United States

This study analyzes the changes in the density of interface traps (DIT) for various atomic layer deposition (ALD) gate dielectric films on gallium nitride (GaN) devices fabricated on bulk GaN substrates. Each film type has been chosen with an optimized insulator/semiconductor interface obtained through chemical cleans or post deposition annealing (PDA). Reduction in dielectric leakage and interface traps in MOSCAP devices may be used in order to understand how these mechanisms affect more complex devices such as vertical GAN MOSFETs. GaN MOSFET and MOSCAP devices are often investigated and implemented in power systems with applications ranging from electric vehicles (EVs) to smart power grids. The unique material properties of GaN such as wide bandgap, high breakdown voltage, and high electron mobility create unique advantages over traditional silicon FETs. Vertical GAN MOSFETs are better suited for higher voltage applications due to current flow and voltage drop perpendicular to the surface while also reducing required chip area for high operating voltages. Although there are many advantages of vertical GaN devices, such an architecture presents unique challenges that must be addressed before such devices may be implemented in high power applications, such as the requirement for a native substrate and the need to improve the gate dielectric interface. The interface between the gate dielectric and the semiconductor is crucial in order to reduce threshold voltage drift, leakage current, and the density charge traps. Previous work has shown the importance of DIT reduction with chemical cleans and PDAs. Accurate calculations of the density of interface traps (DIT) are vital to fully realize device operation and reliability. The choice between surface cleans and dielectric annealing as a leakage reduction technique is dependent on the material and available processing capabilities. While PDAs may reduce the dielectric breakdown strength, they have been found to reduce gate leakage current and improve device performance for most gate dielectrics. Several methods have been proposed for DIT characterization, including the voltage dependencies of low- and highfrequency capacitance (CV), conductance (GV), and surface potential (WY). DIT may be calculated from the surface potential of the dielectric using the quasi-static capacitance and the capacitance of the chosen gate dielectric. This work uses quasi-static CV and surface potential electrical characterization of GAN MOSCAPs to determine DIT of various gate dielectric materials treated with PDAs and chemical cleans. Optimization of such films on MOSCAPs will allow for implementation into vertical GAN MOSFET architectures. Interface trap density may be reduced by decreasing the capacitance associated with defects and surface contamination or by selection of a dielectric material with a lower dielectric constant. Fixed charges at the dielectric/semiconductor interface are manifested in hysteresis between forward and reverse voltage sweeps and may be the result of contamination prior to ALD or point defects in the material. Al,0. is an attractive gate dielectric material due to its dielectric constant, which may improve breakdown strength (Fig. 1), but it may suffer from higher memory charge between continuous bias sweeps. SiO, ALD films have lower breakdown and significantly lower leakage currents but show the lowest density of remaining fixed charges between sweeps, suggestive of the lowest DIT of the tested dielectrics (Fig. 2) with the addition of PDA. Results were verified by altering bias conditions during testing in order to accumulate and deplete interface charges. Minor plateaus in capacitance in forward bias indicate the presence of negative fixed interface charges. Additional characterization may be necessary to determine the optimal conditions for DIT reduction for HfO., films, including reduction in PDA temperatures in combination with improved substrate chemical cleans.

EMC 2022

Interface Trap Density Characterization of ALD Gate Dielectrics for GaN Power MOSFETs

gate

interface traps

mosfet

After two years of virtual conferences, we're excited to host EMC 2022 in person at The Ohio State University in Columbus, Ohio. In addition to cutting-edge technical presentations given in multiple parallel sessions, opportunities to network with colleagues will be present throughout the conference. The size and format of the meeting makes networking easy, whether you have been coming to the conference for many years or are a first-time attendee.

EMC is coordinated with the Device Research Conference (DRC), which will be held the same week, from June 26-29. This recognizes the strong interaction between device and electronic materials research and provides fruitful exchange of information between attendees of both Conferences.

For information on the 80th Device Research Conference, please visit the DRC 2022 website.



