Manufacturing Process Control Research Articles

Neural network driven sensitivity analysis of diffraction-based overlay metrology performance to target defect features

Overlay (OVL) is one significant performance indicator for the lithography process control in semiconductor manufacturing. The accuracy of the OVL metrology is extremely critical for guarantee the lithography quality. Currently, diffraction-based overlay (DBO) is one of the mainstream OVL metrology techniques. Unfortunately, the accuracy of the DBO metrology is largely affected by the defect features of the OVL target. Therefore, there is a strong need to investigate the impacts of these target defects on the DBO metrology performance. However, efficiently investigating the statistical and interactive impacts of various DBO target defects remains challenging. This study aims to address this issue through proposing an intelligent sensitivity analysis approach. A cumulative distribution based global sensitivity analysis (GSA) method is utilized to assess the nonlinear influences of multiple defects in the OVL target on the DBO inaccuracy. The scenarios with both known and unknown distributions of the OVL target defects are considered. For the former, a neural network driven forward model is constructed for fast calculating the optical diffraction responses to accelerate the GSA process. For the latter, another neural network based inverse model are built for efficiently estimating the distribution of the target defects. Finally, a series of simulation experiments are conduct for typical DBO targets with multiple common defect features. The results demonstrate the effectiveness and robustness of the proposed approach as well as give valuable insights into the DBO defect analysis. Our study provides a strong tool to assist the practitioners in achieving intelligent and efficient DBO analysis and thus in enhancing OVL metrology performance.

Development of Sustainable 3D Printable Ternary Composite

3D printing technology has the potential to reduce construction waste through its controlled manufacturing process and optimized material consumption, making it a promising sustainable construction approach with reduced environmental impact. However, current 3D printed structures often have a high carbon footprint due to the large amounts of fine-grained primary materials used in mortar composites. Most efforts to develop 3D printed structures are associated with concrete and Portland cement-based materials.This research presents an innovative approach for the reuse of gypsum to develop a material suitable for 3D printing, with the goal to industrialize the usage of waste gypsum in civil engineering. The objective of this study is to develop a sustainable gypsum-cement-pozzolanic (GCP) ternary binder composite, incorporating recycled industrial waste gypsum materials such as construction demolition waste gypsum (CDG), along with Portland cement and secondary pozzolanic materials. The GCP ternary composite combines the desired properties of gypsum, such as fast setting time, with those of Portland cement, which offers high final strength. The addition of a pozzolanic component is essential to ensure the chemical stability of this ternary system.As part of this study, ternary composites were designed and tested for their mechanical properties, durability, and suitability for 3D printing. Development and optimization of the ternary composite was carried out as a loop of procedures (Fig.1). Through these optimization procedures, a range of mixtures with necessary properties for 3D printing was found. To ensure sustainability and to assess the environmental impact of the mixtures, life cycle assessment (LCA) was performed.The LCA of GCP composites showed at least 1.5 times lower embodied energy and carbon dioxide emissions compared to Portland cement-based mortars. This is without considering avoided emissions from the end-of-life stage of gypsum waste, as well as higher technical properties that make GCP composites superior to regular cement composites. Considering these factors, GCP composites exhibit even lower emissions and are considered a viable alternative to Portland cement-based mortars.

Immersive Digital Twin under ISO 23247 Applied to Flexible Manufacturing Processes

Digital twin (DT) technology provides a path for implementing cyber–physical systems (CPS) and developing smart manufacturing because they are essential tools for monitoring and controlling manufacturing processes. It is considered a vital technology in smart manufacturing and is being widely researched in academia and industry. Furthermore, the combination of DTs and immersive environments has shown great potential for integrating novel capabilities into the new generation of CPS. This research presents an architecture for implementing immersive digital twins under ISO 23247 in flexible manufacturing processes. The proposed system is based on the integration of DT technologies in conjunction with augmented reality (AR) and gesture tracking, and validation was performed in the sorting station of the MPS 500 to increase the interaction and flexibility between physical and virtual environments in real time, thus enhancing the capabilities of the DT. The methodology used for the design and implementation of the DT includes (1) general principles and requirements; (2) models with functional views based on domains and entities; (3) attributes of the observable manufacturing elements; and (4) protocols for the exchange of information between entities. The results show that the integration of these technologies improves the monitoring, control, and simulation capabilities of processes using 3D resources and immersive environments, achieving a higher level of interactivity. In addition, error detection tests were carried out, where a reduction of time was observed in the resolution of errors that may be caused by internal or external disturbances of the process, thus avoiding production delays.

