TECH FOCUS—300 mm

Using strategic planning to prepare for 300-mm fab construction

Terry R. Behrens, IDC

The semiconductor industry is poised on the brink of a new era of technology innovation that is in some ways similar to, yet in many critical ways markedly different from, the technological leaps that have characterized the industry's relatively short history. It may be tempting to regard the pending wave of conversions to 300-mm manufacturing as just the latest in a perpetual string of technology upgrades, such as the moves from 4- to 6-in. or 6- to 8-in. wafer production not so long ago. Manufacturers may assume that they can again change out a few tools, move a wall here and there, and tweak their facilities systems and their fabs will be ready to go until the next technology wave washes through.

The wafer world has changed dramatically since the relatively smooth shift to 6-in. production, however, and there is a growing awareness that the move to 300 mm will not be a business as usual conversion. Although previous technology changes have all resulted in an increase in the size and complexity of fabs and their operations, the advance to 300 mm threatens to wield a disproportionate impact on the industry. Companies are confronting the reality that the risks associated with deficient strategic planning have risen orders of magnitude as a result of increased capital improvement costs as well as the increasing value of lost yields. Among the issues that differentiate the 300-mm conversion from the challenges of past upgrades are the following:

• Automated materials-handling system (AMHS) performance. In the past, use of an AMHS was not essential for manufacturing success. Where such systems existed, they mainly performed bay-to-bay transport. In the world of 300 mm, however, an AMHS will become the tool-to-tool workhorse, largely as a result of the ergonomic complexity of handling 300-mm wafers and carriers.

• Computer-integrated manufacturing. The greater reliance on automated systems for the transfer of materials will need to be accompanied by a greater reliance on automated systems for the transfer of information.

• Tool utilization. The press for maximized productivity will demand optimal equipment performance with no on-line bottlenecks. This new reality will depend on a reliable and comprehensive flow of automated material and information.

• Information turns. A 300-mm environment requires 10 times more data than a 200-mm facility. Thus, automation systems must be capable of delivering more frequent and more detailed data streams. Company survival will depend on management's capacity to make better decisions, and to make them faster than ever before.

• Resource allocation. In the past, fabs have sometimes deployed staff, materials, and tools through intuition and approximation. The high overhead of 300-mm manufacturing makes every resource precious. Resource consumption will need to be precisely measured and allocations will need to be carefully modeled using dynamic simulation techniques and tools.

• Cost of product/profit margins. Where once low product costs were a desirable goal, they will now be required by the rigors of 300-mm operations. The strong profit margins generated in some product sectors can no longer be taken for granted. The good news is that with shrewd strategic planning prior to 300-mm upgrades, desired margins and the means to protect them over the long term can be built into a facility's planning process.

All of the above concerns can be addressed through strategic planning. The key to a successful transition is to do such planning early and thoroughly. In the past, when the stakes of wafer manufacturing were not as high as they are now, true strategic planning was often overlooked or delayed until a facility was retooled and operating, or it was applied as a troubleshooting tool as fab problems required. With 300 mm, effective preconstruction planning is imperative. The remainder of this article reviews some of the primary considerations that make up a strong strategic planning process.

Industry Standards

The international standards for 300-mm production developed by two industry consortia—the International 300-mm Initiative (I300I) and the Japan 300-mm Semiconductor Conference (J300)—represent the first opportunity in industry history for equipment manufacturers to accommodate U.S. and Japanese standards with a single tool.^{1, 2} Both I300I and J300 specify carriers that will accommodate 13- and 25-wafer lot sizes, and the process tool load port specs will be the same in both standards. It has been agreed that the pods called for by the standards are not a manageable size for operators, and that a design for a cart for manual moves will be required. This cart could also be used as a low-cost alternative for an AMHS. The standards will be administered by SEMI.

