctions extension over the positive ions. Thus the strong local electric dipole related field exist which drags electron into the solids and repels the positive ions. The electrons are tunneling conveying the energy into the solid interior. The positive ions are retarded in the same electric field, which allows them to loose excess kinetic energy and to be located smoothly into the adsorption sites. In this way the excess energy is not dissipated locally avoiding melting or creation of defects, in accordance with the experiments. The scenario is supported by the ab intio calculation results including density function theory of the slabs representing AlN surface and the Schrodinger equation for time evolution of hydrogen-like atom at the solid surface. Other nitrides and nitrogen atoms from plasma are also considered. The research was funded by Polish National Science Center by grant: DEC-2015/19/B/ST5/02136.

Authors : Hilmi Ünlü
Affiliations : Department of Physics Engineering, Faculty of Science and Letters İstanbul Technical University, Maslak 34469 İstanbul, TURKEY E-mail: Phone: 90-212-285 3201, Fax: 90-212-285 6386

Resume : Advances in growing semiconductor thin films having differing physical properties with varying composition and layer thickness approaching atomic dimensions have provided new opportunities and challenges for experimental and theoretical scientific studies and fabrication of novel electronic and optical device applications. In this work, we present a semi-empirical second nearest neighbor (2NN) sp3s* tight binding model, which incorporates the chemical disorder and relaxation parameter, to investigate the effect of alloy composition and interface strain on the electronic band structure and core diameter of nanoscale III-N based binary/ternary heterostructure spherical core/shell quantum dots. In this model each atom is described by its outer valence s-orbital, three p-orbitals and a fictitious excited s* orbital to take into account the effect of higher lying d-states. Adding the excited s* state and spin-orbit coupling to sp3 orbitals set on the cation and anion atoms with 2NN interactions and spin-orbit coupling of p-states allows to have a better prediction of conduction band dispersion curves, especially at X and L symmetry points, reproducing the nonlocal empirical pseudopotential bands which cannot be done with NN and 2NN sp3 tight binding models. Interaction matrix elements are obtained by fitting the calculated results to real band structure of semiconductors at high symmetry points. Nonlinear expressions are derived for the atomic energy levels and bond lengths, including the chemical disorder and relaxation parameter, as function of alloy composition and interface strain in spherical core/shell nanostructures. We have applied this method to predict the composition and strain effects on the bandgap energy and core diameter of AlGaN/GaN and InGaN/GaN spherical core/shell quantum dots. The interface strain effects on the conduction band offsets are found to be large when lattice mismatch increases with increase in alloy composition, especially when the conduction band deformation potential is large. The proposed model is not only convenient in computation but is also useful for a more detailed understanding of the effects of alloy disorder on the electronic and optical properties of nanoscale heterostructures.The proposed model has a considerable potential in the design and optimization of the properties of spherical core/shell nanostructures for electronic and optical devices.

Authors : Hussein Mehdi(a), Guillaume Monier(a), Christine Robert-Goumet(a), Vladimir G. Dubrovskii (b,c,d)
Affiliations : (a)Institut Pascal, Université Clermont Auvergne, CNRS, SIGMA Clermont, F-63000 Clermont-Ferrand, France (b)St. Petersburg Academic University, Khlopina 8/3, 194021 St. Petersburg, Russia (c)Ioffe Physical Technical Institute of the Russian Academy of Sciences, Politekhnicheskaya 26, 194021 St. Petersburg, Russia (d)ITMO University, Kronverkskiy prospekt 49, 197101 St. Petersburg, Russia

Resume : The high density of interface and surface states that cause the strong Fermi pinning observed on GaAs surface is reduced by passivating GaN thin film on GaAs. To further develop this passivation, it is necessary to investigate the nitridation phenomena by identifying the different steps occurring during the process and to understand and quantify the growth kinetics of GaAs nitridation under different conditions. Nitridation of the cleaned GaAs substrate was performed using a N2 plasma produced by an electron resonance cyclotron source. Two approaches have been combined. Firstly, AR-XPS study is carried out to determine the chemical environments of the Ga, As and N atoms and the composition depth profile of the GaN thin film. The nitridation process can be summarized in three steps. The first one is the adsorption and the diffusion of the nitrogen through the vacant interstitial sites of the GaAs bulk. The second step is the formation of the strong Ga-N bonds. The final step is the diffusion and desorption of the free arsenic outward of the GaN film. The temperature and time treatment have a significant impact on the formation of the GaN layer. The second approach is a refined growth kinetic model which better describes the GaN growth as function of the nitridation time. This model clarifies the exchange mechanism of arsenic with nitrogen atoms at the GaN/GaAs interface and the phenomenon of quasi-saturation of the process observed experimentally.

Authors : Przemyslaw Niedzielski, Ewa Raj, Zbigniew Lisik.
Affiliations : Lodz University of Technology, Department of Semiconductor and Optoelectronic Devices, 211/215 Wolczanska str., 90-924 Lodz, Poland, E-mail:,

Resume : Metal-Organic Chemical Vapour Deposition (MOCVD) is one of the epitaxial growth technologies used for electronic device manufacturing. The technique is one of the major ones used to produce modern optoelectronic devices based on gallium nitride (GaN) and its compounds. It requires high temperatures and is characterised by complexity of the growth; hence a better understanding of the phenomena emerging in the reactor is needed. One of the most important problems during the growth process is to maintain appropriate temperature field within the MOCVD chamber and in the close proximity to the sample surface. Unfortunately, the temperature control is difficult because of rigorous conditions in the reactor chamber and the limited access to perform measurements. Hence, the numerical investigations of the reactor can be a useful tool to predict conditions during the growth process of semiconductor structures. The study presents numerical analysis of thermal-kinetic phenomena in Close Coupled Showerhead reactor obtained with the aid of reduced 3D model proposed by the authors. The model allows reducing computational efforts while keeping the results accuracy. Conducted simulations give detailed information regarding influence of several important parameters such as type of carrier gases, their flow rates, pressure, heater temperature and substrate type on the temperature profiles and precursors distribution in the tested chamber.

Authors : James T. Griffiths, Siyuan Zhang, Jeremy Lhuillier, Dandan Zhu, Ashley Howkins, Ian Boyd, David Stowe, David J. Wallis, Colin J. Humphreys, and Rachel A. Oliver
Affiliations : 1 Department of Materials Science and Metallurgy, 27 Charles Babbage Road, Cambridge, CB3 0FS, United Kingdom; 2 Experimental Techniques Centre, Brunel University, Uxbridge, UB8 3PH, United Kingdom; 3 Gatan UK, 25 Nuffield Way, Abingdon, Oxon, OX14 1RL, United Kingdom

Resume : The study of cathodoluminescence (CL) in a scanning transmission electron microscope (STEM) offers the potential to reveal the nanoscale optical properties of quantum emitters 1,2; an approach that is now commonly referred to as nano-CL. The approach also offers the opportunity to correlate the luminescence properties with the nanoscale structural and chemical properties to reveal new insights into the emission processes. This approach has emerged as a powerful technique in the study of III-nitride optical devices. Whilst the improved spatial resolution associated with nano-CL can reveal optical characteristics on the nanoscale, the high incident electron energies may induce structural damage. Here we investigate the effect of high energy electrons on the luminescent intensity 3. We show there is an exponential decline in the luminescent intensity over time with exposure to the electron probe, which we calculate is due to the formation of nitrogen vacancy-interstitials (Frenkel defects). We observe the luminescent intensity is more stable at lower accelerating voltages and suggest that the electron induced structural damage may be suppressed below the energy damage threshold. The nitrogen Frenkel defects lead to a non-radiative region that extends over the measured minority carrier diffusion length.