Coating process control in lithium-ion battery manufacturing using cumulative sum charts

Coating process control in lithium-ion battery manufacturing using cumulative sum charts

Aflatoxin and microbial safety in commercial confectionery peanut paçoca produced in São Paulo state – Brazil

Paçoca is a typical Brazilian confectionery of Indigenous origin prepared by mixing ground roasted peanuts, sugar, and salt. A total of 52 samples of paçoca (37 pressed, 15 cooked), produced and sold in the state of São Paulo, were evaluated for aflatoxin content (AFB1, AFB2, AFG1 and AFG2), moisture content and water activity. The microbiological quality was assessed by counting mesophilic aerobic bacteria, fungal and yeast counts, most likely number of thermotolerant coliforms and Escherichia coli and Salmonella sp. The limit of quantification (LOQ) by Ultra Performance Liquid Chromatography coupled to the Mass Spectrometer (UPLC-MS/MS) ranged from 0.29 μg kg−1 for AFB1 and AFG2; 0.36 μg kg−1 for AFG1 and 0.37 μg kg−1 for AFB2. The recovery averages were 89.2, 83.5, 80.5 and 82.8% for AFB1, AFB2, AFG1 and AFG2. 71% of the samples showed aflatoxin contamination, ranging from 0.29 to 38.8 μg kg−1 (mean of 7.06 μg kg−1). Six samples (16%) exceeded the limit value tolerated by the Brazilian guideline for AFT (20 μg kg−1). The mean AFT content in cooked paçocas (19.55 μg kg−1) was higher than the levels in pressed paçocas (9.08 μg kg−1) (p < 0.05). AFB1 had a higher incidence in the two groups of paçocas (75% pressed and 60% cooked). The mean moisture content and water activity of cooked paçocas (8.86% and 0.7081) were higher than those of pressed paçocas (2.08% and 0.4863). Twenty-one samples (56%) showed fungal and yeast contamination ranging from 1.0 to 4.25 log CFU g−1. The total count of aerobic mesophilic bacteria in cooked paçocas (4.85 log CFU g−1) was also higher than in pressed paçocas (2.77 log CFU g−1) (p < 0.05). E. coli was confirmed in these samples in the range of 3.0–160 NMP g−1. Salmonella sp. was detected in 2 samples of pressed paçocas, indicating poor hygiene, inefficient heat treatment or possible cross-contamination. The results of this study demonstrate the need for greater control of the paçoca production chain through the adoption of Good Manufacturing Practices and process control, to prevent exposure of the population to aflatoxins and possible pathogenic microorganisms.

Powder-size driven facile microstructure control in powder-fusion metal additive manufacturing processes

Microstructure control in metal additive manufacturing is highly desirable for superior and bespoke mechanical performance. Engineering the columnar-to-equiaxed transition during rapid solidification in the additive manufacturing process is crucial for its technological advancement. Here, we report a powder-size driven melt pool engineering approach, demonstrating facile and large-scale control in the grain morphology by triggering a counterintuitive response of powder size to the additively manufactured 316 L stainless steel microstructure. We obtain coarse-grained (>100 μm) or near-monocrystalline microstructure using fine powders and near-equiaxed, fine-grained (<10 μm) microstructure using coarse powders. This approach shows resourceful adaptability to directed energy deposition and powder-bed fusion with no added cost, where the particle-size dependent powder-flow preheating effects and powder-bed thermophysical properties drive the microstructural variations. This work presents a pathway for leveraging feedstock particle size distribution towards more controllable, cost-effective, and sustainable metal additive manufacturing.

Temperature control of beer fermentation based on variable domain fuzzy PID and neural network technology and its application analysis

