High Efficiency of III-Nitrides MicroLight-Emitting Diodes by Sidewall Passivation Using Atomic Layer Deposition

8:20 AM KK01 High Efficiency of III-Nitrides Micro-Light-Emitting Diodes by Sidewall Passivation Using Atomic Layer Deposition Matthew S. Wong1, David Hwang1, Abdullah I. Alhassan1, Changmin Lee1, Ryan Ley3, Shuji Nakamura1, 2 and Steven P. DenBaars1, 2; 1Materials Department, University of California, Santa Barbara, Goleta, California, United States; 2Department of Electrical and Computer Engineering, University of California, Santa Barbara, Santa Barbara, California, United States; 3Department of Chemical Engineering, University of California, Santa Barbara, Santa Barbara, California, United States.

As ultra wide bandgap materials, AlGaN and AlN can be implemented in high power devices and as such, conductive n-and p-type regions through implantation are required to operate these devices. Compared to epitaxial doping, ion implantation is a promising technique for doping area selectivity and compensation management such as DX-center formation and hydrogen passivation. So far, high conductivity of ≈1.3 Ω -1 cm -1 has been achieved in Si-implanted AlN at room temperature, which is more than two orders of magnitude higher than that of epitaxially doped AlN by MOCVD. Implantation damage needs to be fully recovered without compensating defect formation during the annealing process to achieve higher conductivities. Thus, this study provides a detailed study on damage recovery by annealing in ion-implanted AlGaN and AlN.Undoped epitaxial Al 0.8 Ga 0.2 N and AlN layers were grown on either a c-plane AlN template/sapphire or an AlN native substrate by MOCVD. Mg and Si with doses varied from 5×10 13 to 1×10 15 cm -2 were implanted into AlGaN and AlN layers with an acceleration energy of 75-100 keV at room temperature, respectively. Depth profile and concentration of implanted impurities were confirmed by secondary ion mass spectroscopy (SIMS). Optical properties were characterized by photoluminescence (PL). High resolution X-ray diffraction (HRXRD) was conducted before implantation, after implantation, and after annealing. Furthermore, recovery behavior was monitored by in-situ XRD measurements within a temperature range varying from room temperature to 1200 °C in nitrogen ambient. Defect structure was studied before and after recovery by transmission electron microscopy (TEM).HRXRD symmetric 2θ-ω measurement on Mgimplanted Al 0.8 Ga 0.2 N and Si-implanted AlN revealed a damage peak at lower angle than that of matrix. The implanted layer showed an expanded lattice constant, suggesting interstitial impurities. In-situ high temperature 2θ-ω measurements indicated that the matrix peak shifted towards lower angle with temperature due to thermal expansion while damage recovery was categorized into four temperature regions: (I) RT-150 °C, stable stage, damage peak showed same thermal expansion effects with that of matrix; (II) 150-300 °C, fast recovery stage, damage peak started shifting towards matrix peak, opposite to the thermal expansion effects, with a strong temperature dependence; (III) 300-600 °C, slow recovery stage, damage peak slowly merged into matrix; (IV) above 600 °C, no damage peak could be discerned. A lower annealing temperature was confirmed to recover the implantation damage than the expected typical condition (1200°C). The details of processes during damage recovery and the resulting consequences will be presented. High critical field strength and saturation electron velocity of gallium nitride (GaN) makes it a promising material for future compact, high voltage, and high-speed switches. Photoconductive semiconductor switches (PCSSs) have advantages in pulsed power applications since it is a device that turns on with fast rise time, low jitter, and allows for high repetition rate. Semi-insulating GaN doped with carbon (GaN:C) reduces the off-state leakage current and extends the blocking voltage, while still allowing for high on-state photocurrent compared to unintentionally doped GaN. However, forming a low resistance contact to highly insulating GaN:C is challenging. Our approach is to leverage ion implantation of Si followed by an activation anneal to form an N+ contact layer. Implantation and activation is a planar process that is promising for formation of contact regions while mitigating high electric field points within the device. In this work, we implant Si with a box profile to a depth of 0.4 μm into GaN:C with carbon concentrations of 5x10 17 and 1x10 18 cm -3 grown on SiC substrates. Then, a sputtered AlN capping layer is introduced before annealing at various temperatures to preserve the surface from decomposition during annealing. The AlN capping layer is then selectively wet etched and Ti/Al/Ni/Au contacts are deposited and then alloyed at 850 °C. Top-to-top lateral PCSSs are fabricated and characterized to identify the impact of the Si implantation and annealing in the contact regions on the contact resistance, off-state leakage, blocking voltage, and photocurrent.

Si-Implanted Contacts in Semi-Insulating Carbon Doped GaN for
III-nitride quantum dots (QDs) have significant potential for single-photon sources or gain media for low threshold and high efficiency visible and UV lasers, among others. "Bottom-up" Stranski− Krastanov growth is widely used, however both the size distribution and densities are difficult to precisely control. Here, we show a top-down fabrication process that can be controlled by the evolving properties of the nanostructures being fabricated. This process, called quantum size controlled photoelectrochemical (QSC-PEC) etching, uses laser excitation at a selected and narrowband wavelength to control the final sizes of the QDs through an etch that self-terminates when the QD band gaps increase due to quantum confinement effects until they exceed the energy of the incident photons. Beginning with epitaxially grown InGaN films, we examine the etch process from large to quantum-scale nanostructures with AFM, scanning TEM (STEM), and photoluminescence measurements. Quantitative analysis of size and density of the ensemble are made after image-post processing techniques and deconvolution of the AFM tip and QD. We further investigate the fabrication of multilayers of QDs from multiple-quantum well structures, which can potentially increase the overall QD density per device. These results show the potential for a combination of unprecedented size uniformity and areal density for InGaN QD devices. We also have investigated the extension of this process to create QDs in other semiconductor materials systems. Results will be presented on the PEC etching of GaAs/AlGaAs planar films using abovebandgap laser light (~790-830 nm) and characterized by AFM and STEM measurements.We have also demonstrated the formation of high density, vertical arrays of very high-aspect ratio (e.g. > 200:1) GaN nanowires via PEC etching of thin GaN epilayers on sapphire. Cross-sectional STEM characterization of these nanowires show that they are not correlated with dislocations from the original GaN epilayer. We propose that lateral etching below a certain diameter is retarded due to carrier depletion in the near-surface region, while allowing for continued vertical etching and very-high aspect ratios. Photoluminescence measurements indicate that their optical quality is equivalent or superior to the starting unetched film, despite the much higher surface area, which is typically a source of nonradiative recombination. Such nanowires could potentially benefit numerous applications such as optoelectronics (e.g. LEDs, lasers, photodetectors, photovoltaics), piezoelectric nanogenerators, and gas sensors.This work was performed, in part, at the Center for Integrated Nanotechnologies, a U.S. Department of Energy, Office of Basic Energy Sciences user facility. Sandia National Laboratories is a multimission Traditional III-Nitride semiconductor materials GaN and AlN have shown excellent optoelectronic properties, enabling light emitting diodes and high-electron mobility transistors. Tremendous interest has recently been generated regarding alloying III-nitrides with the transition metal Scandium (Sc) due to its predicted extremely large solubility in III-nitride crystals, a potential 3X or more increase in the piezoelectric coefficients, and potential ferroelectric behavior. [1]In this work, we report the epitaxial growth studies of single-crystalline ScxGa1-xN alloys with Sc concentrations of approximately 2-10% by Molecular Beam Epitaxy (MBE). ScxGa1-xN alloy layers were deposited on MBE grown GaN layers grown on GaN-SiC templates by standard RF-plasma MBE. Ga metal was supplied from a Knudsen effusion cell, and nitrogen was supplied using a radiofrequency (RF) plasma source from a purified N2 gas line. Sc metal was supplied from a 10kV electron beam evaporator. GaN layers were grown with a Ga cell base temperature of 935C and a nitrogen plasma power of 200 watts. This corresponds to a growth rate of ~390 nm/hour. GaN layers were grown under metal-rich conditions to promote smooth interfaces. ScxGa1-xN layers were then grown with a Ga flux greater than the active nitrogen flux to test for preferential incorporation of Sc into the wurtzite crystal structure. Both layers were grown at a substrate temperature of 750 -800C, measured by a thermocouple. In-situ reflection high energy electron diffraction (RHEED) was used to assess surface evolution during growth.Sc concentrations were measured in the films after growth using out of plane lattice constant measurements in X-ray-diffraction (XRD), and corroborated by X-ray Photoelectron Spectroscopy (XPS). Lattice constant data was correlated with Sc concentration using secondary-ion mass spectrometry (SIMS) with calibrated sensitivity factors from Rutherford Backscattering spectrometry (RBS). Films with surface roughness less than 1.5 nm have been determined by atomic force microscopy (AFM). XRD measurements of symmetric 002 reflections indicate an increase in the c-plane lattice constant, indicating that Sc has incorporated into the crystal. This follows with the fact that the Sc-N bond length is larger than the Ga-N bond length. This is in contrast to the lattice constant values previously predicted for ScxGa1-xN alloys. However, those calculations were based on the emergence of a metastable nonpolar layered hexagonal phase. The obtained lattice parameters here approximately correlate with a Vegard's law extrapolation to a thermodynamically unstable wurtzite structure for ScN with endpoint c lattice constant of 5.58 Å. This data is also consistent with the Sc concentration determined by RBS. The results suggest a solid solution exists with Sc atoms substituting for Ga atoms at the cation sites. AFM images of samples with Sc concentrations from 2-10% indicate the epi-layers are smooth for thicknesses ranging from 20 to 90 nanometers. At higher Sc fluxes, the RHEED pattern changes from hexagonal symmetry to a cubic-twinned pattern. This is indicative of potential phase separation into cubic (111) oriented ScN regions. The (111) lattice constant of cubic ScN is within 0.2 % of the GaN in-plane lattice constant (3.861 Å). The cubic structure of ScN is the thermodynamically stable phase, so the total energy of the system during growth at elevated temperatures may be reduced by forming coherent twin boundaries with (111) out of plane orientation. These results potentially place a limit to the extent which Sc may be incorporated into the crystal structure under the current range of growth conditions. Future work will focus on uncovering fundamental growth mechanisms of the Sc-based III-nitride systems to incorporate larger Sc concentrations while maintaining smooth interfaces and preventing phase separation from the wurtzite structure. Due to high diffusion barriers for nitrogen adatoms on the bare surface of GaN [1], metal-rich conditions are usually employed to obtain smooth GaN layers of any polarity in plasma-assisted molecular beam epitaxy (PA-MBE). Another approach to get smooth GaN layers was proposed by Koblmüller et al. [2], where nominally slightly nitrogen-rich conditions were used in Ga-polar GaN growth at high temperature. In this concept, growth was carried out in so called "wet surface N-rich" growth conditions, where GaN surface decomposes significantly and metal layers is formed due to decomposition. As it was pointed out by B.M. McSkimming et al. [3], presenting full growth diagram for Ga-polar GaN, dry surface in N-rich conditions results in rough surface, so called 3D growth front, of the crystal. On the other hand, theoretical calculations presented by T. Zywietz [1] indicate that lower diffusion barriers can be expected for N-polar surface comparing to more commonly used Ga-polar. Higher diffusivity of adatoms for N-polar surface comparing to Ga-polar one enabled the smooth growth in N-rich conditions for N-polar substrates with miscut angle close to 4 deg [4]. In these work, we investigate the growth of GaN on N-polar high quality GaN substrates in N-rich growth conditions at miscut angles ranging from 0.3 to 4 deg at temperature 750 o C. Meandering atomic step edges have been observed for wide range of atomic fluxes used for the growth. Contradictory to previous reports on meandering (finger-like) morphologies for GaN and cubic crystals [5][6][7], wavelength of those meanders decrease with increasing miscut angle and increase with increasing growth rate. Such observation led us to form the hypothesis that, in this case, finger-like morphology cannot be explained in simple model with non-interacting adatoms. Lower wavelength of meanders for higher miscut angles indicate that diffusion over atomic steps, blocked by Ehrlich-Schwoebel barrier (ESB), is an important process that effectively leads to lower diffusion length for higher step density. Increase in meanders' wavelength for higher growth rate, on the other hand, indicate that Ga adatoms interact with each other leading to diffusion enhancement. We postulate that this effect can be realized by lowering ESB. To prove the hypothesis, kinetic Monte Carlo simulations were used. N-polar GaN surface morphology obtained for N-rich growth conditions was studied. Simulations were performed to test the influence of Ga and N fluxes as well as miscut angle. Model that takes into account ESB at atomic steps that can be lowered by the presence of Ga adatoms at neighboring sites was able to qualitatively reproduce experimentally observed morphologies. Growth diagram indicating transitions between different growth modes for N-rich growth on N-polar surface in PA-MBE will be presented. [1] J. N. Tosja Zywietz, and Matthias Scheffler, Applied Physics Letters 73, 487 (1998 High growth rate molecular beam epitaxy (MBE) of III-nitrides has been reported with growth rates as high as ~10 μm/hour utilizing high flow plasma sources for nitrogen. Growth of low bandgap, high indium content InGaN is essential to realize many types of III-nitride devices, particularly solar cells and efficient green light emitting diodes (LEDs.) High quality, single phase InGaN is achievable with plasma-assisted molecular beam epitaxy (PAMBE). This plasma-assisted style of MBE employs a plasma to efficiently break strong N 2 gas bonds in order to produce active nitrogen species in the form of excited neutral and ionized nitrogen molecules and atoms. High indium content nitrides possess a high sensitivity to the plasma discharge as observed in both atomic force microscopy (AFM) and reflection high energy electron diffraction (RHEED.) The monitoring, control, and optimization of a high growth rate Veeco plasma discharge is detailed in order to minimize damage to the III-nitrides, particularly high indium content films. In-situ plasma discharge monitoring is achieved both with optical emission spectroscopy (OES) and a flux gauge collector pin repurposed as a Langmuir probe. By varying plasma input conditions, primarily gas species, flow rate, and applied power, we can correlate these input conditions to the output plasma discharge observed with OES and a Langmuir probe determined floating acceleration voltage, ion and electron current to finally optimize sensitive InN containing material quality. We present results indicating that increased nitrogen flows increase positive ion content, but reduce the magnitude of the floating voltage between the plasma and the film. Unexpectedly, the results also illustrate that increased applied power to the plasma bulb only slightly increases positive ion content, while having a minimal effect on the floating voltage. AFM results of 1 µm thick InN grown films exhibit reduced pit densities when grown under plasma discharge conditions with low densities of positively charged ions. These combined techniques for plasma optimization should enable improved high indium content III-nitride device performance. Heterostructures of III-nitride semiconductors (GaN, AlN, and InN) have unique electronic and optical properties that make them promising for fabrication novel near-infrared and terahertz emitters and detectors based on intersubband transitions. Importantly, application of III-nitride heterostructures may include frequencies that are fundamentally inaccessible to their arsenide analogs due to the reststrahlen absorption of radiation. One of the major limiting factors in the growth of complex nitride heterostructures is the lattice mismatch between AlN, GaN, and InN. In case of In 0.17 Al 0.83 N, the film is lattice-matched to GaN in the (0001) plane offering the potential for fabrication of thick strain-free heterostructures. However, InAlN is subject to segregation into In-rich regions, deviating strongly from a random alloy distribution. Due to strong differences between its antecedents, AlN and InN, InAlN can form a 'honeycomb' microstructure [1][2][3]. In work by Sahonta et al., the growth of InAlN films by plasma-assisted molecular beam epitaxy (PAMBE) is described by three successive steps: the formation of Al-rich InAlN dynamical platelets, their partial coalescence, and the preferential indium incorporation at platelet boundaries due to tensile strain generated between platelets [1]. Thus, the growth mechanism includes the inevitable quasithree-dimensional stage on the onset of InAN formation and the transient to further two-dimensional growth, which is similar to InN growth mechanism at low temperatures [4]. Recently, several groups reported on the elimination of 'honeycomb' microstructure in metal-polar and nitrogen-polar lattice-matched InAlN films [5][6][7]. It has been assumed that the homogeneous films were achieved through the N-rich PAMBE by lowering the growth rate and increasing indium to aluminum ratio [5,6]. However, understanding of the growth mechanism affecting segregation remains primitive at best, and importantly, the control of the nucleation stage on the onset of InAlN growth has not been extensively studied. In this work, we study the impact of early stages of InAlN growth on the honeycomb microstructure formation. InAlN samples were grown on GaN surface using different schemes of growth initiation: growth on an N-rich (2x2) reconstructed GaN surface, GaN surface exposed to In-flux prior the onset of InAlN growth, and GaN surface cooled to the InAN growth temperature under indium coverage. The ex-situ characterization of the dynamic Al-rich platelets on the onset of the InAlN growth was done on 4-nm-thick layers using atomic force microscopy, whereas plan-view scanning transmission electron microscopy was applied to analyze 'honeycomb' microstructures. We discuss the dependence of the platelets size on the initial growth conditions. It was found that the size of dynamic platelets can be effectively increased by initiating the InAlN growth on GaN covered with metallic indium layer. Importantly, that excess indium on the surface is consumed during the InAlN growth and does not result in a buildup of crystalline indium droplets on the surface. The correlated analysis of high-resolution AFM and plan-view STEM data provides a better understanding of the impact of earlier stages of InAlN growth on the elimination of the 'honeycomb' microstructures. Introduction:Recently, Metal/Oxide/Nitride/Oxide/Si (MONOS) nonvolatile memory (NVM) with high-k gate material is attracting much attention to achieve low voltage operation by reducing the insulator thickness. As the scaling of gate insulator, the interface roughness at the high-k gate insulator/Si is one of the critical issues to realize the high reliability of MONOS NVM. In our previous report, the atomic steps on Si(100) surface was realized by annealing at 1050°C/60 min in Ar/4.0%H 2 ambient utilizing the clean furnace, and the degradation of leakage currents of Hf-based MONOS diodes after 10 3 program/erase cycles were decreased by flattening. However, the flattening process by annealing in Ar/4.0%H 2 ambient still has some issues such as a periodic surface roughness with a few micrometer length. Therefore, to realize ideal atomically flat Si(100) surface, the periodic surface roughness should be decreased. In this study, the use of chemically oxidized Si(100) for stable atomically flattening and the electrical characteristics of Hf-based MONOS diodes were investigated.Experimental procedure:After cleaning of p-Si(100) substrate with off-angle less than 1.0°, some of the substrates were immersed in H 2 O 2 /60 min at room temperature (RT) to form 0.7 nm thick chemical oxide (Ch. Ox.). After annealing at 1050°C/10 min in Ar/4.0%H 2 (3 SLM) utilizing the clean furnace, the unintentional oxide layer which was formed during the annealing process was removed by wet etching (HF:HCl=1:19). Then, HfN 0.5 (M)/HfO 2 (O)/HfN 1.0 (N)/HfO 2 (O) gate stack with thickness of 10/10/3/2 nm, respectively, was in-situ deposited on p-Si(100) by ECR plasma sputtering at RT. After postdeposition annealing was carried out at 600°C/1 min in N 2 (1 SLM), Al top and back electrodes were evaporated. Finally, the post-metallization annealing was carried out at 300°C/10 min in N 2 /4.9%H 2 (1 SLM). The Si surface roughness was observed by non-contact mode atomic force microscopy (AFM). The C-V characteristics of MONOS diodes were evaluated at RT.Results and Conclusion:The scan size dependence was decreased, and uniformity was improved by using chemically oxidized Si in spite of the short flattening duration such as 10 min. Furthermore, the obtained surface RMS roughness of 10 x 10 μm 2 was decreased from 0.60 nm (without Ch. Ox.) to 0.26 nm (with Ch. Ox.). This is caused by suppressing the periodic surface roughness probably because the unintentional oxide formation with thickness fluctuations was suppressed. Large memory window (MW) of 1.9 V was realized by the flattening of chemically oxidized Si compared to the MW of 1.7 V in the case of without flattening. In conclusion, annealing at 1050°C/10 min in Ar/4.0%H 2 ambient with chemically oxidized Si (100)  Multi-gated metal oxide semiconductor field effect transistors (MOSFETs) have recently become important in high-performances CMOS integrated circuits. Multi-gated devices, commonly called FinFETs, have reduced short channel effects, allowing for greater scalability.