The Electronic Materials Conference (EMC) is the premier annual forum on the preparation, characterization and use of electronic materials. This year’s event will feature a plenary session, topical sessions, a poster session and more.

technical paper

Digital neural networks (NNs) and the availability of large training datasets have allowed for rapid progress in the performance of machine-based tasks for a wide range of applications including image analysis, sound recognition, and natural language translation. The enhanced capability has, however, come at a computational cost as increased complexity and accuracy have necessitated the need for ever larger deep neural networks (DNNs). One alternative to DNNs is the use of optical processors that have the advantages of ultra-fast processing times and low energy costs. These systems can be employed as stand-alone processors or as front-end accelerators for digital systems. In either case, optical systems are most impactful when used for the linear matrix-vector multiplications that comprise the convolution operations in DNNs as these are often the most computationally burdensome components typically comprising more than 90% of the required floating-point operations (FLOPs) in popular CNNs. In the case of image analysis, free-space approaches are attractive as spatial multiplexing can be readily achieved as well as the fact that an optical front-end can potentially be integrated directly with an imaging system. The most traditional approach to free-space based optical image processing is the use of 4f optical correlators where spatial filters, either passive or dynamic, are placed in the Fourier plane of a 2-lens optical system. Recorded spatial features are then fed to a lightweight digital NN back-end for classification. An alternative approach is the use of diffractive neural networks which utilize cascaded diffractive elements as convolutional layers. Image classification is realized through redistribution of optical energy on the detector plane requiring minimal digital processing. The tradeoff is the need for several diffractive layers as well as coherent illumination, precluding use with ambient lighting. While these approaches have shown benefits in terms of processing speed and energy consumption, they necessitate enlarged imaging systems. Furthermore, none of these approaches utilize the additional information channels, such as polarization, that are available when utilizing an optical front-end. Here, we demonstrate the use of meta-optic based optical accelerators that serve as the convolutional front-end for a hybrid image classification system. Spatial multiplexing is achieved by using a multi-channel metalens for image duplication and a metasurface-based convolutional layer. This system has the advantage of being compact while the use of metasurfaces allows for additional information channels, in this case, polarization, to be accessed enabling both image and polarization-based classification. The hybrid network utilizes end-to-end design such that the optical and digital components are co-optimized while also incorporating statistical noise resulting in a robust classification network. We experimentally demonstrate the classification of the MNIST dataset with an accuracy of 92% as well as 94% accurate classification of polarized MNIST digits. Due to the compact footprint, ease of integration with conventional imaging systems, and ability to access additional information channels, this type of system could find uses in high-dimensional imaging, information security, and machine vision.

Meta-Optic Accelerators for Object Classifiers

Digital manufacturing is revolutionizing the manufacturing industry. Recent breakthroughs in hybrid 3D printing by the EM Lab at the University of Texas at El Paso have made it possible to build three-dimensional parts composed of any distribution of metals, dielectrics, and other materials. This talk will cover some of the new and exciting applications in circuits, electromagnetics, photonics, and metamaterials that are being enabled by digital manufacturing. Electrical circuits can be made three-dimensional where components are placed at any position xyz and oriented at any angle theta-phi. Interconnects can meaner along smooth paths throughout all three dimensions. This creates new degrees of freedom for layout and routing that can make circuits more compact, smaller, and lighter. Shorter trace lengths reduce parasitic impedance to improve power efficiency and increase bandwidth. 3D circuits can be manufactured into unconventional form factors and exploit physics, such as dielectric anisotropy, not easily accessed by planar circuits. Another breakthrough by the EM Lab allows periodic structures to be bent, twisted, conformed, and otherwise spatially varied without deforming the size, shape and spacing of unit cells. Such deformations would weaken or erase the electromagnetic and optical properties of the lattices. The technology was first applied to photonic crystals where discoveries were made that led to the tightest bend of unguided optical beam, refractionless lenses, and photon funneling. The technology was next applied to metamaterials to develop a tool to generate metamaterial lattices for devices designed by transformation optics. The technology is now being used to generate periodic structures across doubly-curved surfaces.

