Semiconductor Factory Research Articles

Assessment Of Cl 2 /CHF 3 Mixture For Plasma Etching Process On Barc And Tin Layer For 0.21 µm Metal Line: Silterra Case Study

In wafer fabrication manufacturing, aluminum etching process is a dry plasma etching process used as main process for construction of aluminum (Al) interconnects structures. As customer requirement changed for faster, more reliable and lower cost chips, chip manufacturers have learned to reduce the size of component on a chip in order to achieve those requirements(Ibrahim, Chik, & Hashim, 2016). As the geometry of the chip getting smaller, the width of Al line wiring specification also shrinking. To print the smaller geometry pattern requirement, the thickness in masking process also has to be reduced for better resolution. Such a thinner resist will create a challenge during plasma etching to ensure a minimal resist loss process which required new type of equipment but this research insist to sustain similar equipment. The use of oxide film as a hard mask has been evaluated by other researchers but alternative approach still needed to suit specific requirement of semiconductor factory installation base. This approach does require a process integration change and require a full technology qualification and easily take a lengthy qualification procedures especially when to qualify the existing products. It is worth trying at the situation of no other solution available. The challenge of insufficient margin for the metal line etching process for 0.2 µm width has caused the deformed metal pattern formation. This chemistry study of Cl2/CHF3 as a replacement gas to existing Cl2/O2 to address Organic backside anti refractive coating (OBARC) was evaluated and proven novelty where detail discussed in the following content.

Integrated production and maintenance planning method with simulation-based optimization

Abstract Maintenance jobs are an important aspect of manufacturing systems and compete for the same resources of the production jobs. Even though literature has pointed operational benefits of integrating the production and maintenance plans, there is the gap of handling the problem in a simulation-based approach that is responsive to disturbances. This paper aims to propose an adaptive method for integrating production and maintenance planning using simulation-based optimization. The method is tested in a simulation model of a semiconductor factory. Results show improvements on key performance indicators when compared to scenarios with simple dispatching rules.

Application of ANN for Fault Detection in Overhead Transport Systems for Semiconductor Fab

In order to ensure safe and fast transportation of wafers in 300 mm semiconductor factories, overhead transport systems (OHT) are primarily used. These systems consist of a rail network and vehicles. To avoid congestion and delays in production, high availability of individual rail sections is essential. In order to ensure this extensive preventive maintenance is required. In this paper, we focus on automatic checks for faults of the rail network by capturing the rail with optical sensors. Our objective is the identification of faults in real time. We considered the identification using artificial neural networks (ANN). Due to the lack of fixed rules designing an ANN we tested different topologies for our application and covered adaptation of ANN to the real conditions in the fab. As a result, our ANN provides accurate real time fault detection which allows a needs-based, resource-saving and efficient maintenance procedure for a reliable OHT and hence 24/7 semiconductor manufacturing.

TO SURVEY POWER QUALITY OF HIGH VOLTAGE KEY-CUSTOMER BY USING VOLTAGE EVENT RECORD

According to IEEE-1159 definition and classification of power quality of applying at supplyside and demand-side of power system to evaluate the power quality of high voltage keycustomer and using the voltage events record (VER) in low voltage level supply. Which voltage parameter set of VER with specific software Event View software by Fluke and can collect event data, plot graphic of event data sets that can using the data mining skill to analysis the voltage swell, voltage sag, power outage, frequency events and three phase load unbalance of high voltage key-customer. By abnormal voltage event statistics with discriminated by quantization of high voltage key-customer which can evaluate the own power quality and an also provide reference for quality requirements for related semiconductor factories with higher demand that can avoid the significant effect the economic losses.

