Benchmarking Semiconductor ManufacturingRobert C. Leachman and David A. HodgesCompetitive Semiconductor Manufacturing ProgramEngineering Systems Research CenterUniversity of California at BerkeleyBerkeley, CA 94720AbstractWe are studying the manufacturing performance of semiconductor wafer fabrication plants in theUS, Asia, and Europe. There are great similarities in production equipment, manufacturingprocesses, and products produced at semiconductor fabs around the world. However, detailedcomparisons over multi-year intervals show that important quantitative indicators of productivity,including defect density (yield), major equipment production rates, wafer throughput time, andeffective new process introduction to manufacturing, vary by factors of 3 to as much as 5 acrossan international sample of 28 fabs.We conduct on-site observations, and interviews with manufacturing personnel at all levels fromoperator to general manager, to better understand reasons for the observed wide variations inperformance. We have identified important factors in the areas of information systems,organizational practices, process and technology improvements, and production control thatcorrelate strongly with high productivity. Optimum manufacturing strategy is different forcommodity products, high-value proprietary products, and foundry business.IntroductionThis comparative study involved the measurement of manufacturing performance andinvestigation of underlying determinants of performance at 28 wafer fabrication facilities in theUnited States, the United Kingdom, Germany, Spain, Japan, Korea and Taiwan. The companiesoperating these facilities are displayed in Table 1.Table 1Companies Participating in the Main Phaseof the Competitive Semiconductor Manufacturing SurveyAdvanced Micro Devices, Inc. (AMD)National Semiconductor Corp. (2 fabs)Cypress Semiconductor, Inc.NEC Corp.Delco Electronics Corp.Oki Electric Industry, Ltd.Digital Equipment Corp. (2 fabs)Samsung Electronics Co., Ltd.Harris CorporationSilicon Systems, Inc. (SSI)Hyundai Electronics Industries, Ltd.Sony Microelectronics Corp. (2 fabs)Intel CorporationTaiwan Semiconductor Mfg. Corp. (TSMC)Int'l Business Machines, Inc. (IBM)Texas Instruments, Inc.ITT IntermetallTohoku Semiconductor Corp. (TSC)LSI Logic Corp. (2 fabs)Toshiba Corp.Lucent Technologies (2 fabs)United Microelectronics Corp. (UMC)Motorola, Inc.
`
`Applied Materials, Inc. Ex. 1028
`Applied v. Ocean, IPR Patent No. 6,968,248
`Page 1 of 10
`
`

`

`-2-Each participant completes a 100-page questionnaire covering objective data including clean roomsize and class, head counts, equipment counts, wafer starts, die yields, line yields, cycle times,computer systems, etc. over the preceding four years. From this data we calculate technicalmetrics of manufacturing performance for each participant. We then compare performances oneach of the metrics.We observe great variations in the scores of the various participants. To understand whatpractices account for such performance differences, we conduct a two-day site visit with eachparticipant. We tour the manufacturing line, interview a cross-section of the entire fab staff, andhold discussions concerning strategies for improving yields, increasing wafer throughput, reducingcycle times, etc. We survey each firm’s activities to improve computer integrated manufacturing(CIM) and information systems, human resources development, effectiveness of work groups andteams, etc. These more qualitative indicators of participants’ operational practices are thencorrelated with the performance scores to identify those practices that underlie good or badperformance. Performance and practice comparisons are separated into VLSI memory, VLSIlogic, and MSI categories, according to the type and sophistication of devices that are fabricated.The individual identities of the participating fabs are coded. Each participant receives all thecomparative data but knows only its own code identifier.Metrics of manufacturing performanceWe use the following technical metrics to measure manufacturing performance:
`
`•
`
`•
`
`•
`
`•
`
`•
`•
`•
`
`•
`•
`
`•
`
`Average line yield, the percentage of wafers started that are completed properly, normalizedto twenty mask layers.
`
`Defect densities, calculated for major process flows in each fab by using reported die yieldsand die sizes in the Murphy model of defect density. The reported defect densities account forall yield losses, including both spot defects and parametric problems. For memory products,the die yields applied to the defect density formula are final die yields after laser repair.
`Integrated fab and die sort yield, calculated as the product of line yield per twenty maskinglayers and the estimated die yield for a 0.5 sq cm die. This die yield is estimated using theMurphy defect density calculated from reported die yields as described above.
`
`Wafer masking layers completed per 5X stepper per calendar day (considering only layersexposed using 5X steppers).
`
`Integrated 5X stepper throughput, the equivalent number of full-wafer operations per 5Xstepper per day, calculated as the number of 5X wafer operations per day times the integratedyield defined above.
`
`Wafer masking layers completed per operator per working day (considering all maskinglayers, regardless of type of lithography equipment).
`
`Wafer masking layers completed per working day divided by the total head count.For all of these metrics, we encountered a wide range in scores, even though the basic processtechnology in use at the participants was generally similar. Table 2 summarizes scores for eachmetric for the CMOS memory category, considering the latest data points we received from each
`
`Applied Materials, Inc. Ex. 1028
`Applied v. Ocean, IPR Patent No. 6,968,248
`Page 2 of 10
`
`Wafer implant layers completed per ion implanter per calendar day.
`Wafer metal layers completed per metallization machine per calendar day.
`Average cycle time per mask layer.
`