Electromagnetics and Photonics in the Age of Digital Manufacturing

We report our recent results on scalable fabrication of helical slow wave structures (SWSs) applicable in millimeter-through-THz traveling wave tube amplifiers (TWTAs) [1]. The structure is unique in that it has microscale diameter dimensions, a helical geometry, and potentially higher electrical/thermal conductivity than other previously reported self-assembled helices. We fabricate the SWSs by self-assembly of a metal nanomembrane (NM) and electroplating of the helical film, thereby overcoming the inherent trade-off between NM thickness and helix diameter/pitch. Here we focus on our recent efforts on understanding and controlling electrodeposition on freestanding and curved NMs. In particular, we investigate the effect of varying current and temperature on the thickness of the deposited film and its uniformity. In addition, we characterize and justify any changes occurring in the geometry of the helix due to electrodeposition. Fabrication of microscale SWSs started with defining NMs strips by optical lithography, metal evaporation, and lift-off. In our work we used Cr/Au/Cr NMs that have an inherent stress gradient across their thickness. We defined the metal strips onto a bulk Si substrate coated with a sacrificial or release layer such as Ge or GeO, . The released NMs selfassembled into helices with a diameter between 100 and 200 mm upon selective etching of the sacrificial layer by a XeF,, plasma. Helix diameter and pitch are controlled by varying the thickness, the elastic modulus, the residual stress, and the in plane geometry of the NMs. The electroplating procedure employs a two-electrode electrochemical setup with the self-assembled helix and a platinized mesh serving as cathode and anode, respectively. A sulphatebased gold solution is used for the process. We use pulsed current source with variable frequencies to obtain uniform electrodeposition of an Au film with a thickness of 1-2 mm depending on the electrodeposition parameters. In all investigated electroplating conditions, we observed uniformly deposited gold with lower roughness than the skin depth of the metal at THz frequencies, which are crucial attributes to minimize losses and heating in a TWTA based on the helical SWS. Reference [1] Prakash, Divya J., Matthew M. Dwyer, Marcos Martinez Argudo, Mengistie L. Debasu, Hassan Dibaji, Max G. Lagally, Daniel W. van der Weide, and Francesca Cavallo. 2020. “Self-Winding Helices as Slow-Wave Structures for Sub-Millimeter Traveling-Wave Tubes.” ACS Nano, 2021, 15, 1229-1239. doi:10.1021/acsnano.0c08296. Work supported by U.S. AFOSR-Award No. FA9550-19-1-0086

Electroplated Helical Slow-Wave Structures for Millimeter-Through-THz Vacuum Electronic Devices

B-phase MgGaO ternary alloy thin films of different Mg compositions were grown by using oxygen plasma assisted molecule beam epitaxy (MBE). The atomic ratios of Mg and Ga in sample A, B, and C were recorded as 11.15 at.%/88.85 at.%, 23.29 at.%/76.71 at.%, and 36.35 at.%/63.65 at.%, respectively, which were measured by energy-dispersive X-ray (EDX) spectroscopy. The samples were characterized by a scanning electron microscope (SEM), showing that the surface becomes rougher in samples with higher Mg contents. Three typical monoclinic structure diffraction peaks located at 19.12°, 38.50°, and 59.14°, which correspond to the B phases of (-201),(-402), and (-603) were investigated by X-ray diffraction (XRD), and all these diffraction peaks exhibit an obvious blue shift compared with that of reference B-Ga,0.. Raman spectra were investigated by a confocal Raman microscope equipped with a high-resolution laser centered at 532 nm with a power of 60 mW. Four out of ten A_ and four out of five B_ Raman-active optical phonon modes were observed. The reduced peak intensity of A (2)(9) A (10) B (2)- (3), and B 8 indicates that Mg * atoms have replaced the position of Ga * ions in both octahedral and tetrahedral chains. The ultra-wide bandgap absorption and transmittance spectra were measured using a UV-Vis-NIR spectrophotometer. Ultra-wide bandgaps of 4.98, 5.05, 5.20, and 5.30 eV were measured for pure Ga,0x, sample A, B, and C, respectively with all transmittances of light over 80% in the visible spectrum region. Temperature-dependent photoluminescence (PL) spectra were investigated by a 325-nm laser with temperatures ranging from 13 to 300 K, revealing only intermediate level transitions, which decrease with the increase of temperature. A MSM solar-blind ultraviolet photodetector (PD) was fabricated based on MgGaO films with Pt/Au electrodes. Photoresponse has been studied by typical dark and irradiated wavelength dependent I-V curves.

