박정우 교수

Magmas enriched in Cu and Au are likely to be the most prospective for magmatic–hydrothermal deposits of these metals. However, the mechanism that leads to the formation of metal-rich magmas is not well constrained. We report major, trace and platinum group element (PGE) data for the Niuatahi–Motutahi lavas, Tonga rear arc, with the aim of studying their petrogenesis with special emphasis on the evolution of chalcophile elements during magmatic differentiation. Major and trace element contents,

Advances in neural network and machine learning algorithms have sparked a wide array of research in specialized hardware, ranging from high-performance convolutional neural network (CNN) accelerators to energy-efficient deep-neural network (DNN) edge computing systems [1]. While most studies have focused on designing inference engines, recent works have shown that on-chip training could serve practical purposes such as compensating for process variations of in-memory computing [2] or adapting to

Platinum group element (PGE) abundances in the upper continental crust (UCC) are poorly constrained with published values varying by up to an order of magnitude. We evaluated the validity of using loess to estimate PGE abundances in the UCC by measuring these elements in seven Chinese loess samples using a precise method that combines NiS fire assay with isotope dilution. Major and trace elements of the Chinese loess show a typical upper crustal composition and PGE abundances are consistent with

Chalcophile element fertility, the chalcophile metal abundance in the source magma, is likely to be a critical factor for the formation of porphyry Cu ± Au deposits. In this study, we provide evidence to support this hypothesis by comparing the platinum group element (PGE) geochemistry of barren and ore-bearing Cu ± Au granitic suites. We report the PGE contents of three barren volcanic and subvolcanic suites from Argentina and Japan and two Cu ± Au bearing suites from Indonesia and Chile. These

Recent advances in neural network (NN) and machine learning algorithms have sparked a wide array of research in specialized hardware, ranging from high-performance NN accelerators for use inside the server systems to energy-efficient edge computing systems. While most of these studies have focused on designing inference engines, implementing the training process of an NN for energy-constrained mobile devices has remained to be a challenge due to the requirement of higher numerical precision. In

Recent works on mobile deep-learning processors have presented designs that exploit sparsity [2, 3], which is commonly found in various neural networks. However, due to the shift in the machine learning community towards using non-sparse activation functions such as Leaky ReLU or Swish for better training convergence, state-of-theart models no longer exhibit the sparsity found in conventional ReLU-based models (Fig. 9.3.1, top). Moreover, contrary to error-tolerant image classification tasks, mo

Indium oxide (In2O3) thin films were deposited via thermal atomic layer deposition (ALD) to exploit their potential as semiconductors in thin-film transistors (TFTs), using a new liquid type indium complex precursor (In(CH3)3[CH3OCH2CH2NHtBu]). In2O3 films were deposited successfully at lower temperatures and exhibited a satisfactory growth rate (∼0.35 Å per cycle). In addition, we investigated the effect of deposition temperature from 100 to 250 °C on the microstructure and chemical and physica

Recent experimental studies and in situ LA-ICP-MS analysis on natural Cr-spinel have shown that Rh and IPGEs (Ir-group platinum group elements: Ru, Ir, Os) are enriched in the lattice of Cr-spinel. However, the factors controlling the partitioning behaviour of these elements are not well constrained. In this study, we report the Rh, IPGE, and trace element contents in primitive Cr-spinel, measured by LA-ICP-MS, from nine volcanic suites covering various tectonic settings including island arc pic

SnO _x thin films were successfully deposited by the thermal atomic layer deposition (ALD) method using N, N'- tert-butyl-1,1-dimethylethylenediamine stannylene(II) as a precursor and ozone and water as reactants. The growth of SnO and SnO₂ films could be easily controlled by employing different reactants and utilizing different ozone and water concentrations, respectively. The formation of both SnO and SnO₂ films exhibited typical surface-limiting reaction charac