• Journal of Power Electronics
• /
• v.17 no.1
• /
• pp.222-231
• /
• 2017

In this paper, a new architecture for a cost-effective power conditioning systems (PCS) using a single-sourced asymmetric cascaded H-bridge multilevel inverter (MLI) for photovoltaic (PV) applications is proposed. The asymmetric MLI topology has a reduced number of parts compared to the symmetrical type for the same number of voltage level. However, the modulation index threshold related to the drop in the number of levels of the inverter output is higher than that of the symmetrical MLI. This problem results in a modulation index limitation which is relatively higher than that of the symmetrical MLI. Hence, an extra voltage pre-regulator becomes a necessary component in the PCS under a wide operating bias variation. In addition to pre-stage voltage regulation for the constant MLI dc-links, another auxiliary pre-regulator should provide isolation and voltage balance among the multiple H-bridge cells in the asymmetrical MLI as well as the symmetrical ones. The proposed PCS uses a single-ended DC-DC converter topology with a coupled inductor and charge-pump circuit to satisfy all of the aforementioned requirements. Since the proposed integrated-type voltage pre-regulator circuit uses only a single MOSFET switch and a single magnetic component, the size and cost of the PCS is an optimal trade-off. In addition, the voltage balance between the separate H-bridge cells is automatically maintained by the number of turns in the coupled inductor transformer regardless of the duty cycle, which eliminates the need for an extra voltage regulator for the auxiliary H-bridge in MLIs. The voltage balance is also maintained under the discontinuous conduction mode (DCM). Thus, the PCS is also operational during light load conditions. The proposed architecture can apply the module-integrated converter (MIC) concept to perform distributed MPPT. The proposed architecture is analyzed and verified for a 7-level asymmetric MLI, using simulation results and a hardware implementation.

• Journal of Korean Neurosurgical Society
• /
• v.29 no.12
• /
• pp.1650-1656
• /
• 2000

Objective : Little consensus exists concerning which temporal lobe structures need to be resected or how much resection should be done during hippocampal resection. The purpose of this study is to identify whether the extent of hippocampal resection influences seizure after anterior temporal lobectomy. Materials and Methods : The extent of hippocampal resection was assessed in 96 patients who underwent temporal lobectomy for medically intractable complex partial seizures originating from a unilateral seizure focus in the anteromesial temporal lobe. Patients who had structural lesion were excluded from the study. Postoperative magnetic resonance imaging in the coronal and saggital planes were used to quantify the extent of the hippocampal and lateral cortical resection. The patients were divided into two groups. Patients who underwent hippocampal resection to the level of the cerebral peduncle were included in the partial resection group, and those who had resection to the level of the colliculus were assigned to total resection group. Seizure outcomes were defined according to the Engel classification and compared between the two groups. Neuropsychologic outcomes in the selected cases were reviewed. Results : The over-all seizure-free outcome(Engel classification 1) was accomplished in 75%(72/96) of the patients (mean duration of follow-up, 36.8 months). The total hippocampectomy group had a statistically superior seizure outcome than the partial hippocampectomy group(87.3% versus 58.5% seizure-free, p-value=0.001). Also, younger patients had a more favorable outcome. Other variables such as laterality, the extent of lateral cortical resection, age at onset and gender were not significant. The pre- and postoperative memory functions were evaluated in 24 patients. A worse postoperative memory outcome was associated with partial hippocampectomy. However this was not acceptable due to a former bias. Conclusion : The result of this study conforms that aggressive hippocampectomy resulted in a better seizure outcome.

• Proceedings of the Korean Vacuum Society Conference
• /
• 2014.02a
• /
• pp.306.2-306.2
• /
• 2014

