Defects are normally considered detrimental to the efficiency of photovoltaic devices, however, we have found a number of cases where defects are actually beneficial, explaining why the efficiency of polycrystalline CdTe exceeds that of single crystal cells [1]. These conclusions are reached by a direct correlation from atomic structure to electrical activity achieved through a combination of scanning transmission electron microscopy (STEM), electron energy loss spectroscopy (EELS), electron-beam induced current (EBIC), and density-functional theory (DFT) calculations.
STEM/EELS observations show that after CdCl₂ treatment a substantial Cl concentration lies on Te sites within a few atomic spacings of the GBs. DFT calculations show that this Cl is sufficient to invert the GBs to n-type, explaining why EBIC shows greatly improved current collection from grain boundaries after a CdCl₂ heat treatment [2]. EBIC also shows that CdCl₂ and Cu heat treatments improve the cell performance independently [3]. Furthermore, we find that intra-grain partial dislocation pairs do not create gap states but cause band bending, hence assisting the separation of electron-hole carriers and reducing recombination [4]. Te-S interdiffusion at the CdTe/CdS interface has also been observed [5]. Possible strategies for improving cell efficiency through defect engineering will be discussed.
References:
[1] C. Li, J. Poplawsky, Y. Wu, A. R. Lupini, A. Mouti, D. N. Leonard, N. Paudel, K. Jones, W. Yin, M. Al-Jassim, Y. Yan and S. J. Pennycook, Ultramicroscopy 134, 113 (2013)
[2] C. Li, Y. Wu, T. J. Pennycook, A. R. Lupini, D. N. Leonard, W. Yin, N. Paudel, M. Al-Jassim, Y. Yan and S. J. Pennycook, Phys. Rev. Lett. 111, 096403 (2013).
[3] J. D. Poplawsky, N. R. Paudel, C. Li, C. M. Parish, D. Leonard, Y. Yan, and S. J. Pennycook, Adv. Energy Mater., 1400454 (2014).
[4] C. Li, Y. Wu, J. Poplawsky, T. J. Pennycook, N. Paudel, W. Yin, S. J. Haigh, M. P. Oxley, A. R. Lupini, M. Al-Jassim, S. J. Pennycook, and Y. Yan, Phys. Rev. Lett. 112, 156103 (2014).
[5] C. Li, J. Poplawsky, N. Paudel, T. J. Pennycook, S. J. Haigh, M. Al-Jassim, Y. Yan, S. J. Pennycook, IEEE Journal of Photovoltaics, 4, 1636 (2014)
Acknowledgment:
This research was supported by the US DOE Office of Energy Efficiency and Renewable Energy, Foundational Program to Advance Cell Efficiency (F-PACE), (CL, NP, YY, SJP, M. A-J.), the Office of BES, Materials Science and Engineering Division (microscope facilities), and by ORNL&’s Center for Nanophase Materials Sciences (J. P.), which is also sponsored by DOE-BES. This research used resources of the National Energy Research Scientific Computing Center, which is supported by the DOE Office of Science under Contract No. DE-AC02-05CH11231.

CdTe is a widely-used photovoltaic material, currently second only to Si in market share. However, the practical efficiencies of CdTe photovoltaic cells are still well below the theoretical limit, indicating possible room for improvement. One aspect in which a fundamental understanding may lead to efficiency improvements is grain boundaries. Atomistic-level characterization, including microscopy and first principles modeling, is crucial in developing such a fundamental understanding. Select models dislocation cores and twin boundaries and their effects on optoelectronic properties have previously been investigated. But grain boundaries are not always stoichiometric and free of impurities. Point defects and defect pairs, whether intrinsic or extrinsic, on or near grain boundaries modify the structural and electronic properties of the grain boundaries. In this talk, we present the results of first principles density functional theory (DFT) calculations of point defects and pairs of point defects that are present on or near several representative grain boundary models. We discuss the thermodynamics of point defect and complex formation, as well as resultant changes in electronic structures. Effects on photovoltaic performance are also suggested.
ACKNOWLEDGEMENT: Use of the Center for Nanoscale Materials was supported by the U. S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Contract No. DE-AC02-06CH11357.

Cadmium Zinc Telluride (CZT) semiconductor crystal properties have been studied extensively with a focus on correlations to their radiation detector performance. The need for defect-free CZT crystals is imperative for optimal detector performance. Extended defects like Tellurium (Te) inclusions, twins, sub-grain boundaries, and dislocations are common defects found in CZT crystals; they alter the electrical properties and, therefore, the crystal's response to high energy radiation. In this research we studied the extended defects in CZT crystals from two separate ingots grown using the low-pressure Bridgman technique. We fabricated several detectors cut from wafers of two separate ingots by dicing, lapping, polishing, etching and applying gold metal contacts on the main surfaces of the crystals. Using infrared (IR) transmission microscope we analyzed the defects observed in the CZT detectors, showing three dimensional scans and plot size distributions of Te inclusions, twins and sub-grain boundaries observed in particular regions of the CZT detectors. We characterized electrical properties of the detectors by measuring bulk resistivity and detector response to gamma radiation. We observed that CZT detectors with more extended defects showed poor opto-electrical properties compared to detectors with fewer defects.

Starting from GaAs, it is generally assumed that high resistivity in semiconductors is caused by the presence of high concentration of deep defect levels. Experimentally it was shown that high resistivity can exist both in undoped CdTe/CZT samples and in CdTe/CZT samples doped by different shallow donors. This urges one to look either for the deep defects that are not related to the dopant type or to look for the mechanism, which can explain the formation of high resistivity without high concentration of deep defects at all.
In this work we extend the existing defect chemistry model to describe the slow cooling process, which takes place during the preparation of CdTe/CZT-based detectors. The resulting quasi-thermodynamic model describes the change of defect concentrations during slow cooling taking into account the defects freezing-in when the kinetic limitations prevent the particular defect from equilibration at particular conditions. Kinetic parameters of the model were derived from the first principles calculations. The model was used to investigate the dependence of the room temperature resistivity after the crystal cooling at different ambient conditions. For some range of Cd overpressures high resistivity at RT was obtained without involving deep defect levels.
Based on these results, a new mechanism for “cooling-assisted” formation of the high resistivity was proposed. This effect is based on the adjustment of the concentration of ionized V_Cd acceptor defects concentration to the concentration of In_Cd donor. Comparison of the results with the experimental data is performed. Specific CdTe material properties that allows for such compensation are discussed.

The combination of luminescence spectroscopy - in particular at liquid He temperatures - with the high spatial resolution of a scanning transmission electron microscopy (STEM) (δx < 1 nm at RT, δx < 5 nm at 10 K), as realized by the technique of low temperature scanning transmission electron microscopy cathodoluminescence microscopy (STEM-CL), provides a unique, extremely powerful tool for the optical nano-characterization of semiconductors, their heterostructures as well as their interfaces. Our CL-detection unit is integrated in a FEI STEM Tecnai F20 equipped with a liquid helium stage (T = 10 K / 300 K) and a light collecting parabolic mirror. Panchromatic as well as spectrally resolved (grating monochromator) CL imaging is used. In CL-imaging mode the CL-intensity is collected simultaneously to the STEM signal - typically chemical sensitive HAADF Z-contrast - at each pixel. The TEM acceleration voltage is optimized to minimize sample damage and prevent luminescence degeneration under electron beam excitation. In CL-spectral imaging mode, a complete CL spectrum is recorded at every single pixel. Subsequently, by evaluating the complex data matrix I_CL(lambda;,x,y), sets of simultaneously recorded monochromatic mappings I_CL(lambda;_i,x,y), CL peak wavelength mappings lambda;_CL(x,y), local spectrum linescans, local CL spectra, etc. can be processed.
Typical results which will be presented include: formation and annihilation of basal plane as well as prismatic stacking faults in non- and semi-polar GaN; individual GaN Quantum Dots; individual InGaN/GaN nanorods as well as complete core-shell nanorod-LEDs. We focus on the crystalline quality, local In incorporation, n- and p-layer quality and defects of the complete structures.

ZnO nanowires (NWs) have been attracting much interest due to their potential use as building blocks for the fabrication of electronic and optoelectronic devices. ZnO is a II-VI direct wide band gap (3.37 eV) semiconductor with large exciton energy (60 meV). At room temperature, ZnO nanowires present luminescence bands in both UV (near band emission) and visible regions. The visible (400-750 nm) and the near-infrared (750 -900 nm) emission spectra are usually attributed to deep levels such as oxygen vacancies (V_O), oxygen interstitials (O_I), zinc antisites (Zn_O), oxygen antisites (O_Zn) or zinc vacancies (V_Zn) [1]. Nevertheless, the origin of native point defects has been investigated and is still far from full understanding [2,3].
In this work, we report the optical and structural properties of ZnO nanowires grown by vapor-liquid-solid method on GaN, 4H-SiC, sapphire, Si <111> and Si<111>/SiO₂/Si structures using a 2 nm Au film as catalyst. Photoluminescence (PL) measurements using a CW 266 nm are performed to study the native point defects in the nanowires grown on different substrates. The results show that PL emission depends on substrate and on the nanowire polarity, even for the NWs grown on the same conditions. Deconvolution of PL spectra is used to identify the native points defects present in each structure. PL results are compared with the ZnO polarity and, reveal the importance of the substrate on the native defects formation.
[1] A. F. Kohan. Phys. Rev. B61, 15019 (2000).
[2] C. Chandrinou. Microelectronics Journal40, 296-298 (2006).
[3] A. Janotti. Phys. Rev. B76, 165202 (2007).

Auger process is well known as the collision of conduction band-edge electron and valence band-edge hole exciting of the electron at the conduction band-edge or the hole at the valence band-edge to much higher energy state. The excited electron or hole relaxes to the most stable band-edge state with the energy loss thermally, thus, the Auger process is treated as energy loss process generally. In this report, a different type Auger process is discussed. The process is explained three localized electron-hole pairs (shallow donor and deep acceptor pairs) converted to two free excitons (electron-hole pairs). For this case, the energy is up-converted from D-A pair energy to free exciton energy with reducing the number of particle from 3 to 2.
The samples used were GaN with different Si doping levels. The photoluminescence (PL) spectra were obtained at room temperature in vacuum. The exciting light was 325.0 nm He-Cd laser. The PL peaks of near band edge (NBE: 3.2 eV), blue luminescence (BL: 2.8 eV), and yellow luminescence (YL: 2.2 eV) were observed. The PL intensities of NBE, BL, and YL were proportional to the intensity of excited light at the weaker excitation light intensity region. Only the intensity of YL was, however, proportional to the power of 1/3 of the intensity of excited light at the stronger excitation light intensity region. The changing point for the YL intensity related with the excitation light intensity increased with decreasing the amount of Si for Si-doped GaN, but the point was the highest for the undoped GaN.
In order to explain the YL intensity being 1/3-order of excitation intensity, rate equations for the numbers of free exciton and D-A pair were assumed with the life time for the transition process. The results showed that the intensity of YL was proportional to the excitation light intensity when the radiative annihilation was dominant, and the intensity was proportional to the power of 1/3 of the excitation light when the Auger process was dominant. Considering from the D-A pair density, the radiative annihilation is dominant for the lower D-A pair density and the Auger process is dominant for the higher D-A pair density. The PL for NBE is considered not to be much influenced from the generated excition by Auger process due to the number of direct generated exciton by the excitation light is much higher than that by Auger process. These results well explains the YL intensity relationship with the excitation intensity. In addition, the crossing point was proportional to the root of the carrier lifetime of Auger process over the carrier lifetime of radiation annihilation process from the rate equation. The measured carrier lifetime of annihilation process for the lower Si-doped GaN was longer than that for the higher Si-doped GaN, and that for undoped GaN was shortest. This result supports the order of the changing point for YL intensity related with the excited light intensity.

Dislocations lead to spatial variations in electrical and optical functions of a semiconductor crystal, affecting degradation of devices. A great deal of efforts has been made to understand details of their dynamical properties as well as electrical and optical properties to improve yield and efficiency of devices. It should note that intrinsic properties of dislocations differ from those of grown-in dislocations. Here, we report native optical and electrical properties of fresh dislocations in GaN bulk crystals deformed plastically at elevated temperatures.
Fresh dislocations of (a/3)-type edge dislocations on the prismatic plane induced several photoluminescence peaks at around 1.8, 1.9 and 2.4 eV, which implies the formation of radiative recombination centers of the dislocations. Simultaneously, near-band-edge (3.48eV) photoluminescence intensity decreased remarkably for a high-density of non-radiative recombination centers originating in deformation-induced abundant Ga-vacancy related clusters. Variation of optical absorption dependent on the strain in plastically deformed GaN was understood in a model of the Franz-Keldysh effect by the electric fields associated with charged dislocations (~5.8 e/nm). Scanning spreading resistance microscopic images showed many spots with high conductivity around the induced dislocations. Current-voltage spectra indicated a Frenkel-Poole mechanism for the electrical conduction, suggesting an application as a dislocation nanowire.