Impact of Early Stages of InAlN Growth by Plasma-Assisted Molecular Beam Epitaxy on Honeycomb Structures Formation
[1] However, there is little in the literature about the atomic defects at the semiconductordielectric interface in FinFETs. This is significant, as FinFETs have recently become the dominant devices in high-performances CMOS integrated circuits. In this study, we explore traps at the FinFET Si/ dielectric interface with electrically detected magnetic resonance (EDMR). The devices involved in this experiment are on (100) silicon-on-insulator wafers with 90nm Si layers and 125nm buried oxides. The FETs have 1nm SiO 2 and 2nm HfSiON/TiN/polySi-capped gate stacks with an effective oxide thickness of about 1.1nm. The body of the devices are lightly doped p-type at 2x10 15 /cm 3 . Each FinFET is configured as a gated diode with n+/p-/p+ with a fin length of 500nm, fin height of 80nm, and fin width of 50nm. For a single set of contacts, 500 fins are connected in parallel. Extensive electrical measurements on these devices have been reported by Young et al.[2][3] In order to increase the defect density and maximize the size of the resonance response, we have irradiated the FinFETs to 1 MRad via a 60 Co gamma source. Pre-irradiated EDMR spectra are weak to below detectable limits whereas quite strong signal to noise spectra appear after the irradiation. Our EDMR measurements utilized a home-built spectrometer. The X-band (≈9.5 GHz) spectrometer includes a 4-inch Lakeshore electromagnet with a Micro-Now microwave bridge and a TE 102 cavity. Spin dependent device current was measured with a Stanford Research Systems Low-Noise Current Preamplifier. The detection utilized a home-built virtual lock-in amplifier. Measurements were conducted at room temperature. Our results are significantly different from what is observed in both conventional EPR and EDMR measurements of irradiated and unirradiated planar MOSFETs. In planar MOSFETs, the magnetic resonance due to defects at the interface is dominated by defects in the P b family. The g tensor components of the P b centers vary from approximately 2.0012 to 2.0083. The zero-crossing g's of the rather broad lines that we observe are within this range; however, the breadth of the spectra, about 25 Gauss are much larger than one would anticipate for P b centers. Assuming the full range of g values found for the P b centers, one would anticipate an overall linewidth of about 12 or 13 Gauss. The FinFET response is twice this width with g values from 2.0038±0.0002 to 2.0044±0.0002. These results suggest a more complex interface defect structure, which likely involves more than P b centers. Most significantly, this work demonstrates the potential of EDMR to study point-defects in FinFETs. Our results, although preliminary in nature, strongly indicate that although P b -like defects likely play important roles in the Si/SiO 2 interface, the defect spectrum is more complex than is the case in conventional planar MOSFETs. This project is sponsored by the Department of Defense, Defense Threat Reduction Agency under grant number HDTRA1-16-0008. The content of the information does not necessarily reflect the position or the policy of the federal government, and no official endorsement should be inferred.
[1] However, the effects of varying N/Si stoichiometry on defect levels and defect chemistry have not been extensively studied with EDMR. In this study, we expand upon the work by Mutch et al. by comparing the EDMR and near-zero field magnetoresistance responses of stoichiometric, Si rich, and N rich films of a-SiN:H.Low field magnetoresistance (MR) phenomena, [2] the change of device current when a low magnetic field is applied, is also a current topic of interest in many materials systems. The MR phenomena is most commonly studied in "messy" organic semiconductor structures; however, a similar effect is also observed in inorganic semiconductor and semiconductor/insulator structures. Inorganic materials systems that are extensively utilized in integrated circuits, such as a-SiN:H, have been very well characterized, providing advantages over organic systems for the study of the MR phenomena.The device structures under observation consist of Ti/a-SiN:H/p-Si capacitors. In all measurements presented in this study, a slowly varying magnetic field is applied to the thin film under bias. For the EDMR measurements, an oscillating microwave frequency magnetic field (v~9.75GHz) or radio frequency magnetic field (v~250MHz) is applied perpendicular to the quasistatic field. As in conventional EPR, energy is absorbed by paramagnetic sites when the resonance condition is met. In the simplest cases, this resonance condition is expressed by hv=gμB, where h is Planck's constant, v is the frequency of the applied microwave/radiofrequency field, g is an orientation dependent parameter typically close to 2, μ is the Bohr magneton, and B is the magnetic field at resonance. In EDMR, the EPR transition is observed through a change in device current. The MR response is observed without the presence of a perpendicular resonant field as the magnetic field is swept through zero applied field.We study 25 nm a-SiN:H samples with N/ Si ratios of 1, 1.35, and 1.5. High frequency and low frequency measurements provide us with information about defect structure, as they allow some separation of spin orbit coupling and electron nuclear hyperfine interactions. The measurements also directly indicate that these defects are involved in electronic transport through these films. A comparison of EDMR measurements taken at high frequency and field and low frequency and field identify the primary defect responsible for transport through these films as a Si dangling bond such as the K center. An identical central resonance condition is observed in all three stoichiometries, consistent with this Si dangling bond. The high frequency EDMR measurements show that the Si rich samples exhibit a different response than the N rich and stoichiometric samples. A comparison of high and low field EDMR indicates that in the Si rich samples, Si replaces some of the K center backbonded N atoms. A comparison of the EDMR response to the MR response in these films provides insight into the MR response, especially with regard to differences in hyperfine interactions.  Hydrogen treatments are used extensively in established silicon technologies to reduce surface and interface recombination. It has been proposed that hydrogenation could also be used to improve the minority carrier lifetime of silicon where lifetime degradation is due to transition metal contamination. Such contamination is often the case for low cost "solar" silicon where purification and/or crystal growth processes are less rigorous than for electronic grade material. The problem is particularly severe for the so called "early" transition metals (lower atomic numbers) which have low diffusivities and cannot be gettered by conventional methods.Much work has been published on hydrogen in silicon and it is evident that in n-type silicon reactions between hydrogen and some TMs can result in passivation or reduced recombination activity [eg A. R. Peaker et al Hydrogen-related defects in silicon, germanium, and silicongermanium alloys, in: Defects in Microelectronic Materials and Devices, eds D. Fleetwood et al. (Taylor & Francis,Boca Raton, 2009), pp. 27-55]. However in general, the bonding between hydrogen and the TMs is weak (~2 eV) and so the hydrogen is dissociated at temperatures well below those experienced in device processing. Hydrogen is very mobile in silicon but its reaction with TMs (and other defects) is highly dependent on the charge state relative to the defect or impurity. Predominantly, the reaction is driven by long range Coulombic interactions. Atomic hydrogen is an amphoteric impurity with negative-U ordered donor and acceptor levels at E c -0.18 eV and E c -0.65 eV. The neutral state is metastable. Isolated hydrogen occurs in the positive charge state in p-type Si, while it is negatively charged in n-type. As the H (+/-) occupancy level lies in the upper half of the band gap it would be expected that in intrinsic Si (mid-gap Fermi level) the H + charge state will dominate. Experimentally, there is evidence that His the stable form when the Fermi level is >0.3 eV above the mid gap energy.The end result is that in n-type Si reactions between hydrogen and transition metals occur while in p-type Si they do not occur under equilibrium conditions. There are no data on p-type reactions although we have published results on some TM-hydrogen reactions in the depletion region of a Schottky diode on p-type silicon … whether such a reaction occurs or not depends on the charge state of the TM with a near mid-gap Fermi level.Recently important increases of minority carrier lifetime have been demonstrated in multi-crystalline p-type silicon when hydrogenated at elevated temperatures under strong illumination [eg B. Hallam et al, IEEE J. Photovolt. 99, 1 (2013);S. Wenham et al., Advanced hydrogenation of silicon solar cells, U.S. Patent 9,190, 556, Nov 2015]. It has been proposed that this is due to the passivation of TMs. We have undertaken a systematic study into the hydrogenation of p-type single crystal silicon after anneals under illumination. The material has been intentionally contaminated with specific TMs, namely Mo, V, and Fe. In this report we focus on molybdenum. We have used DLTS, Laplace DLTS and capacitance-voltage measurements to study the formation of TM-H complexes upon illuminated annealing at temperatures at or below 200°C. We show that electrically inactive TM-H complexes can be formed in samples to which hydrogen has been introduced via remote plasma followed by annealing under illumination with a 830 nm LED. This creates a population of minority carriers in the material. The complexes are shown not to be present in similarly treated samples annealed in the dark. The initial concentration of electrically active TM is shown to be recovered following higher temperature annealing (~250 °C). This is a temperature known to dissociate TM-H complexes in n-type Si. Possible mechanisms behind TM-H bonding in p-type Si are discussed. The use of scanning tunneling microscopy (STM) probe tip-based hydrogen lithography to deterministically pattern phosphorus dopants on Si (100) surfaces is a promising method to produce atomic-scale devices [1] such as nanowires, quantum dots, and single-dopant transistors. This technique is a powerful tool to create the building blocks of the Si-based Kane quantum computer [2], study the ultimate scaling of conventional semiconductor devices, and engineer two-dimensional (2D) metamaterials. In this technique, an appropriately biased STM tip is used to selectively remove H-atoms from a fully H-terminated Si surface leaving behind exposed regions of silicon where phosphine (PH3) molecules can selectively bind P into the Si. Subsequent annealing is used to incorporate the P dopants which are then encapsulated by low-temperature Si overgrowth. Dopant segregation, diffusion, and surface roughening can occur during the epitaxial encapsulation process which introduces uncertainties in the final locations of the dopant atoms. Thus, it is a serious materials science challenge to move this compelling fabrication approach from scientific concept to practical implementation. To succeed, effective measurements of the dopant positions are needed.We present complementary secondary ion mass spectrometry (SIMS) and quantum transport methods to quantify dopant movement at the atomic scale. By combining segregation/diffusion models with sputtering profiling simulations we developed a robust method to quantify dopant movement at the atomic scale from SIMS depth profiling measurements.
[3] We applied this refined SIMS analysis to UHV fabricated devices containing two-dimensional (2) highly P-doped "monolayers" that include ones made by using a locking layer (LL), and ones made without a LL. The application of thin room-temperature grown LLs followed by encapsulation overgrowth at elevated temperatures is a key development to address the trade-off between low-temperature encapsulation needed for sharp dopant confinement and high-temperature encapsulation for optimum epitaxial quality.A more unorthodox method to determine confinement of the delta-layer dopants in these samples is to measure the quantum mechanical effect, weak localization, in parallel and perpendicular magnetic fields. [4] This approach allows the extraction of the delta-layer thickness from fits to the Hikami-Larkin-Nagaoka equation. These magnetotransport measurements offer the compelling advantage of overcoming the resolution limits of SIMS, and we find good agreement These magnetotransport measurements offer the compelling advantage of overcoming the resolution limits of SIMS, and we find good agreement with our results of the advanced SIMS analysis.References:[1] see, for example, S. Schofield, et al., Phys. Rev. Lett. 91, 136104 (2003 Niobium nitride (NbN) is a refractory compound metal which undergoes a transition to the superconducting state at a temperature (~17K), significantly higher than that of elemental Niobium (9.3K). For that reason, NbN thin fims are currently used in superconducting electronic applications. It has been demonstrated that NbN thin films can be grown epitaxially on silicon carbide (SiC), aluminum nitride (AlN), and gallium nitride (GaN) substrates by a variety of deposition techniques. We present here the growth of heterostructures incorporating superconducting/metallic NbN thin films with III-nitride semiconductor films by plasma assisted molecular beam epitaxy (PAMBE). The growth is performed using an electron beam evaporation source for niobium, an RF plasma nitrogen source, and an effusion cell provides the source of gallium. NbN film growth has been performed for substrate temperatures ranging between 700 o C up to 1150 o C on GaN substrate and above 1300 o C on SiC substrate. We demonstrate optimization of structural and superconducting properties of epitaxial NbN films on GaN and SiC substrates as a function of substrate temperature, growth rate, and the ratio of niobium and nitrogen fluxes. We conclude that the crystalline phase, stoichiometry, and electrical properties of NbN films grown by PAMBE are highly sensitive to substrate temperature and the ratio between the flux of Nb and N. Root mean squared surface roughness of optimized NbN films on GaN or SiC substrate as measured by atomic force microscopy (AFM) for 1um 2 area are approximately 0.5nm. Superconducting critical temperatures in excess of 15K are demonstrated for films as thin as 20nm. The resistivity and superconducting transition are shown to be highly dependent on growth conditions, though NbN thin films reliably exhibit metallic electronic properties above the critical temperature with resistivities of approximately 150μΩ-cm, making NbN significantly more resistive than most metals commonly used in semiconductor devices. We also compare the structural and electronic properties of NbN films grown on SiC and GaN as characterized by x-ray diffraction (XRD), atomic force microscopy (AFM), transmission electron microscopy (TEM), electron back-scatter diffraction (EBSD), I-V measurements of resistivity, and Rutherford backscattering spectroscopy (RBS). Structural differences, symmetry mismatch, and lattice parameter misfit between NbN, SiC and GaN are shown to be responsible for observed differences in the properties and optimized growth parameters between NbN films grown on GaN and those grown on SiC.
