Roland A. - PVD for microelectronics (779636), страница 21
Текст из файла (страница 21)
POWELL AND S. M. ROSSNAGELFIG. 4.6 Bucking magnet hardware used to further confine secondary electrons and reduce operating pressure of a DC magnetron (after ref. 4.9).References4.1. R.W. Wilson and L. E. Terry, "Application of high-rate E x B or magnetron sputtering in themetallization of semiconductor devices," J. Vac.
Sci. & Tech. 13(I): 157-164 (1976).4.2. J.A. Thornton and A. S. Penfold, "Cylindrical Magnetron Sputtering," in Thin Film Processes,pp. 75-113, John L. Vossen and Werner Kern, Eds., Academic Press, New York, 1978.4.3. David B. Fraser, "The Sputter and S-Gun Magnetrons," in Thin Film Processes, pp. 115-129,John L.
Vossen and Werner Kern, Eds., Academic Press, New York, 1978.4.4. Robert K. Waits, "Planar Magnetron Sputtering," in Thin Film Processes, pp. 131-173, John L.Vossen and Werner Kern, Eds., Academic Press, New York, 1978.4.5. Stephen M. Rossnagel, "Glow Discharge Plasma and Sources for Etching and Deposition," inThin Film Processes II, pp.
11-77, John L. Vossen and Werner Kern, Eds., Academic Press,New York, 199 !.4.6. Robert Parsons, "Sputter Deposition Processes," in Thin Film Processes H, pp. 177-208, JohnL. Vossen and Werner Kern, Eds., Academic Press, New York, 1991.4. 7. Donald L. Smith, Thin Film Deposition: Principles & Practice, McGraw-Hill, NY, 1995.4.8. R. E. Demaray, J. C. Helmer, R. L.
Anderson, Y. H. Park, R. R. Cochran, and V. E. Hoffman,"Rotating Sputtering Apparatus for Selected Erosion," U.S. Patent No. 5,252,194 (Oct. 12,1993).4.9. K. Lai, "Design of Magnetron Sputtering Source for Low Pressure Operation," U.S. Patent No.5,593,551 (Jan. 14, 1997).4.10. R. Scholl, "Process improvements for sputtering carbon and other difficult materials using combined AC and DC process power," in Proc. 35th Ann.
Soc. of Vacuum Coaters (SVC) Tech.Conf., pp. 391-394 (1992).4.11. R.A. Scholl, "Advances in arc handling in reactive and other difficult processes," technical report of Advanced Energy Industries, Fort Collins, CO (1994).THE PLANAR MAGNETRON1014.12. L. Anderson, "A new technique of arc control in DC sputtering," in Proc. 35th Ann. Soc. ofVacuum Coaters (SVC) Tech. Conf., pp. 325-329 (1992).4.13. S. Beisswenger et al., "Economic considerations on modem web sputtering technology," inProc. 35th Ann.
Soc. of Vacuum Coaters (SVC) Tech. Conf., pp. 128-134 (1992).4.14. R. L. Cormia et al., "Method for Coating a Substrate," U.S. Patent No. 4,046,659 (Sept. 6,1977).4.15. S. Schiller, K. Goedicke, J. Reschke, V. Kirchhoff, S. Schneider, and F. Milde, "Pulsed magnetron sputter technology," Int. Conf. on Metallurgical Coatings and Thin Films (ICMCTF93),San Diego, CA, April 19-23, 1993.4.16.
S. Schiller, K. Goedicke and Ch. Metzner, "Advances in pulsed magnetron sputtering (PMSprocess)," Int. Conf. on Metallurgical Coatings and Thin Films (ICMCTF94), San Diego, CA,April 25-29, 1994.4.17. G. Este, "A quasi-direct current sputtering technique for deposition of dielectrics at enhancedrates," J. Vac. Sci. & Tech. A: 1845ff (1988).4.18. T. Asamaki, T. Miura, G.
