• Proceedings of the Korean Vacuum Society Conference
• /
• 2012.02a
• /
• pp.434-435
• /
• 2012

Plasma enhanced chemical vapor deposition (PECVD) silicon dioxide thin films have many applications in semiconductor manufacturing such as inter-level dielectric and gate dielectric metal oxide semiconductor field effect transistors (MOSFETs). Fundamental chemical reaction for the formation of SiO2 includes SiH4 and O2, but mixture of SiH4 and N2O is preferable because of lower hydrogen concentration in the deposited film [1]. It is also known that binding energy of N-N is higher than that of N-O, so the particle generation by molecular reaction can be reduced by reducing reactive nitrogen during the deposition process. However, nitrous oxide (N2O) gives rise to nitric oxide (NO) on reaction with oxygen atoms, which in turn reacts with ozone. NO became a greenhouse gas which is naturally occurred regulating of stratospheric ozone. In fact, it takes global warming effect about 300 times higher than carbon dioxide (CO2). Industries regard that N2O is inevitable for their device fabrication; however, it is worthwhile to develop a marginable nitrous oxide free process for university lab classes considering educational and environmental purpose. In this paper, we developed environmental friendly and material cost efficient SiO2 deposition process by substituting N2O with O2 targeting university hands-on laboratory course. Experiment was performed by two level statistical design of experiment (DOE) with three process parameters including RF power, susceptor temperature, and oxygen gas flow. Responses of interests to optimize the process were deposition rate, film uniformity, surface roughness, and electrical dielectric property. We observed some power like particle formation on wafer in some experiment, and we postulate that the thermal and electrical energy to dissociate gas molecule was relatively lower than other runs. However, we were able to find a marginable process region with less than 3% uniformity requirement in our process optimization goal. Surface roughness measured by atomic force microscopy (AFM) presented some evidence of the agglomeration of silane related particles, and the result was still satisfactory for the purpose of this research. This newly developed SiO2 deposition process is currently under verification with repeated experimental run on 4 inches wafer, and it will be adopted to Semiconductor Material and Process course offered in the Department of Electronic Engineering at Myongji University from spring semester in 2012.

• Analytical Science and Technology
• /
• v.28 no.1
• /
• pp.40-46
• /
• 2015

Hydrochar has been generated from food waste via hydrothermal carbonization (HTC) reaction. As a solid product of HTC reaction, hydrochar has a great potential as an adsorbent of pollutants from the various media. The surface area and pore volumes are very important parameters to be served as an adsorbent. It requires an expensive equipment and consumes time to measure those parameter. Therefore, titration methods including iodine and methylene blue adsorption were evaluated to be correlated with that of BET analysis. Even though the absolute values of the computed surface area and pore volumes were not able to be matched directly, the patterns of change were successfully correlated. Among the reaction conditions, the reaction time and temperature at $230^{\circ}C$ for 4 h was determined as an optimization condition, which confirmed by titration method and BET analysis. Titration method for surface area and pore volumes computed by combination of iodine and methylene blue adsorbing values would be a simple and fast way of determining the optimization condition for hydrochar as an adsorbent produced by HTC reaction.

• Journal of Microbiology and Biotechnology
• /
• v.23 no.2
• /
• pp.211-217
• /
• 2013