Authors : John Buckeridge1, Zijaun Xie2, Yu Sui2, C. Richard A. Catlow1, Aron Walsh3, David O. Scanlon1, Alexey A. Sokol1
Affiliations : 1University College London, Kathleen Lonsdale Materials Chemistry, Department of Chemistry, 20 Gordon Street, London WC1H 0AJ, U.K.; 2Department of Physics, Harbin Institute of Technology, 92 Xidazhi Street, Harbin 150001, P. R. China 3Department of Materials, Imperial College London, London WC1H 0AJ, U.K.

Resume : GaN is an important wide-gap semiconductor that is an essential component in blue light emitting diodes. Many experimental observations, both optical and electronic, such as intrinsic n-type conductivity and a broad yellow luminescence band found in a majority of samples, have been attributed to defect structures in the material. The identification of the particular defects involved in such phenomena through first-principles calculations has been a source of considerable controversy over the past few decades. The main challenge is to provide an accurate description of the electronic structure in this wide gap semiconductor. Using a state-of-the-art hybrid quantum mechanical/molecular mechanical embedded cluster approach, whose main advantage is a detailed description of polarisation due to local charges in a crystal, we compute formation and ionisation energies of point defects, i.e. vacancies, interstitials and antisites, as well as complexes, in GaN. Our results explain a broad range of experimental findings, including photoluminescence phenomena and deep level transient spectroscopic measurements. We find that n-type conductivity can be attributed to N vacancies, while the yellow luminescence can be related to the presence of Ga vacancies. As well as experimental work, we compare and contrast our results to other computational studies in the literature and account for the observed differences where present.

Authors : Pawel Kempisty(1,3), Yoshihiro Kangawa(2,1), Kenji Shiraishi(1), Stanislaw Krukowski(3), Michal Bockowski(3), Koichi Kakimoto(2), and Hiroshi Amano(1)
Affiliations : (1) Center for Integrated Research of Future Electronics (CIRFE), Institute of Materials and Systems for Sustainability (IMaSS), Nagoya University, Japan; (2) Research Institute for Applied Mechanics, Kyushu University, Japan; (3) Institute of High Pressure Physics, Polish Academy of Sciences, Poland;

Resume : Using Density Functional Theory (DFT) calculations we analyzed the stability of carbon and oxygen atoms incorporated into GaN layers close to (0001) and (000-1) surfaces. Calculations were done for few different coverages corresponding to the change of conditions in the gas phase during MOVPE and HVPE growth. Due to the multiplicity of cases we focus mainly on the substitutional position of impurity, i.e. ON, CN and CGa. The change of formation energy relative to the bulk value was determined. Our results confirm that the crucial factor is the Fermi level pinning and the accompanying bending of energy bands. With the knowledge of these two parameters one can estimate roughly a tendency in behavior of foreign atoms. Additionally in few layers closest to the surface a strong dependence on the site symmetry is observed. This is a typical quantum effect associated with the interaction of impurities and surface states. It is shown in which cases a combination of these two effects is responsible for facilitating or hindering incorporation of dopants. In most cases, a change of the energy for carbon and oxygen atoms has the opposite sign. From a technical point of view calculations were performed by using SIESTA code. The relaxed geometry and total energy of system was found within GGA-WC approximation. A correction scheme called GGA-1/2 was later used to analyze the energetical positions of the defect levels. The surface was represented by the slab model containing up to 1000 atoms.

Authors : Thoan H. Nguyen*, Do T. Nguyen, Anh T. L. Luu, Trung N. Nguyen, Vinh V. Le
Affiliations : School of Engineering Physics, Hanoi University of Science and Technology, Hanoi, Vietnam

Resume : The classical molecular dynamics (MD) technique was used to investigate the formation of HfO2 thin films from a solid phase on a GaN high electron mobility semiconductor surface. The local atomic structure and the liquid-solid transition of HfO2 were analyzed through potential energy per atom, pair radical distribution functions, bond angle distributions, coordination numbers and the void distribution. The evolution of structural HfOx units (where x=3–7) were examined. The structures and HfO2/GaN interfacial properties were also investigated as a function of the oxide thickness. The density of interfacial dangling bonds (DB), which are well known as interface electron/hole traps in electronics devices, was estimated. In addition, we also considered the DB-passivation efficiency by annealing in hydrogen.

Start atSubject View AllNum.Add
Doping, defects : Kazunobu Kojima
Authors : Bernd Witzigmann, Friedhard Römer
Affiliations : University of Kassel and CINSaT, Electrical Engineering and Computer Science Department Kassel, Germany

Resume : InGaN/GaN multiple quantum well regions constitute the active light emitting structure in most visible LEDs or lasers. For maximum efficiency, the electron and hole density need to be distributed uniformly in the quantum wells at all relevant operating conditions. We analyze the carrier injection and distribution in polar and semi-polar LEDs using a 1-dimensional microscopic carrier transport model. It is based on drift-diffusion currents, and quantum wells are treated as scattering centers, with self-consistent kp-band structure calculation for the quantum well carriers. Tunneling currents are included via an effective potential model. Impurities such as Si, Mg and O are activated according to a Schottky model; their distribution is found from SIMS data. The activation depends both on the impurity density as well as the local electric field (Poole-Frenkel effect). Comparison to experimental data reveals the following findings: the hole injection efficiency depends critically on the proximity of the Mg dopants to the p-side quantum well. At the same time, electron leakage is suppressed by ionized Mg atoms forming an electrostatic barrier, which is reduced with rising current due to saturation of the acceptor levels. Impurities that form shallow donors (e.g. oxygen) also hamper the hole injection efficiency as they reduce the Mg ionization rate. Therefore a high performance LED requires nanoscale control of intentional and unintentional impurity distribution. Maximum efficiency and minimum drooping is achieved for homogeneous carrier distribution in the MQW stack. The ideality factor of the pin-diode contains information on the injection homogeneity of the LED and design criteria for homogeneous carrier injection is presented.

Authors : A. Hoffmann1, G. Callsen1, C. Nenstiel1, F. Nippert1, T. Kure1, S. Schlichting1, N. Jankowski1, M. R. Wagner1, M. P. Hoffmann2, S. Fritze2, A. Dadgar2, A. Krost2, F. Bechstedt3
Affiliations : 1Insitut für Festkörperphysik, Technische Universität Berlin, Berlin, Germany; 2Insitut für Experimentelle Physik, Otto-von-Guericke-Universität Magdeburg, Magdeburg, Germany; 3Insitut für Festkörpertheorie und -optik, Friedrich-Schiller-Universität, Jena, Germany.

