Semiconductor Manufacturing Operations Research Articles

Samsung Uses Data-Driven Approach to Manage Work-in-Process of Bottlenecks in Semiconductor Manufacturing Operations

Samsung Uses Data-Driven Approach to Manage Work-in-Process of Bottlenecks in Semiconductor Manufacturing Operations

Problems and Solution Methods of Machine Scheduling in Semiconductor Manufacturing Operations: A Survey

Machine scheduling problems associated with semiconductor manufacturing operations (SMOs) are one of the major research topics in the scheduling literature. Lots of papers have dealt with different variants of SMOs’ scheduling problems, which are generally difficult to tackle theoretically and computationally. In this paper, the single machine, parallel machines, flow shops, and job shops scheduling problems from SMOs have been reviewed, based on different processing constraints, e.g., batch processing, auxiliary resources, incompatible job families, and reentrant flow, etc., with the cycle time, flow time, and throughput-related performance measures. Given the vast and diverse nature of the current literature, it is urgently needed to make a systematic survey in order to identify the important research problems, research trends, and the progress of the related solution methods, as well as clarify future research perspectives. We hope the findings and observations could provide some insights to the researchers and practitioners in this domain.

Design of Electrochemical Cells for Targeted Metals Removal Using Carbon Electrodes

Selective separations are needed in a wide variety of industrial and commercial applications where discharge to publicly owned treatment works (POTW) requires certain metals concentrations to be sufficiently low to protect public health and the surrounding environment. Metals such as lead (Pb), copper (Cu), chromium (Cr), nickel (Ni), zinc (Zn), and cadmium (Cd) represent a non-exhaustive list of compounds requiring removal for discharge regulations. Typically, coagulants such as iron and aluminum combined with precipitation chemistry or ion exchange processes are used to meet these regulations.1 However, these methods are not particularly selective and can produce sizable sludge waste that must be disposed of properly. Size-selective membranes are one alternative approach, but the pretreatment requirements for these membranes further complicates the water treatment process. Capacitive deionization (CDI) is an emerging water treatment option as well with notable advances in recent years, but it currently lacks the selectivity needed for many industrial streams.2 Therefore, alternative methods are being sought to realize these separations.The use of electrochemical processes offers a number of benefits such as a defined interface for interaction with the metal of interest, the ability to modulate the interface easily and quickly through changes in localized voltage, use of the electrical current to monitor system conditions, and the in situ generation of chemical species that can aid in the separation. Of particular interest in a wide variety of industrial applications is the removal or Cu from water being discharged to the POTW. Cu is found in waste streams emanating from electroplating, electronics, semiconductor, and battery manufacturing operations. While coagulation approaches mentioned above can often be used to meet effluent regulations, metal recovery through an electrochemical process can be highly effective and efficient, reaching current efficiencies in excess of 95% in many applications. The ability to plate Cu at a cathode under highly localized conditions affords the removal of Cu down to levels <100 ppb. Electrowinning has been used for over a century in the creation of purified metals such as Cu, so the concept is not entirely new, but the design of electrode materials and overall cell construction capable of removing Cu to such low concentrations in streams that have conductivities <1 mS/cm opens up new avenues for water treatment in industrial and commercial waste.In this talk, electrochemical cell design and operation as well as feed water conditions will be reviewed towards the development of selective metal removal technologies. Copper removal will be highlighted as an example, but the concept will also be applied to other metals of interest, demonstrating the more ubiquitous nature of the approach.