Corynebacterium glutamicum is one of the well-studied industrial strain that is used for the production of nucleotides and amino acids. Recently, it has also been studied as a possible producer of organic acids such as succinic acid, based on its ability to produce organic acids under an oxygen deprivation condition. In this study, we conducted the optimization of medium components for improved succinate production from C. glutamicum under an oxygen deprivation condition by Plackett-Burman design and applied a response surface methodology. A Plackett-Burman design for ten factors such as glucose, ammonium sulfate, magnesium sulfate, potassium phosphate ($K_2HPO_4$ and $KH_2PO_4$), iron sulfate, manganese sulfate, biotin, thiamine, and sodium bicarbonate was applied to evaluate the effects on succinate production. Glucose, ammonium sulfate, magnesium sulfate, and dipotassium phosphate were found to have significant influence on succinate production, and the optimal concentrations of these four factors were sequentially investigated by the response surface methodology using a Box-Behnken design. The optimal medium components obtained for achieving maximum concentration of succinic acid were as follows: glucose 10 g/l, magnesium sulfate 0.5 g/l, dipotassium phosphate ($K_2HPO_4$) 0.75 g/l, potassium dihydrogen phosphate ($KH_2PO_4$) 0.5 g/l, iron sulfate 6 mg/l, manganese sulfate 4.2 mg/l, biotin 0.2 mg/l, thiamine 0.2 mg/l, and sodium bicarbonate 100 mM. The parameters that differed from a normal BT medium were glucose changed from 40 g/l to 10 g/l, dipotassium phosphate ($K_2HPO_4$) 0.5 g/l changed to 0.75 g/l, and ammonium sulfate ($(NH_4)_2SO_4$) 7 g/l changed to 0 g/l. Under these conditions, the final succinic acid concentration was 16.3 mM, which is about 1.46 fold higher than the original medium (11.1 mM) at 24 h. This work showed the improvement of succinate production by a simple change of media components deduced from sequential optimization.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.17 no.7
• /
• pp.970-976
• /
• 2007

In this study, Polynomial Network Pattern Classifier(PNC) based on Fuzzy Inference Mechanism is designed and its parameters such as learning rate, momentum coefficient and fuzzification coefficient are optimized by means of Particle Swarm Optimization. The proposed PNC employes a partition function created by Fuzzy C-means(FCM) clustering as an activation function in hidden layer and polynomials weights between hidden layer and output layer. Using polynomials weights can help to improve the characteristic of the linear classification of basic neural networks classifier. In the viewpoint of linguistic analysis, the proposed classifier is expressed as a collection of "If-then" fuzzy rules. Namely, architecture of networks is constructed by three functional modules that are condition part, conclusion part and inference part. The condition part relates to the partition function of input space using FCM clustering. In the conclusion part, a polynomial function caries out the presentation of a partitioned local space. Lastly, the output of networks is gotten by fuzzy inference in the inference part. The proposed PNC generates a nonlinear discernment function in the output space and has the better performance of pattern classification as a classifier, because of the characteristic of polynomial based fuzzy inference of PNC.

• Journal of Radiation Protection and Research
• /
• v.26 no.4
• /
• pp.417-424
• /
• 2001

The pseudo-biologic objective function has been designed for the IMRT optimization. The RTP Tool Box (RTB) was used for this study. The pseudo-biologic function is similar to the biological objective function in mathematical shape, but uses physical parameters. The concepts of the TCI (Target Coverage Index) and the OSI (Organ Score Index) have been introduced for the target and the normal organs, respectively. The pseudo-biologic objective function s has been defined using these TCI and OSI's. The OSI's from the pseudo-biological function showed better results than from the physical functions, while TCI's showed similar tendency. These results revealed the feasibility of the pseudo-biologic function as an IMRT objective function.

• Journal of Korea Water Resources Association
• /
• v.51 no.9
• /
• pp.803-812
• /
• 2018

The purpose of this study is to develop Particle Swarm Optimization (PSO) automatic calibration algorithm with multi-objective functions by Python, and to evaluate the applicability by applying the algorithm to the Soil and Water Assessment Tool (SWAT) watershed modeling. The study area is the upstream watershed of Gongdo observation station of Anseongcheon watershed ($364.8km^2$) and the daily observed streamflow data from 2000 to 2015 were used. The PSO automatic algorithm calibrated SWAT streamflow by coefficient of determination ($R^2$), root mean square error (RMSE), Nash-Sutcliffe efficiency ($NSE_Q$), and especially including $NSE_{INQ}$ (Inverse Q) for lateral, base flow calibration. The results between automatic and manual calibration showed $R^2$ of 0.64 and 0.55, RMSE of 0.59 and 0.58, $NSE_Q$ of 0.78 and 0.75, and $NSE_{INQ}$ of 0.45 and 0.09, respectively. The PSO automatic calibration algorithm showed an improvement especially the streamflow recession phase and remedied the limitation of manual calibration by including new parameter (RCHRG_DP) and considering parameters range.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.21 no.7
• /
• pp.744-751
• /
• 2010

