Automation Reduces Contamination and Costs within Semiconductor Fab Cleanrooms
To reduce the potential for contamination and human error, automated material handling systems assume the roles of machine operator, inventory manager and stock person.
By John Chrisos and John Horan
The density associated with today`s more complex and highly sophisticated Very Large Scale Integration (VLSI) semiconductor devices creates major interrelated challenges for the semiconductor industry. Microprocessor CPUs such as Intel`s Pentium chip contain greater than three million transistors in an extremely compact area. As device geometries are reduced, due to advancements in processing technologies, chips become more susceptible to contamination and damage during the fabrication process. Advanced semiconductor device manufacturing also requires more complex process equipment, as well as an increased number of discrete processing steps during fabrication. Today`s most advanced devices require as many as 400 distinct, front-end process steps, each of which can contribute contamination to the wafers, and ultimately kill devices before they are completed. From a financial perspective, the value of one 25-wafer lot can vary from a value-added cost of $20,000 to a realized sales value of well over $500,000. There are generally thousands of lots–representing billions of dollars–in process in a semiconductor fab at any given time.
The trend toward increased device complexity has fueled a dramatic increase in capital spending on equipment–from the 40 percent level in the 1970s to more than 70 percent today–as a percentage of total fab construction, facilities and equipment costs, according to Sematech, the industry consortium. At the same time, increasingly more complex products, combined with a larger number of process steps, increase the probability of device contamination and damage as a result of the manual handling of wafer lots (process tool loading, unloading, and inter-process transport). Cumulative losses incurred through manually mishandling or misprocessing valuable wafers can make a large financial impact on a chip maker`s bottom line.
Faced with the conflicting realities of rising semiconductor manufacturing equipment and production costs vs. end-user demand for lower prices, chip manufacturers are actively seeking out innovative and effective approaches to improve fab throughput, increase manufacturing yields and maximize process tool utilization. To increase revenues, manufacturers must produce a greater number of quality devices more efficiently and at a lower cost. One significant way to improve throughput is to ensure that costly semiconductor fabrication equipment is consistently supplied with product. Factory automation systems can dramatically reduce tool idle time and increase yields by removing the human element and providing product to the process tools in a timely manner (See Fig. 1).
The semiconductor industry has adopted automation at a rapid rate. Proof of this is reflected in the growth in capital spending on material handling equipment. According to VLSI Standards Inc. (San Jose, CA), it has doubled in recent years, from $854 million in 1994 to an estimated $1,738 million for 1996.
The role of the cleanroom
Cleanrooms perform an indispensable role in semiconductor manufacturing by providing an industry-proven method of reducing airborne particle levels, thereby increasing wafer fabrication yields. However, to maintain adequate quality and yields, many semiconductor fab cleanrooms have had to evolve from the Class 100 operations of the 1980s to today`s Class 1 operations. Some leading-edge semiconductor device manufacturers are even now requesting equipment for operation within Class 0.1 cleanrooms (extrapolated from Federal Standard 209E). In fact, the cleanliness level for integrated circuit manufacturing far exceeds the guidelines set for general operating rooms, pharmaceutical manufacturing and medical equipment manufacturing by a factor of 10,000 or more.
Cleanrooms are an effective but costly approach to increasing manufacturing yields through reduced device contamination. The cost of building Class 1 cleanroom space may be upwards of $3,000 per square foot. In addition, operating costs associated with maintaining the required cleanliness level on a daily basis can be significant. Once again, semiconductor manufacturers are being challenged to implement qualified solutions in order to optimize production quality and yields while keeping costs under control.
The case for automated systems
Studies such as the Norcross Project, which studied the effects that automated material handling has on process cleanliness, have demonstrated that Automated Material Handling Systems (AMHS) reduce wafer contamination and damage–major causes of reduced quality and yield in semiconductor fabrication.
People are the primary source of contamination within the cleanroom environment. In fact, a single cleanroom operator can potentially generate hundreds of thousands of airborne particles per minute just standing still. To reduce the potential for contamination and human error, AMHS, which are rated for Class 1 or better environments, assume the roles of machine operator, inventory manager and stock person. When incorporated into semiconductor manufacturing processes, an AMHS effectively manages the inter-process storage, transport and material tracking logistics of silicon wafers associated with complex semiconductor fabrication processes.
Increased production yields are realized by minimizing staff inside the cleanroom and through productivity improvements following the implementation of a clean AMHS system design. The automation system can operate within the cleanroom without adding harmful particulates to the clean environment. Automation empowers fab operators to focus on such value-added activities as improving the operation and performance of the process equipment instead of manual lot searching, tool-loading or inter-process material transport. As an added benefit, AMHS control the precise acceleration, deceleration and velocity of the systems that manage and transport wafer lots throughout the fabrication process. This advantage ensures the safest possible transport of materials while minimizing the risk of wafer contamination through particulates caused by wafer vibration and mishandling from human operators.
In addition to improving quality and yields, AMHS can be extremely valuable in meeting other critical semiconductor fab challenges. By enabling precise but flexible material tracking across the entire fab process, the AMHS–linked to a shop floor controller–can optimize productive uptime for expensive process equipment. This enhances device manufacturing productivity and efficiency while improving the return on investment for process equipment. Take, for example, a new fab costing $1 billion, with approximately $700 million in capital equipment expenditures. According to venture capitalist firm Robertson & Stephens and Co., with a fab revenue-to-cost ratio of about 1.4 to 1, just a one-percent increase in equipment utilization can provide an additional $9.8 million in annual revenues. Likewise, an increase in overall equipment productivity could reduce capital equipment expenditures by providing a means to sustain or improve wafer throughput with fewer process tools.