전자기기의 휴대성과 이동성이 강조되고 있는 현대사회에서 비휘발성 메모리는 메모리 산업에 있어 매우 매력적인 동시에 커다란 잠재성을 지닌다. 이미 공정의 한계에 부딪힌 Flash 메모리를 대신하여 10nm 이하의 공정이 가능한 상변화 메모리(Phase-Change Memory, PRAM), 스핀 주입 자화 반전 메모리(Spin Transfer Torque-Magnetic RAM, STT-MRAM), 저항 변화 메모리(Resistive Random Access Memory, ReRAM)가 차세대 비휘발성 메모리 후보로서 거론되고 있으며, 그 중에서도 ReRAM은 빠른 속도와 낮은 소비 전력, CMOS 공정 호환성, 그리고 비교적 단순한 3차원 적층 구조의 특성으로 인해 활발히 연구되고 있다. 특히 최근에는 질화물 또는 질소를 도핑한 산화물을 저항변화 물질로 사용하는 ReRAM이 보고되고 있는데, 이들은 동작전압이 낮을 뿐만 아니라 저항 변화(Resistive Switching, RS) 과정에서 일어나는 계면 산화를 방지할 수 있으므로 ReRAM의 저항 변화 재료로서 각광받고 있다. 그러나 Cell 단위의 ReRAM 소자를 Crossbar Array 구조에 적용시켰을 때 주변 Cell과의 저항 상태 차이로 인해 전류가 낮은 저항 상태(LRS)의 Cell로 흘러 의도치 않은 동작을 야기한다. 이와 같이 누설 전류(Leakage Current)로 인한 상호간의 간섭이 일어나는 Cross-talk 현상이 존재하며, 공정의 간소화와 집적도를 유지하면서 이 문제를 해결하는 것은 실용화하기에 앞서 매우 중요한 문제이다. 따라서, 본 논문에서는 Read 동작 시 발생하는 Cell과 Cell 사이의 Cross-talk 문제를 해결하기 위해 자가 정류 특성(Self-Rectifying)을 가지는 실리콘 질화물/알루미늄 질화물 이중층(Si3N4/AlN Bi-layer)으로 구성된 ReRAM 소자 구조를 제안하였으며, Sputtering 방법을 이용하여 제안된 소자를 제작하였다. 전압-전류 특성 실험결과, 제안된 구조에 대한 에너지 밴드 다이어그램 시뮬레이션 결과와 동일하게 Positive Bias 영역에서 자가 정류 특성을 획득하였고, 결과적으로 Read 동작 시 발생하는 Cross-talk 현상을 차단할 수 있는 결과를 확보하였다.

• Proceedings of the Korean Vacuum Society Conference
• /
• 2012.02a
• /
• pp.196-196
• /
• 2012

Graphene, two-dimensional one-atom-thick planar sheet of carbon atoms densely packed in a honeycomb crystal lattice, has grabbled appreciable attention due to its extraordinary mechanical, thermal, electrical, and optical properties. Based on the graphene's high carrier mobility, high frequency graphene field effect transistors have been developed. Graphene is useful for photonic components as well as for the applications in electronic devices. Graphene's unique optical properties allowed us to develop ultra wide-bandwidth optical modulator, photo-detector, and broadband polarizer. Graphene can support SPP-like surface wave because it is considered as a two-dimensional metal-like systems. The SPPs are associated with the coupling between collective oscillation of free electrons in the metal and electromagnetic waves. The charged free carriers in the graphene contribute to support the surface waves at the graphene-dielectric interface by coupling to the electromagnetic wave. In addition, graphene can control the surface waves because its charge carrier density is tunable by means of a chemical doping method, varying the Fermi level by applying gate bias voltage, and/or applying magnetic field. As an extended application of graphene in photonics, we investigated the characteristics of the graphene-based plasmonic waveguide for optical signal transmission. The graphene strips embedded in a dielectric are served as a high-frequency optical signal guiding medium. The TM polarization wave is transmitted 6 mm-long graphene waveguide with the averaged extinction ratio of 19 dB at the telecom wavelength of $1.31{\mu}m$. 2.5 Gbps data transmission was successfully accomplished with the graphene waveguide. Based on these experimental results, we concluded that the graphene-based plasmonic waveguide can be exploited further for development of next-generation integrated photonic circuits on a chip.

• Journal of the Korean Magnetics Society
• /
• v.26 no.6
• /
• pp.201-205
• /
• 2016

In this report, we measured the impedance spectrum in TMR device, and the relaxation behavior of the real and imaginary impedance spectrum was analyzed by using the equilibrant circuit of tunneling capacitance ($C_T$) and tunneling resistance ($R_T$). The relaxation frequency was increased with AC voltage in both the parallel and antiparallel alignment of two magnetic layers. The $R_T$ with AC voltage showed the typical bias voltage dependence. However, the $C_T$ showed large value than the expected geometrical capacitance. The huge increase of $C_T$ was affecting as a limiting factor for the high speed operation of TMR devices. Thus, the supercapacitance of $C_T$ should be considered to design the high speed TMR devices.

• Journal of the Korean institute of surface engineering
• /
• v.48 no.6
• /
• pp.360-370
• /
• 2015

As the size of the semiconductor devices shrinks to nanometer scale, the importance of plasma etching process to the fabrication of nanometer scale semiconductor devices is increasing further and further. But for the nanoscale devices, conventional plasma etching technique is extremely difficult to meet the requirement of the device fabrication, therefore, other etching techniques such as use of multi frequency plasma, source/bias/gas pulsing, etc. are investigated to meet the etching target. Until today, various pulsing techniques including pulsed plasma source and/or pulse-biased plasma etching have been tested on various materials. In this review, the experimental/theoretical studies of pulsed plasmas during the nanoscale plasma etching on etch profile, etch selectivity, uniformity, etc. have been summarized. Especially, the researches of pulsed plasma on the etching of silicon, $SiO_2$, and magnetic materials in the semiconductor industry for further device scaling have been discussed. Those results demonstrated the importance of pulse plasma on the pattern control for achieving the best performance. Although some of the pulsing mechanism is not well established, it is believed that this review will give a certain understanding on the pulsed plasma techniques.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.37 no.8
• /
• pp.1-8
• /
• 2000

