No abstract

Nanocrystalline diamond (NCD) films were grown by hot filament CVD and the precursor composition dependence of the structural properties was examined. Films grown at 1 and 2 CH4 Vol% were found to be NCD layers with grain sizes of ~23-25 nm while films grown at 3-5 Vol% were identified as the mixtures of microcrystalline diamond and graphitic phase. The sp2/sp3 bonded carbon ratio in the grown films increased as the CH4 content increased up to 3 Vol% and then decreased beyond 4 Vol%. Microstructure and deposition rate were also found to be affected by the precursor composition and the NCD film grown at 1 CH4 Vol% showed a very dense microstructure and the highest deposition rate of ~3 nm/min.

α-Ga2O3 has the largest bandgap (~5.3 eV) among the five polymorphs of Ga2O3 and is a promising candidate for high power electronic and optoelectronic devices. To fabricate various device structures, it is important to establish an effective dry etch process which can provide practical etch rate, smooth surface morphology and low ion-induced damage. Here, the etch characteristics of α-Ga2O3 epitaxy film were examined in two fluorine-based (CF4/Ar and SF6/Ar) inductively coupled plasmas. Under the same source power, rf chuck power and process pressure, an Ar-rich composition of CF4/Ar and an SF6-rich composition of SF6/Ar produced the highest etch rates. Monotonic increase in the etch rate was observed as the source power and rf chuck power increased in the 2CF4/13Ar discharges, and a maximum etch rate of ~855 Å/min was obtained at a 500 W source power, 250 W rf chuck power, and 2 mTorr pressure. A smooth surface morphology with normalized roughness of less than ~1.38 was achieved in the 2CF4/13Ar and 13SF6/2Ar discharges under most of the conditions examined. The features etched into the α-Ga2O3 layer using a 2CF4/13Ar discharge with 2 mTorr pressure showed good anisotropy with a vertical sidewall profile.

Deep convolutional neural network (CNN) helped enhance image quality of cone-beam computed tomography (CBCT) by generating synthetic CT. Most of the previous works, however, trained network by intensity-based loss functions, possibly undermining to promote image feature similarity. The verifications were not sufficient to demonstrate clinical applicability, either. This work investigated the effect of variable loss functions combining feature- and intensity-driven losses in synthetic CT generation, followed by strengthening the verification of generated images in both image similarity and dosimetry accuracy. The proposed strategy highlighted the feature-driven quantification in (1) training the network by perceptual loss, besides L1 and structural similarity (SSIM) losses regarding anatomical similarity, and (2) evaluating image similarity by feature mapping ratio (FMR), besides conventional metrics. In addition, the synthetic CT images were assessed in terms of dose calculating accuracy by a commercial Monte-Carlo algorithm. The network was trained with 50 paired CBCT-CT scans acquired at the same CT simulator and treatment unit to constrain environmental factors any other than loss functions. For 10 independent cases, incorporating perceptual loss into L1 and SSIM losses outperformed the other combinations, which enhanced FMR of image similarity by 10%, and the dose calculating accuracy by 1–2% of gamma passing rate in 1%/1mm criterion.