`

`-3-of the 28 participants. Similar spreads in the data were found also for CMOS logic fabs. The timeinterval covered is from the middleof 1992 to the middle of 1995.Rates of improvement were studiedfor each participant. Scores for eachtechnical metric were computed foreach quarter over a period of threeto four years. For most metrics, theranking of participants changesslowly, i.e., we found few caseswhere a last-place participantovertook the leader for a particularmetric, although some participantsimproved their rankings considerablyover the period.One of the most striking trends weobserved in our measurementsconcerns the initial defect densitiesfor process flows, i.e., the defectdensities realized in the first calendarquarter after transfer of the processflow into manufacturing. Werecorded a factor-of-ten range ininitial defect densities. Those fabswith poor starting points tend tohave faster rates of improvement, butnot nearly fast enough to overtakethose with good starting points, atleast not for several years, as thosewith good starting points also makesteady if somewhat slower progressreducing defect densities.The integrated stepper throughputmetric is perhaps our best indicatorof overall fab productivity, at leastfor submicron fabs dependent on thistechnology for photolithography.The varying strengths andweaknesses in line yield, die yield(defect density) and stepper throughput among our participants are integrated to see the overallthroughput of good silicon per machine. Even for such an integrated metric, we find a remarkablefactor-of-seven range in performance.Technical Metric Scores: Memory FabsMetricBestAvgWrstLine yield per twenty layers (%)98.89387.1Murphy defect density -0.45 - 0.6 micron CMOS memory0.030.591.34(defects per sq cm after repair)Murphy defect density -0.7 - 0.9 micron CMOS memory0.010.511.81(defects per sq cm after repair)Murphy defect density -1.0 - 1.25 micron CMOS memory0.310.591.08(defects per sq cm after repair)Integrated fab and sort yield (%)0.45 - 0.6 micron CMOS memory91.772.146(0.5 sq cm device)Integrated fab and sort yield (%)0.7 - 0.9 micron CMOS memory92.973.935.9(0.5 sq cm device)Integrated fab and sort yield (%)1.0 - 1.25 micron CMOS memory7766.748.3(0.5 sq cm device)5X Stepper throughput (wafer606463281operations per 5X stepper per day)Ion implanter throughput (wafer1360855339operations per implanter per day)Metallization throughput (wafer27314753operations per machine per day)Integrated 5X stepper throughput479344160(Full-wafer ops/stepper-day)Cycle time per mask layer (days)1.82.94.1Direct labor productivity (mask71.742.618.4layers completed/operator-day)Total labor prod. (mask layers/p-d)51.627.315.1Table 2
`
`Applied Materials, Inc. Ex. 1028
`Applied v. Ocean, IPR Patent No. 6,968,248
`Page 3 of 10
`
`