AbstractTemperature control in the beer manufacturing process is crucial for product quality. Given the gap between China's automation in beer production and the international level, improving the technology in this area has gradually become a core issue in optimizing domestic beer production. This study combines a proportional integral derivative controller with a fuzzy modeling strategy and incorporates a variable‐domain structure to propose a variable‐domain fuzzy proportional integral derivative controller control method. To cope with the challenges of production interaction, the study also introduces neural network technology. The experimental data indicated that the variable‐domain fuzzy proportional integral derivative controller outperforms the conventional proportional integral derivative controller and the fuzzy proportional integral differential controller in terms of overshooting, with a maximum overshoot of only 1.0, compared with 0.50 and 0.70, respectively. The variable‐domain fuzzy proportional integral differential controller exhibited a minimal overshoot of only 0.01 when the model parameter is increased by 20%. In comparison, the other methods reach overshoot values of 0.92 and 1.0. The proposed method maintained superior stability even under the influence of impulse disturbance, step disturbance, and modeling variations. These results demonstrated that the research method is significantly more stable than both the proportion integration differentiation (PID) controller and fuzzy PID controller in complex dynamic parameter environments. The proposed method involved 60 rounds of neural network control, which was successfully implemented. The temperature readings T1 and T2 remained stable within the range of 1.0%–1.02% throughout the experiments. The study demonstrates that the proposed methods have higher accuracy and less fluctuation in actual application, making them more available. Taken together, the above results show that the combination of variable‐domain fuzzy PID controller and neural network technology in beer production has achieved excellent control results. This study not only provides a strong technical support for the progress of beer production technology in China, but also has important industrial application value and wide promotion prospects.

Failure Analysis of Header Flange Cracking in Propane Dehydrogenation Unit

The boiler fault of propane dehydrogenation unit is investigated by macroscopic inspections, mechanical properties testing, chemical composition analysis, metallographic microstructure, X-ray diffraction (XRD), Scanning Electron Microscopy (SEM) and Energy Dispersive Spectroscopy (EDS). The experiments confirm the reason for cracks occurrence of SA182-F347H boiler that the stress corrosion cracking occurs. The sulfur injected into the device reacts with the flange metal to produce sulfide corrosion products. The improper control of flange manufacturing process and the operating temperature in the sensitized range lead to the sensitization of flange materials and reduce the corrosion resistance. During the equipment shutdown, the sulfide corrosion products on the surface of the flange reacted with air and water to form polysulfuric acid, and causing stress corrosion cracking and leakage of the sensitized F347H stainless steel flange.

Deep reinforcement learning-assisted extended state observer for run-to-run control in the semiconductor manufacturing process

In the semiconductor manufacturing process, extended state observer (ESO)-based run-to-run (RtR) control is an intriguing solution. Although an ESO-RtR control strategy can effectively compensate for the lumped disturbance, appropriate gains are required. In this article, a cutting-edge deep reinforcement learning (DRL) technique is integrated into ESO-RtR, and a composite control framework of DRL-ESO-RtR is developed. In particular, the well-trained DRL agent serves as an assisted controller, which produces appropriate gains of ESO. The optimized ESO then presents a preferable control recipe for the manufacturing process. Under the RtR framework, the gain adjustment problem of ESO is formulated as a Markov decision process. An efficient state space and reward function are wisely designed using the system’s observable information. Correspondingly, the gain of the ESO is adaptively adjusted to cope with changing environmental disturbances. Finally, a twin-delayed deep deterministic policy gradient algorithm is employed to implement the suggested scheme. The feasibility and superiority of the developed method are validated in a deep reactive ion etching process. Comparative results demonstrate that the presented scheme outperforms the ordinary ESO-RtR controller in terms of disturbance rejection.

Modeling the selective growth advantage of genetically variant human pluripotent stem cells to identify opportunities for manufacturing process control

Background aimsThe appearance of genetically variant populations in human pluripotent stem cell (hPSC) cultures represents a concern for research and clinical applications. Genetic variations may alter hPSC differentiation potential or cause phenotype variation in differentiated cells. Further, variants may have properties such as proliferative rate, or response to the culture environment, that differ from wild-type cells. As such, understanding the behavior of these variants in culture, and any potential operational impact on manufacturing processes, will be necessary to control quality of putative hPSC-based products that include a proportion of variant threshold in their quality specification. MethodsHere we show a computational model that mathematically describes the growth dynamics between commonly occurring genetically variant hPSCs and their counterpart wild-type cells in culture. ResultsWe show that our model is capable of representing the growth behaviors of both wild-type and variant hPSCs in individual and co-culture systems. ConclusionsThis representation allows us to identify three critical process parameters that drive critical quality attributes when genetically variant cells are present within the system: total culture density, proportion of variant cells within the culture system and variant cell overgrowth. Lastly, we used our model to predict how the variability of these parameters affects the prevalence of both populations in culture.