Resume : Heavy doping of bulk semiconductors is well known to result in the dissociation of excitons due to the screening of the Coulomb interaction, leaving behind only the emission of band-to-band transitions. In contrast to common expectations, we observe pronounced emission signatures in bulk GaN:Ge, mimicking exciton-like behavior at free carrier concentrations up to 9x1019 cm-3. Here, we present our combined experimental and theoretical study on highly Ge-doped bulk GaN that confirms the existence of unconventional, exciton-like quasiparticle complexes that stabilize upon rising free electron concentration. We observe two pronounced emission lines in addition to the band-to-band luminescence which intensify and narrow spectrally with rising doping concentration. Based on our comprehensive experimental data and theoretical treatment, we suggest a model that ascribes the observation of these novel excitonic emission lines to an exciton and a tetron (interband electron-hole pair bound to additional intraband pair) stabilized by a degenerate electron gas in bulk GaN:Ge. A critical prerequisite for this observation is the crystal quality of the GaN:Ge samples that show almost no sign of compensation even at the highest free carrier concentration. Hence, an unperturbed Fermi sea of electrons is filled upon rising doping concentration that enables pair localization in the high density limit - an effect that otherwise is only achieved by e.g. spatial confinement in nanostructures.

Authors : Douglas L. Irving1, Joshua S. Harris1, Jonathon N. Baker1, Kelsey J. Mirrielees1, Brian D. Behrhorst1, Dorian Alden1, Ronny Kirste2, James Tweedie2, Ramón Collazo1, Zlatko Sitar1
Affiliations : 1Department of MSE, North Carolina State University, Raleigh, NC USA; 2Adroit Materials Inc., Cary, NC USA

Resume : While the reduction of extended defects in AlN has led to many advances in the creation of novel electronics, controlled doping remains an outstanding challenge. This is due, in part, to the larger band gaps and the depth of the valence bands in these materials. Because of these and other factors it is not uncommon for desired dopants to be compensated by native or unintentional impurity defects rather than a free carrier during growth. Thus, to obtain desired carrier concentrations, higher concentrations of intentional dopants are often added to the determent of the mobility of the free carriers in these materials. In an effort to develop a mechanistic understanding of compensation of dopants in these materials, we have implemented first principles density functional theory calculations with screened hybrid exchange-correlation functionals to determine the properties of the individual defects. The formation energies of each defect are also used in a grand canonical equilibrium model to investigate the presence of predominant defects as a function of growth conditions. Results from these methods are compared with complementary experimental data that includes below band gap optical absorption and photoluminescence as well as available compositional measurements.

Authors : Pawel Strak 1), Pawel Kempisty 1, 7), Konrad Sakowski 1), Agata Kaminska 2, 3), Dawid Jankowski 2), Krzysztof P. Korona 4), Kamil Sobczak 2), Jolanta Borysiuk 2,4), Mark Beeler 5), Ewa Grzanka 1, 6), Eva Monroy 5) and Stanislaw Krukowski 1)
Affiliations : 1) Institute of High Pressure Physics, Polish Academy of Sciences, Sokolowska 29/37, 01-142 Warsaw, Poland 2) Institute of Physics Polish Academy of Sciences, Al. Lotników 32/46, 01-142 Warsaw, Poland 3) Cardinal Stefan Wyszynski University, College of Science, Department of Mathematics and Natural Sciences, Dewajtis 5, 01-815 Warsaw, Poland 4) Faculty of Physics, University of Warsaw, Pasteura 5, 02-093 Warsaw, Poland 5) CEA Grenoble, INAC-SP2M, 17 av. des Martyrs, 38000 Grenoble, France 6) TopGaN Ltd. Sokolowska 29/37, 01-142 Warsaw, Poland 7) Institute of Materials and Systems for Sustainability, Nagoya University, Nagoya 464-8603, Japan

Resume : Polarization in the solid system is critically analyzed. It is demonstrated that Berry phase polarization corresponds to an unphysical state which may be used for determination of spontaneous polarization by appropriate addition of polarization induced electric fields. These fields are obtained by the use of continuity of the electric displacement field across the whole system in the absence of free charge. The obtained spontaneous polarization is much smaller than Berry phase polarization obtained from geometric phase formalism. The electric fields in III-nitride superlattices (SL) and are consistent with those obtained by Bernardini et al. [1] provided that the appropriate correction of polarization is made. The spontaneous polarization and polarization related electric fields in bulk AlN, GaN and InN were determined using ab initio slab calculations. The obtained spontaneous polarization values are: 8.69·10-3 C/m2, 1.88·10-3 C/m2, and 1.96·10-3 C/m2 for AlN, GaN and InN, respectively. The related Berry phase polarization values are 8.69·10-2 C/m2, 1.92·10-2 C/m2, and 2.86·10-2 C/m2, for these three compounds. These values are in good agreement with the earlier derived values. The GaN/AlN multi-quantum well (MQWs) were also simulated using ab initio calculations. The obtained electric fields in structures are in good agreement with the fields derived from bulk values. GaN/AlN MQWs structures, obtained by plasma assisted molecular beam epitaxial (PA - MBE) growth, were characterized by transmission electron microscopy (TEM) and X-ray measurements. Time dependent photoluminescence (PL) measurements were used to determine optical transition energies in these structures. The PL obtained energies are in good agreement with ab initio data confirming overall agreement between model and experimental data. It was found that decay rates decreases four order of magnitude with increase of the SL period what confirms existence of high electric field. The electric field caused by spontaneous polarization reshapes waves in valence and conductions bands in such a way that it reduces recombination probability.

Authors : U. Wahl1, L.M. Amorim2, V. Augustyns2, A. Costa1, E. David Bosne1, T. A. L. Lima1, G. Lippertz1, J. G. Correia1, M.R. da Silva3, M.J. Kappers4, K. Temst2, A. Vantomme2, L.M.C. Pereira2
Affiliations : 1Centro de Ciências e Tecnologias Nucleares, Instituto Superior Técnico, Universidade de Lisboa, 2695-066, Bobadela LRS, Portugal; 2KU Leuven, Instituut voor Kern- en Stralingsfysica, 3001 Leuven, Belgium; 3Centro de Física Nuclear da Universidade de Lisboa, 1649-003 Lisboa, Portugal; 4Cambridge Centre for Gallium Nitride, University of Cambridge, Cambridge CB3 0FS, U.K.

Resume : Mg-doped p-type GaN is well known to suffer from doping limitations: once the Mg concentration surpasses ~1E19-1E20 cm-3, further introduction of Mg does not lead to an increase in the hole concentration. The reasons for this have been heavily debated. Some theorists suggested that Mg is an amphoteric impurity, i.e. substitutional Mg on Ga sites forms the desired acceptors but interstitial Mg acts as a compensating double donor [1-2], while this possibility has been discarded by others [3]. In this contribution, we provide direct lattice location measurements for Mg in GaN of different doping types using the beta- emission channeling technique, which proof the existence of interstitial Mg and its amphoteric nature. Radioactive 27Mg (t1/2=9.5 min) was implanted at CERN?s ISOLDE facility. Following implantations between room temperature and 800°C, the majority of 27Mg occupies the substitutional Ga sites, however, below 350°C significant fractions were also found on interstitial positions ~0.6 Å from ideal octahedral sites. The interstitial fraction of Mg was correlated with the GaN doping character, being highest (up to 31%) in samples doped p-type with 2E19 cm-3 stable Mg during epilayer growth, and lowest (< 7%) in Si-doped n-GaN. Implanting above 350°C converted interstitial 27Mg to substitutional Ga sites in all doping types (via trapping of mobile Mg by Ga vacancies), which allows estimating the activation energy for migration of interstitial Mg as between 1.3 and 2.0 eV. [1 ] F.A. Reboredo and S.T. Pantelides, Phys. Rev. Lett 82 (1999) 1887. [2] G. Miceli and A. Pasquarello, Phys. Rev. B 93 (2016) 165207. [3] C. G. Van de Walle and J. Neugebauer, J. Appl. Phys. 95 (2004) 3851. [4] U.Wahl et al., Phys. Rev. Lett. (2017), accepted.