[DJ1]The all-epitaxial nitride superconductor/ semiconductor materials system composed of NbN, AlN, GaN, InN and their alloys enables novel interfaces between established semiconductor devices and superconducting devices. We demonstrate a high performance III-nitride high electron mobility transistor (HEMT) incorporating electronic transport through a buried superconducting NbN thin film. In addition, the ability to fabricate all-epitaxial semiconductor/ superconductor heterojunctions enables exploration of new phenomena and devices utilizing the interface between electrons in the superconducting state with the range of electronic phenomena possible in III-N heterostructures. Electronic transport across an epitaxial interface between NbN source and drain contacts and a high mobility twodimensional electron gas (2DEG) state at an AlGaN/GaN interface is demonstrated and characterized in both the superconducting and normal state, utilizing the field effect to modulate the density of carriers in the 2DEG. Epitaxial all-nitride Josephson junctions utilizing novel epitaxial tunneling barrier materials are currently being explored and preliminary results will be presented.
Laterally grown III-V epitaxial layers offer multiple advantages over conventional vertically grown layers for fabricating advanced semiconductor devices. Etching vertical structures to define 30nm thick fins is a challenge and leads to considerable damage of the device structure, while lateral devices with pre-fabricated templates can be easily used to achieve lower thicknesses. Growing lateral devices on top of dielectric films also eliminates leakage currents through conducting substrates. A promising approach to laterally grow epitaxial films is by using dielectric templates to confine the growth and its direction. This technique was originally developed for Si[1,2] and later used for integrating III-V semiconductor device structures with conventional Si technology [3,4]. Structural properties of these epitaxial layers grown in confined structures is required for optimizing growth parameters to improve device performance. Owing to the sub-micron scale of these embedded epitaxial layers, it is challenging to characterize them. Here we present scanning electron microscopic (SEM) and transmission electron microscopic (TEM) studies of InP Confined Epitaxial Lateral Overgrowth (CELO) structures grown by chemical beam epitaxy (CBE) and metal organic chemical vapor deposition (MOCVD) on patterned InP (001) and InP (110) substrates. We discuss the influence of growth conditions on parasitic nucleation, evolution of facets and nature of defects. The dependence of facet planes on the growth direction and substrate orientations, and the effects of patterning procedures and geometry of the confined structures on crystal quality and yield is also explored. We further show growth and characteristics of CELO heterostructures. For CBE, the growths were performed in a VG-Semicon V80H system using phosphine(PH 3 ) and trimethylindium (TMI) as precursors with temperatures varying from 470-520 °C. The results show a clear dependence of selectivity and facet formation on growth. Etching the top oxide layer and performing tilted SEM, shows that the facets are dependent on the growth directions. Vertical facets form in structures growing in the <1-10> direction, while <001> oriented overgrowths show slanted {111}B facets for a InP(110) substrate. The facets are consistent with those expected from a general zincblende structure. Cross sectional TEM images confirm these facets, while also revealing stacking faults in the {111} plane propagating throughout the growth in the structures oriented along the <100> direction. However, the growth interface at the seed hole shows excellent crystal quality and appears to be a continuous defect free crystal lattice, suggesting that the stacking faults do not form until a {111} facet forms during the growth. Preliminary analysis show both the (110) and the (001) oriented InP wafers exhibiting similar stacking fault densities, while a reduction of the fault density is observed at higher temperatures. The TEM results consistently show a depression of the CELO box ceiling above the seed hole which has been solved in subsequent fabrications by using a spin-on resist-based process. Scanning TEM imaging carried out on a MOCVD grown CELO InP/InAs heterostructures show that extremely sharp heterojunctions, up to a few monolayers thin, can be realized in a lateral direction using this templatebased growth technique. This provides a path to fabricate highly efficient lateral tunnel FETs  Embedded air voids in III-V semiconductors have demonstrated a wide range of enhanced photonic functionality such as beam-shaping, polarization control, and enhanced light extraction/absorption in optoelectronic devices. From a materials perspective, the formation of high-quality well-faceted embedded micro-scale air voids in III-V materials is the primary goal, with challenges arising from material type, anisotropic growth fronts, and fabrication. Air voids are commonly formed in GaN Pendeo-epitaxy due to highly anisotropic growth afforded by the Wurtzite crystal structure [1]; however, applying similar growth techniques to conventional zincblende III-V materials such as GaAs remains limited, requiring other methods such as etch substrates to form nano-voids [2].Recent investigations of epitaxially-integrated dielectric silica gratings in MBE-GaAs have yielded high-quality planar encapsulation on (001) substrates [3]. Expanding the investigation to other silica nanostructures, here we report the formation of complete air voids immediately above masked dielectric regions in hexagonal pillar arrays. Conical-like in shape, these air voids are self-formed as a consequence of the directional-dependence of adatom surface migration during GaAs overgrowth of the dielectric pillar array. The primary benefits of forming air voids in this fashion are that the method is fully-scalable, entirely monolithic, and does not require additional fabrication steps (e.g. postgrowth selective etch and/or several regrowth steps).Samples were grown by solid-source molecular beam epitaxy in an EPI Mod. Gen II system on n-doped (001) GaAs substrates. Prior to growth, the substrate was patterned with 1.9µm hexagonal pillars arrays of 1.1µm silica pillars. Growth of GaAs was performed at 635°C using a cyclic selective growth technique known as periodic supply epitaxy (PSE), depositing 7.5ML of GaAs per growth cycle and annealing for 30s per cycle for decomposition of any poly-GaAs on the silica pillars to desorb. A total of 540 cycles were needed to fully encapsulate the air void. Planview and cross-sectional SEM imaging was performed to characterize air void coalescence and trench formation. Optical characteristics will be reported at the conference. Patterned air gaps integrated with high-quality epitaxial semiconductors have wide ranging potential applications, including high-contrast nanophotonics, gas sensing, and opto-fluidics [1, 2]. For example, encapsulated air gratings can be exploited to replace Bragg reflectors in VCSELs, create high-Q resonators, create lab-on-a-chip sensing systems, or utilize Fano resonance for all-optical switching. However, it is difficult to integrate these structures with standard III-V optoelectronic devices without compromising material quality as conventional III-V growth techniques offer limited lateral control and other techniques, such as wafer bonding of patterned structures, can introduce interfacial defects. Previous attempts to create air gaps in zincblende III-V semiconductors have been limited in shape and size [3] and silicon-based processes damage the surface of the material.Recently, we showed that tailored molecular beam epitaxy (MBE) can be used to encapsulate patterned silicon dioxide structures into a high-quality crystalline III-V matrix, yielding monolithically integrated high contrast photonic structures [4]. Here, we increase the achievable refractive index contrast by post-growth selective etching of the silicon dioxide. Specifically, mesas were defined that extend through the silicon dioxide layers, followed by highly-selective lateral wet etching of the silicon dioxide with buffered oxide etch (BOE). This approach is capable of yielding interconnected air cavities of arbitrary shape and size, embedded in single-crystal III-V materials.We confirmed the successful fabrication of nanometer-scale air cavities in GaAs with scanning electron microscopy (SEM) and Fourier transform infrared spectroscopy (FTIR). Complete etching was confirmed by monitoring the silicon dioxide phonon absorption resonance at 1100 cm -1 . Finitedifference time-domain (FDTD) simulations were used to evaluate potential applications of these structures. This work establishes a basis for improved high-contrast photonics and opens up new potential applications, including lab-on-a-chip sensing via seamless integration of micro-fluidics with III-V semiconductor photonic devices. This work was supported by the National Science Foundation (NSF-ECCS-1408302).References Heteroepitaxy of lattice-mismatched materials has garnered renewed interest in recent years, as new mechanisms are sought for substrateinvariant growth of high-quality epitaxial layers with repeatable, reliable device performance. Several methods of relaxed heteroepitaxy have been proposed [1] including a variety of graded buffer schemes, aspect ratio trapping, and strained superlattice dislocation filters. One technique that has demonstrated particular promise is epitaxial lateral overgrowth (ELO) [2,3], a growth approach where small openings in dielectric-patterned substrates act as seeds for heteroepitaxial growth. Using a highly selective lateral growth process, propagating defects are localized to the small seed regions resulting in large relatively defect-free regions located immediately above the patterned dielectric layers. ELO has been previously demonstrated on several III-V binaries such as GaN/Sapphire and GaAs/Si substrates on patterned SiO 2 and/or SiN x dielectric masks using MOCVD and LPE crystal growth techniques, respectively. Recently, we developed a high-quality, all-MBE, growth approach for the planar encapsulation of dielectric microstructures in homoepitaxial GaAs on (001) oriented substrates [4]. Drawing on this capability, we extend this work to planar encapsulation of dielectric masking layers in large latticemismatch heteroepitaxial III-V MBE growth. As a test bed system, we explore the defect reduction in the device regions on InAs/GaAs from ELO of silica masking layers. To this end, we show the first demonstration of ELO metamorphics using an all-MBE approach, achieving in planar coalescence of ELO InAs on (001) GaAs substrates and resulting in a ~1.3x improvement in photoluminescence (PL) and ~2-3x reduction of threading defect densities (TDD) relative to an equivalent InAs/GaAs control. Samples were grown by solid-source molecular beam epitaxy in an EPI Mod. Gen II system on n-doped (001) GaAs substrates. After an initial growth of 300nm GaAs, a 500nm InAs metamorphic buffer layer was relaxed using the interfacial misfit array (IMF) technique. Then, a 25nm thin silica dielectric mask was deposited and patterned in the form of gratings at a 50% duty cycle, with pitches ranging between 1.4-2.2µm. Using MBE growth under highly selective growth conditions, InAs was grown until planar coalescence was achieved, resulting in a total growth over the patterned dielectric of ~2.25µm. For comparison, an InAs/GaAs control was grown using the identical growth layer thicknesses. Under the best grown conditions, PL of the ELO-InAs showed a ~1.3x improvement in relative PL intensity compared to control. Additionally, surface TDD was estimated using electron channeling contrast imaging (ECCI), which showed a ~2-3x reduction in dislocation density in ELO-InAs as compared to the control. This work was supported by the National Science Two-dimensional (2D) materials have generated a lot of interest in the scientific community due to their high tunability via surface sensitivity, strain, thickness, and exposure to external electric fields. Since the discovery of monolayer graphene, there have been scientific breakthroughs in other classes of 2D materials for electronic and opto-electronic applications such as monolayer transition metal dichalcogenides (e.g. MoS 2 ). Therefore, monolayer and few-layer magnets provide a route to achieve highly tunable magnetic properties, which is desirable for low-power magnetoelectronic devices for memory and logic applications. Recently, there has been tremendous interest in developing 2D magnetic van der Waals crystals with some recent pioneering efforts that have achieved intrinsic ferromagnetism in the monolayer limit (e.g. exfoliated crystals of CrI 3 and CrGeTe 3 ). However, the magnetic ordering in these reports has been limited to low temperatures (below 60 K) and realizing the full potential of monolayer magnets for applications will require ferromagnetic ordering up to room temperature. In this work, we report room temperature ferromagnetism in stabilized layers of MnSe 2 grown by molecular beam epitaxy (MBE) on GaSe/GaAs(111) surfaces. Structural and magnetic properties are confirmed through scanning tunneling microscopy (STM) images and superconducting quantum interference device (SQUID) magnetometry measurements. The demonstration of ferromagnetic ordering at room temperature in a large-area 2D magnet holds a great promise for 2D spintronics. In the past few decades, magnetic metal thin films have played the indispensable role for the development of spintronic devices and as experimental platforms for the fundamental study of two-dimensional magnetism. However, at least two drawbacks exist in the epitaxially grown thin films when they are thinned down to a few atomic layers. Firstly, the uniform thickness and single-crystallinity typically occur in crystallites of tens of nanometers size. Thus, any measurement with a lateral spatial resolution larger than this length scale acquires smeared signals averagely of differing crystallites, including edges and grain boundaries as well. Secondly, structural and electronic properties of thin films are drastically altered by substrates, on which epitaxy of thin films occurs. Therefore, intrinsic magnetic properties of 2D materials purely arising from quantum confinement effect of their 3D counterparts are hardly approachable in these conventional experimental platforms. We recently discovered the first 2D magnetic van der Waals crystal Cr 2 Ge 2 Te 6 which constitutes an unprecedentedly ideal material platform that hosts intrinsic 2D ferromagnetism. The close-to-ideal 2D Heisenberg ferromagnet allows us to access a critical physical regime, in which intrinsic magnetic anisotropy is so tiny that a small external magnetic field can substantially modify the spin wave excitation spectrum and effectively control Curie temperatures. Furthermore, the observed prominent dimensionality effect highlights the significant interlayer magnetic coupling despite of the van der Waals spacing. Also, I may present our advance in pushing the boundary of Curie temperatures to make a 2D magnetic van der Waals crystal suitable for room temperature applications. Meanwhile, relevant work on spin-valley and spin-phonon coupling in magnetic van der Waals crystals and heterostructures may be discussed. Our work highlights 2D magnetic van der Waals crystals as a class of emergent 2D magnetic materials with unprecedented physical properties and materials/devices opportunities. 9:00 AM OO03 Electronic Structure of Layered In 2 Se 3 and (In,Bi) 2 Se 3 Alloys Anderson Janotti, Wei Li, Felipe D. de Lima and Fernando P. Sabino; Materials Science and Engineering, University of Delaware, Newark, Delaware, United States.