Characterization of β-Phase MgGaO Thin Films Grown by Molecular Beam Epitaxy

Gallium oxide (Ga2o3.) is an ultra-wide bandgap semiconductor (4.85-5.3 eV) with emerging applications in power electronics because of its high thermal stability, large breakdown voltage and high dielectric constant. Ga, 03 occurs in the a, B, y and k(€) polymorphs having trigonal, monoclinic, cubic normal spinel and orthorhombic structures, respectively. B- “Ga, 03 is thermodynamically stable and produced in melt-grown substrates. Recently, k- -Ga,0. has accrued increased interest because it has demonstrated stability at temperatures 2 700 °c! and its large spontaneous polarization could be exploited in heterostructures” for high electron mobility transistors. In this study, metal organic chemical vapor deposition (MOCVD) was used to grow epitaxial Ga,0. films under different growth conditions including temperature, growth rate, and diluent gas flow rate. The growth experiments were conducted in a vertical, low-pressure, cold-wall MOCVD reactor using triethylgallium (TEGa) and 0, precursors and commercially available 0.15 off-axis (toward m-plane) c-plane sapphire. No was used as both the carrier gas for TEGa and the diluent gas. The chamber pressure was held at 20 Torr. The growth rate was controlled by the TEGa flow rate which ranged from 0.29 sccm to 2.1 sccm. The diluent gas flow rate was varied from 4 slm to 8 sim. The O flow rate was maintained at 500 sccm. The temperature was constant for each growth but varied from 470°C to 570 C. Phase composition and microstructural evolution of the films were investigated using different analytical tools including the X-ray diffraction (XRD), scanning electron microscopy (SEM), high-resolution transmission electron microscopy (HRTEM) and scanning transmission electron microscopy (STEM). XRD and SEM revealed that the top layer varied in phase composition between ~100% k-Ga,0, and ~100% B-Ga,0.; the surface microstructure ranged from poorly coalesced to completely coalesced grains as a function of growth temperature, growth rate, or diluent gas flow rate. In general, it was found that the k-phase tends to grow at lower temperatures and higher growth rates (higher TEGa flow rates). The growth of nominally single- phase k-Ga 90 within the top layer was observed in a temperature range between 500° C and 530° C. Below 470° C, only amorphous Ga 90,was obtained; above 570° Cc only the B-phase was deposited. HRTEM and STEM of a film grown at 530° C revealed the initial pseudomorphic growth of 3-4 monolayers of a-Ga,0., a 20-60 nm transition layer that contained both B- and y-Ga,0., and a top ~700 nm thick layer of phase-pure k-Ga,0,. H.Y. Playford, A.C. Hannon, E.R. Barney, and R.|. Walton, Chem. Eur. J. 19, 2803 (2013). R. Roy, V.G. Hill, and E.F. Osborn, J. Am. Chem. Soc. 74, 719 (1952). Y. Oshima, E.G. Villora, Y. Matsushita, S. Yamamoto, and K. Shimamura, J. Appl. Phys. 118, 085301 (2015). J. Lee, H. Kim, L. Gautam, K. He, X. Hu, V. Dravid, and M. Razeghi, Photonics 8, 17 (2021). P. Ranga, S.B. Cho, R. Mishra, and S. Krishnamoorthy, Appl. Phys. Express 13, 061009 (2020). On K. Jiang, J. Tang, M. J. Cabral, A. Park, L. Gu, R. F. Davis and L. M. Porter, J. Appl. Phys. (2022)