Wednesday poster : Juergen Christen
Authors : R. Mohamad1, A. Béré2, H. Ben Ammar1, A. Minj1, P. Gamarra3, C. Lacam3, J. Chen1 and P. Ruterana1
Affiliations : 1CIMAP, UMR 6252, UNICAEN, CNRS, ENSICAEN, CEA, 6 Boulevard Maréchal Juin, 14050 Caen and Pôle Universitaire d'Alençon, Campus de Damigny, 61250 Damigny 2Laboratoire de Physique et de Chimie de l’Environnement, Université de Ouagadougou, 03 BP: 7021 Ouagadougou 03, Burkina Faso 3III-VLab, 1 Avenue Augustin Fresnel, Campus Polytechnique, 91767 Palaiseau, France

Resume : For the last twenty years, great progress has been achieved in the development of optical and electronics devices based on group III nitride compounds InN, GaN, AlN, and their alloys. Still, in order to obtain a high performance and highly reliable devices, it is necessary to reduce the defect density in the III nitride quaternary alloys such as InAlGaN which are excellent candidates for the fabrication of barrier layers in high mobility transistors. It has been shown that for instance growing thick InAlN, or InGaAlN epitaxial layers lead to the formation of very defective layers and the underlying mechanisms do not appear to be well understood. In this work, our aim is to make a detailed theoretical investigation of the miscibility of quaternary alloys in correlation with the experimental work that is carried out using transmission electron microscopy. To this end, we stay in the regular solution model and for determining the interaction parameter from the formation energy (enthalpy) calculated using a modified Stillinger-Weber for III-nitride ternary alloys. We thus obtain their metastable, unstable and stable regions as deduced from phase diagrams. We present the results on miscibility of III nitride quaternary alloys, mainly InxAlyGa1-x-yN. Our calculations show that the phase separation temperature increases with the increase of indium concentration. However, for In concentration lower than 9.4%, the quaternary alloy is miscible whatever Al and Ga compositions.

Authors : Qing Shi1, Ying-Chih Chen1, Faqrul A. Chowdhury2, Zetian Mi2, Vincent Michaud-Rioux1, Hong Guo1
Affiliations : 1Department of Physics, McGill University, Montréal, QC, H3A 2T8, Canada 2Department of Electrical and Computer Engineering, Montréal, QC, H3A 0E9, Canada

Resume : We report the band gaps and band alignments calculation for the GaNSb alloys in the low Sb range by density functional theory (DFT) with the HSE06 hybrid exchange-correlation functional. The valence band edge is found to drop dramatically as Sb concentration goes below 1% (Fig. 1, 4) due to quantum interactions between Sb impurity states and host GaN states (Fig. 2, 3). Based on the position of valence band edge, the GaNSb alloy with ∼0.5% Sb is predicted to be suitable for photoelectrochemical water splitting applications. This level of doping introduces negligible lattice strain hence significantly reduces defects and dislocations in the material. The HSE06 level DFT calculation for systems over 1,000 atoms was carried out by the RESCU electronic package.

Authors : Bei Ma, and Yoshihiro Ishitani
Affiliations : Chiba University

Resume : Although excitonic photoluminescence of GaN has been observed at room-temperature, electron-hole plasma state becomes dominant in GaN laser devices at temperatures higher than 150K. We believe that the transportation of the energy in non-thermal-equilibrium state is a key issue for improving the exciton stability in devises. In this paper, we discuss the dynamics of carrier, exciton, LO and LA phonons, and investigated the exciton stability.In our carrier-dynamics simulation, the interactions of Fröhlich, deformation potential, and the piezo-electric were included. The dissipation, association and dissociation processes of phonons were also included. The carriers were excited instantaneously at time zero, and the following dynamics were simulated. The energy transport among exciton-carrier-phonon system was focused. Varied parameters in the simulation were the excitation photon energy or the amount of generated phonons, phonon escape time from the excited region, excited carrier number, and LO phonon dissociation rate. Thermal phonon was neglected. It was found that the delay of phonon escape in the order of 1-10 ps induces the delay of PL rise to approximately 75 ps. The decay of the total kinetic energy of excitons follows but slower than the decay of the total LO phonon energy, and far faster than the decay of LA phonon energy. Our simulation reveals that the energy transport among LO phonon, carriers, and excitons is the critical issue for exciton stability.

Authors : Sandeep Sankaranarayanan, Swaroop Ganguly, Dipankar Saha
Affiliations : Reasearch scholar at department of electrical engineering IIT Bombay

Resume : A method to calculate the width-dependent piezoelectric polarization of single quantum well lateral InGaN nanowires from the exciton binding energies is proposed. With a decrease in nanowire width exciton binding energy increases and the lattice tends to become more strain relaxed which decreases the piezoelectric polarization field. Thus an increase in the quantum confinement results in a decrease in the polarization field and a higher exciton binding energy. This relationship between exciton binding energy and the polarization field is used to predict the polarization field strength for a particular wire width. An analytical model based on fractional dimension for GaN/InGaN/GaN system is used to calculate exciton binding energy as a function of nanowire width. The nanowires of widths as small as 15 nm and as high as 100 nm are fabricated. Low temperature (10 K) photoluminescence (PL) is done on the samples. Exciton binding energy is the difference between the PL peak energy and the first bound state energy. The first bound state energy for a nanowire is a function of wire width and the polarization field. It can be seen that both the first bound state energy and the expression for exciton binding energy from the theoretical model depend on the polarization field. By solving the two cases self-consistently we can find the polarization field strengths for each nanowire widths such that the exciton binding energies from both theory and experiment match.