Fire Department Building Regulations without Fire Dampers for Fire Compartments in Semiconductor Factories

Fire Department Building Regulations without Fire Dampers for Fire Compartments in Semiconductor Factories

An uncommon outbreak of irritant contact dermatitis caused by rubber accelerators: a historical cohort study

Our aim was to study the association between the rubber accelerators and predisposing factors of irritant contact dermatitis (ICD) among the semiconductor factory workers. The research design was a historical cohort study. The study population was divided into two groups: (1) the rubber accelerator exposure group, including workers who wore natural rubber finger cots containing 2,2 dibenzthiazyl disulfide (MBTS), zinc diethyldithiocarbamate (ZDEC), and zinc dibutyldithiocarbamate (ZDBC); and (2) the non-rubber accelerator exposure group, including the workers who wore finger cots with less accelerators and without ZDBC (i.e., the non-exposed group). The samples were calculated, and each group comprised 99 workers. The content of rubber accelerators in the finger cots was determined using high-performance liquid chromatography technique. Cases of ICD and the respective associated factors were collected from the medical records as well as information derived from the questionnaire. Inferential statistics using multiple regression analysis were applied and the results presented as relative risk (RR), adjusted relative risk (adjusted RR), and the 95% confidence interval (CI). The demographics of the exposed and non-exposed group were statistically different in age, duration of employment, duration of contact to rubber accelerators, underlying atopic diseases, underlying allergic rhinitis, family history of atopic diseases, and family history of allergic rhinitis. The results showed that the ICD occurred in 27 (27.3%) of workers in the exposed group compared to 4 (4.0%) of workers in the non-exposed group. The adjusted RR of rubber accelerators containing MBTS, ZDEC, and ZDBC was 9.84 (95% CI 3.70, 26.15) and the predisposing factors which associated with ICD were finger cot contact duration more than 4 years, underlying allergic rhinitis, and family history of atopic diseases [1.78 (95% CI 1.03, 3.08), 1.50 (95% CI 0.85, 2.64) and 1.97 (95% CI 1.12, 3.46), respectively], while having underlying atopic dermatitis was found to be a protective factor adjusted RR 0.62 (95% CI 0.10, 4.13). Rubber accelerators MBTS, ZDEC, and ZDBC were the causes of ICD. The possible agent was ZDBC. Other predisposing factors play less important role in this epidemic.

Semiconductor Work and Adverse Pregnancy Outcomes Associated with Male Workers: A Retrospective Cohort Study.

A hazardous work environment in semiconductor factories is a threat to the workers' health. Semiconductor manufacturing characteristically requires young workers, and reproductive toxicity is an important issue. Studies investigating reproductive toxicity among individuals working in the semiconductor manufacturing industry have primarily focused on outcomes in women. Information on the reproductive health of male semiconductor factory workers is limited. This study aimed to evaluate the association between workplace exposures among male workers in a Korean semiconductor company and adverse pregnancy outcomes. Based on the data from the 2015 Semiconductor Health Survey (SHS), which evaluated the workplace exposures, pregnancy outcomes, and general health of 21 969 employees of the semiconductor industry in South Korea, we included 3868 male workers with 7504 pregnancy outcomes identified by self-reports for this retrospective cohort study. Data regarding the pregnancy outcomes, order of pregnancy, and the years of the outcomes were collected via the SHS questionnaire. Adverse pregnancy outcomes were defined as preterm labor, spontaneous abortion, and stillbirth. Workplace exposures were classified as fabrication, assembly, others, lab, and office work (reference group). A generalized estimating equations model including repeated events of individuals and producing relative risk (RR) and 95% confidence interval (CI) was used to estimate the association between workplace exposure and adverse pregnancy outcomes. Analyses were adjusted for work location, spouse's employment in semiconductor production work, educational level, marital status, risky alcohol drinking, smoking status, body mass index, order of pregnancy, and age and year of pregnancy outcome, which were based on a priori decisions. The adjusted risk for adverse outcomes was higher [RR (95% CI): 1.47 (1.04, 2.07)] among assembly process workers compared with the office workers. Adjusted risks for adverse outcomes among workers in assembly and fabrication, whose spouses also worked in semiconductor production, were 1.60 (95% CI: 1.04, 2.46) and 1.74 (95% CI: 1.18, 2.57) times higher, respectively, compared with the office workers with spouses not working in semiconductor production. Based on these findings, semiconductor work might be considered a risk factor for reproductive toxicity among male workers, especially for those whose spouses have the same job.