Phase Composition and Microstructure of κ-Ga2O3 Heteroepitaxial Films Grown by MOCVD

Parametric expressions that describe the complex dielectric function of polycrystalline Mg 2n 1 films as a function of the bandgap E_ have been applied in the metrology of cadmium telluride-based solar cell structures. Valid over the range 3.264 < 3.826 eV, these parametric expressions are applied to facilitate mapping spectroscopic ellipsometry of device structures consisting of Mg Zn, _O deposited on TEC™-15 glass. Mapping spectroscopic ellipsometry is shown to provide maps for the Mg 2n,_,0 effective thickness, or volume per planar substrate area, and Mg 2n,_,0 bandgap within confidence limits of + 1 nm and + 0.003 eV, respectively. As a second application of the parametric expressions, an as-deposited cadmium telluride-based device structure has been analyzed by through-the-glass spectroscopic ellipsometry. Such an analysis is also shown to provide the Mg,2n, 0 effective thickness and bandgap. The outcome of the through-the-glass spectroscopic ellipsometry analysis for this device structure enables simulations of the external quantum efficiency spectra of the resulting solar cell assuming different recombination losses within the individual layers of the structure. A comparison of these simulations with corresponding experimental spectra reveals high current collection from the front of the device with the almost complete collection from the cadmium sulfide components of the Mg,2n _O/cadmium sulfide and cadmium sulfide/cadmium telluride interface layers of the structure.

An Optical Response Analysis of Cadmium Telluride-Based Solar Cell Structures Incorporating High Resistivity Transparent Layers of Magnesium-Zinc Oxide

Transition metal multinary chalcogenides are a unique class of material for photovoltaic (PV) and quantum dot light emitting (QLED) applications. They offer a higher level of flexibility from a materials perspective to achieve tunable band gaps through structural and compositional control. We aim to realize new and more sustainable direct band-gap semiconductor materials from multinary compositions of earth-abundant elements. To this end, detailed ab initio calculations are first performed for the parent binary Cu-chalcogenide composition. A series of new p-doped, n-doped, and degenerate semiconductors based on multinary Cu-chalcogenide nanocrystals are predicted using the theoretical investigations of the band gap and density of states. We experimentally realized the new multinary Cu-chalcogenide compositions via novel solution chemistry approaches. A detailed material characterization based on x-ray diffraction, electron microscopy, and ultraviolet-visible spectroscopy is reported for the new Cuchalcogenide nanocrystals and thin films. These new results will provide key theoretical and experimental insights for achieving more sustainable absorber layer materials for inorganic solar cells.

Designing New Sustainable Semiconductor Materials with Cu-Chalcogenide Nanocrystals

A new class of magnetic metamaterials can be produced by placing magnetic molecules between two ferromagnetic electrodes. Magnetic molecules can be designed to create strong exchange coupling between two ferromagnetic electrodes. Magnetic molecule-induced strong exchange coupling that can be harnessed as the means to initiate an automatic spin filtering process to shuffle the spin density of states in the two ferromagnetic electrodes. Our recent preliminary experimental studies showed that iron and nickel atom-based organometallic molecules (FeNiOMC) produced a strong spin filtering effect between NiFe and Co/NiFe ferromagnetic electrodes (Tyagi, P.; Riso, C., Molecular spintronics devices exhibiting properties of a solar cell. Nanotechnology 2019, 30 (49), 49540). Fe-Ni-OMC dramatically modified the magnetic properties of ferromagnetic electrodes present in the magnetic tunnel junction testbed at room temperature. This paper talks about the resulting Fe-NiOMC and magnetic tunnel junctionbased magnetic metamaterial. This metamaterial showed an intriguing solar cell effect. Further experimental studies are conducted to design several other forms of novel magnetic metamaterials by combining a wide range of Singlemolecule magnetic (SMM) and porphyrin-like magnetic molecules with various permutations of magnetic electrodes. To produce a magnetic tunnel junction testbed that is to be used to test various molecules, we made extensive experimental optimization efforts at each stage of device fabrication to address device fabrication challenges associated with molecule-ferromagnet interactions. In this study, we have systematically optimized several factors using multiple Taguchi Design of Experiment approaches to improve the fabrication of the bottom electrode of the magnetic metamaterial. We systematically varied photolithography parameters and plasma treatment process to yield a bottom ferromagnetic electrode with