`

`-4-Practices underlying manufacturing performanceOur main objective is to identify those operational practices that underlie leading-edgemanufacturing performance. Summarized below are the operational practices that distinguishthose fabs achieving best or near-best scores in one or several of the metrics described above.(For the sake of brevity, we refer to such fabs as the “leading” fabs.) But before summarizing ourfindings in that regard, it is only fair to acknowledge that our analysis does not account for severalstrategic factors concerning product design and fab design that may strongly influencemanufacturing performance.First, the restrictiveness of product design rules can have a strong influence on observed die yieldsand hence on our calculated defect densities. We made no attempt to normalize defect densityscores for potential differences in design rules among the participants.Second, the range of sizes of fabs in our survey, in terms of wafer starts, spans a factor of almostfifty. Small fabs generally have inferior labor and equipment productivity scores, because of theindivisibility of machines and personnel, and because of the tendency to install extra equipment toavoid situations in which a particular process step must be performed by a one-of-a-kindequipment type. In the tables of metric scores, we do not adjust productivity scores to accountfor fab size. For a general assessment, we define as large fabs those that make more than 7,000wafer starts per week, medium fabs that make 2,500 - 7,000 wafer starts per week, and small fabsat less than 2,500 wafer starts per week. Large fabs lead almost every one of our labor andequipment productivity metrics, although fab size above 7,000 wafer starts per week does notimprove performance. In the yield and defect density metrics, small and medium fabs arecompetitive with the large fabs.Third, using older-generation processing equipment on newer process flows may make theachievement of world-class defect densities much more difficult than if newer equipment is used.While yields may be lower when employing older processing equipment, capital costs are lower aswell, and so the strategy might turn out to be economically competitive or even superior to thestrategy that employs solely new processing equipment. We made no attempt to adjust defectdensity scores for the generations of equipment applied.With these strategic factors aside, we now summarize the various operational practices we foundto be correlated with good manufacturing performance (in terms of the manufacturing metrics wehave defined). We define eight basic themes for key practices that underlie leading performance.In short, these themes are:1. Make manufacturing mistake-proof2. Integrate process, equipment and product data, and analyze it statistically3. Automate information handling and step-level material handling4. Develop a problem-solving organization5. Reduce the division of labor6. Secure the requisite technical talent7. Manage new process introductions8. Schedule manufacturing activity
`
`Applied Materials, Inc. Ex. 1028
`Applied v. Ocean, IPR Patent No. 6,968,248
`Page 4 of 10
`
`

`

`-5-Proper execution of the very complex manufacturing process is essential. Some participants witha narrow product mix and very disciplined, well-trained operators achieve high line yields withlittle or no automation. But other leading participants have applied very effective forms ofinformation automation that make manufacturing very mistake-proof. Such automation includesprocedural checks that require the right production lot and the right machine to be selected beforeprocessing activity may be initiated, and automated download of the machine recipe (i.e., theprocessing parameters) to the processing machines.Good process control systems do not make manufacturing strictly mistake-proof, but they serveto contain losses to minimal levels. All fabs in our survey apply SPC to their processes andequipment. The leading fabs make considerable use of sensors and computers to monitorequipment performance, and provide automated notification of out-of-control conditions and on-line assistance for trouble-shooting.The leading fabs utilize computerized tracking systems to achieve excellent data collection andexcellent data analysis capabilities. They collect large amounts of data concerning process andproduct conditions (an activity termed engineering data collection, or “EDC”), equipmentmaintenance and operation history, lot production history, and yield results. They integrate thesedata in a single relational database. Statistical tools are routinely applied to these data by processengineers, enabling them to expeditiously pinpoint causes of low die yields and make rapiddeployment of counter-measures to contain losses.The leading fabs rigorously measure the overall equipment efficiency (OEE) of their keyprocessing equipment, identifying losses in throughput and prioritizing needed improvements. Inthe best fabs, equipment status is automatically captured from machine logs using SECS-IIinterfaces. Actual processing time is automatically monitored and compared against engineeringstandards; alarms are triggered when elapsed times are excessive.Automation of information handling and step-level material reduces the overhead surrounding theperformance of processing steps. Automation of information handling includes procedural checksand auto-recipe download as described above. It also includes automated capture of engineeringdata and equipment tracking data using bar codes and sensors, as well as automated notificationof operators or technicians when machines are about to become idle or when they requiremaintenance or attention.Material handling automation efforts may be divided into three types: interbay automation,intrabay automation, and step-level automation. Interbay automation concerns the movement ofproduction lots between equipment bays using automated guided vehicles (AGVs) or overheadtracks to transport lots between stockers serving the bays. Intrabay automation concerns themovement of lots between stockers and processing machines in the bay using AGVs or travelingrobot arms. Step-level automation involves the use of robot arms or tracks to handle wafers orcassettes of wafers between lot box and processing chamber, or between consecutive processingchambers. We find that step-level automation has the greatest positive impact on fab performanceamong our participants. For instance, fabs that have linked up coat, expose and develop steps inphotolithography into a single automated sequence achieve higher yields and lower cycle timeswith no reduction of equipment throughput.Fabs that have developed a strong problem-solving organization are very good at problemrecognition, problem solving, and elimination of repetitive problems. Semiconductormanufacturing is characterized by immature processes and immature equipment, and by
`
`Applied Materials, Inc. Ex. 1028
`Applied v. Ocean, IPR Patent No. 6,968,248
`Page 5 of 10
`
`

