Roland A. - PVD for microelectronics (779636), страница 35
Текст из файла (страница 35)
Timewhen the tool is not productive (e.g., tool is down for scheduled maintenance or is being used to run qualification wafers) would not be includedin the calculation of M T B FAs a simple illustration of PPvD cost-of-ownership, consider that the selling price of a high-end PVD cluster tool in 1997 was on the order of $5M.Assuming this tool were processing 200-mm product wafers 16 hrs/day(2 production shifts) for 5 days/wk at a throughput of = 45 wafers per hour(e.g., for an AI slab interconnect application), then ~ 190,000 wafers peryear are produced, which over a depreciated tool lifetime of 5 years translates into ~ 1 x 10 6 wafers.
Dividing the initial tool cost by the totalwafers processed gives $5 per wafer, providing an extremely crude estimate of the cost-of-ownership for a PVD tool - - and one that is absolutelya lower bound.This simple calculation underestimates the real cost of owning and operating a PVD tool in several respects.(1) Occupancy Costs. Even if the tool is not in use it occupies space,and one needs to consider the fixed costs associated with the areataken up by the frontend of the tool in the clean room (face print) aswell as the footprint of the tool in the adjoining, equipment maintenance bay.(2) Consumables. When the tool is running, one needs to include thecost of consumables such as sputter targets and process gases. Foran integrated process such as slab A1, where one could be depositinga Ti/TiN/AI/TiN stack, a mix of targets and gases need to be factoredinto the equation.(3) Maintenance and Repair The cost of spare parts such as shields orreplacement pumps will also add to the operating cost over the lifeof the tool.(4) Labor There will also be labor costs (salary and fringe benefits) associated with running and maintaining the tool, which can be significant.
For example, two 8-hour shifts/day x 200 work days/year x5 years = 16,000 hours of labor, assuming one shift operator per tool.180R. POWELLAND S. M. ROSSNAGEL(5) Utilization. Tool utilization will be less than 100% because ofscheduled downtime for preventive maintenance and target andshield changes; unscheduled downtime to correct operator errors, repair broken equipment, etc.; running test wafers for process requalification after the tool has been brought back to operational status;and carrying out processes such as "Ti-pasting" in a TiN depositionchamber (see Section 11.8) that can reduce Ti target life by 20-30%.This means that instead of 16 hrs/day, the tool may really be utilizedto run product wafers 12.8 hrs/day (80% of the time).(6) Yield.
Finally, since no process has 100% yield, some of these product wafers will contain die that fail electrically when probed due toone or more defects introduced by the PVD tool (e.g., an open interconnect line caused by a large particle that landed on the waferduring PVD and masked the subsequent deposition).Sophisticated modeling software, such as the widely used SEMATECHCost of Ownership Model, take these and other economic considerationsinto account.
The cost per wafer calculated using a sophisticated CoOFIG. 5.48 Cost-of-ownership sensitivity analysis of an advanced PVD cluster tool showing the importance of throughput and preventive maintenance (PM) time. (Source: P. Singer in SemiconductorInt'l., p. 113, July 1995).SPUTTERING TOOLS181model could be much higher (perhaps 50-100% more) than the $5 perwafer calculated by simply dividing the selling price of the tool by the totalwafers run through it. Equally important to C o 0 is the sensitivity of thevalue to changes in input parameters, since this identifies where engineering or process improvements will have the greatest return on investment.Figure 5.48 shows a CoO sensitivity analysis reported for an advancedPVD cluster tool; of the variables considered, wafer throughput and preventive maintenance time had the greatest impact on cost per wafer, orCPW [5.55].
For example, a 10% increase in throughput might lower overall CPW by 6% (sensitivity index o f - 0 . 6 ) , while a 10% increase in preventive maintenance time might raise CPW by 2% (sensitivity index of+0.2) This type of analysis is also useful in evaluating interrelated economic impacts of a given engineering improvement. For example, using alarger diameter target might allow one to achieve an erosion profile leading to a greater target life for a given PVD film uniformity; however, thisreduction in CPW might be largely offset by increased target cost.References5.1.
R Clarke, "Sputtering Apparatus: Low-Pressure Operation," U.S. Patent No. 3,616,450 (Oct.26, 1971).5.2. R. W. Wilson and L. E. Terry, "Application of high-rate E • B or magnetron sputtering in themetallization of semiconductor devices," J. Vac. Sci. & Tech. 13( I): 157-164 (1976).5.3. Anonymous, "Sputtering deposition trends," Semicond.
Int., 43-55 (1979).5.4. P. E. Guise and R. Blanchard, Semiconductor and Integrated Circuit Fabrication Techniques,p. 127, Reston Publishing, Reston VA, 1979.5.5. T. G. O'Neill, "Evaporation systems remain competitive," Semicond. Int. 3(8): 85-100 (1980).5.6. R S. Burggraaf, "Magnetron sputtering systems," Semicond. Int., 37-52 (Oct.
1982).5.7. R Burggraaf, "Advances in metallization technology," Semicond. Int., 73-79 (Nov. 1985).5.8. G. Birkmaier, A. Tampon, and H. Grunes, "Ultrahigh vacuum in production applications,"Semicond. Int., 108-109 (Apr. 1991 ).5.9. I. Hashim, I. J. Raajimakers, S.-E. Park, and K.-B. Kim, "Vacuum requirements for next wafersize physical vapor deposition system," J. Vac. Sci. & Tech., A15(3): 1305-1311 (1997).5.10. A.