Layered indium selenide, with formula unit In 2 Se 3 , is a semiconductor and can found in different crystal structures with quintuple Se-In-Se-In-Se layers that are bonded through van der Waals interactions. It is a promising material for a series of high technological applications such as for phasechange memory, thermoelectrics, and photodetectors. It has been reported that few layers of In 2 Se 3 have strong photoconductive response into ultraviolet, visible, and near-infrared spectral regions. The layered β-In 2 Se 3 phase share the same crystal structure as the topological insulator Bi 2 Se 3 , and has been mixed in (In,Bi) 2 Se 3 alloys for band gap engineering. However, the electronic structure of the alloys and the parent compound In 2 Se 3 are still quite controversial, possibly due to the presence of competing phases of In 2 Se 3 . Using density functional theory with the Heyd-Scuseria-Ernzerhof hybrid functional (HSE), we investigate the electronic and optical properties of In 2 Se 3 and (In,Bi) 2 Se 3 alloys bulk, and the evolution of their electronic structure with the number of layers. We also discuss the position of the band edges with respect to other semiconductors, and compare our results to available experimental data and previous calculations. Engineering exotic topological states motivates extensive contemporary research in topological materials, including research of magnetic topological insulators as a materials platform for the study of interactions between topological surface states and symmetry-breaking magnetic ordering. We report the results of a study that identifies several topologically nontrivial phases in Bi 2 MnSe 4 that depend on the magnetic ordering and on the number of layers in a slab configuration. In stoichiometric bulk, the material is predicted to be an antiferromagnetically ordered strong (Z2) topological insulator, which transforms into a magnetic Weyl semimetal when the spins are aligned ferromagnetically. In a magnetic Weyl semimetal, 2D slices of k-space between the pair of Weyl points have a non-trivial Chern number, calculated from the integral of the Berry curvature over a closed manifold in the Brillouin zone, leading the material to display anomalous Hall conductivity proportional to the separation between the Weyl points, as well as unusual features like surface Fermi arcs. When the material is reduced to 1-2 layers it is topologically trivial, while increasing the thickness above three layers is predicted to result in a band inversion and a 2D Chern insulating phase, which displays quantized anomalous Hall conductivity. Furthermore, the magnetic ordering and topological behavior can be modulated by, for example, adjusting the separation between Bi 2 MnSe 4 layers by layers of Bi 2 Se 3 .Bi 2 MnSe 4 is formed by the intergrowth of {111} planes of rock-salt MnSe with quintuple layers of Bi 2 Se 3 , and has been previously demonstrated to grow in a self-assembled multilayer heterostructure with layers of Bi 2 Se 3 when grown by molecular beam epitaxy (MBE), where the number of Bi 2 Se 3 layers separating the single Bi 2 MnSe 4 layers is approximately defined by the relative arrival rate of Mn ions to Bi and Se ions during growth [1]. The compositional, structural, and electronic properties of these Bi 2 MnSe 4 /Bi 2 Se 3 selfassembled heterostructures are presented. We support a model for the epitaxial growth of Bi 2 MnSe 4 in a near-periodic self-assembled layered heterostructure with Bi 2 Se 3 with corresponding theoretical calculations of the energetics of this material and those of similar compositions. Computationally derived electronic structure of these heterostructures demonstrates the existence of topologically nontrivial edge states at sufficient thickness. We report optical and electronic properties of this material with different concentrations of Mn, and therefore different average separations between Bi 2 MnSe 4 layers, to explore the physics of this topologically non-trivial material.J. A. Hagmann et al., New J. Phys. 19, 085002 (2017)  Phase transitions along with electronic property changes in twodimensional (2D) materials demonstrate potentials in novel devices for beyond-CMOS applications. One interesting 2D material is 1T-Tantalum Sulfide (1T-TaS 2 ), which exhibits a series of charge density waves (CDW) and periodic lattice distortions. One of its phase transitions (known as metal-insulator transition, MIT) is from nearly commensurate CDW (NCCDW) phase to commensurate CDW (CCDW) phase with about 10-100x resistivity increase. Harnessing the physics of electron correlation and electronic phase transitions in this material could potentially help enable field-effect devices with performances beyond the limitations of Silicon. We present our work on directly synthesizing 1T-TaS 2 flakes on different substrates (SiO 2 /Si, sapphire, graphene etc.) by powder vapor deposition. With this route, we are able to directly introduce exotic dopants into the material and tune its phase transition properties. The dopants have been evidenced by a complementary characterization tools including XPS, TEM etc. We studied the effects of dopants on 1T-TaS 2 phase transitions by both vibrational spectroscopy (Raman) and transport measurements. We found that at an optimized amount of dopant level, we could stabilize 1T-TaS 2 CCDW phase towards much higher temperatures (100°C higher than the reported transition temperature). The characteristic peak associate with its CCDW phase has been observed by Raman technique at room temperature. This dopant's stabilization effect has been further supported by our DFT calculations. The two-dimensional materials toolbox continues to expand and has recently grown to include two-dimensional (2D) nitrides, heralded by the discovery of graphene-stabilized 2D-gallium nitride (2D-GaN) 1 . This novel phase of 2D-GaN is composed of bilayer Ga bonded to N in a buckled R3m structure (with stoichiometry Ga 2 N 3 ; direct E g = 5 eV), and it is formed at the epitaxial graphene (EG)/SiC interface via a two-step Ga intercalation and nitridation process. In addition to the synthesis of 2D-GaN, we have also successfully synthesized inherently stable, graphenecapped 2D-gallium and indium layers in a similar fashion. We have optimized the large-area, uniform intercalation of atomically thin gallium and indium films via a simple thermal vaporization process. We have identified the formation of crystalline 2-3 atom-thick layers via crosssectional transmission electron microscopy (TEM), electron energy loss spectroscopy (EELS) and energy dispersive x-ray spectroscopy (EDS) mapping, among other techniques.The epitaxial graphene/SiC interface provides researchers with a unique and robust platform for the intercalation of various light and heavy elements to form confined, crystalline layers at the two-dimensional limit. In addition to the natural 2D-confinement of these intercalated structures, they are also protected from the environment and oxidation due to the pre-existing graphene overlayers. Thus, this system is ripe for the investigation of low-temperature electronic transport, phase transitions, magnetoresistance, and other physical phenomena in these pristine minimal-disorder systems. With all this in mind, we have carried out preliminary temperature-dependent resistivity measurements on graphene-encapsulated 2D-Ga (2-3L) heterostructures and observe a steep "partial" superconducting drop (R ≠ 0Ω) at a transition temperature ~ Tc = 4.2 K, roughly 4X higher than that of bulk α-Ga (Tc = 1.08 K) 2 . At constant temperature T = 2 K, we also observe a sharp increase in resistance with increasing magnetic field, further proof of a superconducting transition in 2D-Ga. While superconductivity has been observed recently for hexagonally-arranged bilayer-Ga films grown on GaN(0001) substrates by MBE, the interfacial influence of the highly polar GaN substrate is unknown 2 . Investigations into similarly thick Ga layers in different environments could help shed light on the interface effect on superconductivity in ordered 2D-metal films and heterostructures.Ongoing work is focused on the continued electrical characterization of this graphene/2D-Ga/SiC system including the elimination of parasitic resistances from our measurements (i.e. graphene/2D-Ga vdW gap) through the direct-contact of the underlying Ga layer and device fabrication on single-domain 2D-Ga/SiC terraces. Ultimately, we strive to determine the impact of Ga layer thickness (1, 2, 3L) on the superconducting critical temperature and fields. Future work is aimed at detecting the superconducting phase in other intercalated 2Dmetal (bilayer indium) and 2D-nitride (InN, GaN) films. Lastly, in addition to superconductivity in 2D-metals, we are actively pursuing electrolytegated transistor measurements of 2D-GaN films in order to assess its semiconducting properties in both lateral and vertical device geometries. We study a new material system formed when n and p type twodimensional conduction channels are separated by a thin dielectric barrier allowing strong electron-hole Coulomb correlation and formation of indirect excitons. In the system formed by two separately gated monolayers of MoS 2 separated by a few monolayers of BN the gate voltages can allow to have n and p channels with the effective e-h Bohr radius on the order of a few nanometers. The band structure parameters are obtained using density functional theory. In the case of two monolayers of MoS 2 separated by the three monolayers of BN the variational values of the radius and binding energy of the indirect exciton are 3 nm and 75 meV, respectively. In order to include strong electron-hole correlation effects we use real-time Green's functions evaluated in the self-consistent T-matrix approximation. The screening of the Coulomb interaction by the carriers in both monolayers was included in the plasmon pole approximation of RPA. The chemical potential was evaluated as a function of carrier density for different temperatures. As the density of the electron-hole plasma increases the indirect exciton states are formed and at lower temperatures the system of carriers consist mostly of bound pairs. At higher densities the Mott-type transition occurs as the screening and fermion statistics turn the system into electron-hole plasma. We defined the degree of the ionization as a relative density of free carriers evaluated in the assumption of the thermal equilibrium of excitons and the electron-hole plasma. At lower temperatures the degree of ionization can be as low as 10 -7 , and the system is the excitonic insulator. At the densities higher than the Mott density the system changes into conducting electron-hole plasma. The results can be useful for the design of novel electronic devices based on heterojunctions of the few-monolayer materials. Glass possesses a number of properties, such as transparency, high resistivity, ability to be worked into many shapes, and potential for tuning of chemical properties through doping, which make it a substrate of choice for optoelectronic applications. In recent years, ultra-thin glass has also been developed, allowing glass to serve as a substrate for flexible optoelectronics as well, particularly for flexible displays. Incorporating next-generation semiconductors such as two-dimensional transition metal dichalcogenides (TMDs) onto glass is a necessary step towards realizing such optoelectronic devices. This work demonstrates such deposition of MoS 2 and WSe 2 at low temperatures on glass substrates.Two-dimensional materials are inherently flexible and transparent due to their incredibly small dimensions. Some TMDs, including such materials as MoS 2 and WSe 2 , also possess direct bandgaps when synthesized as a single monolayer; this coupled with superior mobility and thermal stability as compared to organic semiconductors make TMDs the best option for glass-based optoelectronics.While glass is an ideal substrate in many ways, one obstacle to realizing glass-based optoelectronics is its tendency to soften at temperatures typically used for deposition of semiconductors. Thus, this work focuses on exploring the deposition of MoS 2 and WSe 2 on glass substrates at temperatures below the softening point of flexible glasses (Corning's Willow and Eagle XG glass); more specifically, below 600 o C. In particular, metalorganic chemical vapor deposition (MOCVD) is explored as a scalable, potentially high-throughput technique that deposits electronic-grade materials with a pristine interface, thus preserving the phenomenal optoelectronic properties of the TMDs.Beyond depositing device-quality TMDs on glass, this work also explores the effect that tuning glass surface chemistry has on deposition, and incorporates a variety of surface-chemistry-based studies investigating plasma treatment, chemical treatment, surface coatings, and different glass compositions.

11:20 AM OO09
Interlayer Charge Transport in 2D Molecular Structures Elad Koren; Materials Science and Engineering, Technion -Israel Institute of Technology, Haifa, Israel.
Weak interlayer coupling in 2-dimensional layered materials such as graphite gives rise to rich mechanical and electronic properties in particular in the case where the two atomic lattices at the interface are rotated with respect to one another. A lack of crystal symmetry leads to anti-correlations and cancellations of the p z orbital interactions across the twisted interface, which gives rise to low friction behavior and low interlayer electrical transport. Using our recent nanomanipulation technology 1 , based on atomic force microscopy, we studied the interlayer electrical conductivity as a function of twist angle between two misoriented graphene layers with unprecedented angular resolution of ~ 0.1 deg. The angular dependence indicates that the electrical transport across the interface is dominated by a phonon assisted channel which conserve the momentum of conduction band electrons, tunnelling across the twisted Dirac bands. Most intriguingly, the conduction is significantly enhanced within a narrow angular range of less than 0.5 deg at pseudocommensurate angles of 21.8 deg and 38.2 deg. This provides the first experimental evidence for the existence of a 2-dimensional interface state originating from the coherent coupling of electronic states in the twisted sheets due to commensurate superlattices 2 . Finally, we show that combined electro-mechanical characterization techniques of mesoscopic graphite structures can be uniquely address open fundamental question related to the dielectric interlayer interactions and electronic charge transport through stacking faulted structures 3 .ReferencesE. Koren et al., Science, 6235 (2015) 679.E. Koren et al., Nature Nanotech., 9 (2016)  Two-dimensional (2D) layered semiconductors have attracted much interest because of their unique crystallographic structure and potential applications. For example, GaSe and InSe are promising for both optical and electronic applications [1,2]. The layered crystals combine thin sheets by out-of-plane van der Waals interactions. For practical applications that use 2D layered semiconductors, the mechanical strength needs to be high. For example, device fabrication processes, such as dicing and wire bonding, require the crystal to be mechanically strong. However, no direct experimental determination has been carried out up to now for van der Waals bonding energy. Therefore, quantitative and direct measurements of the bonding energy are urgently required in order to obtain crucial understandings of 2D semiconductor materials. In this study, a tensile testing machine was constructed for the first quantitative determination of the interlayer van der Waals bonding force, where two stainless steel L-shaped sliders were combined, with a strain gauge between them.As a result, the interplanar binding strength of layered GaSe 1-x Te x and InSe crystals were directly determined using our tensile testing machine, respectively. These crystals were grown by a low and fixed growth temperature liquid phase method under a controlled Se vapor pressure [3]. The stoichiometry-controlled GaSe 1-x Te x crystal has the ε-polytype structure of GaSe, where the Te atoms are occupied with some Se sublattice in the GaSe crystal. InSe crystal has γ-polytype structure. The effect of adding Te on the bonding strength with respect to GaSe layers was determined from direct measurements of the Van der Waals bonding energy [4,5]. The interlayer bonding strength of GaSe 1-x Te x (x=0.106) was about 7 times larger than that of GaSe. The bonding strength was discussed for GaSe and InSe, respectively. Semi metallic rare earth-V (RE-V) nanoparticles (NPs) epitaxially embedded in lattice matched III-V semiconductor materials attracted considerable research in recent years for their potential applications in thermoelectric, plasmonic, terahertz, and tunneling devices. Hybrid nanostructures are known to elicit an enhanced optical response due to plasmonic coupling effects. It was predicted that InAs quantum dots (QDs) coupled to ErAs NPs in GaAs matrix would get the emission and absorption significantly enhanced by plasmonic energy transfer. To realize the coupling, the frequency match and maximized plasmonic field overlap are both critical. Recently, progress has been made on QD-NP self-alignment during growth where an inverted InAs QD can be located in the divot formed by GaAs overgrowth of an ErAs NP. However, it is highly desired to get a NP grown directly on top of a QD in order to independently optimize the QD and NP growth. In this work, we use a novel way to direct the alignment of ErAs metal NP and InAs QD using molecular beam eptaxy (MBE) in a GaAs matrix. Strain driven nucleation mechanism allows lightly Er doped InAs QDs self-aligned to underlying InAs QDs. Annealing uncapped Er:InAs QDs at high temperature desorbs the indium atoms in QD. However, due to much lower volatility, Er atoms condense out and nucleate as ErAs NPs that align with the underlying InAs QDs. It was found that the amount of Er incorporated into InAs QD depends on the solubility of Er in InAs, and the Er inhomogeneous distribution in QD and non-uniform indium desorption from QD surface affects NP-QD alignment, and density of ErAs NP is lower than that of InAs QD due to coarsening effects in both InAs QDs and ErAs NPs. The defects are attributed to the rock-salt structure formed by Er atoms taking interstitial sites in InAs lattice and can be eliminated by thorough indium evaporation. With further optimization, the measure opens a new path in precise control of the position of NPs or any nanostructures that starts with formation of NPs in specified locations. The production of ErAs powders via nanosecond pulsed laser ablation in an inert environment allows for the growth of nanoparticles of a desired morphology and quantity for future use in thermoelectric film growth. Targets consisting of pressed Er and As elemental powders demonstrate incongruent ablation due to differences in vaporization enthalpy of the two elemental species and mechanical removal of target material. SEM and XRD are used to monitor the elemental and crystalline composition of the powder during growth, showing an As:Er ratios that initially reach at least 22:1 within the first 10 minutes of growth, before leveling off to approximately 2:1; suggesting self-limiting vaporization characteristics. The influence of target composition and collection distance are investigated as a means with which to control powder composition and crystallite size. Vapor interaction is controlled by inert gas pressure and laser spot size, and is shown to have a significant impact on the overall ErAs composition of the resulting powders; such that increasing vapor interaction increases overall ErAs composition. These growth conditions are then used to demonstrate improved control over nanoparticle morphology and composition. Epitaxial integration of rare-earth monopnictides with III-V semiconductors has been of tremendous interest over the past few decades due to their numerous potential applications including buried metallic contacts in semiconductors, infra-red and terahertz optoelectronic devices, thermoelectrics and solar cells. In recent years, these compounds have also been shown to exhibit remarkably large magnetoresistive behavior, with a pressure-temperature phase diagram that bears resemblance to other extreme magnetoresistance materials indicating a similarity in the underlying physics responsible for such a behavior in disparate material systems.Possessing a simple rock-salt structure with a lattice constant that is closely matched to a variety of III-V semiconductors, rare-earth monopnictides offers an exciting materials platform to both understand and tune their remarkable magnetoresistive properties. To that end, we have successfully synthesized (001) LuSb and LuBi thin films on (001) GaSb substrates that show large magnetoresistance (1.8×10 3 and 1.1×10 2 at 14 Tesla, 2K for LuBi and LuSb, respectively). Our films are smooth, single-phase, epitaxial and shows 1×1 reconstruction in Reflection High Energy Electron Diffraction (RHEED).Using a combination of magnetotransport studies and ab-initio calculations we establish that a perfect electron-hole compensation in these semi-metallic films is the likely mechanism that results in their remarkable magnetoresistance behavior. Additionally, we have also synthesized epitaxially strained pseudomorphic thin films on lattice mismatched substrates that allows us to break the cubic symmetry of the rock-salt structure. By doing so, we can control the relative occupation of the electron and hole bands that brings about a dramatic change in their magnetoresistive properties. I will show how our ability to controllably tune the electronic structure via strain engineering helps us gain important insights into their emergent magnetoresistance behavior.Furthermore, our ability to synthesize these rare-earth monopnictides in a thin film form also allows us to dimensionally confine these electronic states in the out-of-plane direction that results in the formation of quantum well states with quasi 2-dimensional transport properties.Recently, it has been proposed that it is possible to bring about a band-inversion between the rare-earth d and pnictogen p states that will make these materials topologically non-trivial with strongly spin-orbit coupled surface states. We have observed strong weak anti-localization in transport that can be fitted well with expectations from Hikami-Larkin-Nagaoka theory, which has also been seen in other topologically nontrivial thin films such as Bi 2 Se 3 . We will discuss our experimental results considering the possibility of topologically non-trivial surface states in LuSb and LuBi thin films.