In this paper, the design of multilayer radome within, the insertion loss, －0.3 dB in X-band with PSO was carried out based on two cases. The first is that, deciding material constant of skin and core, each layer thickness of c-sandwich radome with PSO is found and the second is that, deciding material constant and thickness of the skins of both sides, the material constant and thickness of three layers between skins of both sides using PSO is decided. The performance of the designed radome almost agreed with the required performance. It was showed that the radome design applying PSO algorithm is easy and fast and the optimum radome is also designed in combination of the various parameters of radome. From these results, the radome having various performance can be designed except the tedious calculation and also be applied to various radome structure.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.12 no.5
• /
• pp.879-886
• /
• 2008

This paper addresses the analysis and the design optimization of differential interconnects for Low-Voltage Differential Signaling (LVDS) applications. Thanks to the differential transmission and the low voltage swing, LVDS offers high data rates and improved noise immunity with significantly reduced power consumption in data communications, high-resolution display, and flat panel display. We present an improved model and new equations to reduce impedance mismatch and signal degradation in cascaded interconnects using optimization of interconnect design parameters such as trace width, trace height and trace space in differential flexible printed circuit board (FPCB) transmission lines. We have carried out frequency-domain full-wave electromagnetic simulations, time-domain transient simulations, and S-parameter simulations to evaluate the high-frequency characteristics of the differential FPCB interconnects. The 10% change in trace width produced change of approximately 6% and 5.6% in differential impedance for trace thickness of $17.5{\mu}m$ and $35{\mu}m$, respectively. The change in the trace space showed a little change. We believe that the proposed approach is very helpful to optimize high-speed differential FPCB interconnects for LVDS applications.

• Journal of Intelligence and Information Systems
• /
• v.29 no.1
• /
• pp.107-119
• /
• 2023

The technology that recognizes a soldier's motion and movement status has recently attracted large attention as a combination of wearable technology and artificial intelligence, which is expected to upend the paradigm of troop management. The accuracy of state determination should be maintained at a high-end level to make sure of the expected vital functions both in a training situation; an evaluation and solution provision for each individual's motion, and in a combat situation; overall enhancement in managing troops. However, when input data is given as a timer series or sequence, existing feedforward networks would show overt limitations in maximizing classification performance. Since human behavior data (3-axis accelerations and 3-axis angular velocities) handled for military motion recognition requires the process of analyzing its time-dependent characteristics, this study proposes a high-performance data-driven classifier which utilizes the long-short term memory to identify the order dependence of acquired data, learning to classify eight representative military operations (Sitting, Standing, Walking, Running, Ascending, Descending, Low Crawl, and High Crawl). Since the accuracy is highly dependent on a network's learning conditions and variables, manual adjustment may neither be cost-effective nor guarantee optimal results during learning. Therefore, in this study, we optimized hyperparameters using Bayesian optimization for maximized generalization performance. As a result, the final architecture could reduce the error rate by 62.56% compared to the existing network with a similar number of learnable parameters, with the final accuracy of 98.39% for various military operations.

• Journal of the Korea Society for Simulation
• /
• v.26 no.4
• /
• pp.35-41
• /
• 2017

With the advances in Internet over Things, the demand in diverse electronic devices such as mobile phones and sensors has been rapidly increasing and boosting up the researches on those products. Semiconductor materials, devices, and fabrication processes are becoming more diverse and complicated, which accompanies finding parameters for an optimal fabrication process. In order to find the parameters, a process simulation before fabrication or a real-time process control system during fabrication can be used, but they lack incorporating the feedback from post-fabrication data and compatibility with older equipment. In this research, we have developed an artificial intelligence based simulator, which finds parameters for an optimal process and controls process equipment. In order to apply the control concept to all the equipment in a fabrication sequence, we have developed a prototype for a manipulator which can be installed over an existing buttons and knobs in the equipment and controls the equipment communicating with the AI over the Internet. The AI is based on the deep learning to find process parameters that will produce a device having target electrical characteristics. The proposed simulator can control existing equipment via the Internet to fabricate devices with desired performance and, therefore, it will help engineers to develop new devices efficiently and effectively.