Nakamura, K. Hotate, and S. Yonaiyama, "High-vacuum planar magnetron discharge," J. Vac. & Sci. Tech. A10(6): 3430-3433 (1992).4.19. T. Asamaki, T. Miura, K. Hotate, S. Yonaiyama, G. Nakamura, K. Ishibashi and N. Hosokawa,"High-vacuum planar magnetron sputtering," Jpn. J.
Appl. Phys. 32 (part I, no. 2): 902-906(1993).This Page Intentionally Left BlankChapter 5 Sputtering ToolsCAVEAT: The purpose of this book is educational and not commercial, andthe specific equipment used to illustrate pedagogical points should not beconstrued as a product endorsement. Furthermore, while we have tried todepict essential equipment features, neither the dimensions nor the detailsof a given piece of hardware are intended to be an engineering drawing.Finally, PVD tools are continuing to evolve with new and/or improvedmodels introduced every few years.
Therefore, the reader is encouraged toconsult the actual suppliers' product brochures or contact the suppliers directly to obtain accurate, up-to-date details on tool construction and product specifications.5.1 Evolution of PVD Tools for MicroelectronicsPVD equipment for microelectronic fabrication has evolved greatly sincethe early 1970s both in terms of technical performance and productivity.This evolution is inexorably linked to the phenomenal growth in complexity of Si devices over the same period (e.g., 1975 = 4K DRAM on 3-inchwafers; 1998 = 256 Mb DRAM on 8-inch wafers) and to the increasinglystringent requirements placed on both the PVD process and the PVDprocess tool to allow these devices to be manufactured in a cost-effectiveway.
Fortunately, PVD tool designers were able to take advantage of concurrent improvements in such supporting technology as vacuum pumping,gas delivery, robotic wafer handling, microprocessor-based control, andprocess automation. As a result, PVD equipment was able to progress relatively rapidly from manual-loaded, stand-alone batch tools to fully automated, vacuum-integrated, single-wafer cluster tools. In this regard, thedevelopment of deposition tools based on PVD was similar to the development of dry etching tools based on plasma technology. Pattern definitionby dry etching began to replace wet chemistry in the late 1970s and, notsurprisingly, today's state-of-the-art cluster tools for PVD and plasmaetching look quite similar.Figure 5.1 summarizes the evolution of microelectronic manufacturingfrom 1975-1998, and the following sections present a brief discussion ofPVD hardware evolution during this time frame, with representative illustrations of equipment.103R.
POWELL AND S. M. ROSSNAGEL104Microelectronic Manufacturing 1975-1998FIG. 5.1The evolution of microelectronic manufacturing and PVD manufacturing tools.5.1.1 PRE-1975" THE STAGE IS SETPrior to 1975, the preferred method of metallization for semiconductorproduction was vacuum evaporation with heating provided by e-beambombardment or resistance-heated filaments. Although sputtering hardware was available (e.g., Western Electric used sputtered Ta and TaN thinfilm resistors in hybrid circuits of the first Touch-Tone phones), PVD wasnot deployed in mainstream A1 metallization. Instead, sputtering was utilized to deposit insulators and refractory metals that were difficult to evaporate (see Fig.
5.2 for a typical RF diode configuration of the time).As device size decreased, the increased aspect ratio of features challenged the limited step coverage of e-beam evaporation, which is basicallya line-of-sight deposition. Also, the move away from pure A1 and towardSPUTTERING TOOLSFIG. 5.2105Early RF diode configuration of PVD showing batch substrates.AI-Si alloys to deal with junction spiking and toward AI-Cu to deal withelectromigration was a complication for evaporation technology due to thedifferent vapor pressures of the alloy constituents. Although sputtering offered a solution to both these problems, it could not be implemented intomanufacturing because RF diode sputtering of AI had relatively low deposition rates and poor electrical and optical film quality. The solution turnedout to be the development of DC magnetron sputtering based on the pioneering invention of the 3-inch sputter gun by Peter Clarke in 1968 [5.1].By confining secondary electrons near the surface of the sputter target,argon ion generation could be increased considerably with the magnetronconfiguration over the diode with resulting higher deposition rate, andelectron bombardment heating of the substrate was reduced [5.2].
