Search

Non-linear output characteristics and the related turn-on voltages of GaN-based current aperture vertical electron transistors (CAVETs) are investigated experimentally. The resistive components are systematically analyzed in dependence of the device layout to determine the dominant resistances in the devices. Current–voltage (IV) and capacitance–voltage-characteristics (CV) are compared to a proposed planar-doped barrier diode (PDBD) model, and the influence of the bound sheet charge density and drift layer carrier concentration is discussed. The observed CV characteristics are in contrast to the PDBD model as a clearly voltage-dependent capacitance was observed and dopant-diffusion forming a p-type aperture was ruled out. Thermionic emission was verified by temperature-dependent IV characteristics indicating interface states causing a potential barrier. Transient drain current measurements revealed a single dominating trap level with an activation energy of EA = 1.086 ± 0.015 eV. This activation energy was attributed to carbon-related acceptor states present at the regrowth interface and the drift layer. Additional test structures revealed that the interface potential barrier and the space charge in the drift layer limit the initial charge transport causing a turn-on voltage in the devices. The results point out the significance of a precise control of the regrowth interface properties and the effective carrier density in the drift layer to enable efficient, high-power devices based on the CAVET technology. [ABSTRACT FROM AUTHOR]

Although highly desirable, the nature of wood-cellulose in its native state has remained difficult to comprehend. Contrary to the traditional acceptance of wood-cellulose being crystalline, in 2016, the authors' research found that the cellulose was not crystalline. Here, additional evidence is presented that further supports the non-crystalline model. One of the key pieces of evidence was obtained by 64% H2SO4 hydrolysis of tension- and opposite-aspen woods (TW and OW, respectively). The TW (G-layer rich) yielded significant amount of CNCs (TW-CNCs, 20.7%), the OW yielded a much lower amount (OW-CNCs, 5.4%). Although a higher yield of TW-CNCs was expected due to the presence crystalline cellulose in the G-layer, the lower yield of the OW-CNCs was a surprise because, assuming absence of G-layer, based on the authors' earlier findings no CNCs were expected to be generated. To explain this anomaly, anatomical examination of the woods using stains was carried out which showed that some OW fibers also contained the crystalline G-layer and therefore, provided an explanation as to why the OW-CNCs were produced. The results clearly showed that the acid hydrolysis did not destroy the crystalline cellulose and therefore, in the case of a normal (G-layer free) wood which, as previously reported had not generated CNCs, the cellulose must have been non-crystalline. An additional indication of the wood's S2 cellulose being not crystalline was the absence of the 93 cm−1 Raman band in the low frequency spectrum of the TW S2 layer. Further evidence was obtained by comparing low frequency Raman spectra of TW-CNCs, TW-holopulp, and aspen-holopulp as well as the mixture-samples of crystalline cellulose and xylan at the concentration levels of their occurrence in these holopulps. Overall, these findings provided further support to the contention that the native wood-cellulose is non-crystalline. [ABSTRACT FROM AUTHOR]