Health Risk Assessment of Photoresists Used in an Optoelectronic Semiconductor Factory.

Photoresist materials are indispensable in photolithography, a process used in semiconductor fabrication. The work process and potential hazards in semiconductor production have raised concerns as to adverse health effects. We therefore performed a health risk assessment of occupational exposure to positive photoresists in a single optoelectronic semiconductor factory in Taiwan. Positive photoresists are widely used in the optoelectronic semiconductor industry for photolithography. Occupational exposure was estimated using the Stoffenmanager® model. Bayesian modeling incorporated available personal air sampling data. We examined the composition and by-products of the photoresists according to descriptions published in the literature and patents; the main compositions assessed were propylene glycol methyl ether acetate (PGMEA), novolac resin, photoactive compound, phenol, cresol, benzene, toluene, and xylene. Reference concentrations for each compound were reassessed and updated if necessary. Calculated hazard quotients were greater than 1 for benzene, phenol, xylene, and PGMEA, indicating that they have the potential for exposures that exceed reference levels. The information from our health risk assessment suggests that benzene and phenol have a higher level of risk than is currently acknowledged. Undertaking our form of risk assessment in the workplace design phase could identify compounds of major concern, allow for the early implementation of control measures and monitoring strategies, and thereby reduce the level of exposure to health risks that workers face throughout their career.

Applying Two-Stage Differential Evolution for Energy Saving in Optimal Chiller Loading

In Taiwan, over 45% of the energy in common buildings is used for the air-conditioning system. In particular, the chiller plant consumes about 70% of the energy in air-conditioning system. The electric energy consumption of air-condition system in a clean room of semiconductor factory is about 5–10 times of that in a common building. Consequently, the optimal chiller loading in energy saving of building is a vital issue. This paper develops a new algorithm to solve optimal chiller loading (OCL) problems. The proposed two-stage differential evolution algorithm integrated the advantages of exploration (global search) in the modified binary differential evolution (MBDE) algorithm and exploitation (local search) in the real-valued differential evolution (DE) algorithm for finding the optimal solution of OCL problems. In order to show the performance of the proposed algorithm, comparison with other optimization methods has been done and analyzed. The result shows that the proposed algorithm can obtain similar or better solution in comparison to previous studies. It is a promising approach for the OCL problem.

Optimal Operation Strategies for a Large Multi-Chiller System Based on Cooling Load Forecast

In semiconductor and electronics factories, large multi-chiller systems are needed to satisfy strict cooling load requirements. In order to save energy, it is worthwhile to design the chilled water system operation. In this paper, an optimal flexible operation scheme is developed based on a two-dimensional time-series model to forecast the cooling load of multi-chiller systems with chiller units of different cooling capacities running in parallel. The optimal integrity scheme can be obtained using the Mixed Integer Nonlinear Programming method, which minimizes the energy consumption of the system within a future time period. In order to better adapt the change of cooling load, the operation strategy of regulating the chilled water flowrates is employed. The chilled water flowrates are set as a design variable. When the chillers are running, their chilled water flowrates can vary within limits, whereas the flowrates are zero when the chillers are unloaded. This forecasting method provides integral optimization within a future time period and offers the operating reference for operators. The power and advantages of the proposed method are presented using an industrial case to help readers delve into this matter.

Solving the Mask Data Preparation Scheduling Problem Using Meta-Heuristics

Mask data preparation (MDP) is a part of the mask data process for fabricating semiconductors, and its importance has commonly been neglected. This paper proposes an integer linear programming model and two meta-heuristics, a genetic algorithm (GA) and simulated annealing (SA), for solving the MDP scheduling problem (MDPSP). The proposed meta-heuristics are empirically evaluated using 768 simulation instances of MDPSP based on the characteristics of a real technology company and compared with the most commonly used first-come, first-served method. The experimental results reveal that the proposed GA and SA algorithms can critically improve the manufacturing schedule for semiconductor factories.