24. CRITICAL IMPORTANCE OF NOZZLE TEMPERATURE TO BIOPRINT VASCULAR GRAFTS WITH SHEAR-THINNING HYDROGEL

Abstract Introduction The development of vascular grafts adapted to the nature of the patient and their treatment is constantly evolving. Fine control of the bioprinting manufacturing process is a key issue to obtain proper personalized vascular grafts. The aim is to check the influence of the nozzle temperature during bioprinting a Gelatin/Gellan Gum/Alginate hydrogel vascular grafts. Methods Vascular grafts were printed using a Gelatin/Gellan Gum/Alginate (8:3:0.5 %w/v) hydrogel in an BioX bioprinter with the printhead and bed set up to 65 °C. Two configurations were set: 1) a standard conical tip (nozzle) and 2) the same nozzle thermally insulated with aluminum foil to uniformly distribute the temperature. A temperature control using a 100 kΩ thermistor was performed in four points: chamber, nozzle, printhead, and bed. Four heating/cooling cycles were performed in each control point. Results The grafts printed with the standard nozzle exhibited irregular wall thickness and an uneven surface, attributed to lower temperatures in the conical tip. In contrast, results from the aluminum-covered nozzle indicated temperature distributions of 33.13±0.87, 49.03±3.92, 52.38±3.53, and 64.98±0.10 °C for chamber, nozzle, printhead, and bed, respectively. This improved temperature distribution enhanced the printability of Gelatin/Gellan Gum/Alginate vascular grafts, resulting in a more regular wall thickness and surface profile. Conclusion The optimization of temperature in the nozzle during the printing process is a critical parameter for bioprinting vascular grafts using this shear-thinning Gelatin/Gellan Gum/Alginate hydrogel.

Fault diagnosis in semiconductor manufacturing processes using a CNN-based generative adversarial network1

In the current industry, quality inspection in semiconductor manufacturing is of immense significance. Significant achievements have been made in fault diagnosis in fabricated semiconductor wafer manufacturing due to the development of machine learning. Since real-time intermediate signals are non-linear and time-varying, the signals undergo various distortions due to changes in equipment, material, and process. This leads to a drastic change in information in intermediate signals. This paper presents a fault diagnosis model for semiconductor manufacturing processes using a generative adversarial network (GAN). The study aims to address the challenges associated with efficient and accurate fault identification in these complex processes. Our approach involves the extraction of relevant components, development of a paired generator model, and implementation of a deep convolutional neural network. Experimental evaluations were conducted using a comprehensive dataset and compared against six existing models. The results demonstrate the superiority of our proposed model, showcasing higher accuracy, specificity, and sensitivity across various shift tasks. This research contributes to the field by introducing a novel approach for fault diagnosis, paving the way for improved process control and product quality in semiconductor manufacturing. Future work will focus on further optimizing the model and extending its applicability to other manufacturing domains.

Process Monitoring, Diagnosis and Control of Additive Manufacturing

Process Monitoring, Diagnosis and Control of Additive Manufacturing

Development of a rapid solidification model for additive manufacturing process and application to Al-Si alloy

Accurate prediction of the solidification microstructure produced by additive manufacturing (AM) process is indispensable for the precise manufacturing process control. In the present study, a simple solidification model that can be applied to the powder bed fusion – laser beam (PBF-LB) process has been developed by integrating conventional and rapid solidification theories in literature. Based on the solidification parameters calculated by Rosenthal equation, the present solidification model can predict the primary dendrite arm spacing (PDAS), primary cell fraction, and solute profile using solute trapping, dendrite tip undercooling, and eutectic undercooling. Single track experiments for Al-12Si binary alloy were also performed to validate the accuracy of the model.

Quantitative study of the thickness-dependent stress in indium tin oxide thin films

A quantitative analysis of stress in indium tin oxide (ITO) films with varying thicknesses was conducted using high-precision stylus profilometry. The experimental results revealed a transition of the stress type from tensile to compressive as the ITO film thickness increased. Further investigation of the surface morphology of the ITO films indicated that the tensile stress originated from the impingement and coalescence of newly deposited equiaxed grains, while the observed compressive stress was attributed to the incorporation of excess material in the boundaries of columnar grains. The ability to monitor stress evolution in sputtered ITO films on glass substrates provides valuable insights for process control in ITO device manufacturing.