`

`-6-continuing increases in complexity. Opportunities to improve yields and/or wafer throughput arealways present. Thus manufacturing has a major engineering aspect as well as the expectedoperational character. This means a fab must continually develop technical competence of itsorganization and continually foster a teamwork approach to recognize problems, deviseinnovative solutions, and implement them quickly and successfully. The leading fabs have instilledproblem-solving skills in their technicians and operators through extensive training, mentorshipand participation in continuous-improvement teams organized under the TQM and TPMparadigms. TQM (Total Quality Management) focuses on product quality, while TPM (TotalProductive Maintenance) focuses on equipment productivity.Leading fabs have strong internal technical talent complementing support from vendors toexpeditiously modify product, process, and equipment to implement changes that have beenidentified by problem solving efforts as desirable or necessary. In particular, leading fabs haveconsiderable in-house equipment engineering talent, identifying and implementing usefulmodifications to process equipment that improve performance or ease maintenance. In contrast,weak-performing fabs have process engineering organizations that are virtually devoid ofequipment engineering skills.Reducing division of labor involves training efforts and job expansions to reduce response time toproblems and to promote more effective formulation of engineering solutions. Operators atleading fabs are trained to perform basic equipment maintenance and trouble-shooting, enablingthem to be pro-active instead of waiting for a technician when the need for maintenance orcorrection arises. In such fabs, each equipment bay on every operating shift is staffed with atechnician or operator designated as the “key man” for that type of processing equipment. Thisperson trains and focuses the rest of the staff for improved equipment operation and maintenance,responds immediately to problems that may arise, and serves as a technical resource and mentorfor other operators.Technicians and operators work together in leading fabs on continuous improvement teams,directed and supported by engineers. Such teams identify and research process and equipmentproblems, find root causes, and then devise, test and implement permanent fixes. The teams serveto expand the knowledge, skills, confidence and job scope of all personnel.The division-of-labor theme also applies to engineers. Process and equipment engineering groupsare merged in leading fabs, broadening the skills of engineers and promoting quicker identificationand implementation of effective solutions to process and equipment problems. Rather than beingexclusively the domain of statisticians or of yield and integration engineers, statistical analysis ofyield is practiced regularly by process engineers at leading fabs.Leading fabs have effective procedures for managing the introduction of new process flows. Theeconomic life of many process flows is three to four years, with unit prices for products of theflow declining rapidly over this period. Thus it is economically important to realize highproductivity quickly after process introduction, and to fairly frequently introduce new processflows. A poor start with a new flow may leave the fab too far behind to catch up before themarket value of the output has declined.The leading participants strategically stagger transitions to new product generations and newprocess generations, and control the number of new modules in each process generation to keepthe difficulties at each step to a tractable level.
`
`Applied Materials, Inc. Ex. 1028
`Applied v. Ocean, IPR Patent No. 6,968,248
`Page 6 of 10
`
`