High-rate partial nitrification of semiconductor wastewater: Implications of online monitoring and microbial community structure

The aims of this study were to investigate the feasibility of partial nitrification (PN) in a sequencing batch reactor (SBR) fed with real wastewater discharged from the washing process at a semiconductor factory, to evaluate the efficacy of online control for nitrite buildup, and to examine the microbial community composition. To this end, online monitoring of NH4+ and NO3− concentrations in an SBR was implemented to track the extant activities of ammonia-oxidizing (AOB) and nitrite-oxidizing bacteria (NOB). The results showed that the rate of NH4+ oxidation by AOB increased during the first 20 days of the operation, and were maintained as high as a maximum volumetric rate of NH4+ removal of 2.17 kg-N/m3/day with an NH4+ removal efficiency of 94.1–99.8%. Successful nitrite buildup (> 99.9%) has been achieved after day 20 when AOB activity was 8.11-times higher than NOB activity. High-throughput sequencing of 16S rRNA genes manifested that the relative abundance of cluster 7 Nitrosomonas species in the family Nitrosomonadaceae increased from 0.2% to 50% over total bacteria. On the other hand, the relative abundance of Nitrospira spp. as the predominant NOB decreased from 1.0% to 0.05%, indicating NOB out-selection during SBR operations. A positive correlation was noted between the abundance of the dominant AOB and NH4+ oxidation rate, suggesting that a simple nitrifier community regulated high-rate PN.

The Estimation of Total Cost Through the Financial Assessment of Customized High Purity Industrial Water

The purpose of this study is to estimate the total cost including the capital operation costs for the production of customized high purity industrial water, and to analyze the benefits and evaluate the financial performance according to the cost and benefits. It is important to obtain and apply objective data when calculating approximate costs. The capital operation costs of pretreatment process and wastewater treatment process were calculated by quoting data from the Ministry of Environment. Rough costs for pure water and ultrapure water facilities were quoted by professional companies and calculated by each project cost. For the benefit/cost ratio = 1.03, the net present value = 6,165 million won and the internal rate of return = 3.09%, the unit price for 10,000 m3/d capacity was estimated at 3,517 ￦/m3. as a result of the financial evaluation with a discount rate of 2.21% based on the ultrapurewater treatment system of semiconductor factory. This is the total price including the pre-treatment and ultra-pure water of facility construction cost and etc. Compared with the current high-purity industrial water business, it is a considerably high total cost. The comparing cases of high purity industrial water for semiconductors, the realistic aspects of the industrial water business process should be considered. The total cost is expected to be ranged from 2,284 ￦/m3 to 3,517 ￦/m3. Key words: Ultrapure Water (UPW), Total Cost, Financial Assessment, Customized Industrial Water, Approximate Cost

Current understanding of the electrostatic risk to reticles used in microelectronics and similar manufacturing processes

This paper explains how an electric field and a reticle interact and describes the different kinds of damage that can be caused to a reticle through its exposure to electric field. It is shown why electrostatic reticle damage has changed from ESD damage (which causes yield to suddenly drop precipitously) into a gradual and cumulative form of degradation that is very difficult to diagnose. It is explained why some of the approaches that have been taken to reduce ESD damage in the semiconductor factory, such as equipotential bonding and the use of static dissipative plastics for making reticle pods, actually increase the risk of this cumulative type of electrostatic degradation in reticles. When assessing the risk to reticles and designing an effective protective strategy for reticle handling, it is shown why one must take into account the temporal characteristics of a reticle’s interaction with electric field—including the effect that the reticle’s immediate surroundings will have on that interaction—as well as considering the strength of any electric field in the reticle handling environment. Solutions are presented that would allow the electrostatic risk to reticles to be reduced significantly, without requiring major changes to operating procedures in semiconductor manufacturing facilities.