A Reinforcement Learning Approach to Robust Scheduling of Semiconductor Manufacturing Facilities

As semiconductor manufacturers, recently, have focused on producing multichip products (MCPs), scheduling semiconductor manufacturing operations become complicated due to the constraints related to reentrant production flows, sequence-dependent setups, and alternative machines. At the same time, the scheduling problems need to be solved frequently to effectively manage the variabilities in production requirements, available machines, and initial setup status. To minimize the makespan for an MCP scheduling problem, we propose a setup change scheduling method using reinforcement learning (RL) in which each agent determines setup decisions in a decentralized manner and learns a centralized policy by sharing a neural network among the agents to deal with the changes in the number of machines. Furthermore, novel definitions of state, action, and reward are proposed to address the variabilities in production requirements and initial setup status. Numerical experiments demonstrate that the proposed approach outperforms the rule-based, metaheuristic, and other RL methods in terms of the makespan while incurring shorter computation time than the metaheuristics considered. Note to Practitioners —This article studies a scheduling problem for die attach and wire bonding stages of a semiconductor packaging line. Due to the variabilities in production requirements, the number of available machines, and initial setup status, it is challenging for a scheduler to produce high-quality schedules within a specific time limit using existing approaches. In this article, a new scheduling method using reinforcement learning is proposed to enhance the robustness against the variabilities while achieving performance improvements. To verify the robustness of the proposed method, neural networks (NNs) trained on small-scale scheduling problems are used to solve large-scale scheduling problems. Experimental results show that the proposed method outperforms the existing approaches while requiring a short computation time. Furthermore, the trained NN performs well in solving unseen real-world scale problems even under stochastic processing time, suggesting the viability of the proposed method for real-world semiconductor packaging lines.

Distance matrix based heuristics to minimize makespan of parallel batch processing machines with arbitrary job sizes and release times

Batch scheduling is prevalent in many industries such as burn-in operations in semiconductor manufacturing and heat treatment operations in metalworking. In this paper, we consider the problem of minimizing makespan on parallel batch processing machines in the presence of non-identical job sizes and arbitrary release times. Since the problem under study is strongly NP-hard (non-deterministic polynomial-time hard), we develop a number of efficient heuristics. In the proposed heuristics, several distance matrix based approaches are first developed to form batches and a scheduling technique is devised to assign the batches to parallel machines. The performance of the proposed heuristics is evaluated by comparing their results to a set of existing heuristic and meta-heuristic algorithms in the literature. The computational experiments show that the proposed heuristics are competitive with respect to solution quality. Moreover, the computational costs of the proposed heuristics are very little.

Evolutionary fuzzy ARTMAP neural networks for classification of semiconductor defects.

Wafer defect detection using an intelligent system is an approach of quality improvement in semiconductor manufacturing that aims to enhance its process stability, increase production capacity, and improve yields. Occasionally, only few records that indicate defective units are available and they are classified as a minority group in a large database. Such a situation leads to an imbalanced data set problem, wherein it engenders a great challenge to deal with by applying machine-learning techniques for obtaining effective solution. In addition, the database may comprise overlapping samples of different classes. This paper introduces two models of evolutionary fuzzy ARTMAP (FAM) neural networks to deal with the imbalanced data set problems in a semiconductor manufacturing operations. In particular, both the FAM models and hybrid genetic algorithms are integrated in the proposed evolutionary artificial neural networks (EANNs) to classify an imbalanced data set. In addition, one of the proposed EANNs incorporates a facility to learn overlapping samples of different classes from the imbalanced data environment. The classification results of the proposed evolutionary FAM neural networks are presented, compared, and analyzed using several classification metrics. The outcomes positively indicate the effectiveness of the proposed networks in handling classification problems with imbalanced data sets.

Stability of alumina, ceria, and silica nanoparticles in municipal wastewater

Inorganic oxide nanoparticles (NPs) are used in semiconductor manufacturing operations such as wafer chemical-mechanical planarization (CMP). Understanding the stability of NPs in municipal wastewater is essential for the evaluation of the fate of NPs released to municipal wastewater treatment plants (WWTPs). This study aimed to evaluate the stability of Al(2)O(3), CeO(2), and SiO(2) NPs and CMP waste effluents containing these NPs in municipal wastewater. Al(2)O(3) and CeO(2) NPs were destabilized by wastewater constituents, as indicated by the formation of large agglomerates. However, the same NPs in the CMP waste slurries showed high stability in wastewater, probably due to additives present in the slurry that modify the surface chemistry of the particles. Likewise, both the commercial SiO(2) NPs and the CMP waste slurry containing SiO(2) NPs showed substantial stability in wastewater since this NP has a very low point of zero charge, which suggests that this NP could be the hardest one to remove in conventional WWTPs by aggregation-sedimentation. In summary, the results indicate that wastewater may destabilize NPs suspensions, which would facilitate NP removal in WWTPs. However, some chemicals present in real CMP slurries may counterbalance this effect. More research is needed to completely understand the surface chemistry involved.