Authors : Xianghua Kong, Wei Ji, Zetian Mi, Hong Guo
Affiliations : Department of Physics, McGill University, Montreal, QC H3A2T8, Canada; Department of Physics and Beijing Key Laboratory of Optoelectronic Functional Materials and Micro-Nano Devices, Renmin University of China, Beijing 100872 , P. R. China; Department of Electrical & Computer Engineering, McGill University, Montreal, QCH3A0E9, Canada; Department of Physics, McGill University, Montreal, QC H3A2T8, Canada;

Resume : Artificial photosynthesis, as a promising viable technology for the development of a clean, cheap, renewable and abundant energy source to replace our reliance on fossil fuels, has received considerable attention. In such process, the solar energy is converted and stored in the form of energy-rich solar fuels (e.g. hydrogen, methanol) via chemical reaction such as water splitting and CO2 reduction. Metal nitride semiconductors, for example, Ga(In)N, have recently emerged as highly promising photocatalysts for solar fuel applications. In this talk, we focus on the unknown interaction between a metal nitride surface and CO2 molecule, and peculiar N-terminated wurtzite nonpolar GaN surfaces. Firstly, we studied the preliminary but vital mechanism of photochemical CO2 reduction on GaN (wurtzite) nanowire, and a detailed investigation of CO2 molecules’ adsorption and activation on pristine, Mg- doped and Ge- doped GaN(100) surface is conducted. We found that CO2 molecules can be spontaneously activated on the clean nonpolar surfaces of wurtzite metal-nitrides. And the photocatalytic activity for CO2 reduction can be dramatically enhanced by incorporating a small amount of Mg-dopant. The Mg-dopant in GaN can not only make the surface band bending downward which facilitate the photo-induced electrons effectively moving to the interface of GaN-CO2, but also improve the nanowire's ability to adsorb and activate CO2 molecules, including larger CO2 adsorption energy (-2.48 eV) and deformation energy (2.86 eV). This study reveals the extraordinary potential of III-nitride semiconductor nano-structures in solar-powered reduction of CO2 into hydrocarbon fuels. Secondly, the atomic structures of wurtzite GaN nonpolar m-plane surfaces with different levels of Ga and N surface coverage are investigated and corresponding formation energies are calculated. The relaxed geometries for various degrees of N and Ga atom surface coverage showed similar features that N atoms on the top surface and in the layer beneath tend to aggregate and form clusters. And the the formation energy for N-terminated surfaces decreases drastically with increasing chemical potential of N atom (μN), for example, the surface where 25% surface N atoms are missing, in addition to the absence of all the Gasurface atoms becomes energetically more stable than the conventional m-plane with equal number of Ga and N atoms on the surface for μN ≈ −6.045 eV, or larger, which is in excellent agreement with the experimental observation of N-rich m-plane surfaces of GaN nanowire arrays grown by MBE under N-rich conditions. Such N-terminated surfaces not only passivate the GaN nanowires against attack by air/aqueous electrolytes, but their negative surface charge is also essential for inducing the “reverse” polarization that is crucial for p–GaN to efficiently separate and transport photogenerated charges for high performance in overall water splitting in pure water. Furthermore, these kinds of N-terminated m-plane surfaces of GaN are expected to do good to molecules’ adsorption and activation.

Authors : Chien-Chih Chen, Wen-Yen Chang, Yang Kuo, Yean-Woei Kiang, C. C. Yang
Affiliations : Institute of Photonics and Optoelectronics, National Taiwan University

Resume : It has been shown that the efficiency droop effect of an LED can be reduced through the coupling of its quantum wells with surface plasmon (SP) resonance induced on a metal nanostructure. The basic idea is that through SP coupling the radiative recombination rate in a quantum well can be increased due to the Purcell effect. However, the dissipation in the metal nanostructure may lead to the decrease of the internal quantum efficiency (IQE) of the LED. We modify the carrier-density rate equation of the so-called “ABC” model to include these two factors for evaluating the effects of efficiency droop reduction under various SP coupling conditions. Based on the previously developed numerical algorithms for SP coupling simulation, we calculate the radiated power of a radiating dipole and the metal dissipation of a nearby metal nanoparticle. With these numerical data as well as assumed maximum IQE and the corresponding current density of a reference LED sample, we can evaluate the IQE of an SP-coupled LED as a function of injected current density. The efficiency droop effect of the LED with a metal nanoparticle for SP coupling can thus be investigated. Individual current components for SRH, radiation, Auger recombination, and metal dissipation can also be identified. The experimental data of SP-coupled LEDs with surface Ag nanoparticles are used for comparing with the numerical simulation results.

Authors : J.-M. Lentali,1 M. Filoche,1 W. Hahn,1 J. S. Speck,2 A. Alhassan,2 L. Martinelli,1 C. Weisbuch,1,2 Y. Lassailly,1 and J. Peretti1
Affiliations : 1 Laboratoire de Physique de la Matière Condensée, CNRS-Ecole Polytechnique, Université Paris-Saclay, 91128 Palaiseau, France 2 Department of Materials, University of California, Santa Barbara, California 93106, USA

Resume : GaN-based alloys are characterized by important spatial composition inhomogeneities resulting from the random distribution of Indium atoms. These variations can induce carrier localization and strongly influence the performances of the devices. In this work, we report on a mathematical model aiming at describing electronic transport and recombination in a disordered medium representing an InGaN/GaN quantum well (QW) heterostructure. This model is based on a recent theory of localization called the localization landscape. The main result of this approach is the derivation of an effective potential whose basins define the localization regions of carriers [1,2], and from which we can determine local effective band gaps and electron-hole overlap. Interpreting the light emission from an InGaN QW additionally requires to consider the possibility for carriers to escape the potential minima in which they are trapped. We thus propose a phonon-assisted hopping model, using the concept of Agmon's distance [2] to evaluate the coupling between localized states and escape probabilities. This model is used to interpret Scanning Tunneling Luminescence measurements performed on single QW structures [3]. Calculations explain well the experimental results which evidence localized recombination at a few nanometer scale. 1. M. Filoche and S. Mayboroda, Proc. Natl Acad. Sci.USA 109, 14761 (2012) 2. D. Arnold et al., Phys. Rev. Lett. 116, 056602 (2016) 3. P. Polovodov et al., this conference

Authors : Hendrik Spende 1,2,*, Johannes Ledig 1,2,3, Klaas Strempel 1,2 , Hergo-Heinrich Wehmann 1,2, Sönke Fündling 1,2, Andreas Waag 1,2
Affiliations : 1 Institute of Semiconductor Technology and Laboratory for Emerging Nanometrology, Braunschweig University of Technology, 38092 Braunschweig, Germany 2 epitaxy competence center ec2, Hans-Sommer-Straße 66, 38106 Braunschweig, Germany 3 Physikalisch-Technische Bundesanstalt (PTB), Bundesallee 100, 38116 Braunschweig, Germany

Resume : Cathodoluminescence (CL) inside a scanning electron microscope (SEM) is a powerful tool to probe optical properties of semiconductor microstructures with high spatial resolution. However, in comparison to photoluminescence (PL) or electroluminescence (EL), the quantification of the generation rate inside the semiconductor material is difficult. The high energy of the primary electrons in the beam exceeds the band gap of the semiconductor causing spatial scattering and distributed energy transfer of several kinds. These also lead to a generation of eh-pairs which subsequently undergo relaxation, drift, and diffusion. As a result, the radiative recombination inside the material is spatially inhomogeneous and strongly depends on material parameters. As the local generation rate of eh-pairs has a significant influence on the emission spectrum in regard to parallel recombination channels, a determination of the charge carrier distribution gives deeper insight to the effective CL mechanism. Here, we analyse the possibilities to quantify the excitation inside a semiconductor using Monte-Carlo-Simulations by means of the CASINO program. The simulations will be compared to data obtained by planar and 3D InGaN/GaN MQW test samples inside a Tescan Mira 3 GMH FE-SEM using a Gatan MonoCL 4 setup for light collection and a LN2 cryo stage for controlling the sample temperature. The influence of spot size and beam current on the excitation density will be discussed with respect to CL spectra.