In order to realize the high reliability of GaN-based light emitting device such as laser diode, it is important to reduce the dislocation density. Consequently, homoepitaxial GaN growth by using GaN substrate is now recommended. Here we present unique mechanical response of homoepitaxially grown high quality GaN layers when nano-deformed under contact loading of the (0001) or (10-10) surface denoted as the c- and m-planes. We clarify the mechanism of plastic deformation under stress with the defect formation.
The nanoindentation experiments were performed on the surface of c-plane (0001) and m-plane (10minus;10) GaN layers on bulk GaN. Penetration depth h was registered, while the load P was applied up to Pmax (9 mN (0.92 gf)). Young&’s modulus and critical resolved shear stress (CRSS) tau; were also obtained. tau; is the maximum load of transition from elastic to plastic deformation, i.e., pop-in load. The threading dislocation density was observed by cathodoluminescence (CL). Photoluminescence, cross-section TEM observation and molecular dynamics (MD) simulation at right after the plastic deformation were also done.
Obtained Young&’s modulus was 284.2 GPa for m-plane GaN. In contrast, the value in c-plane (0001) GaN was higher such as 323.8 GPa. Cathodoluminescence observed dislocation density. By analyzing the correlation between dislocation density and yield shear stress, we derived a relational equation that reprised the inverse Hall-Petch relation.
Load-depth curves for c- and m-plane GaN were obtained with various threading dislocation density. The maximum loadtau;(CRSS) of transition from elastic to plastic deformation, i.e., pop-in load, depended on dislocation density. The relation between ρ_T.D. and tau; was obtained as
tau;= 25 minus; 0.025#12539;ρ_T.D.^1/4
If there are four dislocations at the four corners of a grain, the grain size d is estimated as d asymp;ρ_T.D.^-1/2 because the number of dislocations on a line of unit length is ρ_T.D. By calculating grain size d and yield shear stress tau;, inverse Hall-Petch relation, i.e., tau; decreases with#12288;decreasing d, was observed.(tau;=25-0.025#12539;d^-1/2) with not only m-plane GaN but also c-plane GaN despite the large anisotropy in mechanical properties. From these results and calculated Peierls-Nabarro stress, we propose that the formation of twin is preferentially occurred on r {-1012} plane induced by indentation rather than dislocation nucleation or slip.
Cross-section STEM was performed in order to clarify those phenomena, and several kinds of dislocation multiplications and dislocation loops were observed including with r-plane dislocation. Molecular dynamics simulation also clearly showed r-plane dislocation and supported the mechanism with obtained r-plane {-1012} slip line right after plastic deformation. It is possible that the nucleation of the local metastable twin boundary along the r -plane concentrated the indentation stress, leading to an r -plane slip.

InGaN alloys are widely used as the active material in high efficiency GaN-based light emitting diodes (LEDs). However, this material contains a high density of threading dislocations. The impact on light emission of such defects has been widely investigated, and it has been shown that dislocations act as non-radiative recombination centres [1-4]. Conversely, in InGaN/GaN QW structures threading dislocations have been observed to open as V-pits [5] which are thought to prevent carrier recombination at dislocation cores [6]. The impact of dislocations on III-Nitrides is therefore twofold and is still poorly understood. In this study we employ a multi-microscopy analysis study of dislocations in InGaN in order to correlate the emission and structural properties of the same defect.
The sample investigated here consists of a thick 150 nm In_0.09Ga_0.91N layer grown by metal-organic vapor phase epitaxy. We thus conducted a multi-microscopy analysis, by performing atomic force microscopy, scanning electron microscopy coupled with cathodoluminescence and transmission electron microscopy on exactly the same dislocations. In this study we were particularly interested in investigating the correlation between the dislocation type, the intensity and wavelength of light emitted by different parts of the V-pit forming at the dislocation apex, and the structure of the dislocation core.
Using this approach we show that irrespective of the dislocation type, a V-pit can emit more intense light than its surrounding material. This observation challenges our current thinking about the general behavior of dislocations in semiconductor materials. The analysis also reveals that the light emitted from the V-pits is at a longer wavelength than its surrounding material. This result suggest that indium may segregate around the core of the dislocations and on the facets of the V-pit. Finally we highlight a correlation between the wavelength shift of the emission from the facets of a V-pit and the intensity of light emitted from the centre of the V-pit. We show that for a dislocation with poor emission properties, the shift is inversely related to the intensity, while for a dislocation with enhanced emission properties, the shift is unrelated to the intensity. This correlation suggests that different core structures may be at the origin of the various optical properties of dislocations in InGaN.
[1] Sugahara et al., Jap. J. Appl. Phys. 37, 398 (1998)
[2] Cherns et al., Appl. Phys. Lett. 78, 2691 (2001)
[3] Dai et al., Appl. Phys. Lett. 94, 111109 (2009)
[4] Armstrong et al., Appl. Phys. Lett. 101, 163201 (2012)
[5] Wu et al., Appl. Phys. Lett. 72, 692 (1998)
[6] Hangleiter et al., Phys. Rev. Lett. 95, 127402 (2005)

In this presentation the UC Berkeley Materials Science Department Chair, Dr. Mark Asta will speak about Prof. Jack Washburn&’s many contributions to the science of structural defects, studied through transmission electron microscopy, and his contributions in education and mentoring of students, both domestic and international. Some remarks from Prof. Wasburn&’s former students are also expected.

Befitting this symposium in honor of TDS&’ late advisor, Professor Jack Washburn, this presentation describes a scientific journey that began while TDS was a student of materials science at UC Berkeley in the early 1980&’s, and culminated in the recent achievement of a true metal/semiconductor superlattice by TDS&’ last Ph.D. students at Purdue University, including Dr. Saha. The nucleus of this thirty-year effort emerged as a result of conversations with Jack Washburn in the early 1980&’s on the topic of defects in semiconductor heterostructures.
At about the same time, device researchers were beginning to experiment with metal-base and permeable-base transistor concepts that required direct or through-pore epitaxial growth of semiconductor/metal/semiconductor double-barrier heterostructures, the earliest of which were realized in the Si/CoSi₂/Si system. Later work at Bellcore by TDS, C.J. Palmstrom and others explored the growth of GaAs/NiAl/GaAs, GaAs/ErAs/GaAs, and related heterostructures. Although these double-barrier heterostructures were demonstrated, interfacial defects originating from space group symmetry mismatch, lattice mismatch and growth morphology prevented the epitaxial growth of multi-period superlattices. This obstacle was recently overcome with the realization of rocksalt superlattices comprising lattice-matched mononitrides, best exemplified by the HfN/ScN and TiN/(Al,Sc)N systems. ScN and (Al_xSc_1-x)N (x < 0.82) are rocksalt semiconductors in film and bulk form that can be doped n-type or p-type. (Al,Sc)N can be stabilized for high AlN mole fractions by lattice-matched epitaxy. TiN, HfN and similar nitride films can be good metals with carrier concentrations approaching 10²² cm^-3. The growth of these rocksalt nitride metal/semiconductor superlattices as well as their potential applications as optical and thermionic metamaterials will be highlighted.

Cadmium oxide (CdO), a direct bandgap II-IV semiconductor with a rocksalt crystal structure, recently has regained attention by offering a variety of potential optoelectronic applications, particularly as transparent conducting oxide (TCO) layers in the next generation solar cells, flat-panel displays, etc. and also as optically active components in photovoltaics (e.g., in n-CdO/p-Si structures demonstrating up to 8.8% efficiency [1, 2]). In contrast to comprehensive theoretical studies of the band structure of CdO [3, 4], very few spectroscopic results are available for direct comparison of the calculated and experimental electronic transitions, so that even such fundamental issues as the actual bandgap values still remain controversial, in part due to uncertainties related to structural quality of the investigated materials. Recent advances in the epitaxial growth provide an opportunity to minimize such uncertainties and to improve the understanding of electronic properties by offering high-quality CdO crystals. In the present work, we report on the developments of structural and electronic properties of CdO films grown on r-plane (1-102) sapphire substrates by metal organic vapor phase epitaxy (MOVPE) under the varied from cadmium-rich to oxygen-rich conditions. By combining the X-ray diffraction, optical emission (photoluminescence) and absorption spectroscopy measurements, we show that a gradual change towards oxygen-rich conditions during the MOVPE synthesis leads to a build-up of strain in the films and also causes a red-shift of the fundamental absorption edge, the latter being attributed to the formation of spurious Cd_xO_2-x. The direct bandgap of ~2.5 eV along with the onset of two indirect bandgaps, at ~0.95 eV and ~1.1 eV, were estimated from the comparative analysis of the emission and absorption spectra. These results agree well with the band structure calculations predicting existence of two indirect energy gaps, (Σ₃-Γ₁) and (L₃-Γ₁), in addition to the direct gap (Γ₁₅-Γ₁). The growth temperature is found to be the key factor in controlling structural and optical properties of CdO, primarily via affecting the process of precursor pyrolysis. As a final point, we discuss possible routes and methods for stabilization of other CdO phases, wurtzite and zincblende, which is an important practical implication of the present study.
References
[1] M. Zaien, N.M. Ahmed, and Z. Hassan, Superlattices and Microstructures, 52, 800 (2012).
[2] R.A. Ismail, and O.A. Abdulrazaq, Sol. Energy Mater. Sol. Cell, 91, 903 (2007).
[3] K. Maschke, U. Rossler, Phys. Stat. Sol. 28, 577 (1968).
[4] S. Tewari, Solid State Commun. 12, 437 (1973).

A comprehensive microscopy analysis has been undertaken to study three-dimensional quantum dot (QD) ordering in multilayered In0.4Ga0.6As/GaAs structures grown with an As2 flux at different substrate temperatures. Atomic force microscopy, transmission electron microscopy, and photoluminescence measurements were employed to fully understand the formation of these extended dot structures. Changes in the lateral pattern of QD ordering are correlated with their vertical alignment. These correlations are analyzed in light of the inherent transformation of the wetting and spacer layers, as well as changes in the shape, strain, and composition of individual QDs. The experimental results are attributed to the anisotropy in the thermally activated surface mass transport and the relaxation of elastic stresses.

Dislocations are important defects in semiconductors with strong effects on their electronic properties. A common approach to describe the effect a dislocation has on the electronic structure is to consider two separate regions: (i) the core region where bonds are broken, i.e. localized dangling bonds are present and (ii) the local strain-field around the core where atoms are fully (bulk-like) coordinated but slightly displaced with respect to their bulk positions. The core region - due to the presence of localized dangling bond states - is expected to induce deep states within the band gap (making the dislocation electrically active) while the strained region around the dislocation shifts the band-edge due to the deformation potential.
Dislocations in III-Nitrides were believed to behave different. It was supposed that that due to the strong iconicity of the bonds they would not act as non-radiative recombination centers. In this presentation, combined structural and optical and theoretical studies on dislocations in GaN will be summarized. It will be shown, that while a-type edge dislocations induce non-radiative recombination independent on the specific core structure of the dislocation, a-type screw dislocations exhibit an unusual strong luminescence at 3.35 eV. According to theory in both cases the high shear strain close the dislocation plays the most important role. In case of the a-type edge dislocations the shear strain induces empty metallic bond states between Ga atoms close to the dislocation core acting as nonradiative recombination centers. In case of a-type screw dislocations the strain field causes a mixing of the s-type state at the conduction band minimum with the next highest state that has p-character and is thus susceptible to the shear strain induced by the dislocation. In consequence the dislocation represenst a kind of quantum wire and localize holes as well as electrons. This is in stark contrast to dislocations in conventional III-V and group IV semiconductors. Photoluminescence experiments show the expected strong polarization along the dislocation line.