Minimising makespan on a single batch processing machine with dynamic job arrivals and non-identical job sizes

Batch scheduling is a prevalent policy in many industries such as burn-in operations in semiconductor manufacturing and heat treatment operations in metalworking. In this paper, we consider the problem of minimising makespan on a single batch processing machine in the presence of dynamic job arrivals and non-identical job sizes. The problem under study is NP-hard. Consequently, we develop a number of efficient construction heuristics. The performance of the proposed heuristics is evaluated by comparing their results to two lower bounds, and other solution approaches published in the literature, namely the first-fit longest processing time-earliest release time (FFLPT-ERT) heuristic, hybrid genetic algorithm (HGA), joint genetic algorithm and dynamic programming (GA+DP) approach and ant colony optimisation (ACO) algorithm. The computational experiments demonstrate the superiority of the proposed heuristics with respect to solution quality, especially for the problems with small size jobs. Moreover, the computational costs of the proposed heuristics are very low.

Modelling and implementation of manufacturing direct labour allocation: a case study in semiconductor production operations

In many current semiconductor manufacturing operations, headcount is manually allocated periodically based on man-machine ratio. Attributed to non-optimised allocation of direct labour to operations/machines, considerable productivity loss occurs. The problem is further complicated by some dynamic and uncertain factors such as constantly changing production targets and work in progress, overlapped labour skills, and variability in manufacturing operations and labour absenteeism rates. Motivated by the needs in real practice, this study aims to develop a model for allocating a direct workforce among semiconductor manufacturing operations to meet production targets and maximise labour productivity. This paper presents a two-stage goal programming model for the headcount allocation problem. To enhance the model's pragmatic use, a queueing module is introduced to account for the interferences among the multi-machine operations. Computational experiments are carried out to evaluate the performances of the proposed algorithms and pilot runs are implemented in a factory. Finally, a prototype system is developed and has been proved to be useful in practice.

A survey of problems, solution techniques, and future challenges in scheduling semiconductor manufacturing operations

In this paper, we discuss scheduling problems in semiconductor manufacturing. Starting from describing the manufacturing process, we identify typical scheduling problems found in semiconductor manufacturing systems. We describe batch scheduling problems, parallel machine scheduling problems, job shop scheduling problems, scheduling problems with auxiliary resources, multiple orders per job scheduling problems, and scheduling problems related to cluster tools. We also present important solution techniques that are used to solve these scheduling problems by means of specific examples, and report on known implementations. Finally, we summarize some of the challenges in scheduling semiconductor manufacturing operations.

Optimal Preventive Maintenance Scheduling in Semiconductor Manufacturing Systems: Software Tool and Simulation Case Studies

This paper presents the architecture and implementation of a preventive maintenance optimization software tool (PMOST), based on algorithms for the optimal scheduling of preventive maintenance (PM) tasks in semiconductor manufacturing operations. We also present results from four complex simulation case studies, based on real industrial data and employing full fab models, to illustrate the use, data needs and outcomes produced by PMOST. These results demonstrate significant improvements in tool production and consolidation of PM tasks. We give a description of the different software modules that compose PMOST, to provide guidelines as well as a template for other implementations of the PM optimization algorithms utilized by PMOST.

