Semiconductor manufacturers face significant challenges, including rising customer demand, increasing product complexity, and expanding product portfolios. These challenges create formidable operational hurdles and emphasize the need for semiconductor manufacturers to innovate and ramp up to deliver new products and processes swiftly.
Each front-end wafer manufacturing fab represents billions of dollars of investment in facilities, equipment, and personnel. These operate around the clock in strictly controlled clean room environments, utilizing highly precise and accurate workflow. A single manufactured product already involves multiple layers and hundreds of production stages across sophisticated tools and machines, and the future portends even greater challenges. There will be a greater product mix going through the fabs; more recipes, reticles, and mask sets to manage; more wafer experimentation; and more use of non-production wafers in planning, testing, and operations.
First generation manufacturing execution systems (MES) are no longer equipped to meet the demands of today’s manufacturing landscape because they typically isolate work-in-progress, track and trace, equipment integration, maintenance, scheduling, reporting, and analytics. Today’s advanced semiconductor environments demand real time integration of production data and control systems. The modern MES must continuously expand its scope to provide front-end master data management, operations control, and equipment management.
Master Data Management
Achieving success in today’s semiconductor manufacturing requires seamless integration and management of data related to production flows, recipes, recipe programs, and recipe parameters as well as reticles, BOMs, data collections, specification limits and sampling strategies. A modern MES provides a common platform that integrates all process data with product lifecycle management (PLM) or enterprise resource planning (ERP) systems, ensuring effective change control and long-term maintainability.
A key capability of the modern MES is facilitating visibility across the process of record (PoR) and the ability to re-use sub-flows and sub-flow blocks, accommodating small changes in recipe, reticle or limit control (Figure 1). It enables the flexible reuse of data collections, recipes, and other common elements, without restrictive hard linking between product or lot specific flows. The modern MES must also coordinate new revisions and new versions of each master data object, incorporating flow revisions or lot assignments. It automates the assignment of a layer number or relevant context, enabling the resolution of conflicts among various manufacturing elements accurately.
Process Planning and Design
In the semiconductor industry, experimentation is necessary to understand the influences of different factors on the measured outputs. The modern MES comes into the production process well before the first recipe goes online. It helps design test recipes and reticles and design the flow of both production and non-production wafers.
Experiments Management
Experiments are increasingly vital for developing or improving products, processes, equipment, and materials and successful experimentation depends on precise data management as well. An integrated MES empowers seamless experiment design and execution of experiment data, combining multiple process variations within a single lot and tracking wafers individually. It provides integrated experiment design tools that support wafer group management, process variations, splits, and merges. It captures all history and traceability data to analyze the effects of different variations on each wafer. Finally, a modern system collects, centralizes, and helps analyze real-time data from equipment, sensors, operators, and other sources and then documents parameters and conditions for each experiment, controlling versions and tracking changes.
Recipe and Reticle Management
Centralized recipe data management is a key function of the future-ready MES. This includes managing programs, sequences, parameters, and sub-recipes, some of which may require specialized editors or software development kits (SDKs). The MES enforces accurate recipe versioning and parameter control, integrating with automation systems to maintain process integrity and facilitating the uploading and downloading from integrated equipment. If a recipe changes once it is uploaded, the MES flags it. It also oversees changes in static and dynamic parameters and the use of sub-recipes to support modularity and reusability (Figure 2).
When recipes specify reticles, the MES ensures the use of the correct mask for each lot, tracks reticle location, and controls maintenance, cleaning and repair. The MES also enforces mask set linking, tracks usage and maintenance and schedules reticle allocations for each job. It also ensures that the correct mask is used at the right step. Additionally, it tracks and enforces the reticle maintenance lifecycle, including its usage counters. For traceability, the system logs all utilized masks. During schedule generation, it assigns available masks to the corresponding jobs based on current needs and availability.
Process Control and Optimization
The MES manages wafers through all front-end phases, often beginning with send-ahead wafer testing and managing run-to-run controls, and sorter and queue management.
Send-Ahead Wafer Testing
Before running a full lot of wafers, in some cases is more prudent to send a wafer ahead to assess process quality at selected equipment locations. The MES tracks these wafers in real time, coordinates integration with control systems, and manages workflow. They might use send-ahead, non-production wafers to detect the presence of contaminants that could impact production wafers during the actual runs.
