Search for: All records

We evaluate the ability of large language models (LLMs) to provide PropBank semantic role label annotations across different realizations of the same verbs in transitive, intransitive, and middle voice constructions. In order to assess the meta-linguistic capabilities of LLMs as well as their ability to glean such capabilities through in-context learning, we evaluate the models in a zero-shot setting, in a setting where it is given three examples of another verb used in transitive, intransitive, and middle voice constructions, and finally in a setting where it is given the examples as well as the correct sense and roleset information. We find that zero-shot knowledge of PropBank annotation is almost nonexistent. The largest model evaluated, GPT-4, achieves the best performance in the setting where it is given both examples and the correct roleset in the prompt, demonstrating that larger models can ascertain some meta-linguistic capabilities through in-context learning. However, even in this setting, which is simpler than the task of a human in PropBank annotation, the model achieves only 48% accuracy in marking numbered arguments correctly. To ensure transparency and reproducibility, we publicly release our dataset and model responses. 

β -Ga2O3 is actively touted as the next ultrawide bandgap material for power electronics. To fully utilize its high intrinsic critical electric field, development of high-quality robust large-barrier height junctions is essential. To this end, various high-work function metals, metal oxides, and hole-conducting oxides have been deposited on Ga2O3, primarily formed by sputter deposition. Unfortunately, reports to date indicate that measured barrier heights often deviate from the Schottky–Mott model as well as x-ray photoelectron spectroscopy (XPS) extractions of conduction band offsets, suggesting significant densities of electrically active defects at these junctions. We report Schottky diodes made from noble metal oxides, IrO2 and RuO2, deposited by ozone molecular beam epitaxy (ozone MBE) with barrier heights near 1.8 eV. These barriers show close agreement across extraction methods and robust to high surface electric fields upward of 6 MV/cm and 60 A/cm2 reverse current without degradation.

 

Here, we report that a source of Si impurities commonly observed on (010) β-Ga2O3 is from exposure of the surface to air. Moreover, we find that a 15 min hydrofluoric acid (HF) (49%) treatment reduces the Si density by approximately 1 order of magnitude on (010) β-Ga2O3 surfaces. This reduction in Si is critical for the elimination of the often observed parasitic conducting channel, which negatively affects transport properties and lateral transistor performance. After the HF treatment, the sample must be immediately put under vacuum, for the Si fully returns within 10 min of additional air exposure. Finally, we demonstrate that performing a 30 min HF (49%) treatment on the substrate before growth has no deleterious effect on the structure or on the epitaxy surface after subsequent Ga2O3 growth.

 

We demonstrate epitaxial lattice-matched Al0.89Sc0.11N/GaN 10 and 20 period distributed Bragg reflectors (DBRs) grown on c-plane bulk n-type GaN substrates by plasma-assisted molecular beam epitaxy. Resulting from a rapid increase in in-plane lattice coefficient as scandium is incorporated into AlScN, we measure a lattice-matched condition to c-plane GaN for a Sc content of just 11%, resulting in a large refractive index mismatch Δn greater than 0.3 corresponding to an index contrast of Δn/nGaN = 0.12 with GaN. The DBRs demonstrated here are designed for a peak reflectivity at a vacuum wavelength of 400 nm, reaching a reflectivity of 0.98 for 20 periods. It is highlighted that AlScN/GaN multilayers require fewer periods for a desired reflectivity than other lattice-matched Bragg reflectors such as those based on AlInN/GaN multilayers.

 

Using low-temperature cathodoluminescence spectroscopy, we study the properties of N- and Al-polar AlN layers grown by molecular beam epitaxy on bulk AlN{0001}. Compared with the bulk AlN substrate, layers of both polarities feature a suppression of deep-level luminescence, a total absence of the prevalent donor with an exciton binding energy of 28 meV, and a much increased intensity of the emission from free excitons. The dominant donor in these layers is characterized by an associated exciton binding energy of 13 meV. The observation of excited exciton states up to the exciton continuum allows us to directly extract the Γ5 free exciton binding energy of 57 meV. 

Abstract We report the growth of α -Ga 2 O 3 on m -plane α -Al 2 O 3 by conventional plasma-assisted molecular-beam epitaxy and In-mediated metal–oxide-catalyzed epitaxy (MOCATAXY). We report a growth rate diagram for α -Ga 2 O 3 ( 10 1 ¯ 0 ), and observe (i) a growth rate increase, (ii) an expanded growth window, and (iii) reduced out-of-lane mosaic spread when MOCATAXY is employed for the growth of α -Ga 2 O 3 . Through the use of In-mediated catalysis, growth rates over 0.2 μ m h −1 and rocking curves with full width at half maxima of Δ ω ≈ 0.45° are achieved. Faceting is observed along the α -Ga 2 O 3 film surface and explored through scanning transmission electron microscopy. 

A D-band (110‒170 GHz) SiC substrate-integrated waveguide (SIW) is characterized on-wafer by two different vector network analyzers (VNAs): a 220-GHz single-sweep VNA and an 110-GHz VNA with WR8 (90‒140 GHz) and WR5 (140‒220 GHz) frequency extenders. To facilitate probing, the SIW input and output are transitioned to grounded coplanar waveguides (GCPWs). Two-tier calibration is used to de-embed the SIW-GCPW transitions as well as to extract the intrinsic SIW characteristics. In general, the two VNAs are in agreement and both result in an ultra-low insertion loss of approximately 0.2 dB/mm for the same SIW, despite stitching errors at band edges. 

We report controlled silicon doping of Ga2O3 grown in plasma-assisted molecular beam epitaxy. Adding an endplate to the Si effusion cell enables the control of the mobile carrier density, leading to over 5-orders of magnitude change in the electrical resistivity. Room temperature mobilities >100  cm2/V s are achieved, with a peak value >125  cm2/V s at a doping density of low-1017/cm3. Temperature-dependent Hall effect measurements exhibit carrier freeze out for samples doped below the Mott criterion. A mobility of 390  cm2/V s is observed at 97  K.

 

Neutron production in antineutrino interactions can lead to bias in energy reconstruction in neutrino oscillation experiments, but these interactions have rarely been studied. MINERvA previously studied neutron production at an average antineutrino energy of ∼3 GeV in 2016 and found deficiencies in leading models. In this paper, the MINERvA 6 GeV average antineutrino energy dataset is shown to have similar disagreements. A measurement of the cross section for an antineutrino to produce two or more neutrons and have low visible energy is presented as an experiment-independent way to explore neutron production modeling. This cross section disagrees with several leading models’ predictions. Neutron modeling techniques from nuclear physics are used to quantify neutron detection uncertainties on this result.