Half Cycle Discrete Transformation for Voltage Sag Improvement in an Islanded Microgrid using Dynamic Voltage Restorer

Growing demand for clean and green power has increased penetration of renewable energy sources into microgrid. Based on the demand supply, microgrid can be operated in grid connected mode and islanded mode. Intermittent nature of renewable energy sources such as solar and wind has lead to number of control challenges in both modes of operation. Especially islanded microgrid throws power quality issues such as sag, swell, harmonics and flicker. Since medical equipments, semiconductor factory automations are very sensitive to voltage variations and therefore voltage sag in an islanded microgrid is of key significance. This paper proposes a half cycle discrete transformation (HCDT) technique for fast detection of voltage sag in an islanded microgrid and thereby provides fast control action using dynamic voltage restorer (DVR) to safe guard the voltage sensitive equipments in an islanded microgrid. The detailed analysis of simulation results has clearly demonstrated the effectiveness of proposed method detects the voltage sag in 0.04 sec and there by improves the voltage profile of islanded microgrid.

Emerging Occupational Health Needs at a Semiconductor Factory Following the 2016 Kumamoto Earthquakes: Evaluation of Effectiveness and Necessary Improvements of List of Postdisaster Occupational Health Needs.

We sought to share our experiences with occupational health efforts at a manufacturing site after a large earthquake event, and to evaluate the effectiveness of our "List of Occupational Health Needs Following Disasters" (Needs List). We implemented occupational health efforts following the 2016 Kumamoto earthquakes in Japan, with reference to the Needs List and a related manual, and analyzed the needs that arose. The instruments used enabled occupational health personnel to predict what needs would arise and facilitate their countermeasures. Among such needs that arose, 87.7% were in the Needs List, but seven undocumented needs were also observed. We confirmed that the Needs List could predict as many occupational health needs as we had expected at the disaster, though it is necessary to improve them by reflecting on our experiences.

Scheduling of Dual Resource Constrained Lithography Production: Using CP and MIP/CP

A dual resource constrained (DRC) scheduling problem arises at the photolithography area in semiconductor manufacturing wherein reticles are required as an auxiliary resource. Reticles are transferred from one place to another. While this shared resource gives the manufacturer a flexibility, it certainly gives rise to a complex DRC scheduling problem surrounded by jobs, machines, and reticles, restricted by sequence-dependent setup-time and location-dependent reticle transfer-time. This study proposes a constraint programming (CP) approach and investigates methods for improving our initial CP using a hybrid method and variable ordering heuristic. Experiments indicate the variable ordering heuristic brings a significant improvement over a pure CP. Note to Practitioners —There has been yet no successful exact approach that addresses litho scheduling problem, covering jobs, machines, and reticles. This study demonstrates CP is able to find an optimal solution for most of realistic industry-size test problems. In a future smart semiconductor factory, seamlessly orchestrating machines, reticles, and jobs would not be the end state. The next level is to integrate material handling system to this already overcrowded production scheduling problem. The specialized keywords of CP, as demonstrated in this study, will enable practitioners to develop a concise code of integrating vehicles, jobs, machines, and reticles.

A Study on the Estimation of the GHGs Emissions to the Reuse of De-ionized Water Production Process in Semiconductor Factory

A Study on the Estimation of the GHGs Emissions to the Reuse of De-ionized Water Production Process in Semiconductor Factory

Indoor SLAM application using geometric and ICP matching methods based on line features

This study presents an autonomous guided vehicle (AGV) with simultaneous localization and map building (SLAM) based on matching method, and extended Kalman filter SLAM. In general, the AGV is a large mobile robot that is used in transportation to carry cargoes, and it is guided using wired or wireless guidance systems. The guidance system based AGV accounts for a majority of robots in the mobile robot industry. However, in semiconductor factories, landmarks are unavailable; hence, the existing system has not been used in the mentioned environments. Therefore, the SLAM technology is applied in the environments, and can guide the AGV without landmarks. However, the accuracy of the SLAM can be low owing to measurement error of sensors and a cumulative calculation caused by localization sensors. Therefore, the accuracy is frequently assumed to be incorrect; moreover, the accuracy of the built map is low. In order to solve the problems, this study proposes the AGV with the SLAM based on matching methods; two matching method; geometric matching method and iterative closest point algorithm. The performance of the proposed method is compared with typical methods such as singular value decomposition / RIGID transformation based technologies using feature-point-based SLAM and is compared with the aforementioned two methods using the extended Kalman filter SLAM. The proposed method is more efficient than the typical methods used in the comparison.

