Roland A. - PVD for microelectronics (779636), страница 40
Текст из файла (страница 40)
In addition, since the sputtered atoms are spatially filtered toarrive at the surface at normal incidence, their kinetic energy is depositedin a small region around the impact sight. Therefore, increasing the aspectratio of the collimator may result in increased film stress in the compressive direction; this has been observed experimentally (Fig. 6.21 [6.22]).Cost For manufacturing applications, the net cost per layer as well asthe long-term reliability of the system are very important. Many of theseissues are discussed in Chapter 5 on system design.
Collimated sputteringresults in higher cost per layer for many reasons, although the most significant is the reduced deposition rate.Tool Issues In any application of collimated sputtering, an eventualproblem will be the result of thick deposits on the collimator itself. Sincethe transmission of the collimator can be as low as a few percent, the restof the sputtered atoms remain on the collimator. As a rough measure ofthis problem; consider a typical sputtering cathode, which has a workingthickness of about 1 cm; i.e., the high-purity target is at least 1 cm thick.Typical collimator diameters are on the order of 1-2 cm. So it can easilyDIRECTIONAL DEPOSITION207FIG.
6.20 (a) Sketch of microstructure of sidewall of a via deposited with collimated sputteringshowing columnar microstructure, (b) SEM of sidewall.R. POWELLAND S. M. ROSSNACEL0213Aspect Ratlo1.IThe stress in deposited Ti films asu Cunct~ono fcollim~~loraapecl ratio 16.221be expected that if the majority of the atoms from the cathode land on thecollimator, thicknesses of many millimeters can be deposited during thelifetime of the cathode, and these deposits will induce problems.As the deposits on the collimator build up, the effective aspect ratio ofthe cell increases.
which not only changes the angular profile of the transmitted atoms but also leads to even faster build up on the collimator wallsand a net reduction in the deposition rate (Fig. 6.22). (This effect is described in detail in ref. 6.23). An example of the change in the angular profile of the transmitted atoms is shown in Fig. 6.23.
The time before a collimator must be removed due to concerns about flaking and/or clogging istypically 25 to 50% of the target lifetime.The eventual flaking of the deposited film results in contamination ofthe wafer. Flaking can occur either when the films simply become toothick or due to thermal cycling that might occur during a deposition,waferlchamber heating, or chamber venting.Another concern with collimator clogging is that it does not occur uniformly across the collimator. As a collimator clogs, its effective diameterdecreases and its aspect ratio increases.
This results in even lower trans-DIRECTIONAL DEPOSITIONIi"--Fi~ililili~li~ iIillI IIIII209IIIIrr"c0.9"ooQ.s13No_.g 0.80zSimulated Data,.0.7 0Experimental Data_200__400600800Waters ProcessedFIG. 6.22 Calculated and experimental deposition rates for a 1.5:l collimator as a function of thenumber of waters processed [6.23].mission and faster clogging. During this whole process the collimator cellsin the center of the wafer tend to clog first, resulting in a net change in theuniformity of deposition across the wafer (Fig.
6.24).Collimator Construction The design and construction of a collimatormay have an effect on its operation and/or lifetime. The earliest collimators were constructed by clamping together arrays of short tubes. Thismoved rapidly to a machined approach, in which close-packed arrays ofround holes were machined into solid plates of AI or Cu. This approachwas designed for water cooling, although it was only possible to extract theheat at the perimeter of the collimator plate. Under high power ( > 10 kW),the center of a 30-cm-diameter AI collimator plate could be 60-100~ hotter than the edge. With no water cooling, collimator temperatures couldeasily reach 400~ due to the combined effects of deposition on the collimator and energetic electrons and photons from the plasma.
Later, platetype collimators were milled with hexagonal holes to minimize the amountof geometrical blocking by the walls of the collimator holes. These werestill cooled, though, and had similar thermal performance to the round-holecollimators.R. POWELL AND S. M. ROSSNAGEL2100.40......-~''i'~-- ~ - ~.................................I.......,.'....0.27............ 0.27. . . . 0.67--0.67.!0.30Pa,Pa,Pa,Pa,New CollimatorOld CollimatorN e w CollimatorOld Collimatorr( 3ET(!.)0.20i/>.1-,cr0.100.00-90.09_-60.0-30.0O. 0L30.060.090.0Angle (degrees)FIG. 6.23 Angular flux distributions of depositing atoms at the wafer surface as a function of collimator filling and also of operating pressure. Note: 0.27 Pa = 2 mTorr, 0.67 Pa = 5 mTorr.
It wouldbe unusual to operate a collimated sputtcring system at pressures much above 2 mTorr 16.13, 6.141.In parallel to the plate-collimator approach, work at Varian Associatescentered on the assembly of collimators from thin sheet metal [6.20]. Themetal strips (approx. 0.2- to 0.5-mm thick) were spot-welded into a hexagonal array. Since the heat transfer across these arrays of thin stainless steelstrips is very low, there was no attempt to cool this type of collimator, andduring high-power operation the temperature of the collimator could approach 500~ or more. This high temperature, when used during the sputtering of a low melting point material such as A1, could actually reduce thenet deposition on the collimator due to reevaporation.