We have achieved a high aspect ratio track-following microactuator (TFMA) which is capable of driving 0.3 ${\mu}m$ magnetic head for hard disk drive (HDD). it was fabricated on silicon on insulator (SOI) wafer with 20 ${\mu}m$ trick active silicon and 2 ${\mu}m$ thick thermally grown oxide and piggyback electrostatic principle was used for driving TFMA. The first vibration mode frequency of TFMA was 18.5 kHz which is enough for a recording density of higher than 10 Gb/in$^2$. Its displacement was 1.4 ${\mu}m$ when 15 V dc bias plus 15 V ac sinusoidal driving input was applied and its electrostatic force was 50 N. The fabricated actuator shows 7.51 dB of gain margin and 50.98$^{\circ}$ of phase margin for 2.21 kHz servo-bandwidth.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.22 no.2
• /
• pp.199-207
• /
• 2011

A Ku-band low noise amplifier has been designed and fabricated by using 0.25 um SiGe BiCMOS process. The developed Ku-band LNA RFIC which has been designed with hetero-junction bipolar transistor(HBT) in the BiCMOS process have noise figure about 2.0 dB and linear gain over 19 dB in the frequency range from 9 GHz to 14 GHz. Optimization technique for p-tap value and electro-magnetic(EM) simulation technique had been used to overcome the inaccuracy in the PDK provided from the foundry service company and to supply the insufficient inductor library. The finally fabricated low noise amplifier of two fabrication runs has been implemented with the size of $0.65\;mm{\times}0.55\;mm$. The pure amplifier circuit layout with the reduced size of $0.4\;mm{\times}0.4\;mm$ without the input and output RF pads and DC bais pads has been incorporated as low noise amplication stages in the multi-function RFIC for the active phased array antenna of Ku-band satellite VSAT.

• Journal of Magnetics
• /
• v.14 no.1
• /
• pp.36-41
• /
• 2009

Magneto-optical Kerr effect (MOKE) magnetometry was used to investigate magnetization reversal dynamics in 30-nm NiFe/15-nm FeMn, 15-nm FeMn/30-nm CoFe bilayers, and 30-nm NiFe/(2,10)-nm FeMn/30-nm CoFe trilayers. The in-plane magnetization components of each ferromagnetic layer, both parallel and perpendicular to the applied field, were separately determined by measuring the longitudinal and transverse MOKE hysteresis loops from both the front and back sides of the film for an oblique incident s-polarized beam. The magnetization of the FeMn/CoFe bilayer was reversed abruptly and symmetrically through nucleation and domain wall propagation, while that of the NiFe/FeMn bilayer was reversed asymmetrically with a dominant rotation. In the NiFe/FeMn/CoFe trilayers, the magnetic reversal of the two ferromagnetic layers proceeded via nucleation and domain wall propagation for 2-nm FeMn, but via asymmetric rotation for 10-nm FeMn. The exchange-biased ferromagnetic layers showed the magnetization reversal along the same path in the film plane for the decreasing and increasing field branches from transverse MOKE hysteresis loops, which can be qualitatively explained by the theoretical model of the exchange-biased ferromagnetic/antiferromagnetic systems.

• Proceedings of the Korean Magnestics Society Conference
• /
• 2011.12a
• /
• pp.7-8
• /
• 2011

Although yttrium iron garnet (YIG) has provided a great vehicle for the study of spin waves in the past, associated difficulties in film deposition and device fabrication using YIG had limited the applicability of spin waves to practical devices. However, microfabrication techniques have made it possible to characterize both the resonant as well as the travelling characteristics of spin waves in permalloy (Py). A variety of methods have been used for measuring spin waves, including Brillouin light scattering (BLS), magneto-optic Kerr effect (MOKE), vector network analyzer ferromagnetic resonance (VNA-FMR), and pulse inductive microwave magnetometry (PIMM). PIMM is one of the most preferred methodologies of measuring travelling spin waves. In this method, an electrical impulse is applied at one of two coplanar waveguides patterned on top of oxide-insulated Py, producing a local disturbance in the magnetization of the Py. The resulting disturbance travels down the Py in the form of waves, and is inductively picked up by the other coplanar waveguide. We investigate the effect of the pulse width of excitation pulses on the generated spin wave packets using both experimental results and micromagnetic simulations. We show that spin wave packets generated from electrical pulses are a superposition of two separate spin wave packets, one generated from the rising edge and the other from the falling edge, which interfere either constructively or destructively with one another, depending upon the magnitude and direction of the field bias conditions. A method of spin wave amplitude modulation is also presented by the linear superposition of spin waves. We use interfering spin waves resulting from two closely spaced voltage impulses for the modulation of the magnitude of the resultant spin wave packets.