A genetic algorithm based approach for scheduling of jobs containing multiple orders in a three-machine flowshop

This paper investigates a new scheduling problem of multiple orders per job (MOJ) in a three-machine flowshop consisting of an item-processing machine, a lot-processing machine and a batch-processing machine, for a semiconductor manufacturing operation that must form a layer on the wafers. The three-machine flowshop MOJ scheduling problem deals with sequencing customer orders, assigning orders to jobs, and then combining the formed jobs into batches. Genetic algorithms are presented for this scheduling problem to minimise the total weighted tardiness (TWT), in the presence of non-zero order ready times, different order due dates, different order weights and unequal order sizes. Extensive experiments were performed to compare the proposed genetic algorithm (GA)-based approach with the benchmark heuristics presented in previous studies. The experiments led to the conclusions that the GA-based approach is significantly superior over other heuristics in terms of TWT and can find near-optimal solutions in an acceptable amount of computation time.

Semiconductor manufacturing scheduling of jobs containing multiple orders on identical parallel machines

This paper investigates the identical parallel machine multiple orders per job (MOJ) scheduling problem to minimise total weighted order completion time, as motivated by 300 mm semiconductor manufacturing operations. Multiple orders per job scheduling is required when multiple customer orders are grouped together in one of a limited number of jobs for simultaneous processing – a challenge as customers measure order-level delivery performance, while manufacturers must schedule at the job level. We investigate two kinds of processing environments, serial and parallel processing, in the presence of non-zero order ready times and sequence-independent setup times between jobs. First, a mixed-integer program is proposed for the scheduling problem. As this problem is NP-Hard, a number of polynomial-time heuristic approaches are proposed. Experimental results identify appropriate heuristic techniques for analysing this practically motivated scheduling problem.

Automated S/TEM Sample Preparation for Semiconductor Process Support

Abstract The semiconductor industry has unquestionably entered the realm of nanotechnology. Critical dimensions of many features are specified in nanometers. Gate oxides are only a few nanometers thick. Barrier and seed layers for copper processes are not much more. Gate lengths are forecast at less than 20 nm by the end of the decade. Additionally, the drive to increase device density is leading to the adoption of FinFET and other new transistor designs that include complex three-dimensional structure. Even conventional planar CMOS designs now incorporate processes such as damascene interconnects that are inherently three dimensional. The combined need for higher spatial resolution and cross-sectional imaging of complex structures has led to a significant increase in the demand for scanning transmission electron microscopy (STEM) and (TEM) in semiconductor manufacturing operations.

A mathematical programming approach to the machine interference problem with fuzzy parameters

Machine interference is an important problem frequently encountered in production and manufacturing operations such as semiconductor manufacturing and maintenance operations. Due to uncontrollable factors, parameters in the machine interference problem may be fuzzy. This paper proposes a mathematical programming approach to construct the membership function of the performance measure of the machine interference system with the machine breakdown rate and the service rate being fuzzy numbers. On the basis of the extension principle, a pair of mathematical programs is formulated to calculate the lower and upper bounds of the fuzzy performance measure at possibility level α. By enumerating different values of α, the membership function of the system performance measure is constructed. A numerical example is solved successfully to demonstrate the validity of the proposed approach. Since the performance measure is completely described by a membership function rather than by a crisp value, the fuzziness of input parameters is conserved completely. Thus more information is provided for designing manufacturing systems. The successful extension of machine interference models to fuzzy environments permits machine interference models to have wider practical applications.