Run-to-Run Control
The MES handles run-to-run control feedforward and feedback loops to optimize parameter settings and improve the overall performance of the lots. It captures key data across lots, wafers, and manufacturing stages, dynamically calculating parameters and optimizing recipe settings for the next runs. The system should allow integration with advanced processing applications, for more advanced statistical calculations, to ensure consistent execution of feedback and feedforward loops.
Sorter Management
During production, the MES helps in sorting wafers based on defined characteristics. It ensures the accuracy of the lot-wafer-carrier-slot information and integrates with the sorter equipment for automatic processing. It also enables sorter job creation and scanning of wafer IDs, and facilitates smooth sorter execution, ensuring real-time updates of MES wafer-carrier position mapping. Table 1 summarizes the range of softer functionality the modern MES supports.
Time Constraint Management
The MES controls sensitive processes that require precise, step-by-step time constraint management. It helps define and enforce these constraints and triggers warnings or corrective actions as needed.
When multiple time constraints apply to a single lot simultaneously, the MES issues proactive alerts to prompt timely action. If any of the constraints are exceeded, the system automatically enforces business rules such as pausing the lot or routing it for rework as necessary.
Sampling
Measuring every lot or wafer in a production run to ensure adherence to specifications would be cost-prohibitive, so manufacturers must rely on sample lots and individual wafers. The modern MES can help with lot-based and wafer-based sampling, defining strategies for various process contexts and metrology steps. For lot-based sampling, the MES supports the definition of sampling evaluation steps and lot-sampling strategies for different process contexts.
Following the lot-sampling strategies, the MES system will automatically evaluate and assign the sampling steps for each lot. For wafer-based sampling, the MES system enforces a wafer-sampling strategy at designated metrology steps, so that when a lot arrives at that step, the system determines exactly which wafers to measure.
Equipment Management
The modern MES ensures that production or non-production wafers reach only equipment that is qualified, capable and available exactly when needed.
Equipment Qualification
The MES keeps a list of qualification procedures for each piece of equipment, together with the events that would trigger it, such as maintenance procedures. If there is a pending qualification for a piece of equipment, the MES ensures that no production material reaches that machine until it has been properly qualified. The MES also assists with the creation of qualification lots and their associated qualification procedures. If qualification procedures require mixed lots of non-production wafers, the MES will manage a sorter in composing them, measuring, running the qualification process, and measuring them again. It then guides dismantling based on type and measurement results. The MES also runs qualification procedures and tracks equipment status periodically.
Equipment Dedication
When certain processes require specific equipment for critical steps, the MES manages equipment assignments. In some process steps, such as lithography, processing of certain critical layers must take place on the exact equipment as the previous lithography steps. Likewise, during experiments, it’s common to ensure the processing of a lot or a set of wafers on the same equipment as a particular process step.
The MES sets equipment targets ahead of time, either dynamically for critical steps like lithography or statically, by planner or engineering user requests.
Multi-Chamber Equipment Management
In some essential semiconductor frontend operations such as deposition and etch, machines typically have several chambers or process modules that can be set up for different processes, which requires an additional level of dispatching rules to govern the sub-component states. In traditional MES platforms, this level of rules is often lacking or underdeveloped, requiring human intervention and manual WIP adjustments and crippling the MES ability to schedule operations efficiently. Advanced MES, however, manages multi-chamber WIP automatically, synchronizing main machine availability with the required inner process chamber module.
Conclusion
The modern MES has evolved into a comprehensive solution that does far more than track and trace and WIP management. It provides a common platform for data integration and improves data cost-efficiency by managing re-usable objects. From early-stage planning and experimentation to recipe and reticle coordination, the system supports every phase of manufacturing, enabling send-ahead wafer validation, run-to-run optimization, sorting, and targeted sampling. It also helps coordinate equipment qualification, dedication, and matching multiple chambers to recipes.
Modern semiconductor MES provides the flexibility and capabilities to handle extremely complex advanced manufacturing scenarios and improve the speed of learning while increasing quality and lowering costs. The modern MES provides expanded functionality in a single system, allowing fabs to keep pace with rapid changes while providing a foundation to accelerate the future of semiconductor manufacturing.
Replacing an outdated solution with a next-generation MES can radically improve user experience and eliminate obsolete IT languages, operating systems, and hardware platforms. It can expand functionality in a single system and keep pace with change, supporting the entire processing lifecycle.
This article was written by Rúben Pereira, Product Manager for Industrial Segments at Critical Manufacturing (Porto, Portugal). For more information, visit here .