Solar Cell Effect with Magnetic Metamaterials Produced by Combining Magnetic Tunnel Junction and Magnetic Molecules

Reliable and controllable growth of two dimensional hexagonal boron nitride (h-BN) is essential for its wide range of applications. Here, we carried out molecular beam epitaxy (MBE) growth of monolayer h-BN on nickel (Ni) substrates by using ammonia (NH) and boron oxide (B04) as the sources. It was found that compared with normal growth of h-BN by introducing the NH, and B,0., source at the same time, the pretreatment with B,0, greatly inhibited the following epitaxial growth of h-BN. With the increase of B,0, pretreatment time and then followed by 15 min h-BN growth, the coverage of h-BN exhibited an exponential decay relationship with the treatment time. The samples were analyzed using depth-profiling X-ray photoelectron spectroscopy (XPS) and reflection high-energy electron diffraction (RHEED). The oxygen/Ni peak intensity ratios derived from the XPS showed that the oxygen concentration increased with the increase of B,0, pretreatment time from the surface while decreased gradually with the increase of XPS sputtering depth. Meanwhile, boron incorporation was negligible regardless of treatment time. The epitaxial relationship was studied by using h-BN coverage-dependent RHEED analysis, obtained on samples including Ni substrate after annealing only, Ni substrate with BAO pretreatment, h-BN samples with the B,03 pretreatment and subsequent 50%, and 100% film coverage. Only two sets of patterns were observed and separated by 30, confirming that there is no in-plane rotation between the h-BN and Ni (111) surface. The oxygen incorporated into the Ni substrate not only suppressed h-BN nucleation by passivating the Ni surface active sites, but also decreased the growth rate of single crystal h-BN domains, shifting from larger sizes to smaller shapes. Detailed structural, optical and electrical characterizations of these monolayer h-BN films will also be presented. Our work reveals the influence of oxygen on the catalytic effect of Ni in the MBE growth of monolayer h-BN and provides an efficient way to control the morphology and quality of the 2D h-BN films.

Molecular Beam Epitaxy Growth of Monolayer Hexagonal Boron Nitride on Nickel Substrates Incorporated with Oxygen Atoms

Exchange bias, a magnetic phenomenon emerging at the interface between a ferromagnet (FM) and an antiferromagnet (AFM), originates from the pinning of spins in the FM layer by the adjacent AFM layer, effectively stabilizing the FM against the environmental stray magnetic fields. This phenomenon has been implemented in a wide range of technological applications including magnetic memories, magnetic read heads, and sensors, and plays the foundational role for advanced spintronics. Here, we report an approach to establishing strong exchange bias in the 2D magnet/oxide heterostructure by a single-step process — atomic layer deposition (ALD). The exchange bias we created is controllable in the range of ~300 to 1500 Oe by replacing different oxides in the heterostructure. Hence, this systematic work elucidates the essential role of ALD-enabled oxidization of layered magnets in the formation of exchange bias and establishes ALD of oxides as a facile, controllable, and generally effective approach to creating giant exchange bias in vaW magnets, representing an integral advance towards practical vdW spintronic devices.

Premium content

Interface Trap Density Characterization of ALD Gate Dielectrics for GaN Power MOSFETs

Downloads

Next from EMC 2022

Meta-Optic Accelerators for Object Classifiers

Similar lecture

Process-Induced Variability of SiC DMOSFET Short Circuit Capability