9:20 AM PP04
An Atomistic Approach to Studying Phonon Scattering of Embedded Nanoparticles Joseph P. Feser and Rohit R. Kakodkar; Mechanical Engineering, University of Delaware, Newark, Delaware, United States.
Nanoparticle-in-alloy material systems are promising candidates for high efficiency thermoelectric materials, due to their greatly reduced lattice contribution to thermal conductivity. In this talk, we use a recently developed frequency-domain perfectly matched layer computational technique to calculate the scattering cross sections of embedded nanoparticles across the entire Brillouin zone and for all phonon modes including transverse acoustic phonons and optical phonons for the first time. For acoustic modes, we compare the computational results against previously used results from continuum mechanics and find excellent agreement so long as the Mie regime is accurately represented within the continuum framework. Interestingly, we find that the interaction of optical phonons is remarkably different compared to its acoustic counterparts, with scattering efficiencies of optical phonons in the "Raleigh" regime scaling independent of vibration frequency and with orders of magnitude higher scatttering efficiency. We are that optical phonons should be viewed in the context of zone folding to be a folded continuation of very short wavelength acoustic phonon spectrum. Furthermore, we directly show that an interdiffused nanoparticle/matrix is more effective at scattering phonons compared to solids nanoparticle with the same net impurity concentration, with scattering efficiencies 2-fold higher in the dominant heat carrying regions.
Due to their significant bandgap narrowing, dilute nitride alloys are promising for concentrating photovoltaics expected to power the next generation. However, N-related point defects often lead to degraded minority carrier transport properties and optical efficiencies. Co-alloying of GaAsN with larger elements, such as indium, antimony, and/or bismuth, allows lattice-matching to GaAs or Ge substrates. Of particular interest is GaAsNBi, which is expected to provide the largest bandgap reductions, with tunable valence band and spin-orbit energy level splittings, which are also promising for temperature insensitive lasers and high selectivity spin valves. In the literature, the magic ratio for lattice matching of GaAsNBi with GaAs is predicted to be x N /y Bi = 0.59, based upon a computed value of the GaBi lattice parameter. In addition, the relationship between the compositions and the bandgap values are most often determined using x-ray rocking curve (XRC) measurements of strain to determine the alloy compositions, assuming the computed value of the GaBi lattice parameter, with films are fully strained to GaAs substrates. Here, we use a combination of direct measurements of alloy compositions, via ion beam analyses of the N and Bi compositions, in conjunction with direct measurements of the out-of-plane misfit via XRC measurements of "x-ray strain", and measurements of bandgap using photoreflectance spectroscopy, to determine a new magic ratio for lattice-matching and a new map of the quaternary bandgaps. The implications of these findings on future device design will be discussed. GaAs 1-x Bi x grown on GaAs and InP substrates is an attractive material for small band gap semiconductor and semimetal applications. Due to the valence band anti crossing effect (VBAC), the band gap of GaAs 1-x Bi x is reduced up to 84 meV/%Bi [1], allowing for new band gap lattice constant combinations that were previously unavailable through epitaxial growth. Most work on GaAsBi has looked at fully pseudomorphic layers grown on GaAs substrates and have reached Bi incorporation up to 22% (GaAs 0.78 Bi 0.22 ) [2]. Although high incorporation percentages have been achieved in GaAsBi, the resulting small band gap materials have a long way to go to be implemented in high quality, small band gap devices. Previous work on high Bi content GaAs 1-x Bi x has relied on samples grown thin to prevent strain relaxation and were plagued by surface droplets and defects that prevented high quality, thick films from being created. In this work, we explored strained and relaxed interlayers between the substrate and the epitaxial film to study the effects on the crystalline quality and bismuth incorporation in thick films of GaAsBi (>250nm). We propose that strained under-layers may affect the incorporation of bismuth atoms in GaAs by providing a way to alleviate the local strain induced by the large atomic size difference between the Bi and As atoms.We grew our samples on a Veeco GENxplor MBE using a valved As 4 source and an effusion cell for our bismuth source. To determine the bismuth content in our films, we used 004 and 224 HRXRD scans to analyze the lattice constant and spectroscopic ellipsometry to determine the band gap. We fit these parameters to the predicted band gap lattice constant curves determined experimentally [1,3]. We characterized the degree of strain relaxation by examining reciprocal space maps around the 224 asymmetric reflection. We additionally used TEM to identify defect centers and through-film compositional variation. We examined structures of GaAs (1-x) Bi x /InGaAs/ GaAs with InGaAs layers of varying thickness and indium content. This allowed us to look at the effects of both tensile and compressive strain on our bismide layers. Lessons learned from these techniques can be translated to higher Bi content materials in order to achieve smaller band gap semiconductors with reduced defect densities on GaAs and InP substrates.  Recently, we demonstrated a method to grow tensile-strained Ge nanowires (NWs) embedded in an In 0.52 Al 0.48 As (hereafter InAlAs) matrix using surface-mediated phase separation; the lattice constant of InAlAs is 3.7% higher than Ge. [4] Here, we demonstrate that tensile-strained Ge quantum dots (QDs) can be grown through a similar growth mode. Comparing tensile Ge QDs with NWs grown under similar conditions, we find that Ge QDs are larger and more anisotropic in shape than NWs, while exhibiting a similar Raman shift and brighter room-temperature photoluminescence (PL). We conducted growth of tensile-strained Ge nanostructures using a III-V molecular beam epitaxy (MBE) system equipped with a Ge effusion cell. All growths started with a latticematched InAlAs buffer grown on an InP (001) substrate. A 300 nm nanocomposite layer consisting of either Ge QDs or NWs in InAlAs was then grown by co-deposition of Ge, In, Al and As 2 . For NW growth, all shutters were opened simultaneously, while for QD growth the Ge shutter was periodically opened and closed to form QD superlattices; the Ge content of the nanocomposite layers ranged from 1.4-10%. High-angle annular dark field scanning-transmission electron microscopy and energy dispersive x-ray spectroscopy (EDX) revealed that Ge nanostructures had phase segregated in both the NW and QD nanocomposites. Plan-view electron channeling contrast imaging (PV-ECCI) showed a 10× difference in the planar density of NWs and QDs at 6×10 10 cm -2 and 6×10 9 cm -2 , respectively. The wide disparity in density can be partly explained by the fact that NWs tend to branch throughout growth, effectively multiplying their density. In contrast, QD growth is interrupted every 10 nm, which prevents such branching from occurring. For similar growth conditions, QDs showed a strongly anisotropic shape lengthened along the direction, while NWs were more isotropic. Thus, QDs form lower density and more anisotropic nanostructures with a larger in-plane size compared to the NWs. Raman spectroscopy showed that both Ge QDs and NWs exhibit an identical Ge-Ge peak at 284.1cm -1 corresponding to a shift of 16.0 cm -1 from the bulk Ge-Ge phonon mode at 300.1 cm -1 . The shift in the Ge-Ge mode for both types of nanostructures corresponds to a 3.6% biaxial tensile strain, which is close to the 3.7% lattice mismatch between Ge and InAlAs. The Raman signal for QDs was found to be highly polarized along the direction, which is consistent with the anisotropic shape observed in PV-ECCI. In contrast, the Raman signal from the more isotropic NWs did not show such strong polarization effects.QD and NW samples grown under similar conditions both exhibited room-temperature PL, and the integrated intensity from the QDs was ~4× stronger than from the NWs. Along with the higher PL intensity, QDs emit at 1176 nm, which is slightly blue-shifted compared to the 1230 nm emission of NWs. By altering the growth rate and hence the size of the nanostructures, we could tune the peak emission wavelength over 100 nm. Taken together, this work expands the range of strained Ge nanostructures that can be formed by surface-mediated phase separation while providing new insight on their basic growth mode.
Three-dimensional (3D) nanostructures like, InAs quantum dots grown on GaAs(001) are one of the most extensively explored semiconductor systems. In contrast, attempts to grow 3D structures on other GaAs surfaces remained without success long time: while on GaAs(001) 3D growth of InAs is preferred, on other low-index GaAs surfaces such as (110) the InAs deposition always results in a two-dimensional (2D) growth and the misfit relaxes plastically. On the other hand, {110} facets often form the sidewalls in self-assembled GaAs nanowires. The growth of 3D nanostructures like quantum dots on such sidewall surfaces is of interest, especially for high efficiency single photon sources [1]. Recent investigations show that the presence of Bismuth (Bi) as a surfactant induces 3D growth on GaAs(110) by reducing the surface energy. Furthermore, Bi exposure on already grown 2D InAs layers can cause a morphological phase transition, resulting in a rapid re-organization of the 2D layer into 3D nanostructures. These so-called 3D islands have optical properties of quantum dots and open the possibility to generate linearly polarized single photons [2].In this contribution Bi induced InAs 3D islands formed within InAs monolayers on GaAs(110) are investigated structurally, using cross-sectional scanning tunneling microscopy (XSTM), for the first time. Sample growth was performed using molecular beam epitaxy. The investigated sample contains four different sets of growth parameters, namely two pairs of 1.1 monolayers deposited with and without Bi and two pairs of 2.1 monolayers subsequently exposed to Bi flux for different durations.XSTM is a powerful tool for investigation with atomic resolution in order to determine the growth mechanisms and the even more possible changes during capping, the latter been necessary for device applications. For this purpose, we cleaved the samples in ultrahigh vacuum perpendicularly to the growth [110] direction. The XSTM images with atomic resolution of the InAs layers show the formation of quantum dots. The XSTM images allow the characterization of geometrical structures in terms of size, density, and atomic composition, all depending on the presence of Bi. Furthermore, we are able to carry out stoichiometric analyzes of the chemical composition by analyzing the variation of local lattice parameter [3]. In order to explore the influence of Bi, we compare the resulting structures upon both growth regimes, mentioned above. Hence, we are able to present a correlation between the amount of Bi in the layers and the 3D island formation.This work was supported by the DFG, SFB 787, Project A4.References:[1] P. Corfdir et al. Phys. Rev. B 96, 045435 (2017)[2] R. B. Lewis et al. Nano Lett,17,4255-4260 (2017)[3] A. Lenz et al. J. Vac. Sci. Technol. B 29, 04D104 (2011).

Influence of QD Morphology on the Electronic States of GaSb/GaAs
Multilayers Christian Greenhill 1 , Eric Zech 1 , Alexander Chang 2, 1 , S. Clark 3 , Sadhvikas J. Addamane 3 , Ganesh Balakrishnan 3 and Rachel S. Goldman 1 ; 1 Materials Science and Engineering, University of Michigan, Ann Arbor, Michigan, United States; 2 Northwestern University, Evansville, Illinois, United States; 3 University of New Mexico, Albuquerque, New Mexico, United States.
Due to the predicted composition and strain dependence of type I versus type II band offsets, GaSb/GaAs quantum dots (QDs) have been identified as promising for a variety of optoelectronic applications. It has been shown that the nucleation of 2D layers (i.e. wetting layers) versus 3D islands (i.e. quantum dots) can be tuned by varying the Sb/Ga beam equivalent pressure ratio during molecular beam epitaxy. For GaSb/GaAs multilayers, atomic structures ranging from dots to rings to clusters have been observed. Although photoluminescence (PL) emissions ranging from 0.9 to 1.3eV are often reported for multi-layered GaSb/GaAs, the association of these emissions with specific nanostructures remains elusive. Here, we examine the structural and optical properties of GaSb/ GaAs multi-layers containing 2D GaSb layers, with and without 3D islands. We use a combination of cross-sectional scanning tunneling microscopy and scanning transmission electron microscopy, in conjunction with atom-probe tomography (APT), to determine the compositions and dimensions of the GaSb wetting layers (WLs), quantum dots (QDs), and QD rings. For both samples, ~2nm thick WLs are observed. Since the Sb fractions, x Sb , in the WLs are ≤ 2%, we attribute the 1.32eV emissions to the GaAsSb alloy. For samples containing 3D islands, QDs and rings of smaller QDs, with typical diameter of 25nm and height of 7nm, are observed. In both cases, Sb-rich QD cores are apparent, with x Sb up to 0.40 and 0.25 for the individual QDs and QDs within rings, respectively. Therefore, we attribute the 1.08eV and 1.2eV emissions to the QDs and rings of QDs, respectively. Local measurements of the electronic states using scanning tunneling spectroscopy will also be presented. We recently demonstrated the successful growth of Al x In 1-x As y Sb 1-y digital alloys on GaSb with period thicknesses between 10 and 20 monolayers (ML) and Al fractions ranging from 0 to 80% [1] and employed them to realize the first working staircase avalanche photodetectors [2]. Additionally, we have successfully used the digital alloy approach on materials lattice matched to InP and GaAs, such as InGaAs, InAlAs, and AlGaAs. Importantly, digital alloys exhibit an extended cutoff wavelength, as compared to their random alloy counterparts, enabling for a broader range of accessible cutoff wavelengths. This is extremely useful for potentially translating III-V mid-infrared photonics from GaSb and InAs substrates to the more mature InP platform, which would enable midinfrared photonic systems that leverage the established photonic integrated circuit infrastructure.Here, we study the effects of digital alloying on the optical, structural, and device properties of various III-V digital alloys, focusing on extending the cutoff wavelength in mid-infrared devices. All digital alloys were grown on either GaSb, semi-insulating (SI) InP, or SI GaAs substrates by solid-source molecular beam epitaxy. In each sample, the digital alloy was strain-balanced to the corresponding substrate by varying the relative thicknesses of the individual (often strained) binary layers. We have successfully grown high quality digital alloys on a number of substrates by varying several growth parameters, such as substrate temperature, surfactant usage, and group-V overpressure. On InP substrates, we have leveraged the digital alloying technique to extend the cutoff wavelength from approximately 900nm to 1100nm with an 8ML period digital alloy in InAlAs. Similarly, we observed an extension of the InGaAs cutoff wavelength from ~1700nm to greater than 1900nm with a 10ML period. Extension of the cutoff wavelength into the mid-infrared with further increase of the period thickness could enable lattice-matched mid-infrared detectors on InP substrates, with enhanced sensitivity compared to their metamorphic counterparts. The use of digital alloys creates the potential to design avalanche photodetectors that have superior noise and bandwidth characteristics to current materials as well as accessing a broader range of operating wavelengths. Further comparisons between random and digital alloys and device growths are ongoing and will be reported at the conference. This work was supported by the Army Research Office (W911NF-17-1-0065  Acceptor based qubits in Si have been proposed due to potential benefits from large spin-orbit coupling and the absence of the valley degeneracy. Additionally, Si/Al/Si heterostructures doped above saturation are predicted to superconducting at approximately 1 K, improving prospects of merging superconducting and semiconducting quantum information processing strategies. The realization of an acceptor-based 2D system is an important precursor to lower-dimensional devices such as nanowires, quantum dots, and single atom devices which can be used as qubits. We have grown a Si/Al/Si heterostructure by MBE and electrically determined it forms a two-dimensional (2D) hole gas complementary to the P doped 2D electron system in Si.Although several attempts to fabricate a dopantbased 2D hole system have been reported, these studies did not report electrical characterizations. Hall measurements at 4 K give the hole carrier concentrations 0.8×10 13 cm -2 from to 2×10 14 cm -2 , with a hole mobility in the range of 10 cm 2 /Vs to 20 cm 2 /Vs. At the highest measured carrier densities, we estimate 100% of dopants are activated based upon comparison of doping densities observed with scanning tunneling microscopy (STM). Hall effect measurements in a tilted B-field support the conclusion that the dopants are in fact confined to a two-dimensional space. Moreover, magnetoresistance measurements at 2 K show an anomalous behavior near zero field, which is similar to weak antilocalization.Some devices were cooled to temperatures to look for the superconducting transition, but evidence for superconductivity is not yet observed. We are continuing to optimize the MBE growth parameters to improve the carrier mobility and concentration towards realizing superconductivity. We will report on STM characterizations of material synthesis during different stages of the growth, and low temperatures electrical transport measurements on Hall bar and mesa-etched nanowire devices. Quantum information science (QIS) promises communication with unbreakable security and exponentially faster computing. Certain QIS protocols rely on entangled photons, which can be generated in semiconductor quantum dots (QDs) via biexciton decay. QD devices are hence promising candidates for compact, scalable entangled photon sources.