Authors : I.G. Batyrev, N.S. Weingarten, K. A. Jones
Affiliations : US Army Research Laboratory, 2800 Powder Mill Road Adelphi, MD 20783-1138

Resume : We present results of DFT simulations of dislocation core structures of mixed dislocations in the pyramidal plane of wurtzite GaN. We also studied the distribution of n- (Si) and p- (Mg) dopants relative to the dislocation core in the pyramidal plane. First, the dislocation cores were constructed using empirical MD potentials. Initial atomistic configurations of single crystal GaN containing a dislocation dipole are generated via classical molecular dynamics. A Tersoff-type interatomic potential is used on a fully periodic system. For edge dislocations, opposing half planes are removed to create trenches, and the system is compressed during the course of the simulation. This results in full dislocations with opposite Burgers vector upon the closing of the trenches. To create screw dislocation dipoles, the appropriate displacement field is applied to all the atoms in the system, and no compression is necessary since no atoms are removed from the system. Mixed dislocation dipoles can also be generated by combining the two procedures, and specific challenges related to this methodology will be discussed. Then we performed geometry and cell optimization of the models containing 1440, 1448, and 3960 atoms using the quantum mechanical code cp2k. The robust method allows one to perform quantum calculations for relatively large systems, which is ideal for the study of extended defects and allows one to go beyond widely used hybrid or tight-binding methods. It also allows to estimate energy positions of the defect levels in band gap using big unit cells.

Authors : Ivan A. Aleksandrov, Konstantin S. Zhuravlev, Vladimir G. Mansurov, Timur V. Malin
Affiliations : A.V. Rzhanov Institute of Semiconductor Physics, Lavrentieva 13, Novosibirsk 630090, Russia

Resume : GaN/AlN quantum dots (QDs) are promising for ultraviolet and visible light-emitting devices, high-speed photodetectors and switches in near infrared spectral range and single photon sources operating at room temperature. In this work the influence of carrier tunneling between quantum dots and defects on the recombination dynamics in an ensemble of GaN/AlN QDs have been studied. The observed nonexponential photoluminescence kinetics was described using the model of carrier recombination dynamics in an ensemble of GaN/AlN QDs, which takes into account QD shape fluctuations and capture of carriers by defects with their subsequent return to the QD [1]. The model allows to describe the experimental decay curves and their spectral dependence. The interaction of QDs with defects slows down the photoluminescence kinetics. Tunneling probability was calculated in the configuration coordinate model in zero-radius potential approximation [2]. Tunneling rate have thermally activated behavior, which is consistent to the observed contribution of the nonradiative recombination to the decay rate at high temperatures. This work was supported by RFBR (grants 16-32-00765, 17-52-04023 and 17-52-52025). [1] I.A. Aleksandrov, V.G. Mansurov, K.S. Zhuravlev, Semiconductors, 2016, 50, 8, 1038-1042 (2016). [2] I.A. Aleksandrov, V.G. Mansurov, K.S. Zhuravlev, Physica E, 75, 309-316 (2016).

Authors : Yuya Inatomi1, Yoshihiro Kangawa1,2,3, Tomonori Ito4, Tadeusz Suski5, Koichi Kakimoto2, and Akinori Koukitu6
Affiliations : 1 Department of Aeronautics and Astronautics, 2 Research Institute for Applied Mechanics (RIAM), Kyushu University; 3 Institute of Materials and Systems for Sustainability (IMaSS), Nagoya University; 4 Department of Physics Engineering, Mie University; 5 Institute of High Pressure Physics PAS; 6 Department of Applied Chemistry, Tokyo University of Agriculture and Technology

Resume : InGaN/GaN quantum structures are used in light-emitting diodes (LEDs) in the ultraviolet–infrared spectral range. Short-period mIn(Ga)N/nGaN superlattices enable bandgap engineering in the blue–green range of the spectrum. In the present work, influence of lattice constraint on indium incorporation into coherently grown InxGa1-xN on InyGa1-yN thick layers was examined theoretically. First, effective mixing enthalpy of InxGa1-xN/InyGa1-yN systems were obtained by empirical interatomic potential calculations. In general, mixing enthalpy curve of an alloy has a parabolic shape. In case of coherently grown InxGa1-xN, however, effective mixing enthalpy curves were deviated from a parabolic shape because of the influence of lattice constraint. These results suggest that the regular solution model cannot be applied to this system. Secondly, we modified the conventional thermodynamic analysis to investigate this issue. Specifically, a formula of activity in the analysis was generalized to expand the applicable system. By the thermodynamic analysis, we obtained the phase diagram for deposition of InxGa1-xN on InyGa1-yN thick layers. The calculation results well reproduce composition pulling effect. The obtained knowledge is useful for accelerating the development of InxGa1-xN/InyGa1-yN system for bright green-orange LEDs.

Authors : Kensuke Oki, Kentaro Nomachi, Bei Ma, Ken Morita, Yoshihiro Ishitani
Affiliations : Chiba University

Resume : GaN has exciton binding energy comparable to room temperature. Thus, the exciton largely affects the optical property of the material. Dynamics of the exciton and free carrier has been widely studied using the rate equation method. However, the rate coefficients were not rigorously calculated. Also, there was a small number of analysis including excitons with principal quantum numbers greater than 1. In this study, we theoretically calculate the rate coefficients of transitions among free carrier and excitons with principal quantum numbers of 1 to 5. The radiative and electron-collisional rate coefficients are calculated using the hydrogen plasma model. The rate coefficients of phonon emission and absorption are calculated using Fermi’s golden rule. Rate equations are solved to obtain the population densities in the steady state. We show the dependence of the densities on temperature and excitation rate. Time evolution of the densities in decay phase is also calculated. We have found that the decay time of 1s exciton density possibly becomes over 10 times longer than the recombination time which is inverse of averaged recombination probability of one exciton. The ratio of the former to the latter time increases as the excitation rate decreases or the temperature increases. Although the decay time has often been considered to be the same as the recombination time in experimental studies, it depends on conditions such as temperature and excitation density.

Authors : K. Kawakami, Y. Yamamoto, K. Yoshimatsu, N. Okamoto, Y. Kangawa, K. Kakimoto and K. Shiraishi
Affiliations : Nagoya Univ., Graduate School of Engineering, Nagoya Univ., IMaSS, Nagoya Univ., RIAM, Kyushu Univ.

Resume : Numerical simulations, such as thermodynamic analysis of driving force, the first principles calculations, and computational fluid dynamics, of GaN metalorganic vapor-phase epitaxy (MOVPE) play significant roles in prediction of appropriate conditions of GaN crystal growth and design of the MOVPE apparatuses. The thermodynamic analysis of the driving force of GaN deposition well finds the appropriate conditions and the crystal growth modes. The analysis furthermore predicts the substrate orientation of the crystal surface with the first principles calculations. A natural question is how to produce the conditions by flow. The flow transport gases and heat, and thus can give rise to their non-uniform and unsteady distribution at the crystal surface. To address this question, we developed a methodology for multiphysics simulation of GaN MOVPE. It is based on the thermodynamic analysis and fluid dynamic analysis of multicomponent flow. These analyses are coupled by boundary conditions in terms of mass fractions produced or consumed by a surface reaction. The boundary conditions are obtained by a simple application of the conditions in terms of mass conservation. We here consider compressible flow of homogeneous mixture gases, whose chemical species are Ga(CH3)3, NH3, H2 and N2, with an equilibrium reaction in the crystal surface. Examples of numerical simulation using the developed method is also presented. A two-dimensional flow in a channel is simulated.