Andthough RF magnetron sputtering was also possible, DC power supplieswere much less expensive than RF and film quality was better. By 1975,the DC magnetron had been refined from a small-scale laboratory curiosity to a production-worthy source (Fig. 5.3).R. POWELL AND S. M. ROSSNAGEL106FIG. 5.3 Production-worthy DC magnetron sputter sources were introduced into microelectronicmanufacturing circa 1975 (an S-Gun T M PVD source is shown).5.1.21975-1979:EARLY YEARS OF PRODUCTION P V DThis period saw the implementation of DC magnetron source technologyinto a variety of PVD system configurations from about a dozen suppliers,including Airco Temescal, Leybold-Heraeus, MRC, Perkin-Elmer Ultek,Sputtered Films, and Varian [5.3]. However, it was not until the late 1970sthat PVD began to displace e-beam evaporation.
For example, the chapteron IC metallization in a well-known "how-to" book published by FairchildSPUTTERING TOOLS107Semiconductor in 1979 [5.4] refers only to the diode configuration of sputtering, noting that "[sputtering] can be accomplished using both DC andRF voltages, and it can be used to deposit almost any material, althoughthe deposition rate is often extremely low."Hardware development was characterized by a strong effort to lowerthe cost per wafer through source designs that gave more uniform targeterosion and thus more effective percent utilization of target material(25-50% was typical).
A variety of target shapes were utilized (rectangular, circular, wedge-shaped) with sputtering being done with the waferabove (face-down), below (face-up), or at right angles (side sputtering) tothe source.Representative of hardware from this period are the Varian batch sputtering system shown in Fig. 5.4 in which wafers were manually loadedonto a planetary assembly that was rotated during deposition to improveuniformity, and the Perkin-Elmer system shown in Fig. 5.5 that used a rotating wafer table to improve uniformity.
The latter system is noteworthyin that it had a vacuum loadlock and an optional heater for outgassingwafers before deposition, anticipating future trends in vacuum cleanliness.Given the large installed base of evaporation equipment, an upgrade market also developed whereby existing e-beam and filament evaporation systems could be retrofit with a suitable PVD source (see Fig. 5.6).FIG. 5.4Batch sputtering system c. 1978 (Varian model 3125 coater is shown).108FIG.
5.5R. POWELL AND S. M. ROSSNAGELBatch sputtering system c. 1978 (Perkin-Elmer model 4410 coater is shown; see ref. 5.3).5.1.3 1980-1984: PVD STAND-ALONETOOLSSemiconductor International could report in 1980 that ~'In many semicon-ductor metallization applications, high vacuum evaporators are being challenged by DC magnetron sputtering systems. However, in no danger of extinction, high vacuum evaporators are a competitive alternative, featuringcost effective operation and high deposition rates" [5.5]. By 1985, however, DC magnetron s p u t t e r i n g - both planar and conical magnetron conf i g u r a t i o n s - was regarded as the state-of-the-art method of depositingthin films for interconnect applications, including AI-Si and AI-Si-Cu fromalloy targets and refractory metal silicides such as MoSi 2, WSi 2, and TiSi 2from composite targets [5.6, 5.7].
The reason for this paradigm shiftin metallization technology was the development of production-orientedPVD systems with high vacuum quality (for improved film purity), bettertemperature control and uniformity (for improved repeatability and uniformity of film properties), reduced particle levels (for higher device yield),and automated substrate transport that eliminated wafer breakage due tooperator mishandling and that supported production-worthy throughputs of60 wafer per hour.SPUTTERING TOOLS109FIG. 5.6 Users of evaporation equipment in the late 1970s could upgrade to PVD by replacingevaporation sources with one or more sputter sources (retrofit of an e-beam system for PVD operation c.
![](https://s.studizba.com/z.php?f=/templates/si/i/noavatar.png&w=100&h=100&t=1"){kind=avatar}