Dislocations are one-dimensional topological defects which occur frequently in functional thin film materials and which typically degrade the performance of optoelectronic devices. However, InGaN-based devices with a low In content can emit light with high efficiencies despite high dislocation densities. This is still not well understood but may relate to the distribution of indium atoms on metal sites in InGaN, as this may influence carrier localization in InGaN quantum wells and affect radiative recombination rates. Previous theoretical work predicted that indium should segregate to c-type (screw) threading dislocations[1]. However, only a small proportion of the threading dislocations found in most InGaN films are c-type and the extent of In segregation to the majority a-type and (a+c)-type dislocation cores has remained unknown.
Here, we present a theoretical study predicting compositional segregation around a full range of dislocation core structures in InGaN alloys over a range of technologically relevant low indium-content InGaN alloys. These simulations were performed using classical atomistic modeling techniques, based on Stillinger-Weber-like potentials developed for InGaN and widely validated in the literature[2], combined with a Metropolis Monte Carlo sampling scheme. This method allowed the prediction of equilibrium dislocation cores at a range of growth temperatures.
Local enrichment of indium in areas of tensile stress and depletion in areas of compressive stress was found in cores with an a-type component, along with an enrichment around cores with a c-type component. Exact predicted composition profiles will be presented, along with typical core configurations with show the exact bonding arrangement at the core. These results allow analysis of the expected number and length of In-N-In-N... chains near the core, which differs substantially from that for a random alloy, which is of specific interest since these chains are associated with localised hole states[3].
Theoretical predictions were experimentally verified using aberration-corrected scanning transmission electron microscopy of the dislocation cores, with energy-dispersive X-ray spectroscopy providing elemental analysis. This found substantial indium enrichment in the vicinity of (a+c)-type dislocation cores, specifically within the tensile region of the dislocation core, as predicted by the theoretical simulations.
The results have substantial implications for the electronic properties and efficiencies of devices containing low indium-content InGaN.
[1] Lei, H., Chen, J. & Ruterana, P. Applied Physics Letters 96, 161901 (2010).
[2] Lei, H., Chen, J., Jiang, X. & Nouet, G. Microelectronics Journal 40, 342-345 (2009).
[3] Chichibu, S. F. et al. Nature Materials 5, 810 (2006).

GaN/InGaN thin films are usually grown on the lattice-mismatched substrates, which leaves large number of defects inside film. Enormous amount of works were done to reduce the defects in p-GaN/InGaN/n-GaN/Sapphire system, but it was found that the V-groove defect is one of the inevitable defects coming from the threading screw dislocations. Even though the V-groove defects are commonly observed defects in InGaN/GaN multi-quantum well (MQW), few luminescence characterization were done, mostly through the cathodoluminescence in scanning electron microscopy and near-field scanning optical microscopy. In this study, we carried out 1:1 mapping of the microstructure and the luminescence in cathodoluminescence transmission electron microscopy (CL-TEM) to find luminescence characteristics in the V-groove defect region. V-groove defects are known to be developed on (1-101) when GaN was grown on basal plane of Sapphire and the screw dislocation is the source of the V-groove development. It was confirmed that the dead luminescence centers were located at the core of threading dislocations and the corner of c-plane and side walls of v-grooves. Most of MQW region showed luminescence of ~460 nm, while the V-groove region showed shift toward the wavelength of 390 nm. Also some of the V-groove region showed extra luminescence at the corner of (0001) and (10-10) plane with the wavelength of ~ 430nm, which was believed to be the existence of the quantum dots, or localized high indium concentration.

Trench defects are a commonly occurring flaw in InGaN/GaN quantum well (QWs) structures, where a region of material is enclosed by a V-shaped trench [1,2]. Nevertheless their impact on light emitting diodes (LEDs) has been largely overlooked. Here we provide evidence of the negative impact of trench defects on LED emission, and their potential contribution to the "green gap".
A first set of two five-period blue InGaN/GaN QW structures with similar photoluminescence peak wavelength were grown by metal-organic vapor phase epitaxy, with different trimethylindium (TMI) fluxes, with the higher TMI flow yielding a tenfold increase in trench defect density [3]. Two LED structures were grown under identical conditions as the initial QW structures, with a p-doped GaN layer deposited on top of the QWs. A second set of samples was grown with five QWs grown at a temperature ranging from 690^oC to 780^oC, and no p-GaN deposited. The samples were investigated by atomic force microscopy (AFM), high angle annular dark field scanning transmission electron microscopy (HAADF-STEM), photoluminescence (PL) and electroluminescence (EL).
In the first set of samples, HAADF-STEM showed that during the growth and anneal of the p-doped GaN layer active region degradation occurs. Besides the presence of indium platelets (as reported by Li et al. [4]) in the sample grown at higher TMI flow, we show that partial or complete desorption of the QWs enclosed by the trench defects occurs. The presence of high densities of trench defects in the LEDs was found to relate to a two order of magnitude reduction in PL and EL emission efficiency, for a range of excitation power densities and drive currents. This reduction in emission efficiency was attributed to an increase in the density of non-radiative recombination centers within the QW stack, believed to be associated with the stacking mismatch boundaries which form part of the sub-surface structure of the trench defects.
In the second set of samples we show a two decade increase in the density of trench defects between blue- and green-emitting QW structures, resulting in (45 ± 2)% of the surface of the green-emitting structure being covered by trench defects. This result implies that the efficiency of green-emitting LEDs may be strongly affected by the presence of trench defects. Our results are consistent with a model that the "green gap" problem might relate to localized strain relaxation occurring through defects [5].
In conclusion we show that trench defects have a negative impact on GaN LEDs. Our data support the hypothesis that trench defects could more severely degrade the efficiency of green LEDs.
[1] Massabuau et al.,Appl. Phys. Lett.101 212107 (2012)
[2] Bruckbauer et al., Appl. Phys. Lett.98 141908 (2011)
[3] Massabuau et al.,Phys. Status Solidi A,211 740 (2014)
[4] Li et al.,Appl. Phys. Lett.103 152109 (2013)
[5] Langer et al.,Appl. Phys. Lett.103 022108 (2013)

GaN and the corresponding ternary compounds AlGaN and InGaN are promising materials for LEDs in the UV and visible spectral range. The occurrence of structural defects strongly a#64256;ects the e#64259;ciency of these LEDs. Therefore we focus on the investigation of the optical properties of such defects and on the assignment of speci#64257;c spectral features to distinct defect types. Our investigations include the direct correlation of localized emission bands measured by cathodoluminescence (CL) in a scanning electron microscope (SEM) with defects found in high resolution transmission electron microscopy (TEM) at the identical sample spot. Using theses results, we are able to model the electronic structure of basal plane stacking faults (BSFs) applying a quantum well model regarding the defects as cubic inclusions in a wurtzite matrix. I₁-, I₂-, and E-type BSFs di#64256;er in the speci#64257;c thickness of this inclusion.
Applying this model to AlGaN with low Al content we found hints that BSFs in semipolar AlGaN layers cause local changes of the Al content. This strongly a#64256;ects the usability of AlGaN as an electron blocking layer in nitride based LEDs. We are able to verify this conclusion by correlation of results from SEM-CL, electron beam induced current (EBIC), and high resolution TEM on a 3D semipolar LED structure. With these #64257;ndings, the increased leakage current of this LED can be explained.

The control of defects both during the growth and subsequent processing of oxide films is vital to obtaining useful film properties. Different approaches to strain engineering will be discussed in the context of the epitaxial growth of lead-free ferroelectric BaTiO₃ thin-films. The challenge is especially great for the case of growing this system epitaxially on Si (100) substrates, where large lattice-constant and thermal-expansion mismatches can lead to large residual strains. We report on the epitaxial growth of room temperature BaTiO₃ thin (~1050 nm) films on Si (100) substrate by pulsed laser deposition technique using the domain matching epitaxy paradigm. By introducing epitaxial buffer layers composed of SrTiO₃/MgO/TiN we have been able to prepare almost fully relaxed cube-on-cube epitaxial BaTiO₃ on Si (100) as evidenced by the in-plane and out-of-plane x-ray diffraction, and transmission electron microscopy. X-ray photo emission spectroscopic (XPS) measurements show that Ti is in the 4+ valence state. Polarization hysteresis measurements together with Raman spectroscopy and temperature-dependent x-ray diffraction confirm the room temperature ferroelectric nature of the BaTiO₃ and its expected diamagnetic properties. Interestingly, laser irradiation of the BaTiO₃ thin film is found to induce ferromagnetic-like behavior, although largely at the expense of its ferroelectric properties. The coexistence of strong ferromagnetism and ferroelectricity will be discussed. The transition to the ferromagnetic is believed to originate from the creation of oxygen vacancies, and can be reversed by subsequent high temperature oxygen annealing.
Acknowledgement: SRS acknowledges support from the National Academy of Science (NAS), USA as a NRC postdoctoral research associate fellowship. The authors are pleased to acknowledge the support of the Army Research Office under Grant W911NF-04-D-0003. Also, the authors acknowledge the use of the Analytical Instrumentation Facility (AIF) at North Carolina State University, which is supported by the State of North Carolina and the National Science Foundation.

In topological insulators (TIs) the energy states are fundamentally modified from ordinary insulators by strong spin-orbit interactions, giving rise to a topologically distinct state of matter with a gapped insulating bulk and a gapless metallic surface. Bi-based chalcogenides are confirmed as a prototypical TI due to their simple surface Dirac cone and relatively large bulk energy gap, such as Bi₂Se₃ and Bi₂Te₃. The possibility to grow Bi₂Se₃ epitaxial thin film on different substrates such as Si (111), SrTiO₃ (111), GaN (0001), GaAs(111)B, and InP (111) has been extensively investigated within the past few years. In this work, epitaxial Bi₂Se₃ thin films were successfully grown by pulsed laser deposition on different platforms; which are Al₂O₃ (0001), Cr₂O₃ (0001)/Al₂O₃ (0001) and MgO (111)/Al₂O₃ (0001), accompanying by the lattice misfit ~13%, 16.3% and 28.12%, respectively by domain matching epitaxy paradigm which can address epitaxy across the misfit scale. Relatively low repetition rate (0.1 Hz) and low deposition temperature (150 ^oC) are key elements to achieving high quality Bi₂Se₃ epitaxial films. The XRDPhi;-scan reveals that the presence of six peaks corresponding to Bi₂Se₃ (01-15) planes are observed in all cases, indicating the formation of twins should be considered more as an intrinsic crystal feature rather than be directly attributed to the choice of substrate. The Bi₂Se₃ shows 60^o rotation with respect to Al₂O₃ (0001) substrate and 30^o rotation in MgO (111) and Cr₂O₃ (0001) templates. The transport properties measured by physical property measurement system and electronic structures collected by angle resolved photoelectron spectroscopy are systematically studied to study the substrate dependence related to TI properties.

Positron annihilation spectroscopy is a very powerful technique for the detection, identification and quantification of vacancy-type defects in semiconductors [1]. In the past decades, it has been used to reveal the relationship between opto-electronic properties and specific defects in a wide variety of materials - examples include parasitic yellow luminescence in GaN [2], dominant acceptor defects in ZnO [3] and broad-band absorption causing brown coloration in natural diamond [4]. In typical binary compound semiconductors, the selective sensitivity of the technique is rather strongly limited to cation vacancies that possess significant open volume and suitable charge (negative of neutral). On the other hand, oxygen vacancies in oxide semiconductors are a widely debated topic. The properties attributed to oxygen vacancies include the inherent n-type conduction, poor p-type dopability, coloration (absorption), deep level luminescence and non-radiative recombination, while the only direct experimental evidence of their existence has been obtained on the crystal surface.
We will present recent advances in combining state-of-the-art positron annihilation experiments and ab initio computational approaches. The latter can be used to model both the positron lifetime and the electron-positron momentum distribution - quantities that can be directly compared with experimental results. We have applied these methods to identify vacancy-type defects in SnO₂, In₂O₃ and Ga₂O₃. We will show that cation-vacancy-related defects important compensating centers in n-type material in all these three oxides. In addition, we will show that O vacancies can be detected, in cases where they are complexed with cation vacancies.
[1] F. Tuomisto and I. Makkonen, Rev. Mod. Phys. 85, 1583 (2013)
[2] K. Saarinen et al., Phys. Rev. Lett. 79, 3030 (1997).
[3] F. Tuomisto et al., Phys. Rev. Lett. 91, 205502 (2003).
[4] J.-M. Mäki et al., Phys. Rev. Lett. 107, 210743 (2011).