9:00 AM RR02 (Student) Self-Assembly of (111)-Oriented Quantum Dots for
[1] However, for robust entanglement, fine-structure splitting (FSS) between a QD's bright exciton states must be vanishingly small. [1] Until recently, forming suitable QDs by MBE required a trade-off. Traditional (100)-oriented QDs self-assemble in a single-step, but suffer from high FSS due to QD asymmetry. [2] In contrast, (111)-oriented QDs have low FSS, but rapid relaxation of compressive strain limits synthesis to droplet epitaxy, which is prone to defects, or complex processing and regrowth steps. [2,3]To overcome these issues, we have developed an alternative self-assembly method. [4] Tensile-strained QDs (TSQDs) form spontaneously on (111) surfaces in a single step, and bear the MBE hallmarks of high purity and precise structural control. [4] GaAs/ InAlAs(111)A TSQDs are highly symmetric with low FSS, and are defectfree and optically active.
[5] Additionally, tensile-strain reduces TSQD band gaps, which is of interest for infrared emission. Tensile self-assembly is consistent with Stranski-Krastanov growth and island scaling theory, providing a bridge from well-established (100) QDs to these new (111) TSQDs. [4]To build on this discovery, we grew several sample series to explore the MBE parameter space for GaAs TSQD self-assembly. We will show how to control TSQD characteristics with deposition amount, growth temperature, growth rate, and V/III flux ratio. We note clear correlations between QD growth, structure, and properties; e.g., QD volume and PL wavelength, and QD density and PL intensity.Given their novelty, TSQD research is still in its infancy. This comprehensive study enables us to optimize TSQD properties for entangled photon generation and will underpin future applications of these unique nanostructures. This presentation describes fabrication of a mid-IR light emitting diode (LED) from IV-VI semiconductor layers grown by molecular beam epitaxy (MBE) on silicon. Device fabrication involved eutectic bonding of epilayers to copper and removing the silicon growth substrate to allow low resistance electrical contacts and improved heat dissipation. This is the first known demonstration of a IV-VI semiconductor mid-IR light emitting device fabricated using a silicon growth substrate and epilayer transfer. LED devices were fabricated from IV-VI semiconductor layers grown by MBE on 3 inch diameter (111)-oriented silicon substrates. PbSe, Se, Sr, and Bi 2 Se 3 effusion cells were used to grow the device structure, which consisted of a 1 μm thick Pb 0.93 Sr 0.07 Se layer followed by a 3.5 μm thick PbSe/Pb 0.93 Sr 0.07 Se multiple quantum well (MQW) layer with 160 PbSe wells. The 7 nm thick PbSe wells were separated by 15 nm thick Pb 0.93 Sr 0.07 Se barriers. The first 2.75 microns of this 4.5 μm thick device structure were doped with bismuth to achieve n-type conductivity with an electron density in the range of 3 x 10 18 cm -3 . The last 1.75 microns were grown under a Se overpressure to achieve p-type conductivity with a hole density in the range of 4 x 10 17 cm -3 . Well and barrier thicknesses along with the absence of any significant bismuth diffusion were confirmed by ex situ secondary ion mass spectroscopic (SIMS) characterization. The IV-VI semiconductor device structure was grown on a 4 nm thick CaF 2 release layer, which was grown on the silicon substrate immediately following oxide desorption as confirmed by observing a 7x7 silicon surface reconstruction in the reflection high energy electron diffraction (RHEED) pattern. Following MBE growth, the wafer was cleaved into 1x1 cm 2 chips, which were flip-chip bonded to copper using InSn eutectic metallurgy. The silicon growth substrate was then removed by immersion in deionized water to dissolve the CaF 2 release layer. Wet etched square mesas, defined by gold evaporated through a shadow mask, were contacted with a probe tip for current versus voltage characterization and wire bonded using conductive epoxy for light emission measurements using a Fourier transform infrared (FTIR) spectrometer.A 500 x 500 μm 2 mesa device at room temperature exhibited a dark current density of 41 mA/cm 2 at a reverse bias voltage of -26 mV. Based on known doping levels, measured charge carrier mobilities (300 cm 2 /vs for both electrons and holes), and intrinsic carrier densities calculated from known effective masses, this dark current is consistent with minority carrier lifetimes of about 1 ns and diffusion lengths greater than 1 μm. A large portion of the MQW layer will thus be populated with injected non-equilibrium minority carriers, which can recombine radiatively, during forward bias operation. Steady state forward bias operation of a fabricated device at room temperature showed strong LED emission between 300 meV and 350 meV. The peak emission at 320 meV is blue shifted by 45 meV relative to the 275 meV band gap of PbSe, an energy difference consistent with the quantum size effect created by 7 nm thick PbSe wells. Strong LED emission was observed even though the mesa surface was completely metalized and light could only be emitted from the etched mesa sidewalls. Demonstration of a room temperature mid-IR LED with continuous wave (CW) room temperature emission at 3.9 microns represents an improvement over commercially available mid-IR LEDS, which can only be operated in quasi-CW mode (50% duty cycle). Further development of IV-VI materials and device fabrication methods offers the promise of a new class of low cost mid-IR light emitting devices for applications such as chemical sensing.
Utilizing self-assembled InAs nanostructures as the active material for semiconductor lasers have been drawing great attentions for their unique properties. Low threshold current density, high defect tolerance, and less temperature sensitivity have granted InAs nanostructures semiconductor lasers better performance over quantum well lasers. Great results have been obtained with the InAs/GaAs quantum dots operating at 1.3m. On the other hand, the InAs/InP material system targeting at 1.55 m for telecommunication is less developed due to the low lattice mismatch between InAs and InP (3.2%). Here, we demonstrate InAs quantum dash narrow ridge-waveguide lasers grown on n-InP native substrate with a low continuous-wave threshold current density of 510 A/cm 2 and an output power of 25 mW.InAs nanostructure photoluminescence (PL) samples were grown on InAlGaAs quaternary alloy buffer, lattice matched to InP substrate, to investigate the growth conditions that gave the highest PL intensity at around 1550 nm. It was determined that, for the InAs nanostructure growth, a growth temperature of 485 C, 3.25 monolayer (ML) of InAs, a growth rate of 0.4 ML/s, and a V/III ratio of 18 during the nanostructure growth result in an overall better PL characteristic. Further tuning of the PL spectrum depends on post-nucleation ripening process, where the structure was held under As 2 overpressure for a given amount of time. Due to the low lattice mismatch between InAs and InP lattice constant, the surface diffusion of In adatoms are highly anisotropic, with the fast diffusion path parallel to the surface reconstruction dimer row. As a result, the nucleated islands tend to elongate along the direction, known as quantum dashes (Qdash). By holding the structure at growth temperature, it approaches the thermodynamically more stable configuration, with the As 2 pressure adjusting the surface diffusivity by changing the surface reconstruction pattern. A combination of As 2 pressure of 1e-6 mTorr and 60 sec of ripening time result in the highest PL intensity.The full laser structure was then grown on n-InP (001) substrate using the optimized growth conditions for the active region. n-and p-type InAlAs, lattice matched to InP, was used as bottom and top cladding layers, respectively. InAlGaAs digital alloy with alternating lattice matched InAlAs/InGaAs layers were used for both separated confinement heterojunction layers (SCH) and grading layers between SCH and cladding layers. A 100-nm thick p-doped InGaAs layer was finally deposited as the p-contact layer. The epi structure was then fabricated into standard Febry-Perot (FP) ridge lasers. The threshold current density is as low as 510 A/cm 2 , which is much less than the previously reported values [1,2]. An output power of 25 mW is also obtained. Since the Qdashes are asymmetric, it is expected that the FP ridge laser will perform differently depending on whether the ridge is parallel or perpendicular to the Qdash elongation direction. Measurements have shown that lasers with ridge perpendicular to the Qdashes have lower threshold and higher output power. The difference is more prominent at wider ridges, where the side wall effect is of less importance. The laser results are very promising for the potential to transfer the high-quality laser structure onto Si substrate. The entanglement of the charge, spin and orbital degrees of freedom can give rise to emergent behavior especially in thin films, surfaces and interfaces. Often, materials that exhibit those properties require large spin orbit coupling. We hypothesize that the emergent behavior can also occur due to spin, electron and phonon interactions in widely studied simple materials such as Si. That is, large intrinsic spin-orbit coupling is not an essential requirement for emergent behavior. The central hypothesis is that when one of the specimen dimensions is of the same order (or smaller) as the spin diffusion length, then non-equilibrium spin accumulation due to spin injection or spin-Hall effect (SHE) will lead to emergent phase transformations in the non-ferromagnetic semiconductors. In this experimental work, we report spin mediated emergent antiferromagnetism and metal insulator transition in a Pd (1 nm)/Ni 81 Fe 19 (25 nm)/MgO (1 nm)/p-Si (~400 nm) thin film specimen. The spin-Hall effect in p-Si, observed through Rashba spin-orbit coupling mediated spin-Hall magnetoresistance behavior, is proposed to cause the spin accumulation and resulting emergent behavior. The phase transition is discovered from the diverging behavior in longitudinal third harmonic voltage, which is related to the thermal conductivity and heat capacity. With a crystal structure and lattice parameters similar to III-V compound semiconductors, the possibility of h-H/III-V heterostructures with unique properties is achievable. Recent work demonstrated a semiconductor to ferromagnetic metal transition in thin films of CoTiSb substitutionally alloyed with Fe, namely CoTi 1-x Fe x Sb and Co 1-y Fe y TiSb [4]. Here electrical and magnetic properties that depended strongly on the Fe concentration, as well as Fe site occupancy, were demonstrated. In addition, half-metallic behavior was predicted for films x≤0.5. The quaternary Heusler compound system's compatibility with other Heusler and III-V compounds and its expected half-metallic behavior, make it promising for the development of future spintronic heterostructures and devices. However, understanding how these distinct electronic and magnetic properties are linked to the interactions between metallic atoms at the nanometer scale will be important to utilizing these thin films in devices. In this work, atom probe tomography (APT) and scanning transmission electron microscopy (STEM) are employed to investigate the presence and distribution of nanometer scale Fe-rich domains and their dependence on Fe content in CoTi 1-x Fe x Sb thin films. CoTi 1-x Fe x Sb samples with Fe fraction (x) of 0.2, 0.3 and 0.5 were grown by molecular beam epitaxy (MBE) on lattice matched In 0.52 Al 0.48 As buffer layers grown on InP(001) substrates. 140 nm-thick CoTi 1-x Fe x Sb layers were grown in a dedicated VG V80 metal MBE system. The films are epitaxial and single crystalline as measured by reflection high-energy electron diffraction and X-ray diffraction. An abrupt interface between the h-H film and III-V layer is observed in STEM. For APT analysis, the samples were evaporated in a LEAP 3000X HR operated in voltage-pulse mode with a pulse to base voltage fraction of 25%. The 3D reconstructions were optimized to visualize the atomic planes in the Z direction [5]. Radial distribution analysis [6] was performed to obtain the self-correlation curves between Fe atoms. Clear Fe-rich domains were evidenced for x=0.2 and x=0.3 but are no longer observed for x=0.5. Cross-and self-correlation curves were also plotted in between Co, Ti, Fe and Sb which did not reveal any other clustering in any samples. These behaviors were confirmed using statistical distribution analysis [7]. For compositions x=0.2 and x=0.3, the APT experimental Using a spin Hall material to switch the spin orientation of a neighboring magnetic layer could serve as a key element in future low-power magnetoresistive random access memory (MRAM) devices. The spin Hall angle is defined as the ratio between the induced transverse spin Hall conductivity due to spin-orbit interactions and the longitudinal electrical conductivity. Recent experimental studies have shown that β-W (A15 phase, tetrahedrally close packed crystal structure) has a very large spin Hall angle (~0.30) [1]. In contrast, the common α-W (bcc phase) has a negligible spin Hall angle ( < 0.07) [1]. While the large spin Hall angle in β-W is thought to be an intrinsic feature of the material, so far there have been (1) no studies that have shown a direct link between the β-W electronic structure and the high spin Hall angle and (2) few studies that even examine the electronic structure of β-W. In this work, we calculate and confirm the fully relativistic band structure (including spin-orbit coupling) and Fermi surfaces of α and β-W using several first principle approaches, including a plane wave pseudopotential formalism (VASP [2]), numerically truncated localized orbitals (QuantumWise [3]), and a relativistic Korringa-Kohn-Rostoker (KKR) Green's function approach [4]. Analysis of the band structure indicates significant band splitting near the Fermi energy due to spin-orbit interactions, including regions near the Γ point, and along the Γ-X and Γ-M high symmetry lines. The spin Hall conductivity in both α and β-W is determined using the Kubo formula based on the Berry curvature [5]. We determine spin Hall angles for α-W ranging from 0.008 (bulk) to 0.021 (film), in agreement with experiment [1]. For β-W, we predict spin Hall angles ranging from 0.1 (bulk) to 0.18 (film). Since O and N are often used to stabilize the metastable β-W crystal structure, we also discuss the potential impact of these dopants on the overall electronic structure, spin Hall conductivity and spin Hall angle. Alternating layers of epitaxial TiO 2 and TiO x superlattices of rutile structure, ~1-nm thick per layer by RF-sputtering of a TiO 2 target for the TiO 2 layers, and by DC-sputtering of another TiO target for the TiO x layers at 570 o C using pure argon on c-Al 2 O 3 substrates. From high resolution transmission electron microscopy (HR-TEM), the periodically alternating layers are well-defined. X-ray photoelectron spectroscopy (XPS) analyses based on the binding energy of Ti 2p 3/2 peaks suggest the co-existence of Ti +3 and Ti +4 , thus verifying the mixed-valence nature of the TiO 2 /TiOx superlattices. The M(H) curves measured at room temperature using superconducting quantum interference magnetometry (SQUID) surprisingly showed hysteretic loops typical of ferromagnetism, although none of the constituting layers showed diamagnetism. Electrical transport measurements of such superlattices done at zero field demonstrate transition of charge ordering at low temperatures, reminiscent of what was found in Ti-rich Ti 1+x O 2 single-layer thin films, made by Ti ion implantation into TiO 2 crystals, in which randomly distributed TiO 2 , Ti 2 O 3 and TiO were found to coexist. Preliminary First-principle (ab initio) calculations to understand the roles of oxygen vacancies showed that locations and amounts of oxygen vacancies as a whole in various TiO 2 super-cells could indeed lead to spontaneous magnetizations. We thus argue that mixed-valence titanium ions are responsible for the ferromagnetism. , has generated interest in β-Ga 2 O 3 devices. A breakdown voltage of 755V and high drain current and on/off ratio of 10 9 has already been demonstrated [3]. There is also increasing interest in β-Ga 2 O 3 for RF applications, which would enable the monolithic integration of power switch and RF devices. Green et al. have recently demonstrated RF performance with a power output of 0.23W/mm with PAE of 6.3% [4]. However, β-Ga 2 O 3 has low thermal conductivity (27 W/m.K in the [010] direction @300K) and self-heating induced catastrophic failure was observed in Ref. [4]. We assess the thermal resistance of β-Ga 2 O 3 of MOSFETs using Raman thermography and simulation. Possible thermal management solutions are investigated. Single finger β-Ga 2 O 3 MOSFETs with a 2 μm gate length (1 μm-long field plate), 200 μm gate width, 5 μm gate-source spacing and 25 μm gate-drain spacing were studied, having a saturated drain current of 58mA/mm and threshold voltage of -30V. Used device structure is illustrated in Fig. 1; more details are in Ref. [3]. Raman thermography measurements were performed following a phonon temperature shift calibration of bulk β-Ga 2 O 3 . A 0.5 NA objective lens was used for the Raman, with a lateral and depth resolution of 0.5μm and ~6-8 μm, respectively, as illustrated in Fig. 1. The relatively large depth resolution is due to the transparency of β-Ga 2 O 3 . Figure 2 shows that the highest measured temperature occurs close the gate edge, where the peak electric field and associated Joule heating occurs in the device channel. The thermal resistance measured at the hottest location is 128 K.mm/W (Fig. 3), which is a lower limit value because of the spatial averaging described above. A 3-D finite element thermal model was used to predict the device temperature using temperature dependent, anisotropic thermal conductivities of β-Ga 2 O 3 reported Ref. [5]. A peak channel temperature of >1000°C was predicted by initially assuming that all heating occurs within a 0.5μm-long region at the drain edge of gate. Such high value is unrealistic because resistive heating in the access regions will become significant as mobility decreases at high temperature. Compared to Raman, simulation underestimates the temperature by ~30% (when averaged over the same volume). This discrepancy is attributed to uncertainty in the depth probed optically. A lower channel temperature limit can be determined by assuming uniform resistive heating in the channel, resulting in a simulated peak temperature of 465°C. More accurate channel temperature prediction will require detailed drift diffusion simulations to obtain the Joule heating distribution. Figure 4 shows the simulated transient thermal response, which can be used to evaluate temperature rise during pulsed operationsevere thermally induced current slump can be avoided by using a short duty cycle and pulse lengths less than a few μs. The results highlights need for thermal management for reliable β-Ga 2 O 3 RF operation. The simulated local lateral heat flux through the gate metal is larger than through the β-Ga 2 O 3 itself, suggesting that top-side heat extraction could be an effective thermal management strategy. Replacing the SiO 2 with a 200 W/m.K diamond thin film (similar to in Ref. [6]) reduces the thermal resistance of the investigated device by 30%. Combining thin film heat spreaders with flip chip mounting onto a high thermal conductivity carrier could result in a significant reduction in thermal resistance. REFERENCES:[1] M. Higashiwaki et al., APL 2012.[2] M. Higashiwaki et al., J. Phys. D: Appl. Phys. 2017.[3]M. H. Wong et al., IEEE EDL 2016. [4] A. J. Green et al., IEEE EDL 2017.[5] A. Guo et al., APL 2014.[6] M. J. Tadjer et al., IEEE EDL 2012

9:00 AM TT02 Thermal Strains in Wafer Bonding and Heteroepitaxial Structures of Monocrystalline β-Ga 2 O 3 and Other IV and III-V Semiconductor
Substrates Chao Li and Mark Goorsky; Materials Science & Engineering, University of California, Los Angeles, Los Angeles, California, United States.