Towards data driven quality monitoring: alignment and correlation of photodiode-based co-axial melt pool monitoring signals to part quality in laser powder bed fusion

Online quality control of advanced manufacturing processes often utilize process monitoring as a qualification method to reduce or eliminate time consuming and costly destructive and non-destructive tests. The need for process monitoring based qualification becomes more apparent with the advent of laser based additive manufacturing for complex geometries and lot-size-one production. However, the usability of the process monitoring solution heavily relies on the accuracy of the established alignment and assignment of process signals to specific part and defects herein. This research proposes a data fusion strategy for data pre-processing to structure and align multi-source sensor data for improving correlation accuracy in process monitoring. A laser powder bed fusion system equipped with a photodiode-based co-axial melt pool monitoring is used for the study. First, the monitoring system collects photon emission data from the melt pool using photodiodes with high frequency (>100 kHz) which are aligned spatially and temporally using position data (xy-coordinates) of the galvanometer scanner and the laser ON/OFF signals. Next, a clustering approach is used to assign each photodiode signal with the individual parts on the build platform. Finally, correlation between melt pool monitoring signals and part level porosity is studied using micro-computed X-ray tomography (μCT) from a build job containing parts produced under varying process conditions.

DRL-dEWMA: a composite framework for run-to-run control in the semiconductor manufacturing process

DRL-dEWMA: a composite framework for run-to-run control in the semiconductor manufacturing process

Anti-crosstalk absolute phase retrieval method for microscopic fringe projection profilometry using temporal frequency-division multiplexing.

In microscopic fringe projection profilometry (MFPP), the traditional absolute phase retrieval method using composite frequency fringe has the shortcomings of low accuracy and poor robustness due to mutual crosstalk of harmonic from the different channels of frequency-division multiplexing. In this study, an absolute phase retrieval method that avoids the inter-channel crosstalk is proposed. By introducing guard bands to accommodate the frequency channels corresponding to the second harmonic that dominate the high order harmonics, the aliasing between the second harmonic and the fundamental is eliminated. Consequently, phase maps without crosstalk can be demodulated using appropriate phase-shifting algorithms. The proposed method is well-suited for high-precision three-dimensional shape measurement scenarios in many fields such as integrated circuit manufacturing process control and micro-electro-mechanical system quality inspection. The experiment results demonstrate that the anti-crosstalk method is effective and can realize three-dimensional reconstruction for discontinuous planar surface and spherical surface.

Residual stress and distortion control in wire-arc additive manufacturing process through novel modular substrate

Wire arc additive manufacturing (WAAM) is a proven technology in metal additive manufacturing (AM), which can produce large-scale components at a faster production rate. Since WAAM adopts a welding heat source for metal deposition over the substrate, severe thermal gradients develop around the deposition resulting in residual stresses in the substrate, thereby distorting the substrate. In the case of area filling in metal AM processes, heat accumulation is progressive and a large amount of heat is accumulated at the end of the deposition. This leads to thermal imbalance over the substrate and an increase in the shrinkage forces during cooling, thereby causing higher distortion. This paper mainly deals with the mitigation of distortion in the substrate for area filling caused by the thermal imbalance over the substrate. This thermal imbalance is addressed in two ways; a) by selectively insulating and b) by selectively conducting the heat from the substrate. Simulations are conducted to study the thermal evolution during area filling and counterbalance the heat by insulating and conducting the substrate selectively during experimental deposition. A comparative assessment of distortions of the full conduction, full insulation, selective conduction and selective insulation cases is detailed. The selective thermal management strategy showed lower thermal imbalance over the substrate resulting in lower distortion.

Automated Interlayer Wall Height Compensation for Wire Based Directed Energy Deposition Additive Manufacturing.

Part quality monitoring and control in wire-based directed energy deposition additive manufacturing (w-DEDAM) processes has been garnering continuous interest from both the academic and industrial sectors. However, maintaining a consistent layer height and ensuring that the wall height aligns closely with the design, as depicted in computer-aided design (CAD) models, pose significant challenges. These challenges arise due to the uncertainties associated with the manufacturing process and the working environment, particularly with extended processing times. To achieve these goals in an industrial scenario, the deposition geometry must be measured with precision and efficiency throughout the part-building process. Moreover, it is essential to comprehend the changes in the interlayer deposition height based on various process parameters. This paper first examines the behaviour of interlayer deposition height when process parameters change within different wall regions, with a particular focus on the transition areas. In addition, this paper explores the potential of geometry monitoring information in implementing interlayer wall height compensation during w-DEDAM part-building. The in-process layer height was monitored using a coherent range-resolved interferometry (RRI) sensor, and the accuracy and efficiency of this measurement were carefully studied. Leveraging this information and understanding of deposition geometry, the control points of the process parameters were identified. Subsequently, appropriate and varied process parameters were applied to each wall region to gradually compensate for wall height. The wall height discrepancies were generally compensated for in two to three layers.