Interesting physical properties have been found in AlB₂-type analogous “tiling” compounds, composed of 2D boron atomic sheets (based on hexagonal graphitic structure) sandwiching rare-earth and transition metal atoms [1]. Ubiquitous building defects in the “tiling” patterns were discovered to exist in even single crystals and proposed to be the origin in anomalous properties in superconductivity, electronic properties, and magnetism [2]. A striking feature is that the defect formation was recently found to be able to be controlled and thus used as a way to control physical properties. Namely, with a counterintuitive approach, single crystals of α-TmAlB₄, which were indicated from TEM and advanced XRD analysis to be defect-free, were successfully grown [3]. The large effect of the building defects on the physical properties could be directly confirmed, such as the origin of “missing entropy”. As a new direction of application of the defects we present results on thermoelectrics, Thermoelectric materials are being actively developed now, utilizing new concepts, state-of-the-art nanotechnology, and nanomaterials [4]. Time-domain thermoreflectance (TDTR) measurements on microcrystals revealed that a percent or so of building defects significantly lower the thermal conductivity by 30% [5]. Coupled with the ubiquitous nature of these building defects in layered compounds, and the recent synthesis control of defect formation, this is shown to be another interesting method to control thermal conductivity and enhance thermoelectric conversion efficiency. These building defects are quite subtle and may in some cases be unperceived, and might possibly be the origin of anomalous physical properties behavior in other layered systems also.
[1] T. Mori, in: The Rare Earth Elements: Fundamentals and Application, ed. D. Atwood (Wiley, Chichester, 2012).
[2] e.g. Phys. Rev. B76, 064404 (2007), Nature Phys.4, 603 (2008), J. Appl. Phys105, 07E124 (2009), 109, 07E111 (2011). Phil. Mag. 93, 1054 (2013).
[3] J. Appl. Phys.111, 07E127 (2012).
[4] Thermoelectric Nanomaterials, ed. K. Koumoto and T. Mori, Springer Series in Materials Science (Springer, Heidelberg, 2013) pp.1-373.
[5] APL Mat.2, 046113 (2014).

P-type transparent conducting materials (p-TCMs) have potential applications in optoelectronics, particularly in photovoltaics as top contacts to thin film solar cells and as interconnect layers in tandem structures. However, current state-of-the-art p-TCMs have hole conductivities orders of magnitude lower than those of standard n-type TCMs (ITO, AZO, etc.), and, as a result, p-TCMs are not yet suitable for practical commercial devices and an efficient device using a p-type TCM has not yet been demonstrated.
With the aim of aiding commercial TCM implementation, we have developed a room temperature deposition process that yields high-performing amorphous p-type Cu_xZn_1-xS films by pulsed laser deposition (0 < x < 0.5, with best transparency and conductivity at x ~ 0.3). All films (x > 0) have a positive Seebeck coefficient, confirming p-type conduction. Optimized films have an optical band gap > 3.1 eV, transparency in the visible of >70%, and hole conductivities up to 30 S/cm. The TCM figure of merit (0.004 Omega;^-1) for these films exceeds all prior literature reports for room temperature processed p-TCMs.
Hole conductivity can be further enhanced by deposition at elevated temperature. Deposition of Cu_xZn_1-xS at 5500C resulted in p-type crystalline films in the sphalerite phase. Optimal films have near-record conductivities > 100 S/cm, sheet resistance as low as 300 Omega;/#9744;, hole concentrations up to 5×10²⁰ cm^-3, and transparency >70%. These crystalline films have a near-record TCM figure of merit of 0.01 Omega;^-1. For comparison, the figures of merit for typical n-TCMs range from 0.2 to 7 Omega;^-1 and the highest value reported for a crystalline p-TCM is ~0.02 Omega;^-1 (Zn_1-xAl_xO:Cu₂O). X-ray photoelectron spectroscopy (XPS) measurements of the valence band (VB) and work function are in agreement with the predicted ab initio crystalline band structure and show that the VB is aligned with the Fermi level. Conductivity measurements performed from 15-450 K are consistent with the band conduction of holes expected from degenerate doping.
Simple heterojunctions of Cu_xZn_1-xS with n-type silicon showed rectification in the dark and photovoltaic activity under illumination. The application of Cu_xZn_1-xS as a transparent contact for n-type InP and other n-type absorbers and as a transparent interconnect in various tandem architectures will be discussed, as well as manufacturability and scalability considerations.

Relevant examples of the importance of interfacial engineering on the solar energy conversion and photocatalytic efficiency of large bandgap oxide semiconductor will be presented. They include the electronic structure origins of interfacial losses in hematite photoanodes and the cause of the literature-reported order-of-magnitude photocurrent increase upon short high-temperature annealing, the size dependence on the interfacial chemistry of maghemite, the c-axis growth of stochiometric and defective tin oxide nanorods, the quantum size effect on the orbital character of the conduction band of anatase nanocrystals and its in-situ characterization of the lateral displacement, complex impedance spectroscopy, and electrical resistivity as a function of temperature and oxygen partial pressure within controlled environments consistent with a Frenkel-based defect disorder model.

Strontium titanate (SrTiO₃) is a wide-band-gap semiconductor with a variety of novel properties. In this work, bulk single crystal SrTiO₃ samples were heated to 1200 C, resulting in the creation of point defects. These thermally treated samples showed large persistent photoconductivity (PPC) at room temperature. Illumination with sub-gap light (>2.9 eV) caused an increase in free-electron concentration by over two orders of magnitude. After the light is turned off, the conductivity persists at room temperature, with essentially zero decay over several days. The results of electron paramagnetic resonance (EPR) measurements suggest that a point defect is responsible for PPC because the photo-induced response of one of the EPR signals is similar to that seen for the PPC. Due to a large barrier for recapture, the photo-excited electron remains in the conduction band, where it contributes to the conductivity. Possible applications, such as holographic memory and optically defined circuits, will be discussed.

Controlling defects in colloidal semiconductor quantum dots (QDs) is a tremendous challenge, due to their inherently large surface area and complicated surface chemistry. We will discuss our recent discoveries about the origin of defects in QD systems. With this knowledge, we are able to tune the charge transport properties of QD solids by surface chemistry treatments to tune the amount of impurities. We will especially focus on the beneficial effects of defects in dark transport and photoconductivity, using scanning probe techniques combined with device characterization. Practical implications on large area electronics and optoelectronic device design will also be discussed.

High mobility conduction at the interface of two dissimilar materials and the ability to manipulate the carrier density and mobility of the interfacial charge by electrostatic gating have lead to many fascinating phenomena in the past decade. In this presentation I will give an overview of our recent work on interfacial reconstruction and formation of two-dimensional electron-gas in strongly-correlated multilayer oxides investigated by means of high-resolution electron microscopy, electron energy-loss spectroscopy, electron holography, and DC electrical transport measurements. Focus will be on the interfacial electronic structure and charge transfer that are related atomic scale defects such as oxygen-hole depletion, valence-state variation, chemical diffusion, and interfacial strain. Examples include SrTiO₃/RO/SrTiO₃ (R=La, Pr, Nd, Sm, Y) and (LaMnO₃)_2n/(SrMnO₃)_n superlattices as well as Heusler alloy (Co₂MnSi and Co₂FeSi)-SrTiO₃ interfaces. The impact of interfacial charge transfer on competing ferroelectric and ferromagnetic order, metal-insulator transition, superconductivity and other magnetoelectric properties will be discussed. Collaborations with M.G. Han, L.J. Wu, D. Su. C.B. Eom, I. Bosovic, R.C. Budhani and C. Ahn are acknowledged. The work was supported by US DOE/BES, under Contract No. DE-AC02-98CH10886.

CZT is a semiconductor material that promises to be a good candidate for uncooled gamma radiation detectors. However, to date, technological difficulties in production of large size defect-free CZT crystals are still prevalent. The most common problem is accumulation of Tellurium precipitates as microscopic inclusions due to an unstable melt interface during crystal growth. These inclusions influence the charge collection through charge trapping and electric field distortion thereby leading to poor detector performance. Competing with the Te inclusions defects are edge effects leading to surface leakage currents. In this work, we characterized the oxide layer formed on as-prepared and Pt-ion irradiated surfaces of CdZnTe crystals. CdZnTe crystal surface was interrogated with mono-energetic Mg K-alpha X-rays (1253.6 eV). XPS peaks corresponding to elemental Cd, Zn, Te and their oxides are observed. Photoelectron lines are observed for Te in CdZnTe crystal at 572.5 eV (3d5/2) and 582.5 eV (3d3/2) and the corresponding Te oxides at 576.5 eV (3d5/2) and 586.5 eV (3d3/2) lines. Likewise we observed photoelectron lines corresponding to Zn and Zn-oxides at 3p3/2 and 3p1/2; and Cd at 3d5/2 (405 eV) and 3d3/2 (412 eV) as well as oxides of Cd, CdO2 at 404 eV and CdOx at 411 eV. Carbon photoelectron line at 282 eV probably from Carbide was used to calibrate the spectra. The spectrum also shows higher Te:Te-Oxide ratios than Cd-Cd-Oxides and Zn-Zn-Oxides. This is explained by the oxidation of Te inclusions present in the crystal. Our results show the presence of oxide layers, and the results are important for metal-semiconductor interface engineering to lower the barrier height.

In this research work, we have demonstrated that amorphous ZnO nanobranches (a-ZnO NBs) could spontaneously react from the crystalline ZnO NWs (c-ZnO NWs) at specific humid environment. We can make the c-ZnO NWs spontaneous reaction happen at different humid environments which due to the interaction between the H₂O molecules and the surface of ZnO NWs. Otherwise the a-ZnO NBs spontaneous reaction can be suppressed by oxygen/hydrogen plasma surface passivation. The spontaneous reaction mechanism and result can be analyzed by humidity controlling and optical microscope (OM)/scanning electron microscope (SEM)/Kelvin probe force microscopy (KPFM)/transmission electron microscopy (TEM) system. We also use the hydrogen plasma surface defect passivation to improve the UV sensing sensitivity more than two fold. This work provides the mechanism and methods of the a-ZnO NBs spontaneous growth and offers the passivation treatment for strengthening and enhancing ZnO-based nanodevice application in humid environment and UV light detection, respectively.

Oxide based HJ (hetero-junction) thin film solar cells, specifically with intrinsic p-type Cu₂O (band gap E_g^Cu₂Oasymp; 2.1 eV) as the absorber layer, show a theoretical efficiency up to 18% ^[1], while currently reported experimental values remain between 2sim;4% ^[2][3]. For further understanding and improvement, both the thin film properties (transport, optical, structural) and interface defects should be investigated. In the present work, a systematic study on the influence of the target power on the PL (Photoluminescence), topographical and structural properties of Cu_xO thin film has been undertaken.
Cu_xO thin films have been deposited on a quartz substrate by reactive rf magnetron sputtering at different target powers P_tar (140sim;190 W) by fixing other process parameters: oxygen mass flow Q(O₂)=3.0 sccm, argon mass flow Q(Ar)=20.0 sccm and substrate temperature T_sub=400 °C. The room-temperature PL spectra show three vacancy-related defects yielding p-type conductivity, including: V_o²⁺, V_o⁺ and V_Cu^[4] for films deposited with P_tar >170 W, while these defect signals are much weaker with P_tar <170 W. In our previous study ^[5], it was observed that films are dominated by the Cu₂O phase with p-type conductivity above 170 W while below that O-rich phases (CuO and Cu₄O₃) dominate, and concurrently films sputtered at higher P_tar demonstrate sharper absorption edges in transmittance spectra. At P_tar =190 W, particularly, the PL spectrum exhibits fine features (phonon-assisted transitions) in the exciton emission band with peaks of ± Γ₁₂^-and +Γ₁₅^-, and the peak intensities of ±Γ₁₂^- slightly exceeds +Γ₁₅^-[6]. Furthermore, in time resolved PL measurements all the films (prepared above 160 W) demonstrate a decreased effective life time from ~380 to ~150 ns with increasing P_tar.

