On the utility of run to run control in semiconductor manufacturing

Musacchio, John; Rangan, Sundeep; Spanos, Costas J.; Poolla, Kameshwar

doi:10.1109/issm.1997.664523

Cited by 7 publications

(10 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There is an essential difference between our EC2C approach and traditional R2R approaches. In semiconductor processes, run-to-run usually means wafer-to-wafer, lot-to-lot or batch-to-batch, which is more of a statistical process control, even though there are a broader classification of R2R including statistics, estimation and artificial intelligence (Musacchio, 1998). In the proposed cycle-to-cycle control, we adopt a smaller scale and smaller time step and focus on a single part fabrication process, that is, to control a single "Run" of the process instead of "Run-to-Run" in a batch process.…”

Section: Methodology Of Evolutionary Cycle-to-cycle Controlmentioning

confidence: 99%

“…The EC2C controller recursively estimates parameters a i and b i in the process model and then invert this model to find a recipe for the desired output. An estimation algorithm is essential for a complete design of EC2C control (Musacchio, 1998).…”

Section: Estimation Methods In Ec2c Time Controlmentioning

confidence: 99%

“…Similar to R2R control, an EC2C process control strategy requires a predictive model which is not necessary to be accurate or detailed (Musacchio, 1998). The process model is the relationship between the outputs and the inputs.…”

Section: Predictive Model In Ec2c Controlmentioning

confidence: 99%

See 2 more Smart Citations

Simulation study on evolutionary cycle to cycle time control of exposure controlled projection lithography

Zhao

Rosen

2016

RPJ

View full text Add to dashboard Cite

Purpose Exposure controlled projection lithography (ECPL) is an additive manufacturing process based on controlled UV photopolymerization. This paper aims to explore an advanced closed-loop control methodology to ECPL. Design/methodology/approach This paper proposes an evolutionary cycle to cycle (EC2C) control method, and started with a reduced order EC2C time control to control only the exposure time for given DMD bitmaps, which correspond to target 3D part cross-sections. A preliminary EC2C time control scheme was developed and followed by two types of EC2C time controllers based on two different parameter estimation methods, recursive least squares and L1 norm minimization (L1Min). Both algorithms were in an exponential weighted form, resulting in EWRLS and EWL1Min, to weight more on recent data to reflect the current process dynamics. Findings EWRLS was found to outperform EWL1Min in terms of computation speed and stability. The simulation study demonstrated that the proposed EC2C time control method was capable of adaptively tracking the ECPL process dynamics and updating online the model parameters with real-time measurements. It could control perfectly the exposure time for each bitmap, achieving the desired height for each layer and resulting in a total cured height conforming to the target 3D part height. Research limitations/implications The accuracy of EC2C time control method relies heavily on fast and accurate measurement, and this research assumes availability of an adequate real-time metrology. Measurement errors are not considered in this paper and will be explored in future. Only simulation study was performed without physical experiments to verify the EC2C controller. Practical implications For implementation, a real-time measurement system needs to be developed and the EC2C control software needs to be programmed and interfaced with the physical system. Originality/value It concludes that EC2C control method is very promising for a physical implementation, and could be extended for the development of a more comprehensive closed-loop controller for both exposure time and intensity to improve the ECPL process precision and robustness.

show abstract

Section: Methodology Of Evolutionary Cycle-to-cycle Controlmentioning

confidence: 99%

Section: Estimation Methods In Ec2c Time Controlmentioning

confidence: 99%

See 1 more Smart Citation

Simulation study on evolutionary cycle to cycle time control of exposure controlled projection lithography

Zhao

Rosen

2016

RPJ

View full text Add to dashboard Cite

show abstract

“…R2R controller uses data from past process runs to adjust settings for the next run as presented for example in [9] and [3]. Note that a R2R controller is associated to one machine and one job family.…”

Section: Introductionmentioning

confidence: 99%

A New CP-Approach for a Parallel Machine Scheduling Problem with Time Constraints on Machine Qualifications

Malapert¹,

Nattaf²

2019

Integration of Constraint Programming, Artificial Intelligence, and Operations Research

View full text Add to dashboard Cite

This paper considers the scheduling of job families on parallel machines with time constraints on machine qualifications. In this problem, each job belongs to a family and a family can only be executed on a subset of qualified machines. In addition, machines can lose their qualifications during the schedule. Indeed, if no job of a family is scheduled on a machine during a given amount of time, the machine lose its qualification for this family. The goal is to minimize the sum of job completion times, i.e. the flow time, while maximizing the number of qualifications at the end of the schedule. The paper presents a new Constraint Programming model taking more advantages of the CP feature to model machine disqualifications. This model is compared with two existing models: an Integer Linear Programming (ILP) model and a Constraint Programming (CP) model. The experiments show that the new CP model outperforms the other model when the priority is given to the number of disqualifications objective. Furthermore, it is competitive with the other model when the flow time objective is prioritized.

show abstract

“…Although research has been performed on this problem, very little has been done to integrate APC constraints. In R2R control for instance, a R2R controller uses data from past process runs to adjust settings for the next run as presented for example in Musacchio et al (1997). Note that a R2R controller is associated to one machine and one job family.…”

Section: Introductionmentioning

confidence: 99%

Scheduling job families on non-identical parallel machines with time constraints

2012

View full text Add to dashboard Cite

International audienceThis paper studies the scheduling of lots (jobs) of different product types (job family) on parallel machines, where not all machines are able to process all job families (non-identical machines). A special time constraint, associated to each job family, should be satisfied for a machine to remain qualified for processing a job family. This constraint imposes that the time between the executions of two consecutive jobs of the same family on a qualified machine must not exceed the time threshold of the family. Otherwise, the machine becomes disqualified. This problem comes from semiconductor manufacturing, when Advanced Process Control constraints are considered in scheduling problems. To solve this problem, two Mixed Integer Linear Programming models are proposed that use different types of variables. Numerical experiments show that the second model is much more effective, and that there is a trade-off between optimizing the scheduling objective and maximizing the number of machines that remain qualified for the job families. Two heuristics are also presented and studied in the numerical experiments

show abstract

On the utility of run to run control in semiconductor manufacturing

Cited by 7 publications

References 7 publications

Simulation study on evolutionary cycle to cycle time control of exposure controlled projection lithography

Simulation study on evolutionary cycle to cycle time control of exposure controlled projection lithography

A New CP-Approach for a Parallel Machine Scheduling Problem with Time Constraints on Machine Qualifications

Scheduling job families on non-identical parallel machines with time constraints

Contact Info

Product

Resources

About