Odor control in semiconductor manufacturing facilities

Throughout the semiconductor industry, the need for controlled atmospheres - free from contaminant gases - for manufacturing processes is no longer just a worrisome concern - it has become a critical reality. Manufacturers have now fully recognized the fact that sensitive electronic components and electrical equipment will be damaged if they are exposed to airborne chemical contamination. More importantly, however, is the recognition that the health and safety of workers can be seriously compromised by over exposure to these same contaminants - many of which are toxic to humans. The total organic chemical concentrations in recirculated air are relatively high in semiconductor clean rooms when comp a red to outdoor or supply air. Low - level exposures to a variety of toxic materials commonly occur from minor leaks and spills as well. Thus, the combined effects of chemical exposures, whether additive or synergistic, are very important. The work - loss rates reported by this industry support many workers' assertions that occupational exposures to chemicals, gases, and other toxic materials commonly occur. Whereas subjective odor measurements are commonly used to initiate many indoor air quality (IAQ) investigations in commercial buildings, this same subjective criteria is many times used as the basis for initiating an evacuation in semiconductor manufacturing facilities. Because of the toxic and hazardous nature of many of the chemicals used in these operations, the detection of even the slightest odor, whether it comes from a leaking process line or a truck parked at the loading dock, may be enough to shut down a manufacturing facility for fear of worker safety. This paper presents odor control issues in semiconductor manufacturing operations from the perspective of employee safety and health, design engineering, and the selection and application of one type of odor control system.

5851293 Flow-stabilized wet scrubber system for treatment of process gases from semiconductor manufacturing operations : Lane Scott; Shah Dinesh; Colman Ronald; Heading Liam R; Simpson Eric Chandler, AZ, United States assigned to ATMI Ecosys Corporation

5851293 Flow-stabilized wet scrubber system for treatment of process gases from semiconductor manufacturing operations : Lane Scott; Shah Dinesh; Colman Ronald; Heading Liam R; Simpson Eric Chandler, AZ, United States assigned to ATMI Ecosys Corporation

A group-technology-coded literature review of semiconductor manufacturing operations publications: the MASMLAB bibliography web site

Modeling and simulation of semiconductor manufacturing operations is a rapidly developing area with a supporting published literature that is exploding in size. In this paper, we present a study of that literature which we use to build a searchable database that is implemented as an Internet site (http://www.eas.asu.edu/masmlab/Biblio.html). The semiconductor manufacturing modeling community needs a focused, electronically accessible, and better structured bibliographic resource. To accomplish this, we have developed a database that uses a group technology coding of articles based upon index searching. The importance of this database is not only in collecting many pertinent articles for users, but also in classifying them. The objectives of this site are to be useful to industrial practitioners, faculty members, students, and software developers, and to be widely available, economically feasible, and easily managed. The user can search the database by problem type, measurement metric, modeling technique, organization decision level, publication year, and publication type to find desirable article lists within a short time. The Modeling and Analysis of Semiconductor Manufacturing Laboratory (MASMLAB) database currently contains 559 coded articles relevant to the modeling and analysis of semiconductor manufacturing.

The implementation of total quality management (TQM) in a semiconductor manufacturing operation

The author reviews his five year experience in the implementation of total quality management (TQM) in the operation of the silicon wafer fabrication (fab) facility in Xerox Microelectronics Center (MEC). First, the TQM basic concepts and its operational implications are summarized. This is followed by a detailed description of the TQM implementation process in the wafer fab. Driven by Xerox corporate commitment to quality, the implementation process includes an organizational system analysis of MEC operations, a restructuring of the wafer fab organization, the formation of a quality council, the launching of a comprehensive education and training program, and the implementation of statistical process control techniques. Also, a cornerstone in the TQM strategy is empowering all employees through participation in problem solving and quality improvement teams, and involvement in an employee suggestion system. It is shown that the TQM strategy proved to be instrumental in achieving significant improvements in customer satisfaction, employee satisfaction, productivity, profitability, product quality, manufacturing costs, and on-time delivery.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Environmental Control in Semiconductor Manufacturing

Semiconductor manufacturing is among the world's most important and technologically advanced industries. Although it once enjoyed a reputation as a “clean” industry, in recent years that reputation has been tarnished by allegations of environmental, worker health, and human reproductive problems stemming from semiconductor manufacturing operations. It is well-documented <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1–3</sup> that making semiconductor devices requires hazardous chemicals that, if not properly used or controlled, could result in serious problems to the facility, its work force, or the community where it exists. This paper first discusses early attempts at semiconductor manufacturing, then tracks technology, early environmental controls, growth in regulatory requirements, and programs implemented to insure regulatory compliance in all areas of environmental concern.