Wafer bonding of monocrystalline β-Ga 2 O 3 to other IV and III-V semiconductor substrates with higher thermal conductivity (typically by 3-20 times) and lower cost and heteroepitaxial structures with β-Ga 2 O 3 as either substrates or epitaxial layers are important for the applications of β-Ga 2 O 3 in high-power devices and GaN-based LEDs. It is necessary to understand thermal strains in bonded β-Ga 2 O 3 and epitaxial structures due to thermal mismatch in order to select proper materials for bonding and epitaxial growth and assess the structural and electrical properties in the bonded and epitaxial structures. In this study, thermal strains along six evenly distributed in-plane directions were determined in wafer bonding structures of (2 0 1), (0 1 0) and (0 0 1) oriented monocrystalline β-Ga 2 O 3 to Si, InP, 3C-SiC, and 6H-SiC substrates, and epitaxial structures of MOCVD (0 0 0 1) GaN/(2 0 1) β-Ga 2 O 3 with an in-plane orientation relation of (1 1 2 0) GaN//(0 1 0) β-Ga 2 O 3 and MOCVD (2 0 1) β-Ga 2 O 3 /(0 0 0 1) sapphire with an in-plane orientation relation of (0 1 0) β-Ga 2 O 3 //(1 1 0 0) sapphire. The thermal expansion coefficients of β-Ga 2 O 3 single crystal wafers were measured over a range of room temperature to up to 1000 °C using X-ray diffraction and a high temperature cell. Measurements that involve single crystals overcome problems associated with previous measurements using powders milled from single crystals, such as strains induced by milling and surface reactions of individual grains. The temperatures used here cover a wider range than had been used earlier in those powder measurements making these measurements more suitable for addressing strains in wafer bonding and high temperature epitaxy. We determined that the 'b' and 'c' axes expand linearly at about twice the rate of the 'a' axis with increasing temperature with expansion coefficients of ~ 7 × 10 -6 K -1 (compared to about 3 × 10 -6 K -1 for 'a'). The angle 'β' between the 'a' and 'c' axes increases slightly (~ 0.1°) with increasing temperature to 1000 °C. Using the measured lattice parameters of β-Ga 2 O 3 single crystal wafer, it is predicted that with a bonding temperature of 600 °C each of the (2 0 1), (0 1 0), and (0 0 1) oriented single crystal β-Ga 2 O 3 bonded to InP, 3C-SiC and 6H-SiC have similar tensile thermal strains (except for that along ~ 1/6-1/2 in-plane directions for the (0 0 1) β-Ga 2 O 3 bonded structures having compressive strains) within a range from ~ 4.0 × 10 -5 to ~ 2.0 × 10 -3 , ~ 1.5-15 times smaller than that of β-Ga 2 O 3 bonded to Si. Additionally, (0 1 0) oriented β-Ga 2 O 3 bonded to other substrates has more uniform thermal strains along different in-plane directions (< 5 times difference between the maximum and minimum values) than (2 0 1) and (0 0 1) oriented β-Ga 2 O 3 . The thermal strains of the (0 1 0) oriented β-Ga 2 O 3 single crystal bonded to InP, 3C-SiC and 6H-SiC range from ~ 2.0 × 10 -4 to ~ 1.0 × 10 -3 , smaller than ~ 6.5 × 10 -4 of GaAs bonded to InP with the same bonding temperature along ~ 1/3 in-plane directions. On the other hand, as for the MOCVD (0 0 0 1) GaN/(2 0 1) β-Ga 2 O 3 grown at 1000 °C, tensile thermal strains were determined along 5/6 in-plane directions while compressive strains along the other in-plane directions. The thermal strains range from 2.5 × 10 -4 to 2.2 × 10 -3 which are smaller than that in GaN grown on sapphire (~ 2.5 × 10 -3 ) while partially (along ~ 1/2 in-plane directions) smaller than GaN on SiC (~ 7.2 × 10 -4 ). In addition, compressive in-plane thermal strains in β-Ga 2 O 3 were determined for the MOCVD (2 0 1) β-Ga 2 O 3 deposited on (0 0 0 1) sapphire substrate at 650 °C with a range from 1.9 × 10 -4 to 2.0 × 10 -3 .
β-Ga 2 O 3 offers great potential for high-power devices with low energy loss owing to its material properties derived from an extremely large bandgap of 4.5 eV. The availability of commercial Ga 2 O 3 substrates with high crystal quality and yet lower cost than GaN and SiC also facilitates the development of vertical power devices. However, the low thermal conductivity of Ga 2 O 3 (11~24 W/m . K) hampers high-power device operation. As for this issue, we consider that heat dissipation from Ga 2 O 3 devices can be improved by directly bonding them to a thermally and electrically conductive substrate. We pursued this idea in a previous work with polycrystalline SiC since these substrates are not only highly conductive both thermally and electrically but also affordable, and successfully demonstrated the formation of a void-free high-quality Ga 2 O 3 / SiC interface by using the surface-activated-bonding (SAB) method [1]. The interface presented a negligibly small thermal resistance and a large bonding strength comparable to the Ga 2 O 3 bulk strength. In this work, we characterized the electrical resistance at the Ga 2 O 3 /SiC bonded heterointerface.Two-terminal electrical test structures were fabricated on (i) a single-crystal β-Ga 2 O 3 (001) substrate (n~3×10 18 cm -3 ), (ii) a polycrystalline SiC substrate (n~1×10 20 cm -3 ), and (iii) a single-crystal Ga 2 O 3 /polycrystalline SiC bonded substrate, where (i) and (ii) were used to extract series bulk resistances. Ti/Au ohmic electrodes were blanket evaporated onto the back sides of all the substrates, and circular ohmic electrodes with a diameter of 200 μm were formed with the same metal stack on the front surfaces. Transfer length method measurement of process monitor wafers yielded specific contact resistances of 4.5~5.4×10 -6 and 2.2~4.2×10 -5 Ω . cm 2 for Ga 2 O 3 and SiC, respectively.All measured current-voltage curves for the test structures fabricated on the Ga 2 O 3 /SiC bonded substrate presented a linear behavior, indicating that the heterointerface was an ohmic junction. The electrical resistance at the interface, estimated by subtracting Ga 2 O 3 and SiC series bulk and contact resistances from the total series resistance of 0.66 Ω, was 0.06 Ω (a specific resistance of 2×10 -4 Ω . cm 2 ). This value was less than 10% of the total series resistance and thus sufficiently small for practical vertical Ga 2 O 3 devices.In summary, we succeeded in fabricating SAB Ga 2 O 3 /SiC substrates with not only large bonding strength at the interface but also high thermal and electrical conductivities through the interface. The Ga 2 O 3 /SiC bonded substrates will be useful for future developments of vertical Ga 2 O 3 power devices.This work was partially supported by Council for Science, Technology and Innovation (CSTI), Cross-ministerial Strategic Innovation Promotion Program (SIP), "Next-generation power electronics" (funding agency: NEDO There is a surge in interest in β-Ga 2 O 3 because of its thermodynamic stability, wide bandgap, and excellent figures of merit for high power devices. Additionally, because β-Ga 2 O 3 consists of a monoclinic network of GaO 4 tetrahedra and GaO 6 octahedra, it is quite similar to structures found in magnetic 3d transition metal oxides. In particular, the Fe 3+ cation exhibits similar ionic radius to Ga 3+ and several Fe 2 O 3 phases are earth abundant. However, none naturally exhibit the monoclinic β-Ga 2 O 3 structure. Here we investigate the possibility to epitaxial stabilize a new form of monoclinic Fe 2 O 3 (called m-Fe 2 O 3 ) by growing it via plasmaassisted molecular beam epitaxy on the (010) face of β-Ga 2 O 3 . After calibrating the Fe-flux, the growth rate of the hypothetical m-Fe 2 O 3 is calculated assuming it has the same lattice parameters as β-Ga 2 O 3 . A preliminary test sample was grown with increasing thicknesses of m-Fe 2 O 3 on β-Ga 2 O 3 followed by 10 nm β-Ga 2 O 3 spacers. Reflection high energy electron diffraction (RHEED) shows the preservation of the β-Ga 2 O 3 overgrowth quality even for quite high m-Fe 2 O 3 thicknesses. High resolution X-ray diffraction of the structure shows distinct thickness fringes and superlattice peaks. High resolution scanning transmission electron microscopy confirms that β-Ga 2 O 3 can still be grown with high quality over subsequent Fe containing layers, and that the high Fe regions shows the same crystal structure as β-Ga 2 O 3 , i.e. m-Fe 2 O 3 . The optical and magnetic properties of this new form of Fe 2 O 3 will be discussed. Because of the large bandgap and the resultant large electrical breakdown strength, β-Ga 2 O 3 emerged as a promising semiconductor which can sustain large voltages, making it attractive for high-power devices. However, high power dissipation in these devices can cause critical challenges, e.g., it can significantly affect the performance and reliability of these devices. It's necessary to understand the thermal transport in β-Ga 2 O 3 to control the hot spot temperature in its active devices and also for the design of packaging and thermal solutions. A few studies have focused on the prediction and the measurements of the thermal conductivity of β-Ga 2 O 3 . However, the computational results don't have a good agreement with the experimental results, and the mechanism of the phonon transport in β-Ga 2 O 3 is not well understood yet. In addition, due to the imperfection in growth processes, the crystal lattice of the β-Ga 2 O 3 contains unintentional localized defects such as vacancies. They could also significantly influence the thermal properties of β-Ga 2 O 3 which is not well understood yet. To better understand the influence of defects on the phonon transport mechanism, we used first-principles density functional theory (DFT) along with the Boltzmann Transport Equations (BTE) to predict the phonon transport properties of pristine and defective β-Ga 2 O 3 . The thermal conductivities of β-Ga 2 O 3 crystals along three different crystal directions are calculated based on the iterative solution of the BTE. Our results have a better agreement with the experimental results compared to the previous studies. The largest thermal conductivity of β-Ga 2 O 3 is observed in the direction of [010]. In addition, the oxygen vacancies indicate a significant influence on the thermal transport of β-Ga 2 O 3 , which results from the phonon scatterings caused by the mass missing of oxygen atoms. Results indicate that at room temperature, 1% and 2% oxygen vacancies decrease the thermal conductivity by 8.5% and 14.3% in [100] direction, 14.9% and 24.1% in [010] direction, 10.7% and 17.4% in [001] direction, respectively. We also find that the optical phonon modes in β-Ga 2 O 3 make a significant contribution to the thermal conductivity compared to the acoustic modes, which is different with most other semiconductors. Furthermore, the oxygen vacancies have more influence on the optical modes than that on acoustic modes, which suppress the contribution of optical modes to the thermal conductivity. In other words, under the influence of defects, the acoustic modes' contribution to the thermal conductivity increases. The results from this work will help us understand the mechanism of phonon transport considering the influence of defects and provide insights for the future design of β-Ga 2 O 3 -based electronic devices. We report polarization dependent photoluminescence studies on unintentionally-, Mg-, and Ca-doped β-Ga 2 O 3 bulk crystals grown by the Czochralski method. In particular, we observe wavelength shifts of the highest energy UV emission dependent on the incident polarization. For 240 nm (5.17 eV) excitation, almost no shift of the UV emission is observed between E||b and E||c, while shift of the UV emission centroid is clearly observed for 266 nm (4.66 eV) excitation. These observations are consistent with the UV emission originating from transitions between conduction band electrons and two differentially-populated self-trapped hole (STH) states. This observation implies that the STHs form primarily at the oxygen involved in the original photon absorption event, thus providing the connection between incident polarization and emission wavelength. The data imposes a lower bound on the energy separation between the self-trapped hole states of 10-30 meV.