However, this effectwith Ti and/or TiN, which are the most widely used cathodes for collimated sputtering, is low.Currently (1998), the sheet metal approach has been adopted by mosttool manufacturers, and there is very little attempt to draw heat from thecollimator. Because most collimator applications use Ti, often the collimator is constructed from Ti sheets. This matches the thermal expansion coefficients of the collimator and the depositing film, which results in lowthermal stress on the films deposited on the collimator during the variousDIRECTIONAL DEPOSITION.==.~..........211.0.75O0.5Z025 0-0125-0.5....0.75....N o r m a l i z e d Radial PositionFIG. 6.24 Normalized clogging rate of a collimator as a function of the radial position of the collimator 16.231.temperature cycling that can occur in manufacturing. The reduction inthermal stress is critical to better adhesion of the films deposited on thecollimator and to less flaking.Collimator Cleaning Historically, there have been two approaches tothe problem of what to do with a heavily deposited collimator.
In the earlydevelopment days, collimator materials were chosen such that the deposited films could be chemically etched and the collimator reused. Thismight be as simple as using an AI machined collimator for the depositionof Cu, which could then be easily removed from the collimator with nitricacid. The cost of a machined collimator for 200-mm wafer applicationswas about $2000-$5000, so cleaning was worthwhile.
Manufacturing applications of collimators have put more stringent requirements on toolingcost as well as the environmental problems associated with cleaning.Currently, the widespread use of sheet metal collimators has allowed amore disposable approach; there is no attempt to clean the collimator afterusage and it is simply replaced.
Sheet metal collimators can be fabricatedmuch less expensively ($500-$1000/collimator), which makes the expenseof cleaning too high.212R. POWELLAND S. M. ROSSNAGELReferences6.1. C.W. Kaanta, S. Bombardier, W. Cote, W. Hill, G. Korszykowski, H. Landis, D. Poindexter, C.Pollard, G. Ross, J. Ryan, J. Wolff, and J. Gonin, "Dual damascene: A ULSI wiring technology," in Proc. IEEE VMIC, Santa Clara, CA 1991, p.
144-152 (unpublished).6.2. S. M. Rossnagel and R. Sward, "Collimated magnetron sputter deposition with grazing angleion bombardment," J. Vac. Sci. & Tech. A13(1): 156 (1995).6.3. J. J. Cuomo and S. M. Rossnagel, "Hollow cathode enhanced magnetron sputtering," J. Vac.Sci. & Tech. A4:393-396 (1986).6.4. S. M. Rossnagel, C. A. Nichols, S. Hamaguchi, D. Ruzic, and R. Turkot, "Thin, high atomicweight refractory film deposition for diffusion barrier, adhesion layer and seed layer applications," J. Vac. Sci. & Tech.
B14:1819 (1996).6.5. J. N. Broughton, C. J. Backhouse, M. J. Brett, S. K. Dew, and G. Este, "Long throw sputter deposition of Ti at low pressure," in Proc. VLSI Multilevel Integration Conf, p. 201. (1995).6.6. I. Wagner, "Sputter deposition of Ti and TiN films with variable target-to-substrate distance,"in Proc. VLSI Multilevel Integration Conf, p. 226 (1995).6.7. A. A. Mayo, S. Hamaguchi, J. H.
Joo, and S. M. Rossnagel, "Across-wafer nonuniformity oflong throw sputter deposition," J. Vac. Sci. & Tech. B15 (1997).6.8. R. S. Robinson, "Energetic binary collisions in rare gas plasmas," J. Vac. Sci. & Tech. 16:179-185 (1979).6.9. S. M. Rossnagel, D. Mikalsen, H. Kinoshita, and J. J. Cuomo, "Collimated magnetron sputterdeposition," J. Vac. Sci. & Tech. A9:261-265 (1991).6.10.
R. N. Tait, S. K Dew, W. Tsai, D. Hodul, T. Smy, and M. J. Brett, "Simulation of uniformity andlifetime effects in collimated sputtering," J. Vac. Sci. & Tech. BI4:679 (1996).6.11. R. V. Joshi and S. Brodsky, "Collimated sputtering of TiN/Ti liners into sub-half-micrometerhigh aspect ratio contacts~lines,'" Appl. Phys. Lett. 61:2613-2615 (1992), and R. V. Joshi andS. Brodsky, in Proc. VMIC, Santa Clara, CA, 1992 (unpublished) p. 253.6.12. T. Janacck, D. Liu, S. K. Dew, M. J.
![](https://s.studizba.com/z.php?f=/templates/si/i/noavatar.png&w=100&h=100&t=1"){kind=avatar}