`

`-7-The leading participants also have effective operational procedures for new process transfer.Identical equipment sets are used in development and in production; development and volumeproduction are often co-located. Transfer of process documentation is electronic. Engineersfrom the recipient manufacturing fab participate in the final stages of development.Rigorous scheduling of production activity aims at reducing cycle time and improving on-timedelivery. The leading fabs utilize automated production planning systems that insure releases ofnew production lots do not overload fab resources and that target out schedules are consistentwith steady flow of work-in-process (WIP) according to target cycle times. Delivery quotations tocustomers are automatically made on the basis of the planned production. Dispatching of lots onthe factory floor is performed to prioritize lots that are behind schedule, supplemented withKanban controls that ensure WIP is kept in balance.Practice trendsEach participant tended to score well or score poorly in most metrics, reflecting the fact thatparticular practices, good or bad, tend to influence several metrics. Yet almost every participantpursued at least one good practice that we did not find elsewhere, whose adoption by the otherparticipants we believe would improve performance.No single fab is the leader in all eight themes described above. The best integrated database andbest statistical yield analysis system we saw was in a Korean fab; the best equipment efficiencyanalysis system was in a Japanese fab; the best information handling and step-level automationimplementation we saw was in a fab in Taiwan; and the best cycle time control and on-timedelivery systems we saw were in fabs in the United States.In cases where a fab scored poorly for certain metrics, the most common reason was that therelevant area was simply not a focus of the fab management. Every participant has certain focusareas that management impresses on the work force as top priorities for improvement; associatedwith each area is a paradigm for data collection, problem-solving, training, etc., that we call aculture. We use the term culture because of the management efforts clearly made to rally theorganization around a unifying theme aimed at improved performance.Manufacturing cultures we encountered include TQM with its focus on product quality andprocess control, TPM with its focus on equipment productivity, intensive in-line data collectionand statistical analysis of the integrated data, automation for mistake-proofing and productivityimprovement, fast cycle time (“time-based competition”), and on-time delivery. These cultures arenot really exclusive, but many times we observed how a fab, stuck in the paradigms of its culture,simply had not placed any focus on certain areas of improvement.We observed varying strengths of different cultures in different parts of the world. TQM waswell-established almost everywhere. Intensive in-line data collection and statistical analysis ofintegrated data is strong in at least two thirds of our participants, representing all parts of theglobe. TPM is very strong in Japan and in some companies in Taiwan and Korea, but generallyweak in the US fabs we visited. On the other hand, cycle time and on-time delivery are strongcultures in US and Taiwanese fabs, but generally quite weak in Japanese and Korean fabs. Nodoubt our participants are working to catch up in their areas of weakness.Since the TPM culture originated in the Japanese machine tool industry, the fact that it hasbecome strong in the Japanese semiconductor industry is perhaps not surprising. What is moresurprising is that the cycle time culture, which first flowered in the Japanese automobile industry,
`
`Applied Materials, Inc. Ex. 1028
`Applied v. Ocean, IPR Patent No. 6,968,248
`Page 7 of 10
`
`

`

`-8-flourishes in many U.S. semiconductor fabs we saw, has not taken hold yet in most Japanesesemiconductor fabs we visited.With the maturity of TQM and the integrated data analysis cultures at most of our participants,defect densities are quite competitive at many of our participants, apart from performancedifferences associated with weaknesses in process development and transfer. Given the weakerpenetration of the TPM culture and of the information handling and step-level automation culture,equipment throughputs and line yields are more prominent discriminators of fab performance inour sample. Unique specific process technologies are a relatively insignificant factordifferentiating manufacturing performances.The limited penetration of cycle time and on-time delivery cultures also ought to discriminateperformance in semiconductor manufacturing. However, the importance of good performance onthese measures varies depending on product category. They are relatively unimportant forcommodity memory products, but at the other extreme, critically important to foundries that servefabless semiconductor firms.Plans for further workCurrently we are extending the benchmarking study to include the 200 mm. generation of fabs.Separately, we are studying in much greater detail the factors contributing to overall equipmentefficiency (OEE). Now that line yield and die yield have reached very high levels, equipmentefficiency has become the area offering best opportunities for increased productivity. On anothertrack, we are studying the factors that contribute to the excellent success of the “fabless-foundry”business model. This business model produces short time-to-market for new products withrelatively low business risks to all parties.BibliographyR. C. Leachman, D. A. Hodges, “Benchmarking Semiconductor Manufacturing,” IEEE Trans. onSemiconductor Manufacturing, TSM-9 (May 1996), pp. 158-169. (Report on first 16 fabs surveyed.)http://euler.berkeley.edu/esrc/csm/index.html (listing of all program reports and personnel, and slides forpresentation of this work)The Competitive Semiconductor Manufacturing Survey: THIRD REPORT on the Results of the MainPhase. R. C. Leachman, Editor (August 1996); The Competitive Semiconductor Manufacturing HumanResources Project: Second Interim Report. C. Brown, Editor (September 1996) Both available fromEngineering Systems Research Center, Berkeley. Reports may be ordered with the form provided at theabove URL.AcknowledgementsWe acknowledge the valuable contributions of many students and of Professors D. C. Mowery(Haas School of Business), C. Brown (Institute of Industrial Relations), M. Borrus (BerkeleyRoundtable on the International Economy) and C. N. Berglund (Oregon Graduate Institute). TheAlfred P. Sloan Foundation of New York has provided grant support since 1991. SEMATECH,the Semiconductor Research Institute of Japan, the Electronic Industry Association of Japan, andseveral of the participating companies have given important additional financial support. Aprevious version of this paper was presented at the European Solid State Device ResearchConference, Stuttgart, Sept. 1997.
`
`Applied Materials, Inc. Ex. 1028
`Applied v. Ocean, IPR Patent No. 6,968,248
`Page 8 of 10
`
`