β-Ga 2 O 3 is an intriguing material because of its huge and direct bandgap of ~4.85 eV, high breakdown field(~8 MV/cm), and excellent thermal and chemical stability. Baliga's figure-of-merit (FOM) and Johnson's FOM of β-Ga 2 O 3 , a quantity used to characterize the potential as power devices, is 3214.1 and 2844.4, respectively, which is much higher than that of GaN (846.0 and 1089.0, respectively) or SiC (317.1 and 277.8, respectively). In particular, although β-Ga 2 O 3 is not a van der Waals material, it can be mechanically exfoliated into quasi-two dimensional(quasi-2D) flakes due to the huge anisotropy of the unit cell. Owing to these characteristics, various studies are undergoing to attract the full potential of β-Ga 2 O 3 as nanoscale power devices.Various techniques have been proposed to enhance the breakdown voltage of power devices. Among them, field-plate techniques to enhance the breakdown voltage of the devices are widely employed owing to the simplicity and efficiency. Preventing breakdown by applying gate and source field-plate structure lead the full potential of β-Ga 2 O 3 as a power device.In this work quasi-2D β-Ga 2 O 3 source fieldplate metal-semiconductor field effect transistor (MESFET) were fabricated by mechanical exfoliation of β-Ga 2 O 3 flakes followed by dry transfer of graphene as a source field material as a source field-plate structure. By applying h-BN as a dielectric material between β-Ga 2 O 3 channel and graphene field-plate, which has ultrahigh breakdown field of 8 ~ 12 MV/cm and atomically flat surface, nanoscale MESFET device by stacking 2D materials was fabricated. Electrical and optical properties, and three-terminal off-state breakdown were analyzed by using semiconductor analysis connected with probe station. Silvaco device simulation framework was employed to study the difference of electric field distributions between field-plated and non-field-plated devices. The details of our work will be discussed. Wide-bandgap semiconductors, such as beta gallium oxide (β-Ga 2 O 3 ), exhibit a number of interesting physical properties. For instance, b-Ga 2 O 3 has a large breakdown field, high transmittance in the UV-vis region, as well as good thermal and chemical stability. This oxide has attracted a lot of attention not only due to its promising potential in power electronics, but also in transparent electronics. In the former application, deep-center dopants, especially Fe and Mg, are used to increase resistivity, while in the latter, shallow centers, usually Si or Sn, are used to decrease resistivity. These centers can be usefully characterized by UV/VIS/NIR reflectance and transmittance measurements which can yield absorption a and reflection R coefficients. We have developed new methodology for accurately determining both large (above band-gap) and small (below band-gap) values of a. In bulk materials, several mid-gap levels typically appear in the a spectrum. To understand these mid-gap features, structural, electronic, and optical properties of β-Ga 2 O 3 with Mg and Fe impurities are studied by employing first-principles calculations based on density functional theory calculations combined with many-body perturbation theory including quasiparticle and excitonic effects. A detailed comparison done between the computed optical absorptions spectra of the systems within and without the Bethe-Salpeter framework, suggests the electronhole interaction significantly modifies the spectra. Also, we look at the role of deep defect levels in the gap found in the modeling in the interpretation of features from experiment. The high electrical conductivity in combination with optical transparency makes In 2 O 3 an interesting material for optoelectronic devices like solar cells, flat-panel displays, or light emitting diodes. Despite the promising applications, the high electrical conductivity in such wide band gap materials is not clear. Recently, the origin of the electrical conductivity has been discussed intensively and studied in order to reach better efficiency for device operation. Scanning tunneling microscopy and spectroscopy (STM/ STS) is a powerful tool to get information about the structural and electronic properties at the atomic scale that is useful for understanding and controlling of the material properties.In this contribution, we present results of clean, freshly-cleaved In 2 O 3 (111) non-polar surfaces, measured by STM and STS. The semiconducting bulk In 2 O 3 single crystals were grown from the melt [1,2]. The measurements were performed on two different types of samples, whereby the one was used as-grown while the other one was annealed in an oxidizing atmosphere, the latter besides other effects reducing the H 2 content in the bulk and also the conductivity [3]. All samples were cleaved in situ at a base pressure below 1×10 -8 Pa.The atomically resolved STM images show a flat cleavage surface with some monoatomic steps. Additionally, one brighter contrast within every surface unit cell was found, being composed of a triangular substructure, corresponding to unoccupied states. The averaged lateral size of the surface unit cell was found as 1.35 nm and the distance between the unoccupied states inside the triangular substructure as 0.38 nm. These values are in good agreement with theoretical calculations leading to 1.43 nm and of 0.37 nm, respectively [4].Scanning tunneling spectra show intrinsic states within the fundamental indirect band gap, being determined to 2.7±0.1 eV with respect to the conduction band minimum (CBM) at the Γ-point. The direct band gap is found to be 3.8±0.1 eV. The first electronic state at an energy of -0.5 eV below the CBM belongs to oxygen vacancies and pins the Fermi level energetically at the surface, which shows the absence of intrinsic electron accumulation. The Fermi level lies energetically approximately -0.5 eV below CBM, close to oxygen vacancyinduced state. The next band gap state was found around -1.4 eV below CBM. It is much stronger visible in the STS spectra of the as-grown sample than in the ones of the annealed sample. Therewith we assign this electronic state to H 2 impurities. Hence, mainly oxygen vacancies but also hydrogen contribute to the electrical conductivity in In 2 O 3 . Upon aging the sample under ambient conditions for a week the electronic states vanish at the surface and a kind of metallic conductivity is observed.This project was supported by the Leibniz Association, Leibniz Science Campus GraFOx, project C2-6. GaN-based semiconductors have attracted extensive interest because of their application in optoelectronic and high-power electronic devices including LEDs and LDs, ultraviolet Schottky barrier photodetectors, heterostructure field-effect transistors, and metal-semiconductor-metal photodetector. As for p-GaN containing a high density of defects, researchers have developed efficient Schottky contacts with low leakage current and high barrier height. In particular, for the fabrication of transparent electronic devices, Schottky contacts need to be transparent on top of reliable Schottky barrier heights (SBHs). In this study, thus, we investigated the electrical and optical properties of high barrier-height and transparent Ti/ITO Schottky contacts on p-GaN for optoelectronic and transparent electronic devices. The SBHs and ideality factors were obtained using current-voltage-temperature (I-V-T), capacitance-voltage (C-V), and barrier inhomogeneity model as a function of annealing temperature. The SBHs and ideality factors estimated using I-V characteristics were estimated to be in the range of 0.36-0.39 eV and 1.74-2.07, respectively, depending on the annealing temperature. On the other hand, the barrier inhomogeneity and C-V methods yielded much larger SBHs of 0.82-1.18 eV. At 560 nm, the Ti/ITO samples transmitted 62.07-94.45%. The XPS Ga 2p core levels obtained from the interface regions of the ITO/Ti/GaN samples shifted toward higher or lower energies, depending on the annealing temperature. The normalized N/Ga atomic ratio showed that N and Ga vacancies were formed at the p-GaN surface region at 300 and 500 °C, respectively. The XPS Ti 2p, N 1s, and O 1s core level results showed the formation of interfacial TiN and TiO 2 phases at 300 and 500 °C, respectively. STEM element mapping results exhibited the outdiffusion of Ga atoms in the sample annealed at 500 °C. On the basis of the XPS and STEM results, the dependence of the SBHs on the annealing temperature is described and discussed.

9:20 AM UU03 (Student) The Role of Hydrogen in Binary and Ternary Transparent
For photovoltaic (PV) devices, the transparent conductive oxide (TCO) properties of interest are electrical conductivity and optical transparency. However, finer control of additional TCO properties has become vital due to the advent of complex PV device architectures, such as silicon heterojunction and tandem devices, where optimal charge extraction is crucial [1]. Needs arise for higher charge carrier mobility, as well as control over band gap (E g ), Fermi level (E F ), and work function (WF). The current commercial TCO of choice is tin doped indium oxide (ITO), which has a typical conductivity of 1-5 x 10 3 S/cm and a transparency of 85% in thin-film form. While ITO meets present device requirements, research efforts are focused on providing inexpensive and environmentally benign TCO alternatives. Record mobilities 3-4 times greater than that of industry-standard ITO have been demonstrated in hydrogenated indium oxide (IO:H), making it an effective ITO-alternative for incorporation into a wide range of PV devices [2][3][4]. We first present our findings on sputtered thin-films of IO:H, which show tunable mobilities, carrier densities, and slightly varying WF, achieved by controlling hydrogen content as shown in Fig. 1 of the extended abstract. Structure is also altered by introduction of H into the IO:H film -higher H content induces a greater amorphous fraction in the as-deposited films. Recently, amorphous structures have become desirable in TCOs due to the discovery of remarkably smooth amorphous phases (surface roughness <0.2 nm) [5], allowing for more conformal coatings that potentially reduce shunting in PV devices with thin stack layers. To move away from indium-based TCOs -necessitated by the material scarcity of indium and resulting price volatility-we next consider a set of low-cost, non-toxic binary and ternary oxides, and the effect hydrogenation has on them, namely tin oxide, zinc oxide, and zinc-tin oxide. We present our findings on control over mobilities, carrier densities, E F , WF, and E g by introduction of hydrogen into these systems. Kelvin probe, spectrophotometer, and Hall measurements are utilized to investigate these properties. This work provides a consolidated picture of the link between composition and optoelectronic properties of promising binary and ternary transparent conductive oxides with intentionally introduced hydrogen. References: Transparent conductors in the visible spectrum are key enablers of countless display and energy technologies. While many transparent conductors have been developed to be transparent in the visible region, few examples exist in the ultraviolet region. A material that is transparent both in the in the visible and ultraviolet range (in particular, 200 -300 nm) would be universally applicable for display, energy, and UV-specific applications, especially disinfection of water and air. We show that the correlated metal strontium niobate (SrNbO 3 ) is suitable for applications in both spectral ranges.Correlated metals offer an alternative design strategy compared to the conventional paradigm of doping wide bandgap semiconductors. Instead of starting with transparent semiconductors and doping them improve conductivity, we start with an conducting material and open a transparency window though tuning the electron effective mass via electron correlation and tuning interband transitions. Adjusting these two parameters allows for a transparency window to extend over a variety of desired spectral ranges. This principle has recently been demonstrated in the correlated metals CaVO 3 and SrVO 3 , which were shown to have a similar transparency and resistivity to indium tin oxide (ITO) 1 .Here, we report on films of the correlated metal, SrNbO 3 , grown by pulsed laser deposition 2 and magnetron sputtering. Transmission for this material is in excess of 80% throughout the visible range and near ultraviolet (270 -600nm) while maintaining a resistivity lower than 2×10 -5 Ω×cm. The fact that this material lends itself to sputtering allows for easy integration into devices. Ongoing research aims to further optimize the absorption edge to improve transparency further into the ultraviolet region through aliovalent doping, using lanthanum and titanium to suppress interband transitions.1. Zhang, L. et al. Correlated metals as transparent conductors. Nat. Mater.15, 204-210 (2016).2. Oka, D., Hirose, Y., Nakao, S., Fukumura, T. & Hasegawa, T. Intrinsic high electrical conductivity of stoichiometric SrNbO 3 epitaxial thin films. Phys. Rev. B 92, (2051).
Transparent conductors are widely used as fundamental part of many day life devices such as solar cells, OLEDs, display technologies and touch panels. Moreover, development of the Internet of Things implies that markets push emerging technologies and lead consumer needs to be geared towards flexible or non-planar devices.So new generation of transparent electrodes exhibiting both a really high transparency, a high electrical conductivity and mechanical flexibility should have to be developed. Up to now, the fabrication of transparent conductive films is currently made up of thin films of transparent conductive oxides such as indium tin oxide (ITO). However the as-made ITO electrodes suffer from limitations like costly fabrication process and brittleness. The use of solution-processable nanomaterials, and especially based on metallic nanowires, appears as a promising alternative since it affords a large area, low-cost deposition method. Among metals, silver as the most conductive metal has been widely reported in the literature for easy nanowires synthesis. The very high aspect ratios of metallic nanowires are required to allow fabrication of percolating random networks with the best trade-off between transparency and electrical conductivity.We will herein present first the synthesis of silver nanowires AgNWs based on polyol process to independentlycontrol nanowire lengths and diameters. Thanks to compatible large scale printing process, elaboration of homogeneous random networks exhibiting excellent performances with sheet resistance less than 20 Ω/sq at 90 % of transmittance will be detailed. Nanowire dimension tuning will strongly modify optoelectrical properties of transparent thin films in terms of sheet resistance, transparency and haze properties. Understanding of percolation pathways in this 2D random networks of AgNWs by visualizing electrical and thermal distributions will be discussed. Prior to integration into devices, electrode stability will be addressed for emphasising good stability over than 3 years (storage in air and in the dark) but also for evaluating failure limits under several environmental stresses. To enhance the interest of macroscopic properties of nanoscale materials, we will finally show results dealing with the use of such electrode as transparent film heaters (TFH) at large scale with very good performances.

10:40 AM UU07 (Student) Comparative Studies of Thermal Transient Responses of
Hotspots in Graphene-Silver Nanowire and Silver Nanowire Transparent Conducting Electrodes Sajia Sadeque 1, 2 , Yu G. Gong 1, 2 , Amir K. Ziabari 1, 2 , Kerry Maize 1, 2 , Amr M.S. Mohammed 1, 2 , Ali Shakouri 1, 2 and David B. Janes 1, 2 ; 1 School of Electrical and Computer Engineering, Purdue University, West Lafayette, Indiana, United States; 2 Birck Nanotechnology Center, Purdue University, West Lafayette, Indiana, United States. Nanostructured random networks have attracted wide attention as transparent conducting electrodes (TCEs), providing optical and electrical properties comparable to conventional transparent conductive oxides along with mechanical flexibility [1,2]. Nanowire (NW)/nanotube networks exhibit percolating conduction, with NW-NW junction resistance as the transport bottleneck. Hybrid networks consisting of a 2D layer stacked on a NW network (e.g. monolayer graphene/silver NW network) can provide low resistance at high transparency, via co-percolation through both subnetworks. While the macroscopic characteristics of NW and hybrid networks have been studied extensively, studies of the local conduction pathways and relative differences between the networks at the microscopic scale have been limited. Thermoreflectance (TR) imaging techniques offer the potential to study microscopic self-heating induced by current flow over large areas of the device simultaneously [3][4][5][6]. In this study, we use a TR imaging technique to compare the thermal transients of hotspots (comprising several NW-NW junctions) in silver NW and hybrid networks (figure 1). Within a time-regime dominated by local self-heating (0-10 µs), we find that hotspots in both networks show similar thermal time constants of less than 1 µs (figure 2) during the heating and the cooling cycle. Compared to the NW network, the hybrid network exhibits ~ 4x fewer hotspots ( figure 3(a)) and a one-third reduction in average hotspot temperature ( figure 3(b)), but ~ 3x more power dissipation per hotspot. These observations can be semi-quantitatively explained in terms of interconnected resistor network models and estimations of local heat spreading from the NW-NW junctions to the substrate via NWs and graphene. We also analyze the temporal evolution of spatial distribution of the hotspots and find that hotspots in hybrid network are more clustered compared to silver NW network. Since hotspots occur at the most resistive links within otherwise conductive pathways, the hotspot distribution provides at least qualitatively information on current pathways. This comparative study captures the essence of percolating/co-percolating transport and illustrates the important role graphene plays in minimizing electrical and thermal resistance of NW-NW junctions making 1D-2D hybrid network an attractive transparent electrode material for flexible applications. Conjugated polymers with electrical conductivity are garnering much attention for applications in organic electronic and optoelectronic devices including such as field effect transistors, organic solar cells and sensors. Polymers based on poly(3,4-ethylenedioxythiophene) (PEDOT) are particularly promising candidates with a semi-metallic range of conductivity, uniform surface planarity and excellent ductility. In this work, homopolymer PEDOT films deposited using oxidative chemical vapor deposition (oCVD) show the maximum conductivity as high as ~3500 S cm -1 at a deposition rate <0.5 nm/min. However, their utility is limited due to the relatively low transmittance and abrupt decrease near the red edge in the visible regime.Here, we demonstrate facile oCVD copolymerization strategies that exhibit a powerful ability to tune the band gap of PEDOTbased copolymers. The oCVD technique provides a single-step process for the synthesis and deposition of functional copolymer films. The crosslinking monomers of biphenyl or anthracene are simultaneously evaporated with EDOT monomer and an oxidant of FeCl 3 during the deposition. Due to the engineered band gap, the optical properties of copolymers are significantly enhanced compared to homopolymer PEDOT. Poly(anthracene-co-EDOT) [p(ANTH-co-EDOT)] shows the superior transmittance of ~ 93% to homopolymer PEDOT (~80%) and poly(biphenyl-co-EDOT [p(BPH-co-EDOT)] (~88%) at the wavelength of 550 nm. In addition, copolymer films show no significant transmission decay in the red edge regime of the light unlike homopolymer PEDOT that presents an abrupt transmission falloff. An improvement in optical transmittance is in agreement with an increase in band gap of materials (p(ANTH-co-EDOT), ~2.3 eV vs. PEDOT, ~1.8 eV).Further, the surface morphology of oCVD copolymers is dramatically changed compared to homopolymer PEDOT. These copolymers show much bigger intermolecular morphological features in atomic force microscopy images, which is expected to be favorable to the carrier transport.To conclude, oCVD-processed band gap tunable PEDOT copolymers with enhanced transmittance and abilities to alter morphology may, therefore, have of great relevance in many organic electronic and optoelectronic devices that require the high optical transparency and better morphologies. 11:20 AM UU09 (Student, LATE NEWS) Role of Native Defects on the Optoelectronic Properties of ZnO Pooneh Saadatkia 1 , Naresh Adhikari 1 , Petr Stepanov 1, 2 , Micah Haseman 1 , Jack Warfield 1 , Gerald E. Jellison 1 , Lynn A. Boatner 3 and Farida Selim 1, 2 ; 1 Center for Photochemical Sciences, Bowling Green State University, Bowling Green, Ohio, United States; 2 Department of Physics and Astronomy, Bowling Green State University, Bowling Green, Ohio, United States; 3 Oak Ridge National Laboratory, Oak Ridge, Tennessee, United States.
ZnO, a wide band gap semiconductor, is considered to be one of the most promising oxides for optoelectronic devices due to its diverse electronic and optical properties. It is also known as a member of transparent conductive oxides that makes it an interesting material for fundamental and applied research. Despite the large number of studies in ZnO, the origin of green luminescence is not well understood. In this study, we combine a wide range of characterization techniques including Hall effect, Photoluminescence spectroscopy (PL), Thermo-luminescence spectroscopy (TL), positron annihilation lifetime spectroscopy (PALS) and digital coincidence Doppler broadening spectroscopy to characterize point defects and investigate the correlation between vacancy type defects, conductivity and green luminescence. The studies were carried out on large number of bulk ZnO single crystals grown by different methods and featuring different defect structures.The measurements showed that conductivity and green luminescence revealed a reverse relationship. Samples with higher conductivity and carrier concentration exhibit weaker green luminescence. Positron annihilation spectroscopy indicated the presence of Zn vacancy related defects, which act as acceptors and source of green luminescence. Absence of Green luminescent as well as low defect concentration were observed in highly conductive samples. Thermo-luminescence spectroscopy indicated that hydrogen act as donors in conductive samples.It is also interesting to note that the ratio between green luminescence emission and near band emission decreases with increasing laser intensity.