• 한국전기전자재료학회논문지
• /
• 제13권10호
• /
• pp.840-846
• /
• 2000

In this work, small-size, high-performance solenoid-type RF chip inductors utilizing a low-loss Al$_2$O$_3$core material were investigated. The size of the chip inductors fabricated in this work were 15$\times$10$\times$0.7㎣, 2.1$\times$1.5$\times$10㎣, and 2.4$\times$2.0$\times$1.4㎣ and copper (Cu) wire with 40 ㎛ diameter was used as the coils. High frequency characteristics of the inductance, quality factor, and impedance of developed inductors were measured suing an RF Impedance/Material Analyzer (HP4291B with HP16193A test fixture). It was observed that the developed inductors with the number of turns of 7 have the inductance of 33 to 100nH and exhibit the self-resonant frequency (SRF) of .26 to 1.1 GHz. The SRF of inductors decreases with increasing the inductance and the inductors have the quality factor of 60 to 80 in the frequency range of 300 MHz to 1.1 GHz. In this study, small-size solenoid-type RF chip inductors with high inductance and high quality factor were fabricated successfully. It is suggested that the thin film-type inductor is necessary to fabricate the smaller size inductors at the expence of inductance and quality factor values.

• Transactions on Electrical and Electronic Materials
• /
• 제17권4호
• /
• pp.217-221
• /
• 2016

In this work, open and short stubs that were fabricated on the silicon substrate and for which the periodically arrayed grounded-strip structure (PAGS) was employed were studied along with their basic RF characteristics for an applicability regarding the RF-matching components. The PAGS-employing open and short stubs showed losses that are much lower than that of the conventional stub on the silicon substrate. Concretely, the Q values of the open and short stubs are 9 and 10.2, respectively, while the Q value of the conventional open stub is 2.5. With the use of the PAGS-employing open and short stubs, a highly miniaturized harmonic-rejection filter was also fabricated on the silicon substrate. The filter exhibited a comparatively sound harmonic-suppression characteristic at n × 13 GHz, and its size is 0.1 mm², which is only 7% of the size of the conventional filter on the silicon substrate.

• Journal of electromagnetic engineering and science
• /
• 제1권1호
• /
• pp.83-87
• /
• 2001

Many communication systems require bandpass filters with sharp skirt frequency characteristics in order to avoid the interferencce, which results in more order in the filter design. However, because of the limited Q values bandpass filters made of small sized ceramic resonators suffer from relatively large ripples at the band edges as the order of the filter increases. In order to compensate the large ripples while maintaining the sharp skirt frequency we propose a new RF amplitude equalizer. The equalizer made of two pole bandpass filter and an amplifier whose amplitude characteristics are the reverse of those of the bandapss filter. At the cellular band 9-pole bandpass filter with 10 MHz bandwidth exhibits 3 dB ripple when 8mm*8mm ceramic coaxial resonators are used. We added the RF equalizer to this filter and the flatness is improved as less than 1 dB.

• 한국진공학회:학술대회논문집
• /
• 한국진공학회 2011년도 제41회 하계 정기 학술대회 초록집
• /
• pp.124-124
• /
• 2011