The first step towards building a high-efficiency solar cell is to develop an absorber with few recombination-active defects. Many photovoltaic technologies have already achieved this (monocrystalline Si, III-V materials grown on lattice-matched substrates, perovskites, polycrystalline CdTe and CIGS); those that have not (a-Si:H, organics) have been limited to low open-circuit voltage. The second step is to develop contacts that both inhibit surface recombination and allow for low-resistance collection of either only electrons or only holes. For most technologies, this is both more difficult and less explored than the first, and a prescribed methodology for selecting materials for contacts to solar cells is still lacking.
We elucidate a unified, conceptual understanding of contacts within which existing contacting schemes can be interpreted and future contacting schemes can be imagined. Whereas a split of the quasi-Fermi levels of holes and electrons is required in the absorber of any solar cell to generate a voltage, carriers are eventually collected through a metallic wire in which no such quasi-Fermi-level split exists. The term “contact” groups then all layers between the bulk of the absorber and the recombination-active interface through which carriers are extracted. The quasi-Fermi levels must necessarily collapse at this interface, and the transition between maximal quasi-Fermi-level splitting (in the absorber) and no splitting occurs entirely in the contact. For standard solar cell architectures, the contact extends from the surface of the absorber to the surface of a metal or transparent conductive oxide layer, and may include doped layers (diffused or deposited, as e.g. in crystalline and thin-film Si cells) and heterostructure buffer layers (e.g., the CdS layer in a CdTe device).
Whereas a passivating layer enables high quasi-Fermi-level splitting in the absorber (large “internal” voltage), where “high” is relative to the splitting dictated by bulk recombination, the additional feature of a carrier-selective contact is to enable a high “external” voltage measured across the contacts, where “high” is relative to the internal voltage. With this definition, a carrier selective contact allows only electrons or only holes to transport from the absorber to any position in the contact that has recombination-active defects. Carrier-selective contacts are then passivating layers that additionally allow for low-impedance flow of either electrons or holes (but not both) to the recombination-active, extracting interface. The most common example is a heavily doped layer in a p-n junction that filters the carriers that may pass to the contact according to the sign of their charge. However, such carrier-selective contact presents fundamental limitations which make other approaches, often involving heterojunctions, of particular interest to improve the external voltage of solar cells, as can be described based on this formalism.

Semi-insulating gallium nitride is an important material for photoconductive switches. This unique high-resistivity material is produced by iron-doping during hydride vapor phase epitaxy growth of GaN; however, very little is known about its electronic structure and charge transport properties, in particular defect type and concentration. This contribution will explore interface and bulk defect states, with an emphasis on impedance spectroscopy analysis.

ZnO is a promising material for application as a transparent electrode or luminescent layer in optoelectronic devices such as solar cells, UV light-emitting diodes and electrochromic devices. The performances of such devices strongly depend on the electrical and optical properties of the films that are affected by the structure and microstructure. Due to the importance of the optical properties of ZnO thin films, it is necessary to have an accurate determination of the bandgap energy (Eg). However, in the literature we find a large scatter in the reported values of the bandgap, suggesting several factors related with the synthesis route and experimental parameters influence the bandgap energy. Changes in Eg are often explained by changes in the carrier density through the Burstein-Moss effect and bandgap narrowing. Besides, it is important to point out that the obtained values can differ significantly depending on the measurement method. Magnetron sputtering is a traditional method for the deposition of ZnO but sputter-deposition of ZnO is often based on RF-sputtering of ZnO targets and uses thermal assistance to grow high quality films. In contrast, the growth of ZnO using reactive DC magnetron sputtering is little considered although it presents some interests from fundamental and applicative point of views for tuning both the defect chemistry and the microstructure.
Thin ZnO films deposited by reactive magnetron sputtering have been deposited in different reactive conditions leading to different oxygen stoichiometries. In zinc-rich conditions, the presence of shallo donor defects lead to an increase of the charge carrier concentration and therefore to an increase of the absorption bandgap energy E_g due to the Burstein-Moss effect while the emission bandgap as remain stable. Moreover, based on the classical models describing the Urbach energy E_u in semiconductors from exciton-phonons and exciton-defects interactions , we show that Eu is a useful parameter to connect the absorption bandgap to the emission bandgap under certain conditions. The results allow a better understanding of the large scatter of the bandgap energy of ZnO reported in literature by proposing a model taking into account the nature of chemical defects.

Recent studies directed towards the fabrication of nanowire arrays have led to the development of techniques for bottom-up type growth of horizontal zinc oxide and other semiconductor nanowires aligned to their crystalline substrate. In the case of ZnO nanowires grown on gallium nitride nanowire heterojunctions have been shown to contain very few intrinsic defects and exhibit strong emission properties under excitation making them ideal candidates for forming p-n junctions. This study focuses on the use of well-defined, quantifiable ion bombardment techniques to further probe the growth behavior of horizontal nanowires. Initial results show the effect of substrate disorder and its distance from interface on the growth direction and crystallinity of horizontal zinc oxide nanowires. Further analysis probes the influence of this disorder on the electronic properties of the substrate-wire interface. Further discussion will be provided on the possibility of use of such nanosystems as model systems for defect analysis of light emitting nanowire devices and modulating the structural properties of nanowires in a scalable fashion using focused ion beams.

Gallium oxide is a promising material for applications in optoelectronics (band gap of 4.6-4.8 eV) and a power semiconductor technology. Ga₂O₃ crystals are transparent up to UV-C range and can be made conductive by doping. This makes them an attractive choice as substrates for epitaxy of wide bandgap semiconductors. Light-emitting diodes on β-Ga₂O₃ substrates and high-voltage β-Ga₂O₃ transistors have been reported [1].
The widespread use of β-Ga₂O₃ crystals attracted great interest in the study of their structural perfection. The most important macrodefects affecting the properties of the crystals are crack, cleavage pores, twining.
We have grown β-Ga₂O₃ crystals by the method of free crystallization in the crucible [2]. And this crystals were explore for the presence of macrodefects.
Crystals were studied under an optical polarizing microscope POLAM P-111. Shown that crystals have four cleavage planes: parallel to the plane pinacoid - perfect; parallel to the plane monohedron - perfect; parallel to the planes of the rhombic prism - imperfect.
Lines parallel to the imperfect cleavage plane were consistently observed in the samples. This is clearly visible by the emerging Becke line arising at the interface due to difference in the refractive indices of the two media, but inside strips, the value of birefringence is the same as that of the main crystal body. Therefore, we make an assumption that the anisotropy of refractive index is related to crystal twinning
Some samples had pores. Their size is from 0,01 to 0,1 mm.
[1] WS Hwang, A. Verma, et al.// Applied Physics Letters 104, 203111.In 2014.
[2] V.N.Maslov, VM Krymov M.N.Blashenkov, A.A.Golovatenko, V.I.Nikolaev // Tech, tom.40, #8470;7, S.56-61. In 2014.

In the last several years, colloidal PbS quantum dots (QDs) have been deposited and crosslinked to form thin films. These thin films are the absorber layer in promising photovoltaic devices and have displayed an impressive rate of increase of reported record efficiencies. It is now well established that these solution-processed QD-films approximate a bulk semiconductor material enabling both heterojunction and p-i-n junction device architectures. However, observed photovoltages are considerably less than expected based on the bandgap of typically employed PbS QDs and thermodynamic considerations. We will present photoelectrochemical characterization of quantum-dot films as an important tool that can yield data that is complementary to the existing literature, and may improve our understanding of the ultimate photovoltage “potential” of these QD-films, and what limits ultimate photovoltages. We will present initial results, which show that electrochemical contacting can be rationally tuned by changing the solution redox potential (E(A^-/A)) to produce both rectifying and ohmic QD-film/liquid junctions. We will present current progress using this tool to study the ultimate photovoltage potential of PbS QD films as a function of QD ligand chemistry.

Thermal double donors (TDDs) are typical defects developed upon annealing at temperatures 300-450°C through aggregation of interstitially dissolved oxygen (O_i) atoms in Si and Ge crystals. Up to now, the atomic structure of TDDs is accepted to be a pair of atomic chains of O_is, called as O_n-2NN model, after the long-term studies in Si [1]. In contrast, Ge is far less known about TDDs due to the low [O_i]. Recently, we reported the segregation kinetics of O_i to form TDDs in an oxygen-enriched Ge crystal at various temperatures [2]. Ge is expected as the next generation semiconductor for high mobility devices, which enhances importance to control the TDD formation critically. Here, we report Sn impurity effects on TDD formation in Ge.
A heavily (Sn+O) co-doped Ge ([O]=5×10¹⁷, [Sn]=7×10¹⁸ cm^-3) and an O sole-doped Ge crystal ([O_i]=5×10¹⁷ cm^-3) were grown by an advanced Czochralski-method [3]. Measuring [O_i] and [TDD] variations against the heat treatment duration were evaluated using infrared absorption and Hall-effect study. TDD formation was suppressed in the (Sn+O) co-doped Ge than in the O sole-doped Ge, dependent on the temperature, implying O_i-trapping effect of Sn impurities. The binding energy between O_i and Sn impurities was evaluated to be 0.37 eV in an analysis of the reaction rate constant of O-dimer formation.

Mg is so far the only acceptor dopant that can be used for creation of p-type GaN material with a hole concentration sufficient for key devices like lasers and LEDs. We have studied the properties of Mg-doped GaN grown by MOCVD on bulk GaN substrates, whereby many defect related issues known to occur for growth on sapphire may be circumvented. Photoluminescence (PL) spectra reveal a specific set of spectral features for the Mg-related bound exciton (BE)and donor-acceptor (DA)-pair spectra, unique for Mg-doping. At high doping levels (> 10¹⁸ cm^-3) structural defects like stacking faults are induced by the Mg-doping, as also revealed in optical spectra. For these doping levels the acceptor related PL spectra are perturbed, showing another deeper BE line as well as low energy broadening of the DAP spectra. We have made an attempt to correlate these properties with recent magnetic resonance data for Mg-doped GaN. The precipitation of Mg into pyramidal defects typical for Mg-doped GaN grown on sapphire seems to be essentially absent for homoepitaxially grown material up to Mg concentrations of 10²⁰ cm^-3. The metastability observed in optical spectra in Mg-doped GaN is assigned to a separate metastable defect in Mg-doped GaN, i e not to the Mg acceptors.

We demonstrate a general approach via spin engineering that is capable of not only deactivating defect-mediated efficient non-radiative carrier recombination channels in a semiconductor that are harmful to photonic and photovoltaic device performance, but also adding new room-temperature (RT) spin functionalities that are desirable for future spintronics and spin-photonics but so far unachievable otherwise. This approach exploits the Pauli Exclusion Principle that prohibits occupation of a non-degenerate defect level by two spin-parallel electrons, thereby providing spin blockade of carrier recombination via the defect level. The success of the approach is demonstrated in the dilute nitride of Ga(In)NAs, which holds promises for low-cost, highly efficient lasers for fiber-optic communications as well as for multi-band and multi-junction solar cell applications. First we identify that Ga_i self-interstitials and their complexes are the most common grown-in defects found in Ga(In)NAs grown by both molecular beam epitaxy (MBE) and metalorganic chemical vapour deposition (MOCVD). They provide a dominant non-radiative shunt path for non-equilibrium carriers, leading to low efficiencies of light-emitting and photon-charge carrier conversion. Spin blockade is shown to lead to a giant enhancement by up to 800% in light emission intensity at RT.
Furthermore we show that via spin engineering these seemingly harmful defects can be turned into advantages by adding unconventional defect-enabled spin functionalities that are highly effective at RT, including some of the fundamental building blocks essential for future spintronics. We demonstrate efficient defect-engineered spin filtering in Ga(In)NAs, which is capable of generating a record-high degree (> 40%) of electron spin polarization at RT [Nature Materials 8, 198 (2009), Phys. Rev. B 89, 195412 (2014)]. We also provide the first experimental demonstration of an efficient RT spin amplifier based on defect engineered Ga(In)NAs with a spin gain up to 2700% [Adv. Materials 25, 738 (2013)]. Such a spin amplifier is shown to be capable of amplifying a fast-modulating input spin signal while truthfully maintaining its time variation of the spin-encoded information [7]. By taking advantage of the spin amplification effect, we show that Ga(In)NAs can be employed as efficient RT spin detectors, with spin detection efficiency well exceeding 100% [8,9]. By combining the spin-filtering effect and hyperfine coupling, we further achieve the first realization of RT nuclear spin hyperpolarization in semiconductors via conduction electrons [Nature Communications. 4, 1751 (2013)], relevant to nuclear spin qubits. We believe that such defect-enabled spin functionalities could potentially provide an attractive, alternative solution to the current and important issues on RT spin injection, spin amplification and spin detection in semiconductors for future spintronics.