Process Tools and Facilities

With the changes in fab operations and the adoption of an AMHS, the 300-mm process tool and facility requirements will also change. The use of an intrabay AMHS means that process tools will require an alternate or nonfrontside user interface/console for maintenance or other non-production-related activities. Process tool vendors will also need to standardize all load/unload ports. Among the changing facility requirements will be an increased demand for subfab space. A good rule of thumb to use for space planning is to allow the same amount of area in the subfab as each tool occupies in the fab. Although more in situ process cleaning will occur and the use of outside cleaning services will develop further, the area required for support rooms will probably remain the same.

Tool layout will be extremely important since it will have a direct impact on the cost-effectiveness of the fab, AMHS requirements and performance, and tool installation strategy. In the past, achieving a dense process tool configuration was not a critical element in the cost of a facility, but with 300-mm technology all resources need to be highly utilized to keep the cost per wafer or good die as low as possible. Process cycle time and throughput also can be affected by a poorly planned fab. Cycles can be lengthened by unnecessary or lengthy product moves or by not using appropriate work-cell and method improvements. In addition, the initial cost of an AMHS may increase if extra track, vehicles, or stockers are needed because of poor planning. Tool installation paths and strategy must be determined early in the development of the process tool layout since clearances need to be maintained (implant and photo tools require 3.3 m; dry etchers, 1.8 m; and all other tools, 2.5 m), and automation and production ramp-up requirements can affect how tool installation will occur.

There are still some general concerns within the industry about process tools. As the technology curve continues, what will be the impact of such tool and process changes as load size, single-wafer processing, in situ cleaning, and the use of multiple chambers for sequential processing? Will space in the fab need to be allocated for additional process techniques such as bump processing? It is too early to tell if the tool areas and specifications established by I300I and J300 can and will be adhered to, and if extra space will be required for currently unknown applications.

Factory Operations

The 300-mm factory will be more complex than current ones, and thus there will be greater reliance on the manufacturing execution system (MES). The MES will need to comprise a well-defined production scheduler, equipment status monitoring, statistical process control, lot-tracking capability, and slot integrity (pods will have dedicated slots for a given wafer). It will also need to be linkable to an enterprise resource planning (ERP) package in the future. When adopted, an ERP system will allow one operating system to control multiple factories and will lead to higher utilization of resources within the enterprise. As MESs become more important, there will be a shift from operators with limited problem-solving skills who just move work in progress to technicians with greater skills and training in automation and tool interface.

Fab Space and Cost Analysis

A model was developed to examine the cleanroom floor space requirements and capital cost per wafer for Sematech's 0.25-µm high-performance logic process flow. Based on I300I, J300, and the equipment database compiled by IDC, the model included the following information:

• I300I-estimated equipment run rates.

• IDC-estimated equipment run rates and footprints.

• I300I space requirements for intrabay automation.

• J300 space requirements for intrabay automation.

• I300I-estimated capital costs per wafer.

The model assumptions were as follows:

• J300 tool space is based on a tool footprint from IDC's 200-mm database with a 1.3x scale factor, plus the requirements for automated guide vehicle (AGV) space and bay clearance.

• I300I tool space is based on a tool footprint from IDC's 200-mm database with a 1.3x scale factor, plus the overhead monorail space and bay clearance requirements.

• The model only considered the fab processing area and did not include any support areas, including the gowning room.

• The Uniform Building Code and OSHA safety standards were used in assessing the space requirements.

• 15% of the cleanroom area was allocated for fab code and circulation.

• 10% of the cleanroom area was allocated for analytical and metrology tools.

• A contingency space of 5% of the cleanroom area also was included.

• Per I300I, AMHS cost was estimated to be about 15% of the total equipment cost.

• Capital cost per wafer was calculated based on the total equipment costs, fixture costs, tool installation costs, and AMHS costs.

Wafer Starts per Week	With I300I Run Rates			With IDC Run Rates
	J300 Space (m2)	I300I Space (m2)	Cost/ Wafer ($)	J300 Space (m2)	I300I Space (m2)	Cost/ Wafer ($)
2,500	3,132	2,845	1,638	4,351	3,848	2,301
5,000	5,530	5,025	1,503	7,921	6,994	2,139
7,000	7,945	7,230	1,458	11,713	10,343	2,112
10,000	10,294	9,367	1,422	15,076	13,298	2,058

Table I: Comparative results of a fab space and cost analysis that used 300-mm equipment run rates estimated by I300I and IDC. Additional data are shown in Figure 1.