This study investigated the interaction of varied plasma power with ultralow-k toluene-tetraethoxysilane (TEOS) hybrid plasma polymer thin films, as well as changing electrical and mechanical properties. The hybrid thin films were deposited on silicon(100) substrates by plasma enhanced chemical vapor deposition (PECVD) system. Toluene and tetraethoxysilane were utilized as organic and inorganic precursors. In order to compare the electrical and the mechanical properties, we grew the hybrid thin films under various conditions such as rf power of plasma, bubbling ratio of TEOS to toluene, and post annealing temperature. The hybrid plasma polymer thin films were characterized by Fourier transform infrared (FT-IR) spectroscopy, atomic force microscopy (AFM), nanoindenter, I-V curves, and capacitance. Also, the hybrid thin films were analyzed by using ellipsometry. The refractive indices varied with the RF power, the bubbling ratio of TEOS to toluene, and the annealing temperature. To analyze their trends of electrical and mechanical properties, the thin films were grown under conditions of various rf powers. The IR spectra showed them to have completely different chemical functionalities from the liquid toluene and TEOS precursors. Also, The SiO peak intensity increased with increasing TEOS bubbling ratio, and the -OH and the CO peak intensities decreased with increasing annealing temperature. The AFM images showed changing of surface roughness that depended on different deposition rf powers. An nanoindenter was used to measure the hardness and Young' modulus and showed that both these values increased as the deposition RF power increased; these values also changed with the bubbling ratio of TEOS to toluene and with the annealing temperature. From the field emission scanning electron microscopy (FE-SEM) results, the thickness of the thin films was determined before and after the annealing, with the thickness shrinkage (%) being measured by using SEM cross-sectional images.

• 한국정보기술응용학회:학술대회논문집
• /
• 한국정보기술응용학회 2002년도 추계공동학술대회 정보환경 변화에 따른 신정보기술 패러다임
• /
• pp.363-370
• /
• 2002

RF에서 동작하는 초미세 공정 MOS 트랜지스터의 기판 효과에 따른 기판회로망과 물리적 의미를 가지는 파라미터 추출법이 고려되었다. 제안된 기판 회로망에는 단일의 저항과 링 -형태의 기판 콘택에 의해 생성된 인덕터가 포함되었다 모델 파라미터는 최적화 과정없이 단절된 게이트와 공통-벌크 구성 을 갖는 MOS 트랜지스터에서 측정 된 S-파라미터로부터 추출된다. 제안된 기술은 다양한 크기 의 MOS 트랜지스터에 적용되어 졌다. 추출된 기 판 회 로망을 이 용한 가상실험 결과와 측정치는 약 30GHz까지 일치함을 검증하였다.

• 한국산업정보학회:학술대회논문집
• /
• 한국산업정보학회 2002년도 추계공동학술대회
• /
• pp.363-370
• /
• 2002

RF에서 동작하는 초미세 공정 MOS 트랜지스터의 기판 효과에 따른 기판 회로망과 물리적 의미를 가지는 파라미터 추출법이 고려되었다. 제안된 기판 회로망에는 단일의 저항과 링-형태의 기판 콘택에 의해 생성된 인덕터가 포함되었다. 모델 파라미터는 최적화 과정 없이 단절된 게이트와 공통-벌크 구성을 갖는 MOS 트랜지스터에서 측정된 S-파라미터로부터 추출된다. 제안된 기술은 다양한 크기의 MOS 트랜지스터에 적용되어졌다. 추출된 기판 회로망을 이용한 가상실험 결과와 측정치는 약 300Hz까지 일치함을 검증하였다.

• 한국습지학회지
• /
• 제25권3호
• /
• pp.167-175
• /
• 2023

강우 데이터는 습지관리, 수문모의, 수자원 관리와 같은 다양한 분야에서 활용되는 필수 입력자료 중 하나이다. 강우 데이터를 활용하여 효율적인 수자원관리를 위해서는 기본적으로 데이터의 결측률을 최소화 시킴으로써 최대한 많은 데이터를 확보하는 것이 필수적이다. 또한 미계측 지역에 대한 강우 데이터를 확보한다면 보다 효율적인 수문모의가 가능하다. 그러나 결측 강우 데이터는 주로 통계학적 기법에 의해 추정되어 왔다. 본 연구의 목적은 데이터 간의 상관관계를 기반으로 새로운 데이터를 예측할 수 있는 머신러닝 알고리즘을 활용하여 결측 강우 데이터를 복원할 수 있는 새로운 방법을 제안하고자 한다. 또한, 기존의 통계적 방법들과 비교하여 머신러닝 기법의 결측 강우 데이터 복원을 위한 활용가치를 평가하고자 한다. 평가를 위해 대표적인 머신러닝 알고리즘인 Artificial Neural Network (ANN)과 Random Forest (RF)을 적용하였다. 강우의 발생 유무를 분류하는 성능은 RF 알고리즘이 ANN 알고리즘보다 강우 발생유무의 분류 정확도가 높은 것으로 나타났다. 분류 모형의 평가 지표인 F1-score나 Accuracy값이 RF는 0.80, 0.77인 반면에, ANN은 0.76, 0.71로 계산되었다. 또한 강우량을 추정하는 성능 역시 RF가 ANN 알고리즘보다 보다 높은 정확도를 보였다. RF과 ANN 알고리즘의 RMSE은 2.8mm/day과 2.9mm/day이고, R²값은 0.73, 0.68으로 계산되었다.