High crystalline quality substrates with controlled electronic transport properties are required to realize a number of high performance devices at high yields. Recently, it was demonstrated that thick high crystalline quality HVPE GaN films can be grown on Ammono substrates, with reduced free carrier concentration [1]. More recently, it has been verified that such HVPE GaN wafers can be used for further homoepitaxial growth. Such verifications are extremely important, because they demonstrate the usefulness of this new type of substrate to fabricate highly efficient optoelectronic and electronic devices. A combination of defect sensitive techniques were employed to evaluate the crystalline and optoelectronic properties of the all three materials, and their potential for device fabrication.
Symmetric and asymmetric x-ray reflection measurements yielded a=3.193 Å and c=5.1856 Å lattice parameters for Ammono substrates; rocking curves of these substrates were measured to have a full width at half maximum (FWHM) of 20 arcsecs, compared to 60 arcsecs reported for typical freestanding HVPE heteroepitaxially grown GaN. From the FWHM value of 20 arcsecs it can be estimated that this Ammono sample has a dislocation density below 10⁵/cm². High resolution x-ray topography (HR-XRT) was used to obtain more precise dislocation density. The lattice parameters of the HVPE freestanding homoepitaxial GaN were measured using symmetric and asymmetric scans (sample A: c=5.1856Å and a=3.1824Å; sample B: c=5.1856Å and a=3.1839Å). These are very close to the bulk lattice constants. The rocking curves of these samples were measured to have FWHM~16 arcsecs, indicating very superior crystalline quality. Selective etching showed that the low concentration of defects observed in the ammonothermal substrates is preserved on the HVPE GaN samples.
High resolution micro Raman scattering measurements were carried out on the front and back surfaces of freestanding HVPE-GaN samples. All E₂² phonon frequencies, at various spots of both faces of the samples, are at 567.6 cm^-1, which is consistent with stress free samples. However, the A₁(LO) phonon frequencies and the linewidths measured on the Ga-polar surface are larger than those measured on the N-polar face; this is consistent with a larger incorporation of donor impurities with increasing growth rate. Low temperature photoluminescence measurements carried out on the Ga-polar faces of these samples are consistent with a higher free carrier concentration than that observed for high quality HVPE GaN [2]. Depth SIMS profile measurements will be carried out to identify and measure the concentration of the background impurities.
[1] Sochacki, et al., JCG 394 (2014) 55.
[2] Freitas, Jr., et al., PRB 66 (2002) 233311.

Zinc oxide is a direct wide bandgap semiconductor, which, when processed as nanocrystals, shows excitonic emission around 3.5 eV. Zinc oxide nanocrystals were prepared via colloidal synthesis methods and characterized using electron microscopies and optical spectroscopies. The nanocrystals show broad luminescence across the visible indicating the presence of defect-related energy states. The visible luminescence of the ZnO is enhanced by annealing the nanocrystals under various atmospheres. In order to assess whether these states could be useful in photocatalysis applications, the energies of the electron and hole states of the as prepared and annealed ZnO nanocrystals are precisely mapped spectroelectrochemically. The effects of the annealing on the defect related photoluminescence are discussed.

Nitride semiconductors are the key materials for solid-state lighting, and they are also increasingly used for power electronics. For all these applications, high-quality material is essential. In the epitaxial layers that provide device functionality, point defects may act as compensating centers, charge traps, or recombination centers. However, unintentional impurities often play an equally important role; for instance, carbon that is unavoidably incorporated during metal-organic chemical vapor deposition can act as a source of yellow luminescence. Still, point defects are likely to affect the radiative efficiency, and we are actively pursuing the microscopic origins of nonradiative recombination. In parallel, the study of point defects is important in the context of bulk crystal growth, and bulk GaN and AlN are increasingly investigated as substrates for nitride devices. In addition to point defects in GaN, we are focusing on identifying the prevailing defects in AlN, which lead to characteristic luminescence and absorption lines. I will briefly discuss the theoretical advances that are enabling us to calculate the energetics as well as electronic and optical properties of point defects with unprecedented accuracy. Overall, I will focus on examples where defects limit device performance.
Work performed in collaboration with A. Alkauskas, C. Dreyer, A. Janotti, J. Lyons, L. Gordon, and Q. Yan, and supported by DOE and NSF.

Effect of relatively low dose of gamma-irradiation on minority carrier transport properties of AlGaN/GaN high electron mobility transistors (HEMTs) was investigated through Electron Beam Induced Current (EBIC) and Cathodoluminescence (CL) techniques. AlGaN/GaN HEMTs were exposed to total dose of 100 Gy to 1000 Gy using a ⁶⁰Co source. Temperature dependent EBIC and CL measurements conducted on the devices allowed for the extraction of activation energies in the band gap, which bears a signature of the defect levels involved in the carrier recombination process. The measurements were conducted prior to gamma-irradiation dose and then continued on devices subjected to various doses of irradiation. Comparing the activation energy before and after gamma-irradiation will identify the defect levels and their dependence on the dose of irradiation. Investigation of gamma-irradiated device gives the important information about the defects in semiconductor heterostructures, especially if additional traps are introduced under external irradiation.

Highly mismatched alloys (HMAs) are semiconductor alloys in which the anions are partially replaced by elements with very different electronegativity and/or atomic size (e.g. N in GaAs or As in GaN) [1]. Electronic band structures of HMAs are strongly modified due to the interaction between localized states of the minority anions and the extended states of the host compound as described by the Band Anticrossing (BAC) model [1]. Consequently, these large changes in the conduction or valence band structure provide a unique means to tailor the properties of HMAs to specific device applications. However, due to the large mismatch of the component materials the synthesis of HMAs is challenging. Recently, we overcame the miscibility gap of GaAs and GaN alloys and synthesized GaN_1-xAs_x alloys over the whole composition range using low temperature molecular beam epitaxy (LT-MBE) [2]. Since the anion mismatch is even larger between Sb and N, BAC calculations predict that GaN_1-xSb_x HMAs are suitable materials for photoelectrochemical water splitting applications. In this talk, I will present our systematic investigation of defects and properties of GaN_1-xSb_x alloys grown by LT-MBE.
The effects of growth temperature, Ga flux and Sb flux on the incorporation of Sb, film structure and optical properties of the alloys are investigated. We found that at the growth temperature of 80^oC, GaN_1-xSb_x with x>6% loses crystallinity and becomes primarily amorphous with small crystallites of 2-5 nm. Despite the range of microstructures found for GaN_1-xSb_x alloys with different composition, a well-developed absorption edge shifts from 3.4 eV (GaN) to close to 2 eV for samples with a small amount, less than 10% of Sb. Luminescence from dilute GaN_1-xSb_x alloys grown at high temperature and the bandgap energy for alloys with higher Sb content are consistent with a localized substitutional Sb level E_Sb at ~1.1 eV above the valence band of GaN. The decrease in the bandgap of GaN_1-xSb_x HMAs is consistent with the formation of a Sb-derived band due to the anticrossing interaction of the Sb states with the valence band of GaN. Our work demonstrates that a large range of direct bandgap energies from 3.4 eV to below 1.0 eV can be achieved for GaN_1-xSb_x HMAs grown at low temperature.
W. Walukiewicz, W. Shan, K.M. Yu, J.W. Ager III, E.E. Haller,I. Miotkowski, M.J. Seong, H. Alawadhi, and A.K. Ramdas, Phys. Rev. Lett. 85, 1552 (2000).
K. M. Yu, S. V. Novikov, R. Broesler, I. N. Demchenko, J. D. Denlinger, Z. Liliental-Weber, F. Luckert, R. W. Martin, W. Walukiewicz, and C. T. Foxon, J. Appl. Phys. 106, 103709 (2009).

Synchrotron-based elemental and spectroscopic characterization has contributed to the evolution of crystalline silicon (c-Si) and thin-film photovoltaics (PV). Modern nanoprobe beamlines enable mapping of elemental composition, chemical bonding, and electrical activity with a beam spot size as small as 40 nm. In this presentation, we review the applications of synchrotron-based analytical techniques to elucidate defect properties in three emerging PV material classes:#8232;
(i) Kerfless c-Si materials have the potential to reduce silicon use by 5x and approach the cost structures of thin-film PV modules, but with proven silicon reliability. Working in close collaboration with a kerfless silicon manufacturer, an improvement in bulk minority-carrier lifetime to the millisecond regime was recently accomplished through rigorous impurity control. Synchrotron-based techniques, coupled to laboratory-based spectroscopy, enabled detection and control of metal point defects and inclusions. This work builds upon two decades of research to develop synchrotron-based nanoprobe techniques for silicon-based materials, pioneered at Berkeley's Advanced Light Source and extended in recent years to an active worldwide community. In closing, we provide a forward-looking perspective of impurity detection in silicon.
(ii) Amorphous metal-oxide carrier-selective contacts: Defect-tolerant disordered materials are an emerging class of materials of relevance for carrier-selective contacts. Ternary buffer layers, including amorphous zinc-tin oxide (a-ZTO), can be composition-tuned to control conduction-band energy. However, the electronic transport properties have been shown to be affected by composition. Extended X-ray absorption fine structure (EXAFS) measurements on a-ZTO elucidate the structure-property relationships in this material class, highlighting the root cause for mobility decrease and suggesting design rules for defect-tolerant, tunable carrier-selective contacts. In closing, we highlight the role of EXAFS in solving future dopant-related challenges in carrier-selective contacts.
(iii) Earth-abundant thin-film absorbers offer a scalable alternative to cadmium telluride (CdTe) and copper indium gallium selenide (CIGS). Control of intrinsic and extrinsic point defects is essential to improving carrier transport properties. X-ray absorption spectroscopy elucidates the dopant states and distributions in novel Earth-abundant thin-film materials, suggesting intrinsic limits to dopant incorporation and second-phase formation. Combining these insights with lab-based techniques, we have achieved record-efficiency thin-film tin-sulfide (SnS) and electrochemically deposited cuprous oxide (Cu₂O) devices. We provide a future perspective of in-situ characterization of thin-film materials using controlled ambients, highlighting the role of the In-Situ Nanoprobe Beamline under development at the Advanced Photon Source

CdZnTe is a semiconductor material that promises to be a good candidate for uncooled gamma radiation detectors. However, to date, we are yet to overcome the technological difficulties in production of large size defect-free CdZnTe crystals. The most common problem is accumulation of Tellurium inclusions and precipitates as microscopic inclusions. These inclusions influence the charge collection through charge trapping and electric field distortion. We employed high energy transmission X-ray diffraction techniques to study the quality of CdZnTe crystals grown by the Bridgman Technique. Crystallinity and defects within two different growth set-ups, i.e. with and without choked seeding, were compared by imaging the crystal orientation topography with white beam X-ray diffraction topography (WBXDT). The X-ray diffraction topography results show a high correlation with large-area infrared transmission images of the crystals. Grain boundaries that are highly decorated with Te inclusions are observed. Characteristic Te inclusion arrangements as a result of growth conditions are discussed. We also measured the electronic properties of the detectors fabricated from ingots grown using two Bridgman processes, and observed a reduction in the electrical resistivity of choked-seeding-grown CdZnTe crystals. Our results show that although choked seeding technique holds promise in the realization of high quality mono-crystalline CdZnTe, current growth parameters must be improved to obtain defect-free crystals. These results are helpful to attain optimal seeding process for Bridgman growth of large single crystals of CdZnTe.

In the past few years, much of the research efforts have been inclined towards the integration of high-k gate dielectric material in ZnO thin film transistors (TFTs) in an attempt to improve the carrier mobility, decrease the gate leakage current, and improve the overall device performance. Among several high-k dielectrics, HfO₂ is promising due to its high dielectric constant (20-25), wide band gap (5.6 eV), and reasonable band-offset with ZnO. However, the performance of the TFTs is largely limited by the quality of the gate dielectric and interface properties between ZnO/HfO₂ interface. Therefore, the study of the interface properties between ZnO and high-k dielectrics such as HfO₂ is essential to improve the TFT performance. Sputtering technique offers advantages in terms of ability to deposit materials at low-temperature on desired substrate such as glass and plastics, large-area scalability. However, low-temperature deposition can also introduce unique defects in high-k and ZnO and their interface which needs to be better understood. Though there have been previous studies of ZnO/HfO₂ TFT fabricated using a combination of Atomic Layer Deposition, Pulsed layer Deposition, and other techniques, our knowledge on the characteristic of interfaces formed between ZnO and HfO₂ when both materials are sputtered is limited.
To bridge this gap in the knowledgebase, we have studied the interface properties of the rf sputtered ZnO (sputtered at room temperature) and HfO₂ (reactively sputtered at 300 ^o C) using a variety of electrical characterization techniques such as capacitance-voltage (C-V) and admittance spectroscopy (G_p/omega; vs. omega;) measurements TFTs structures. TFTs were fabricated using two thicknesses of room temperature sputtered ZnO (50 nm and 70 nm) with sputtered HfO2 dielectric and Ru electrode in bottom electrode structures. Both TFTs demonstrated reasonable transistor switching characteristics. The devices show well-behaved C-V characteristics indicating a good dielectric with low-leakage current. Interestingly, interface state density demonstrated a strong dependence on the thickness of ZnO. Using admittance spectroscopy, the interface states density was estimated to be on the order of 10¹² eV^-1 cm^-2 in TFT with 70 nm of ZnO as compared to 10¹³ eV^-1 cm^-2 in TFT with 50 nm of ZnO thickness. This may indicate intermixing of ZnO with HfO₂ resulting in the formation of an interfacial layer which will impact thinner ZnO more severely than thicker ZnO channel. The intermixing could result from the fact that HfO₂ was deposited at relatively higher temperature (300 ^oC) which is critical to achieve good dielectric properties. Towards improving this interface and the overall TFT characteristics, data will be presented on our efforts to minimize the interface state density between ZnO/HfO₂ by exploring various interfacial layer such as MgO and lower temperature HfO₂, and dopants in ZnO such as Hf, Mn, and Mg.