Figure 1: Comparative fab space and cost analysis results using 300-mm run rates estimated by I300I and IPC. Selective data from this figure are also given in Table I.

Figure 1 and Table I present the results of modeling space and cost requirements using the I300I and IDC run rate estimations, which represent two extremes. The aggressive approach is to use the I300I rates, and the conservative approach is to use those estimated by IDC. Breakdowns of the space requirements by functional area for the two approaches are shown in Figures 2 and 3, while Figure 4 depicts a conceptual fab layout in which the effect on space requirements for the I300I and J300 to fit "while" approaches are compared. The actual space required for 300-mm manufacturing will most likely fall between the two models because run rates for some tools will improve over 200-mm speeds as the result of tool performance improvements. This modeling study emphasizes the discrepancies in available information, particularly run rates. It is also too early in the tool development stage to determine the improvements that can be achieved in tool performance and area requirements.

Figure 2: Space percentage requirements by functional area with I300I run rates.

Figure 3: Space percentage requirements by functional area with IDC run rates.

Computer Modeling and Simulation

Although the modeling study described showed significant variations because of discrepancies in underlying assumptions, it will be extremely important to use computer modeling in planning and operating a 300-mm factory. Modeling will assist the user in predicting the fab's performance under uncertain situations with different operational strategies. Through modeling, the company can project staffing requirements, plan equipment ramp-up, derive estimates of factory cost and performance, optimize overall equipment effectiveness (OEE), and easily perform multiple "what if" analyses. Modeling is also useful for evaluating material-handling strategies and vendors. A static spreadsheet model for simple capacity planning can be used to accomplish these last two tasks, but other types of analyses require dynamic or discrete event simulation modeling software, such as AutoMod/AutoSched (AutoSimulations, Bountiful, UT) or Factory Explorer (Wright, Williams and Kelly, Dublin, CA). Each of these software packages offers modeling programs for fab strategic planning.

Figure 4: Comparison of footprint for conceptual fab using AGV (J300) versus overhead (I300I) transport.

Conclusion

Strategic planning that encompasses the issues described above is extremely important when a company must make a large investment in a 300-mm facility. If the internal resources required to undertake such a planning project are not available, an outside firm may be called on for assistance, but do not expect to receive a standardized approach to 300-mm manufacturing. Because each fab's production and operations requirements are unique, planning has to be a team effort, with company personnel closely involved in the process. All parties need to agree on the assumptions used in the planning process, and a factory handbook needs to be created to capture how the facility will function. The handbook will become an effective way to transfer information to start-up staff, vendors, consultants, and suppliers that will be working in or on the new facility, as well as providing a record of decisions made and system specifications. With a strong strategic planning process, the transition to 300 mm will occur smoothly and the economic risks that new construction entails will be minimized.

References

1. Bass E, and Ghatalia A, "I300I Factory Guidelines," Austin, TX, International 300-mm Initiative (I3001), June 1997.

2. "Global Joint Guidance for 300-mm Semiconductor Factories—I300I and J300," Austin, TX, I300I, July 1997.

Terry R. Behrens is director of manufacturing technology and manager of the industrial engineering department at IDC (Portland, OR). He has performed a number of studies and design services for U.S., European, and Asian semiconductor manufacturers planning for the transition to 300-mm production and is experienced in the areas of long-range planning, construction sequencing, microcontamination control, process flow analysis, automation strategies, facilities planning, workstation concepts, and tool selection and design. Before joining IDC three years ago, Behrens was a senior industrial engineer for Intel, where he was responsible for coordination and budget control of multiple work groups performing fit-ups of the fab, subfab, and primary support rooms, as well as other activities integral to the operation of semiconductor facilities. He has a BS in industrial engineering from Oregon State University and a BS in civil engineering from Portland State University. (Behrens can be reached at 503/224-6040.)

MicroHome | Search | Current Issue | MicroArchives
Buyers Guide | Media Kit