Automated systems typically increase overall production efficiency through effective storage, transport and tracking of cassettes, reticles and other essential components within the fabrication process. With the advent of 300 mm wafers, operator ergonomics have become a more prevalent uncertainty. Automation addresses these concerns, as well as operator safety, by assuming broad responsibility for material transport and lot buffering logistics throughout the entire fabrication process.
Factory automation systems in action
Factory automation systems add the value of material accessibility, convenience and timeliness within the front-end semiconductor fabrication process. Key automation drivers are enhanced asset utilization, the assurance of clean and safeguarded material movement, continuous wafer lot identification and tracking, and timely process availability.
Interbay automation (see Figure 2) designates an automated material transport, storage and control system for managing work-in-process (WIP) between processing bays within a fab. The role of the Interbay AMHS is to expedite each wafer lot to its next process bay. Today, several interbay automation methodologies exist, including overhead monorail, automated guided vehicle (AGV) and conveyor type systems. Overhead monorail transport systems, qualified for better than Class 1 cleanroom environments, are the predominant material handling solution utilized by today`s semiconductor manufacturers. Ideally, the interbay transport system provides the fab with complete AMHS layout flexibility, allowing for easy expansion. However, traditionally, layouts have been rather simplistic, failing to deliver fully on the potential benefits of an AMHS.1 Therefore, flexibility in routing transport and locating storage systems within the process bays reduces walking distances and provides fab operators with a logical location to finding the next lot–within close proximity to supported process tools. Interbay automation also enables the shop floor control system to effectively manage work-in-process so as to reduce inventory and improve process tool utilization by reducing process tool idle time.
Monorail material transport systems assure safe, reliable and contamination-free transport of cassettes, cassette boxes, SMIF Pods and reticles between fabrication processes. Unlike conveyer belts or manual transport systems, the automated monorail approach allows for the direct transfer of open or sealed wafer lot carriers within the AMHS`s domain. As an added benefit, the material transport system is suspended from the ceiling, isolating products from floor-level contamination and other environmental factors that could compromise quality and yield.
Automated Storage and Retrieval Systems (AS/RS), or “stockers,” provide high density storage and control systems to wafer lots awaiting processing within each local process bay. The stocker system should be configured for the most efficient use of valuable cleanroom floor space, i.e., storing wafers vertically rather than horizontally, maximizing valuable cleanroom space. Also rated for Class 1 environments, the stockers must provide material access ports that enable quick and reliable storage and retrieval transactions of designated wafer lots in process. The typical cleanroom stocker may be outfitted with application-specific clean air systems that prevent potential particle contamination from entering the storage environment. Sematech recently completed independent cleanliness testing on a PRI Automation (Billerica, MA) stocker equipped with a new air management system. Test results indicated zero particles within the wafer storage environment 0.1 µm.
The interbay monorail approach can be implemented by an active track, such as a linear induction drive system, or a passive track, where the vehicle is intelligent and self-powered. Magnetically levitated (Mag Lev) suspension and induction drive track systems have recently been introduced, but they can add to the system`s overall complexity. Alternate configurations using aluminum track allow for flexible layouts and features– including turntables, redundant transport pathways and interfloor transport capabilities– enabling the wafers to travel the shortest possible distance between processes and further reducing the chances for contamination.
Intrabay automation entails the application of an AMHS to manage in-bay tracking and transport of wafer lots between lot storage buffers and automated process tools. Several methods can be used. The primary approach is the use of guided vehicles to transport wafer lots between local storage systems and process tools. Some guided vehicles provide a self-contained, Class 1 minienvironment, which guarantees ultra clean wafer transport between automated process tools and stockers.
Combined with the automation of process tools within the bay, guided vehicles reduce contamination caused by human operator intervention. Intrabay automation minimizes wafer damage by assuring gentle and consistent material handling within the process bay.
Factory automation systems are now being used to enhance manufacturing productivity within the photolithography bay, a processing area previously untouched by automation. Within these bays, glass plates or “reticles” project circuitry onto semiconductor wafers. The photolithography process is one of the major gating factors within the entire semiconductor fabrication process, and therefore, the photobay paces the actual wafer output of the fab. The functionality level of an automated photobay can range from a standalone reticle stocker to a completely integrated reticle storage, delivery and stepper loading system.
As shown in Figure 3, one additional defect per wafer due to contamination or damage could potentially result in million dollar losses in annual revenue for semiconductor manufacturers. Automated material handling systems have proven their ability to maximize yields by dramatically reducing semiconductor contamination and damage within the cleanroom. Just as important, these systems do not require the construction of additional cleanroom space. Quite the opposite, AMHS offers the best possible utilization of cleanrooms and other existing elements of the fab infrastructure by greatly reducing the storage space needed for work-in-place. n
1. Davis, Jeff and Weiss, Mitchell, “Addressing Automated Material Handling in an Existing Fab,” Semiconductor International, June 1995, p. 125-128.
John Chrisos is director of marketing at PRI Automation, Inc. in Billerica, MA. He has worked in the semiconductor industry for 15 years and has experience in engineering, marketing and sales.
John Horan is the assistant product manager for Interbay Automated Material Handling Systems in PRI`s Factory Automation Group.
Figure 1. Factory automation can dramatically reduce tool idle time and increase yield by removing the human element and providing product-to-process tools in a timely manner.
Figure 2. Two Process Bay models represent Interbay and Intrabay automation techniques. Interbay automation is depicted with overhead monorail and stockers. It automates WIP transport, buffering, and logistical control between process bays. A Rail Guided Vehicle automates the in-bay transport and delivery of WIP between local bay stockers and automated process tools.
Figure 3. One additional defect per wafer due to contamination or damage could potentially result in million dollar losses in annual revenue for chip manufacturers.