Defects in amorphous silica are studied experimentally for more than fifty years (see for example [1] for E&’_g center in amorphous SiO₂) but the attribution of absorption/photoluminescence bands to a point defect structure is still a major challenge in the domain [1,3]. Indeed experimental identification of defects is complex and requires the combination of numerous experimental characterization means such as Electronic Paramagnetic Resonance (EPR) and absorption and photoluminescence spectroscopy. But each experimental means has its own intrinsic limitations. For example, EPR can only detect paramagnetic defects. This is why, today, atomic scale simulation is an important tool to give access to point defects properties but also to exchange common observables with experiments.
In this study, we performed first principles calculations on defects and dopants in crystalline and amorphous SiO₂ and GeO₂. Calculations based on the Density Functional Theory (DFT) using the Local Density Approximation (LDA) are performed through the PWscf code from the Quantum ESPRESSO distribution [4]. These calculations allow us to have access to the atomic configurations of defects and dopants and to their energy properties, but also to the EPR parameters of the paramagnetic defects by exploiting the gauge including projector augmented wave (GIPAW) method [5]. Starting from the wavefunctions and from the atomic configurations obtained in DFT, we use the SaX code version 2.0 [6] to apply the GW approximation to obtain the band gap in order to evaluate the effects of point defects on the electronic properties of SiO₂ and GeO₂ and the influence of dopants on these properties. Finally, the optical properties (particularly the absorption spectra) of pure and doped SiO₂ and GeO₂ are given including excitonic effects with the resolution of the Bethe-Salpeter equation (BSE). Electronic and optical properties and EPR parameters obtained here will be discussed and compared to the results coming from previous theoretical and experimental studies. Some of the results in this study have been presented in four previous papers [7,8,9,10].
[1] R.A. Weeks, J. Appl. Phys., 27, 1376 (1956).
[2] S. Girard et al., IEEE Trans. Nuc. Sci., 60(3), 2015 (2013).
[3] L. Skuja Journal of Non-Crystalline Solids 239, 16 (1998).
[4] P. Giannozzi et al., J. Phys. Condens.Matter, 21, 395502 (2009).
[5] C.J. Pickard and F. Mauri Phys. Rev. Lett. 88, 086403 (2002).
[6] L. Martin-Samos and G. Bussi, Comp. Phys. Com., 180, 1416 (2009).
[7] N. Richard et al., Journal of Non-Crystalline Solids, 357, 1994 (2011).
[8] N. Richard et al., J. Phys.: Condens. Matter, 25, 335502 (2013).
[9] L. Giacomazzi et al., Phys. Rev. B, 90, 014108 (2014).
[10] L. Giacomazzi et al., Phys. Rev. X, submitted.

The efficiency potential of multicrystalline silicon has recently increased significantly mainly due to improvements of the crystallization procedure which helped to suppress the development of dislocation clusters and small angle grain boundaries. Besides the crystal structure, metals in small concentrations still reduce the attainable solar cell efficiency for this material. These contaminants mainly originate from the silicon feedstock and the crucible system during crystallization.
It is the aim of this presentation to quantitatively assess the impact of occurring metallic contaminants on attainable cell performance. Metals are typically found in different configurations, as dissolved impurities as well as precipitates. Depending on the temperature profiles of crystallization and cell process, dissolved impurities may cluster to precipitates or precipitates may dissolve.
Lifetime measurements, metastable defect imaging and simulations are combined to quantify the role of the mentioned defects for different feedstock and process sequences. Our Fokker-Planck-based precipitation model gives access to both, the dissolved metal concentration and the precipitate size and density distribution. The detailed temperature profiles as they occur during the cooling procedure after crystallization as well as during cell processing steps (e.g. emitter diffusion, contact firing) are considered. External gettering of metals is taken into account using a state-of-the-art model.
These findings are combined with specific recombination models. While dissolved impurities are considered as Shockley-Read-Hall recombination centers, precipitates may act as strong Schottky-type recombination centers. The latter are described by thermionic emission processes and are modeled with input from the precipitate size and density distribution simulation.
It turns out that the limiting path of carrier recombination varies strongly as a function of processes, block position and, particularly, type of base doping (n- and p-type). The limiting role of different recombination centers is discussed in this view and representative experimental results on identical silicon block crystallizations with different type of base doping are discussed.
From comprehensive experimental lifetime analyses the cell efficiency potential of the materials under test is quantified which provides a sound measure for material qualification and optimization strategies.

Mono-like silicon (Si) has been extensively studied as a Si ingot in the next generation for solar cells, replacing conventional multicrystalline Si. We proposed a growth technique for mono-like Si; the concept is to control crystal grain morphology, defects density, impurity concentration and strain distribution by utilizing artificial grain boundaries composed by multi-seed crystals. In other words, we applied grain boundary engineering to the growth of mono-like Si. We named such useful grain boundary “functional grain boundary” [1, 2]. In this paper, we present one of benefits of the functional grain boundary that is suppression of multi-crystallization, which is a problem faced by mono-like Si.
A 70kg ingot of 40 cm on a side and 20cm in height was grown on multi-seed crystals. In this ingot, we intentionally formed two different regions; with and without functional grain boundaries near the crucible side walls. As the functional grain boundary to suppress the multi-crystallization, we chose sigma 5 grain boundary, which shows low electrical property and stably extends in the <100> growth direction [3]. In the region without the functional grain boundaries, almost all the grain boundaries nucleated on the crucible side walls were sigma 3 grain boundaries. They inclined to the <100> growth direction. As crystal growth proceeded, lateral position of them moved inward, which result in an increase in the area of multicrystalline grains. On the other side, where we artificially formed sigma 5 grain boundaries near the crucible side walls by using the special configuration of the multi-seed crystals, the sigma 3 grain boundaries formed on the crucible side walls intersected with the sigma 5 grain boundaries and sigma 15 grain boundaries were consequently formed. The sigma 15 grain boundaries tended to extend in the crystal growth direction to decrease its energy as do sigma 5 grain boundaries, which results in no further increase in the area of multicrystalline grains. Therefore, a large volume of quasi-single crystalline Si up to the top of the ingot was achieved. This suppression increases the yield of the mono-like wafers in an ingot and would enable a large reduction of cost/watt of the solar cells.
In summary, we applied grain boundary engineering to the growth of mono-like Si for solar cells. Functional grain boundaries are useful for suppressing the multicrystallization in mono-like Si.
[1] K. Kutsukake, N. Usami, Y. Ohno, Y. Tokumoto, and I. Yonenaga, Appl. Phys. Express 6, 25505 (2013).
[2] K. Kutsukake, N. Usami, Y. Ohno, Y. Tokumoto, and I. Yonenaga, IEEE J. Photovolt. 4 84 (2014).#12288;
[3] K. Kutsukake, N. Usami, K. Fujiwara, Y. Nose, and K. Nakajima, J. Appl. Phys.101, 063509 (2007).

Grain boundaries (GBs) are inevitably introduced in multicrystalline silicon (mc-Si), which has been applied in photovoltaic devices, and they influence the device performance via impurity contaminations and the defect levels. They act as agglomeration sites, where impurity atoms segregate or precipitate, and the contaminants control optoelectronic properties. Accordingly, comprehensive knowledge of the GB agglomeration mechanism is essential for engineering the distributions and sizes of impurity-related nanostructures at GBs in controlled fashions, for producing cost-effective functional photovoltaic devices. The agglomeration mechanism is, however, far from being understood due to difficulties characterizing both crystallographic and chemical properties of the same GB at atomistic levels. In the present work, we have jointly employed atomic-resolution transmission electron microscopy (TEM) and atom probe tomography (APT), accompanied with ab-initio calculations, to comprehend microscopic mechanisms of GB segregation, and discussed segregation abilities in terms of bond distortions of host atoms at GBs.
High-resolution three-dimensional (3D) impurity distributions at Σ9{114} GBs, that were frequently observed in mc-Si-based solar cells, were determined by APT combined with TEM, with a low impurity detection limit (lower than 0.0005 at.%) about two orders lower than the limit by TEM, simultaneously with a high spatial resolution (about 0.4 nm) comparable to the resolution of TEM. Our analytical method enabled us to determine the exact location of each GB even when its segregation ability was quite low [1]. Arsenic (n-type dopant), gallium (p-type dopant) and oxygen (neutral impurity) atoms segregated at the GBs irrespective of charge states. Therefore, those impurities would segregate so as to reduce the strain energy due to bond distortions nearby GBs, rather than to reduce the electronic GB energy. Also, boron and carbon atoms did not segregate at the GBs, suggesting attractive strain nearby the GBs. Atomic resolution TEM and ab-initio calculations revealed 3D distribution of large bond distortions, inducing high tensile atom hydrostatic stresses, in the GBs. ATP revealed 3D impurity distributions in the same GBs, and the segregation abilities would be correlated with the tensile stresses.
[1] Y. Ohno, et al., Appl. Phys. Lett. 103, 102102 (2013).

In this study, we investigated dislocation formation mechanisms of <112>-oriented core-shell heterostructure Si-Ge nanowires under tension at 893K using molecular dynamics (MD) simulations. We tested Si-Ge nanowires of three different sizes; (Si-core diameter, Ge-Shell thickness) = (5, 2.5), (10, 5), (15, 7.5)nm, with two different interatomic potential models; Tersoff [1] and Stillinger-Weber (SW) [2,3,4]. Tersoff model predicts that dislocations nucleate at the core-shell interface and propagate to the core and to the shell, gliding on two {111} glide planes of highest Schmidt factor. On the other hand, with SW potential, we observed not only dislocation formation but also crack propagation. Cracks nucleated at the core-shell interface and propagate outward in Ge-shell region on the plane normal to the NW orientation direction. We analyzed the failure behavior of Si-Ge core-shell NWs that depends on the sizes of NWs and interatomic potential models.
Acknowledgement
This research was supported by Basic Science Research Program through the National Research Foundation of Korea(NRF) funded by the Ministry of Education (2013R1A1A2063917).
References
[1] J. Tersoff. Modeling solid-state chemistry: Interatomic potentials for multicomponent systems. Phys. Rev. B, 39:5566-5568, 1989.
[2] Stephane Ethier, Laurent J. Lewis. Epitaxial growth of Si_1-xGe_x on Si(100)2 1: A molecular-dynamics study. J. Mater. Res, 7:2817-2827, 1992.
[3] F. H. Stillinger, T. A. Weber. Computer Simulation of Local Order in Condensed Phases of Silicon. Phys. Rev. B, 31:5262-5271, 1985
[4] K. Ding, H. C. Andersen. Molecular Dynamics Simulation of Amorphous Germanium. Phys. Rev. B, 34:6987-6991, 1986

4H-SiC is one of the most mature post-Si material in the field of high voltage/high power devices. Although the quality of epitaxial films has been strongly improved in the last years, several defects still exists in grown films: mainly stacking faults (SF), dislocations and point like defects. The electrical behavior of these defects impacts the overall quality of the SiC devices, for example Single Shockley stacking faults (SSSF) may act as recombination centers in bipolar devices; SF can increase both the resistance and the Schottky-barrier in n-doped SiC films; point defects, resulting from ion implantation, may reduce the minority carrier lifetime and increase the leakage current.
In this paper we have studied the connection between crystalline defects and the local electrical transport properties by simultaneously mapping the photoluminescence (PL) and the photocurrent (PC) induced by a laser beam with wavelength above the bandgap. 4H-SiC Schottky diodes fabricated with a 38 µm thick n^- epitaxial layer have been studied.
The measurements have been performed by focusing the laser beam through a microscope, obtaining images with spatial resolution on the order of 2 mu;m.
The adopted technique allows to determine the absorption properties of an extended defect as well as the carrier diffusion length, and to link the electrical response to the photoluminescence spectrum. As an example, regions characterized by photoemission at 2.6 eV, usually ascribed to (4,4) stacking faults, exhibit very low photocurrent, without significant reduction of the diffusion length, indicating that the absorption is strongly reduced.
On the contrary, in regions exhibiting high photo-emission at 2.9 eV, characteristic feature of single Shockley stacking faults, the photocurrent is comparable to that measured in regions without defects. The measured diffusion length is instead strongly reduced, down to 50%.
The dependence of the photocurrent on the defect orientation has been also investigated.