Authors : Y. Yamamoto, K. Kawakami , K. Yoshimatsu, N. Okamoto, Y. Kangawa, K. Kakimoto and K. Shiraishi
Affiliations : Graduate School of Engineering, Nagoya University; Institute of Materials and Systems for Sustainability, Nagoya University; Research Institute for Applied Mechanics, Kyushu University

Resume : We examine influence of flow on appropriate GaN growth conditions in metalorganic vaper-phase epitaxy (MOVPE). The conditions are well predicted by the thermodynamics analysis of the driving force of the crystal growth. We performed a multiphysics simulation at the crystal growth in chambers, using the thermodynamic analysis and computational fluid dynamics. The source gas species, Ga(CH3)3 (TMG) and NH3, are transported to the substrate together with the carrier gases, H2 and N2. Ga gas is produced by a decomposition reaction of TMG [Nagamatsu et al. physica status solidi b (in press)]. The low Mach number approximation is applied to the basic equations of compressible flow. Then, the resulting equations are spatially discretized by second-order central finite difference method using the staggered grid in two-dimensional channel. The explicit Euler method is employed as the time marching method. The thermodynamic analysis results in a quartic polynomial equation in terms of the equilibrium partial pressure of Ga source gas, which are solved by the Newton method. We will explore a channel configuration, especially inlet configuration, to reduce the amount of the source gases in the GaN crystal growth. Influence of flow on the substrate orientation on the GaN crystal will also be reported, using the thermodynamic analysis coupled with the first principles calculations.

Authors : C. Brimont1, T. Guillet1, D. Scalbert1, B. Jouault1, P. Valvin1, T. Bretagnon1, P. Lefebvre1, L. Lahourcade2, N. Grandjean2, B. Damilano3, M. Vladimirova1
Affiliations : 1Laboratoire Charles Coulomb, CNRS/Université de Montpellier, Montpellier, France; 2Institute of Physics, EPFL, Lausanne, Switzerland; 3Centre de Recherche sur l?Hétéro-Epitaxie et ses Applications, CNRS, Valbonne Sophia Antipolis, France

Resume : Light emission in polar group-III nitride quantum wells (QWs) optically pumped at densities significantly below the lasing threshold is usually interpreted in terms of excitons. Whereas thin GaN QWs with fast radiative lifetimes are employed in light-emitting devices, the excitons in thick QWs (typ. 7nm) are characterized by a non-zero dipole moment and long radiative lifetimes, due to the built-in electric field. These features result in many interesting properties of dipolar excitons: they can propagate over large distances and thus cool down to the lattice temperature before recombination, offering the possibility for studies of cold and dense gas of interacting bosons. In this context, two important issues must be addressed: the strength of the exciton-exciton interaction, and the excitonic Mott transition, that sets the maximum density of excitons. Combining spatially and time-resolved photoluminescence in different excitation geometries, we provide a set of experimental data showing that the emission linewidth is proportional to the emission energy in a wide range of carrier densities and temperatures. The proportionality coefficient is almost temperature independent. These results are compared to the existing data in various material systems and discussed in terms of the Mott transition and its criteria, exciton-exciton interaction and its possible contribution to the linewidth, as well as the role of disorder in the density-dependent emission spectra.

Authors : 1Motoshi Uchino, 1Toru Akiyama, 1Kohji Nakamura, 1Tomonori Ito, 2,3Hideto Miyake, 3Kazumasa Hiramatsu
Affiliations : 1Department of Physics Engineering, Mie University; 2Graduate school of Regional Innovation Studies, Mie University; 3Department of Electrical and Electronic Engineering, Mie University

Resume : AlN has been paid much attentions for use of deep ultraviolet light emitting devices. One of methods to fabricate AlN films on sapphire substrate is sapphire nitridation. However, the AlN thin film formed by sapphire nitridation has a chemically unstable nitrogen polarity. To improve crystal quality, there are several techniques on the basis of polarity inversion, where oxygen atoms trapped at the interface reset the polarity difference. Although a possible atom-scale model for polarity inversion has been proposed, there have been few theoretical studies for detailed structures and role of oxygen atoms on the polarity inversion layer. Here, we theoretically investigate effects of oxygen atoms on polarity inversion of AlN on N-polar AlN buffer layers. The calculations demonstrate that Al-polar AlN layers on N-polar AlN are formed with oxygen atoms substituting nitrogen lattice sites and Al vacancies, resulting in the formation of a diagonally stacked octahedral structure. We also find that the calculated interface energy of this interface structure is lower than that of N-polar AlN with oxygen atoms by ~0.04 eV/ Å2. Since the value of interface energy is positive, flat and smooth interface is preferable to reduce the energy deficit caused by the polarity inversion. These calculated results are qualitatively consistent with the experimental results, suggesting that charge neutrality induced by oxygen atoms are crucial for polarity inversion of N-polar AlN buffer layers.

Authors : Paweł Strak 1), Yoshihiro Kangawa 2,3), Stanislaw Krukowski 1), Paweł Kempisty 1,3), Michal Bockowski 1)
Affiliations : 1) Institute of High Pressure Physics PAS, Sokolowska 29/37, 01-142 Warsaw, Poland 2) Research Institute for Applied Mechanics, Kyushu University, Kasuga, Fukuoka 816-8580, Japan 3) Institute of Materials and Systems for Sustainability, Nagoya University, Nagoya 464-8603, Japan

Resume : Density Functional Theory (DFT) calculations was used to investigate adsorption of molecular oxygen (O2) and water at clean GaN(0001) surface. It was shown that both molecular gases dissociate at contact with GaN(0001) surface that reflects metallic character of the clean Ga(0001) surface with fractional occupation of Ga broken bond state. Density-functional calculations concerning the structure and stability of wurtzite polar Ga-terminated GaN(0001) surface are exposed on water. Water, hydrogen and oxygen adatoms entail states in the bandgap, as well as Ga dangling bonds. Adsorption of water molecule is possible on top position with adsorption energy equal -3.46 eV. Water molecule can dissociate on surface for hydrogen atom and OH hydroxyl group with further energy gain -1.74 eV. Hydrogen atom is located on top of gallium atom saturating single dangling bond. OH hydroxyl group is located in bridge configuration attached to two neighboring gallium atoms. OH group can further dissociate in to H and O atoms with additional energy gain -1.14 eV. The obtained mixed H-O coverage provides full saturation of Ga broken bonds in which all adsorbate bonds are also saturated. The process transforms active metallic surface into inactive Ga-terminated surface observed experimentally. The oxygen atoms are located in H3 site. Adsorption of oxygen molecule proceeds via dissociation on surface with the adsorption energy 13.2 eV. The results confirm that critical factor in the adsorption energy is related to Fermi level pinning at the surface and the electron transfer during the process. At zero level coverage molecular oxygen decomposes to be attached in H3 sites. The research was partially funded by Polish National Science Center by grant: DEC-2015/19/B/ST5/02136.