Applications of Electro-Fenton for Sustainable Water Treatment and Reuse: A Case Study in the Electronics Industry

Electrochemistry has an important role to play in increasing the sustainability and energy efficiency of water treatment and reuse systems. In Singapore for example, the water demand has been projected to more than double by 2060, with non-domestic water demand expected to account for 70% (560 millions of gallons per day). In this context, industrial water reuse is becoming a necessity and the electronics sector has been identified as one of the priority targets due to the extra-pure water requirements. The main issue with industrial water reuse is the removal of “hard-to-treat” contaminants and current interest is focusing on advanced oxidation processes (AOPs), because they produce highly oxidizing species like •OH radicals that can non-selectively mineralize all sorts of organic pollutants. Among them, electro-Fenton has emerged because of its many advantages, including ease of operation and high degradation rates of numerous organic pollutants, thanks to the recent development of improved electrode materials (e.g., carbon nanotubes, graphene, etc.) (1-3). In this work, we showcase the performance at lab-scale and pre-pilot scale of an electro-Fenton prototype to treat a real waste stream from a semi-conductor factory using a novel carbon-fiber brush cathode. First, the brush was fully characterized in terms of its electrical conductivity and surface area. Its interfacial charge-transfer resistance (R ct) – as determined by electrochemical impedance spectroscopy – equaled 1.40 Ω. Cyclic voltammetry was used to determine the electroactive surface area of the electrode, which was found equal to 416 cm2, by using the Randles-Sevcik equation (Eq. 1): Ip = 2.69 × 105 × AD 1/2 n 1/2 γ 1/2 C (Eq. 1) where Ip is the peak current (A), n is the number of electrons participating in the redox reaction (n = 1), A is the area of the electrode (cm2), D is the diffusion coefficient of the molecule in solution (7.60 × 10-6 cm2 s-1), C is the concentration of the probe molecule in the bulk solution (1 × 10-5 mol cm-3), and γ is the scan rate of the potential perturbation (0.01 V s-1). The catalytic activity of the novel electrode was then assessed by its capacity to electrogenerate H2O2. Since H2O2 is an intermediate required to produce •OH in the electro-Fenton reaction (Eq. 2) and is easier to detect, due to its longer half-life, its production arguably constitutes a better indicator of the cathode efficiency than the generation of •OH (4, 5): H2O2 + Fe2+ → Fe3+ + •OH + HO- (Eq. 2) The influence of the current density (normalized to the cathode projected surface area) on H2O2 production through cathodic reduction of O2 was thus evaluated and the best results were obtained at 1.25 mA cm-2, yielding a specific maximum H2O2 concentration of 3.00 ± 0.12 mg-H2O2 L-1 cm-2. Upon addition of Fe2+ to trigger the electro-Fenton reaction following Eq. 2, a mineralization efficiency of 80% was attained within 3 hours using real wastewater from the microelectronics industry in a lab-scale batch setup (Fig. 1) and the energy requirement was as low as 0.14 kWh g-1of total organic carbon (TOC). Following these promising results, a continuous-flow pre-pilot scale prototype was designed and a computational flow simulation allowed optimizing various parameters, from electrode spacing to flowrate and mass transfer rates towards the electrodes (Fig. 2). The performance of this pre-pilot prototype was evaluated in the same conditions as the lab-scale reactor, confirming and even improving the results obtained at lab-scale with over 90% TOC removal reached within 3 hours (Fig. 1). However, the energy consumption increased with the intensity of the treatment from 0.8 KWh g-1 of TOC in the first hour to 1.34 kWh/g-1of TOC, after 3 hours, showing that electro-Fenton would be better used as a pre-treatment technology, over a limited period of time. Overall, these results demonstrate the potential of electro-Fenton as a water treatment and reuse technology and constitute an important step towards full-scale applications. A detailed technical economic assessment of the technology for industrial water reuse will be presented at the conference.