It is widely believed that oxygen vacancies play a major role in controlling properties of complex oxides, but their precise role has been an open issue. Experiments have difficulty detecting them and theory alone has difficulty addressing the consequences of vacancies. This talk will describe how theoretical atomic-scale or multi-scale calculations, performed in the context of experimental observations, can provide a detailed description of the role of vacancies in many diverse phenomena. Many semiconducting oxides are known to exhibit memristive behavior, which is characterized by a hysteresis loop in current-voltage characteristics that defines a high- and a low-resistive state. These materials are investigated for next-generation non-volatile memories and other applications. The origin of memristive behavior through a “vacancy breathing” mechanism will be demonstrated in polycrystalline ZnO and BiFeO₃, with the grain boundaries acting as two-dimensional nanovaristors. In several cases of interfaces between different transition-metal oxides, a combination of atomic-scale calculations, Z-contrast imaging, and electron-energy-loss spectroscopy provide clear evidence that oxygen vacancies are often essential in compensating for interfacial discontinuities in structural, electronic, or magnetic properties. In some cases, ordering of oxygen vacancies plays a key role in accounting for the observed structural, electronic, and magnetic properties of epitaxial oxide thin films. Finally, it will be demonstrated that selenium vacancies play a key role in the nucleation and growth of defect structures in monolayer semiconducting MoSe₂.
This research was supported by NSF and DOE. The author acknowledges many collaborators at Vanderbilt and ORNL.

Visible-light absorptions and luminescence of wide band gap (3.25 eV) strontium titanate (SrTiO₃) have been well-known with the origins, the existence of natural oxygen deficiency. In this study, as the first time, we report that the surface and bulk oxygen vacancies of SrTiO₃ (STO) play different roles in the optical and magnetic properties, using density functional theory (DFT) calculations. We found that the doubly charged state of oxygen vacancy is dominant (V_O²⁺) in bulk SrTiO₃ and does not contribute to the sub-band-gap photoexcitation or intrinsic magnetism of STO. The oxygen vacancy (V_O) on SrTiO₃(001) surfaces terminated with both TiO₂- and SrO- layers induce magnetic moments which are dependent on the charge state of V_O. The calculated absorption spectra show that the 2.5-3.1 eV range is mainly absorbed by the V_O⁰/V_O²⁺ on TiO₂-terminated SrTiO₃(001) surface and mid-infrared range (~0.2eV) is also absorbed by the V_O⁰/V_O²⁺ of SrO-terminated surface. Especially, the neutral oxygen vacancy, V_O⁰ on TiO₂-terminated surface has lower formation energy than on SrO-terminated surface and possesses half-metallicity exclusively, which is related to the strong spin-dependent photoexcitation characteristics and the recently observed X-ray magnetic circular dichroism (XMCD) measurements [W. D. Rice, et al., Nat. Mater. 13, 481 (2014).].

Tungsten trioxide (WO₃) is a promising material for an array of applications, including gas sensors, Li-ion batteries, photocatalysis, and electrochromic devices. As an electrochromic material, it is turns from transparent to blue upon doping. This color change is correlated with a shift in absorption and reflectivity from near-UV to near-IR, and has applications in smart windows for energy efficiency. WO₃ has a perovskite-like structure with a vacant A-site, making it possible to achieve high levels of intercalation. In addition to monovalent ions that occupy the A-site, oxygen deficiencies have also been invoked as a contributor to electrochromism. However, the influence of oxygen vacancies on electric and optical properties of WO₃ remains a topic of debate.
In this work, we examine the properties of oxygen vacancies in monoclinic WO₃ using first-principles calculations based on hybrid density functional theory. We investigate the stability of the vacancy on inequivalent oxygen sites and in different charge states. Our results show that oxygen vacancies are shallow donors, and do not induce gap states. We discuss the similarities with monovalent ion doping, and its implications for device development. Finally, we compare our theoretical findings with experiment to provide an understanding for the mechanisms that govern electrochromic behavior.
This work is supported by DOE and by NSF.

Predicting reliably the formation energies of defects has a challenge over many years, not the least due the band gap problem of DFT approximations commonly used for defect supercell calculations. Computationally more demanding approaches like hybrid functionals allow band-gap corrected defect calculations, but consistency between different functionals remained a concern. Combining density and hybrid functional defect calculations with GW quasiparticle energy calculations for the pure materials, we have recently shown that the results of different functionals can be reconciled when using analogous reference energies for the valence band maximum (VBM) and the elemental chemical potentials [1].
In order to employ ab initio defect calculations for real world problem, it is often necessary to address specific issues in more detail. As an example, we discuss the case of group 15 dopants (P, As, Sb, Bi) in the transparent conducting oxide (TCO) SnO₂, where the multivalence (III or V) of the group 15 elements implies that the dopant can act either as a donor (V) or as an acceptor (III). Ultimately, the electrical activity of the dopant depends on the subtle energy difference between the respective charge transition and energy of the conduction band minimum (CBM). We addressed this problem using hybrid functionals and additional GW defect supercell calculations [2], finding that only Sb is a viable dopant for electron doping. In another application related to TCO, we addressed the defect equilibrium of Ga doping in ZnO, where the above mentioned GW corrections for the VBM energy and the thermodynamics of dopant-defect pairing play important roles with large, but counteracting contributions to the free electron concentrations. Notably, ZnO:Ga thin-film samples exhibit the predicted (T, pO2) behavior only above about 500C, pointing to a pronounced non-equilibrium state of the as-grown samples [3].
Finally, addressing the design of new materials via aliovalent alloying, we extend the traditional dilute defect model to higher concentrations. The interesting aspect of aliovalent alloying is that the modification of band structure and electrical properties become intertwined. Addressing divalent (Mg, Zn, Cd) and chalcogen (S, Se) substitution in Cu₂O, which is of interest as a p-type oxide and a photovoltaic material, we determine the dependence of the band gap and the net doping as a function of composition [4]. The model predicts a wide range of band-gaps and doping levels in these alloys, including type conversion by Cd alloying, which makes these alloys interesting for materials design for optoelectronic applications.
[1] H. Peng, D.O. Scanlon, V. Stevanovic, J. Vidal, G.W. Watson, S. Lany, Phys. Rev. B 88, 115201 (2013).
[2] H. Peng, J.D. Perkins, S. Lany, Chem. Mater. 26, 4876 (2014).
[3] N.H. Perry, T.O. Mason, D.S. Ginley, S. Lany, Appl. Phys. Lett. 103, 232106 (2013).
[4] V. Stevanovic, A. Zakutayev, S. Lany, Phys. Rev. Appl. 2, 044005 (2014).

One of the most important issues in semiconductor physics is to control the charge carriers through doping. This is because the application of semiconductors as electrical and optical devices depends critically on their doping properties. For example, CdTe is one of the leading materials for low cost, high efficiency solar cell absorbers, due to its suitable band gap of 1.5 eV, high optical absorption, and easiness of growth. However, CdTe solar cell has a relatively low efficiency (~19%) comparing with its theoretical limit sim;30%. One of the main reasons is because of the low majority carrier concentration of CdTe and low carrier life time caused by defect-induced carrier recombination. Therefore, to improve its solar conversion efficiency, it is necessary to understand and control its doping properties. Using first-principles band structure methods we systematically studied the defect properties in CdTe, including calculation of the defect formation energies and transition energy levels of intrinsic and extrinsic point defects and defect complexes, grain boundaries, defect diffusion barriers, and carrier concentration as function of atomic chemical potentials and temperature. From the calculated results, we investigate the limiting factors for p-type and n-type doping in CdTe and the potential non-radiative recombination centers. Possible approaches to significantly increase the doping limits and reduce the recombination centers, as well as passivation of grain boundary defect levels are discussed. General understanding of the chemical trends of defect properties in thin-film solar cell absorbers will also be discussed.

Indium tin oxide (ITO) thin-films are heavily relied upon as transparent electrodes in many optoelectronic applications. High conductivities useful for enhanced lateral transport can be achieved due to a combination of high carrier concentrations (on the order of 10²¹cm^-3) and average mobilities (~30 cm^-2V^-1s^-1). However, in applications for many solar cell devices, high conductivity comes at the expense of unacceptably low transparency: significant optical and electrical losses occur due to the well-known trade-off between free-carrier absorption and resistivity in this degenerate semiconductor.
Introduction of hydrogen as a replacement dopant in the indium oxide (IO:H) system has been presented as an alternative to maintain high conductivity at low carrier concentrations by providing large improvements in carrier mobility. Koida et al. first introduced this material and reported mobilities in excess of 100 cm^-2V^-1s^-1 [1]. Barraud et al. demonstrated in a silicon heterojunction solar cell that the resulting increase in transparency translates into a 1 mA/cm² increase in short-circuit current without a loss in fill factor [2].
According to these prior studies, the exceptionally large mobility values stem from hydrogen acting as a donor that minimizes scattering from doubly charged and neutral defects. Density functional theory simulations seem to corroborate this hypothesis, and some authors have calculated the lattice position of hydrogen and relaxation of the surrounding atoms [3].
We present a comprehensive comparison of optical, electrical, and structural properties for intrinsic indium oxide and IO:H with varying levels of hydrogenation. These films of IO:H were deposited on glass via RF sputtering of an In₂O₃ target in an argon atmosphere with small amounts of oxygen and water vapor. Water vapor partial pressure was varied from 4mu;Torr to 8mu;Torr with a total process pressure of 5mTorr. The resulting thin-films display mobilities ranging from 124.7-6.8 cm²V^-1s^-1 for free-carrier concentrations on the order of 10²⁰ cm^-3 and lower.
We disentangle the various scattering mechanisms affecting electron transport using temperature-dependent, Hall-Effect measurements. Finally, we directly probe the local neighboring environment of the indium atom by extended X-ray absorption fine structure (EXAFS). This helps to clarify the position of hydrogen in the lattice and whether significant distortion around the central atom is observed as hydrogenation increases.
Citations:
[1]: T. Koida, H. Fujiwara, M. Kondo, Japanese Journal of Applied Physics 46 (2007) L685-L687.
[2]: L. Barraud & et al., Solar Energy Materials & Solar Cells 115 (2013) 151-156.
[3]: S. Limpijumnong, P. Reunchan, A. Janotti, and C.G. Van de Walle, Phy. Review B 80 (2009) 193202.

Semiconducting CdS thin films have been prepared using chemical bath deposition technique and the effect of thickness on the physical properties of these films have been investigated. The thickness of the films has been varied in the range 1000 Å to 2300 Å. The XRD patterns present a single peak corresponding to the characteristics peak of CdS. The grain size was observed to increase with increase in film thickness, indicating the decrease in strain and improvement in the degree of crystallinity of these films. Decrease in the resistivity and the band gap of the films have been noticed with increasing the film thickness. The observed decrease in the band gap is explained on the basis of red shift of the absorption edge, presence of impurities and decrease in lattice strain.

Thin-film solar cells based on polycrystalline Cu(In,Ga)Se2 (CIGS) and CdTe photovoltaic semiconductors have reached remarkable laboratory efficiencies. It is surprising that the efficiencies of these polycrystalline solar cells can reach so high even these thin-films contain grain boundaries (GBs), which are considered to be nonradiative recombination centers for carriers. In the past decade, we have studied the physics of GBs in polycrystalline photovoltaic semiconductors including Si, CdTe, CuInSe2 (CIS), and Cu2ZnSnS4 (CZTS) using a combination of state-of-the-art density-functional theory (DFT) and advanced high-resolution electron microscopy. We have determined the atomic structures and chemical compositions of grain boundaries in these materials. We have conducted systematical density-functional theory study of the electronic properties of the grain boundaries. We found that clean grain boundaries produce detrimental gap states. However, tgrain boundaries can be passivated by impurity segregations. In CuInSe₂, and Cu₂ZnSnS₄, grain boundaries can be passivated by O segregation. In CdTe, grain boundaries can be passivated by Cl segregation. Our results suggest approaches for improving the performance of polycrystalline thin-film solar cells.