Authors : Marta Gladysiewicz Robert Kudrawiec
Affiliations : Faculty of Fundamental Problems of Technology, Wroclaw University of Science and Technology, Wybrzeże Wyspiańskiego 27, 50-370 Wrocław, Poland

Resume : The electronic band structure and material gain is calculated for polar ALGaN quantum wells grown on virtual AlGaN substrates using the kp method. The calculation were made for different virtual AlYGa1-YN substrates with Y=20, 40, 60, 80, and 100%. This can cover the UV-A, -B, and -C spectral range. The positive material gain for a given carrier concentration very strongly depends on the QW width and is blueshifted in comparison to the emission spectrum due to polarization effects. Calculations were made for different QW width and different carrier concentrations. For polar QW positive material gain is observed for carrier concentration n > 1×1019 cm-3 while for non-polar QWs a positive material gain is observed already for the carrier concentration one magnitude lower (n ~ 1×1018 cm-3). The material gain is larger for narrower QWs. For wider QWs the gain spectra are composed of multiple peaks since the energy differences between electron (hole) levels are smaller for such QWs and higher energy levels are occupied by electrons (holes) at the carrier concentration n > 1×1019 cm-3. In addition it is visible that the gain peak shifts to blue with the increase in carrier concentration. We also simulate luminescence spectra of LEDs. The model of random quantum well was applied to simulate this spectra. It is clearly shown that the emission wavelength from AlGaN quantum wells can be tuned by both the QW width and barriers thickness, but the range of QW width for which an efficient luminescence is expected is rather small (2-4nm), since a very weak electron hole overlap is present for wider QWs due to polarization effects. The emission wavelength and intensity can be significantly tuned by barriers if their thicknesses are comparable with the QW width. However the most promising way of tuning the emission wavelength is changing the QW content engineering.

Authors : B. Eydoux1,2, B. Baris1, S. Gauthier1, X. Bouju1, D. Martrou1
Affiliations : 1-Centre d'élaboration de matériaux et d'études structurales CEMES-CNRS UPR 8011, Nanosciences Group. 29 rue Jeanne Marvig, F-31055, Toulouse, France 2-Université Toulouse III, UPS, 118 route de Narbonne, F-31062 Toulouse, France;

Resume : The polar character of the wurtzite AlN(0001) implies a macroscopic surface dipole where the resulting electrostatic divergence is compensated by the distribution of surface charges. The surface charge must be equal to the bulk nitrogen charged plane divided by four. This criterion should be fulfilled whatever the adsorbate on the AlN(0001) surface. Three main reconstructions of AlN(0001) are accessible by usual technical growth such as NH3 molecular beam epitaxy and MOCVD. Namely, they are the (2x2)-Nad, the (2x2) 3Al-H and the (2x2) Al-H+Nad-H. Due to the high growth temperature, the surface is a statistical mixture of these three reconstructions. The (2x2)-Nad and the (2x2) 3Al-H hold an empty sp2 Al dangling bond (acceptor) that can transit toward a sp3 state when donor species are added. Both of the two non-stoichiometric reconstructions have their Fermi level quite close to the conduction band making them the most reactive ones. The reactivity of these reconstructed surface are explored in the framework of chemical reactions with typical molecules. We have also studied the adhesion of metals (Mg, In, Ag and Au) on the (2x2)-Nad reconstructed surface. In the particular case of Au, 2D monoatomic islands with two moiré patterns resulting from the mismatch between Au and AlN have been identified experimentally. DFT studies show that Al-Au and Nad-Au bond creations are the main driving force of the 2D monoatomic islands stabilization.

Authors : Takashi Hanada, Takashi Matsuoka
Affiliations : Institute for Materials Research, Tohoku University

Resume : The use of N-polar (-c) plane of InGaN is effective to increase In incorporation [Keller et al., APL 90, 191908 (2007)]. Red-to-blue GaN-based LEDs have been demonstrated on the N-polar plane [Shojiki et al., APEX 8, 061005 (2015)]. During the MOVPE growth, H is provided by NH3 dissociation even if it is not used as a carrier gas. Ab initio calculations have been performed to predict stable surface structures on various planes under various growth conditions. They predicted that the N-polar GaN surfaces are covered with H adatoms [Northrup et al., APL 85, 3429 (2004)]. However, group-III adatoms are necessary for growth. In this study, we evaluated the extremely low density of group-III adatoms on the growing surface of the N-polar plane by thermodynamic calculations under the presence of hydrogen. The extremely low density is consistent with the very low nucleation rate on a stepless N-polar surface [Lin et al., APEX 6, 035503 (2013)]. On the other hand, H suppresses the condensation of the adatoms. As a result, hydrogen assists the moderate step-flow growth if the sufficient density of steps are provided on the growing surface. Therefore off-cut substrates, which produce high-density steps, are usually used in the N-polar growth to prevent the hexagonal pyramids formed by the spiral growth. The quantitative evaluation of the group-III adatom density is important to discuss not only the the growth mode but also the incorporation ratio among group-III atoms.

Authors : Yuma Tsuboi, Toru Akiyama, Kohji Nakamura, Tomonori Ito
Affiliations : Department of Physics Engineering, Mie University, Tsu 514-8507, Japan.

Resume : Two-dimensional (2D) semiconductors have been paid much attention as adopting honeycomb structure via van der Waals force [1]. Recently, graphene-like hexagonal (Hex) AlN has fabricated [2], and other group-Ⅲ nitrides have been predicted for consisting Hex structure [3-4]. For fabricating AlN, it has been reported that Hex transform WZ structure when the film thickness increase [2]. However, the deformation processes and stability among different crystal structures in group-Ⅲ nitrides are remained unclear form theoretical viewpoint. In this study, we investigate the crystal structure deformation path in group-Ⅲ nitrides by using density functional calculations. Especially AlN, contour plots of the energy difference with respect to the total energy with WZ structure as functions of structural parameters b/a and c/a demonstrate that c/a mainly increases in transforming from WZ to Hex structure. Additionally, AlN with Hex structure whose lattice constant is larger than that with WZ structure, consistent with the experimental result [2]. The other hand, the energy barrier from Hex to WZ structure in AlN (0.079eV) is smaller than BN (0.25eV) [5]. The difference in the energy barrier originates from both the interaction among bond charges and attractive interaction among ionic charges. These results provide a precious guidance for crystal structure deformation as well as structural stability in group-Ⅲ nitrides. [1] Y. Shi et al., Nano Lett. 12, 2784 (2012). [2] P. Tsipas et al., Appl. Phys. Lett. 103, 251605 (2013). [3] H. Sahin et al., Phys. Rev. B 80, 155453 (2009). [4] H. L. Zhuang et al., Phys. Rev. B 87, 165415 (2013). [5] Y. Takemoto et al., e-J. Surf. Sci. Nanotech. 12, 79 (2014).


No abstract for this day

No abstract for this day