`

`-9-Q & A from the Integrated Reliability Workshop, 10/13/97Q: What correlation do you observe between wafer throughput and yield?A: We saw one extreme case of excessive focus on yield, via heavy use of inline monitoring andrework, that reduced wafer throughput and total output. Raising yield from 85% to 92% is notworth a 20% reduction in wafers processed! In general, however, better performers were morebalanced, focussed on total output.Q: What correlation do you observe between effectiveness of SPC activity and cycle time?A: We have not studied that specific point, though we could do it with our data set.Q: How widespread is the use of automatic recipe download to processing equipment?A: Far from universal. Fabs running only one process, or with only minor variations, don’t needit. However, the best fabs employing a considerable range of process parameters haveimplemented auto download to minimize misprocessing.Q: To what extent to you find use of multivariate statistical analysis in manufacturing?A: Not much in manufacturing. However, advanced statistical methods are heavily utilized inprocess development prior to transfer to manufacturing.Q: How are process and product changes staggered in logic and ASIC fabs?A: Foundaries generally offer several mature fab processes. They encourage inexperiencedcustomers to work with these. New process generations may be introduced in a partnership withselected highly experienced customers who are willing endure the risks and costs of advancedtechnology.Q: What patterns do you observe in process change control?A: There must be a formal, documented procedure for evaluating potential process changes andfor making decisions and assigning responsibility for introducing any process change.Q: How can a firm be ahead in planning and scheduling, yet behind in equipment throughput?A: Firms in this category are well-disciplined in utilizing a defined machine capacity. But thecapacity they define and the total output they achieve is significantly below the best.Q: Do you normalize for number of mask layers in your figures on integrated chip yield?A: Line yield is normalized to 20 layers. Probe yield is not adjusted for layer count, but includesmemory chips that are repaired with redundancy. Integrated yield is the product of these two.Q: Have you seen single fabs that are excellent simultaneously in logic and memory?A: This is unusual, but at least one of the top performers achieves good results with a very largemix of memory and logic products in continuous production. They have invested heavily inautomatic reticle changers and autorecipe download. This is a very disciplined organization.
`
`Applied Materials, Inc. Ex. 1028
`Applied v. Ocean, IPR Patent No. 6,968,248
`Page 9 of 10
`
`

`

`-10-Q: To what extent do you see non-product production wafers devoted to process control, qualitycontrol, etc?A: Everyone is trying to minimize fabrication of full sequence non-product wafers. Processmonitoring is conducted using test structures placed in scribe lines of every production wafer.Introduction of new process modules may involve short-loop processing of test wafers. We didnot collect data on this latter activity.Q: As semiconductor manufacturers make transitions from 6” to 8” and then to 12”, are goodpractices carried over? And where are all the skilled technical and professional people comingfrom to staff all the new fabs under construction?A: In general, companies move experienced people from older fabs into new ones. In most caseswhere we have benchmarked 2 fabs of one company, similar practices are evident in both fabs.Concerning staffing of new fabs, one company that is building new fabs seemed overstaffed withengineers in the fab we benchmarked; but perhaps this is in preparation for anticipated needs.Q: Have you compared the time required for a cycle of learning (production cycle time plus timefor analysis and decision-making) at different companies?A: No. This would become more important if we observed bigger differences in rates ofimprovement at different fabs.Q: Have you studied any fully automated fabs?A: None. One company that built and operated such a fab in the past told us they would not do itagain, because it was too inflexible for making changes and improvements.Q: Your production cycle time seem to be those for normal production lots? How much fasterare “hot lots”?A: We did not collect any such data, but we know that “hot lots” can be considerably faster.Q: Why are the rates of improvement in yield (and other measures) so similar among fabs?A: This is an interesting question. We expected to find bigger differences. Empirically, mostmanufacturing industries observe similar uniformities in improvement rates. Much research hasbeen conducted on possible underlying fundamentals (key words: “learning curve” and “progressfunction”). See http://www.lbsr.com/raccoon/learncurve.html for a short paper and goodbibliography.
`
`
`
`View publication statsView publication stats
`
`Applied Materials, Inc. Ex. 1028
`Applied v. Ocean, IPR Patent No. 6,968,248
`Page 10 of 10
`
`