• Journal of Information Display
• /
• 제12권2호
• /
• pp.109-113
• /
• 2011

This paper summarizes the recent fabrication and characterizations of flexible high-speed radio frequency (RF) transistors, PIN-diode single-pole single-throw switches, as well as flexible inductors and capacitors, based on single-crystalline Si nanomembranes transferred on polyethylene terephthalate substrates. Flexible thin-film transistors (TFTs) on plastic substrates have reached RF operation speed with a record cut-off/maximum oscillation frequency ($f_T/f_{max}$) values of 3.8/12 GHz. PIN diode switches exhibit excellent ON/OFF behaviors at high RF frequencies. Flexible inductors and capacitors compatible with high-speed TFT fabrication show resonance frequencies ($f_{res}$) up to 9.1 and 13.5 GHz, respectively. Robust mechanical characteristics were also demonstrated with these high-frequency passives components.

• 한국전기전자재료학회:학술대회논문집
• /
• 한국전기전자재료학회 2002년도 제4회 영호남학술대회 논문집
• /
• pp.73-79
• /
• 2002

This study evaporates ZnO layer thickness' differently with RF sputtering method on Si Wafer(n-100). This study is performed to examine the characteristics of photon energy and dielectric loss according to the thickness of ZnO and increase the reliability and reproduction of ZnO thin film. It is confirmed that the variation of electric Permittivity by frequency is resulted from the formation of particles within thin film, the particle size and the polarization on grain boundary. Peak of electric Permittivity value of thin film has slower and less value in early low wavelength by the coulomb force involved in carrier combination according to the increase of frequency. Reversal of electric Permittivity values is induced by dipole polarization shown in the dielectric of thin film. Complex electric constant $({\varepsilon}_1,{\varepsilon}_2)$ has larger peak values as it's thickness is thinner and then it is larger according to the increase of frequency. Electric Permittivity by photon energy has large value in imaginary number and is reduced exponentially by the increase of carrier density according to that of photon energy.

• 상하수도학회지
• /
• 제35권4호
• /
• pp.259-275
• /
• 2021

This study explored the usefulness and implications of the Bayesian hyperparameter optimization in developing species distribution models (SDMs). A variety of machine learning (ML) algorithms, namely, support vector machine (SVM), random forest (RF), boosted regression tree (BRT), XGBoost (XGB), and Multilayer perceptron (MLP) were used for predicting the occurrence of four benthic macroinvertebrate species. The Bayesian optimization method successfully tuned model hyperparameters, with all ML models resulting an area under the curve (AUC) > 0.7. Also, hyperparameter search ranges that generally clustered around the optimal values suggest the efficiency of the Bayesian optimization in finding optimal sets of hyperparameters. Tree based ensemble algorithms (BRT, RF, and XGB) tended to show higher performances than SVM and MLP. Important hyperparameters and optimal values differed by species and ML model, indicating the necessity of hyperparameter tuning for improving individual model performances. The optimization results demonstrate that for all macroinvertebrate species SVM and RF required fewer numbers of trials until obtaining optimal hyperparameter sets, leading to reduced computational cost compared to other ML algorithms. The results of this study suggest that the Bayesian optimization is an efficient method for hyperparameter optimization of machine learning algorithms.