McGeown, "'Sputter deposition via single-wafer, multichamber systems," MicroelectronicManufacturing and Testing, 11-13 (Aug. 1989).5.11. S. Dushman, Scientific Foundations o f Vacuum Technique, 2nd ed., J. M. Lafferty, Ed., Wiley,New York, 1962.5.12. A. Roth, Vacuum Technology, 3rd ed., Elsevier North-Holland, New York, 1990.5.13. J. E O'Hanlon, A User ]s"Guide to Vacuum Technology, 2nd ed., Wiley-Interscience, New York,1989.5.14. J. E O'Hanlon, "Ultrahigh vacuum in the semiconductor industry," J. Vac.
Sci. & Tech. A12(4):921-927 (1994).5.15. E. H. A. Granneman, "Film interface control in integrated processing systems," J. Vac. Sci. &Tech., B12(4): 2741-2748 (1994).182R. POWELL AND S. M. ROSSNAGEL5.16. P. Singer, "Vacuum pump technology leaps ahead," Semicond. Int., 52-54 (Sept. 1993).5.17.
R. Heyder, L. Watson, R. Jackson, G. Krueger, and A. Conte, "Nonevaporable gettering technology for in-situ vacuum processes," Solid State Tech., 71-73 (Aug. 1996).5.18. R. Giannantonio, M. Succi, and C. Solcia, "Combination of a cryopump and a non-evaporablegetter pump in applications," J. Vac.
Sci. & Tech. A15(1): 187-191 (1997).5.19. N. Lifshitz, W. Y. C. Lai, and G. Smolinsky, "Water related degradation of contacts in the multilevel MOS IC with spin-on glasses as interlevel dielectric," IEEE Electron Device Lett. 10:562-563 (1989).5.20. T. Tokunaga and N. Owada, "Effects of multichamber processing on reliability of submicronvias," in Proc. Conf. on Multichamber and ln-Situ Processing of Electronic Materials, SPIEvol. 1188, pp. 61-68 (1989).5.21.
Y. Tanaka, H. Suzuki, E Yanagawa, B. Cohen, H. Hanawa, T. Taniguchi, M. Togashi, and K.Watanabe, "Damage free, low via resistance sputtering cleaning technology for ULSI devices,"in Technical Proc. 183rd Electrochem. Soc. Symp. vol. 93-1, abstract no. 314, p. 481 (1993).5.22. R . A . M .
Wolters and W. C. J. Heesters, "Experimental study of metal-metal contact propertiesusing spin on glass," in Proc. VLSI Multilevel Interconnection Conf., pp. 447-449 (1990).5.23. C. T. Gabriel and J. E McVittie, "How plasma etching damages thin gate oxides," Solid StateTech., 81-87 (June 1992).5.24. J. E McVittie, "Plasma charging damage: An overview," in Proc.
1st Int. Symp. on PlasmaInduced Damage, 7-10, San Jose, CA, May 1996.5.25. W. Class and R. Hieronymi, "The measurement and sources of substrate heat flux encounteredwith magnetron sputtering," Solid State Tech., 55-61 (Dec. 1982).5.26. A.N. Pargeilis, "Evaporating and sputtering: Substrate heating dependence on deposition rate,"J.
Vac. Sci. & Tech. A7(i): 27-30(1989).5.27. L.T. Lamont Jr., "Thermal history of substrates during sputtering and sputter etching," SolidState Tech., 107-112 (Sept. 1979).5.28. W. Eckstein and J. P. Biersack, "Rellection of heavy ions," Zeitschr~J't fur Phvsik B63(4):471-478 (1986).5.29. D.
R. Wright, D. C. Hartman, U. C. Sridharan, M. Kent, T. Jasinski, and S. Kang, "Low temperature etch chuck: Modeling and experimental results of heat transfer and wafer temperature," J. Vac. Sci. & Tech. AI0(4): 1065-107()(1992).5.30. D. C. Evans, "A generalized mathematical model for wafer cooling with gas," NuclearInstruments and Methods in Physics Research B21:385-390 (1987).5.31. M.
E. Mack, "Wafer cooling and wafer charging in ion implantation," in Handbook ~?[ hmImplantation Teclmology, pp. 599-646 J. F. Zicglcr, Ed., Elsevier, North-Holland, Amsterdam,1992.5.32. A. Clarke, "Contamination control and thermal management of aluminum sputtering,'"Semicond. Int., 189-196 (June 1997).5.33. G. J. van Kolk, M. J. Verkerk, and W. A. M. C. Brankaert, "Effects of contamination on aluminum films, Part I: Room temperature deposition," Semicond. Int., 224-227 (May 1988);M. J. Verkerk, G. J.
van der Kolk, and W. A. M. C. Brankaert, "Effects of contamination on aluminum films, Part II: Elevated temperature deposition," Semicond. Int., 106-111 (June 1988).5.34. P. S. McLeod and L. D. Hartsough, "High rate sputtering of aluminum for metallization of integrated circuits," J. Vac. Sci. & Tech. AI4(I): 263-265 (1977).5.35. D. R. Wright, L. Chen, E Federlin, and K. Forbes, "Manufacturing issues of electrostaticchucks," J.
![](https://s.studizba.com/z.php?f=/templates/si/i/noavatar.png&w=100&h=100&t=1"){kind=avatar}
