Arthur Sherman - Chemical Vapor Deposition for Microelectronics (779637), страница 27
Текст из файла (страница 27)
al.SST 4/83)TISI.Figure 23: Resistivity versus time for different annealing temperatures and forWSi x and TiSi x .211%), and this value was reduced well below this level after annealin~?2 Theas-deposited film was really a sandwich of three layers. The top layer was thinamorphous silicon covered by several A of silicon dioxide, the intermediateone was an amorphous silicide layer of composition Tio.94Si, and the bottomone was another layer of amorphous silicon of equal thickness. After annealing, the average grain size on a polysilicon substrate was 5000 A, and on asingle-crystal substrate was 1000 A.
Also, dopants from the substrates did notdiffuse up to the silicide layers during annealing.Oxidation behavior of the TiSi 2 on silicon is the final concern.2 2 A gooddeal of surface and interfacial roughness was observed on the films that hadbeen put down on silicon and wet oxidized for one hour at 800°C. The oxidelayer was about 1800 A thick. Further study of the quality of the oxide layerwill be necessary before this film can be considered for device applications.148Chemical Vapor Deposition for Microelectronics5.7 ALUMINUMAgain, in the previous chapter we saw that deposition of thermal CVDaluminum films is possible.
Clearly, they improve the step coverage problem,since they are reasonably conformal. However, they are not completely satisfactory because of the inability to readily dope them with copper to limitelectromigration, and they exhibit grain sizes that are too large.As we saw in the TiSi 2 case just covered, it may be possible to depositan amorphous film and anneal, if necessary, to achieve a desirable grain size.In fact, PECVD of aluminum films composed of small crystals with randomorientation have been obtained 23 with as-deposited resistivities of 5 to 10,un-cm.Depositions were obtained from either AICI 3 or AI(CH 3 )3 mixed witheither Ar or H 2 , or both.
The plasma was excited by a 13.56 MHz source.At a power level of 200 watts, the deposition rate was 250 A/min. Step coverage over 1 micron high steps was at Ieast 50%.No atternpts were made to dope these films (electromigration), but sinceone of the key features of PECVD is the ability to control film composition,it may be possible to do this.REFERENCES1.
Rosier, R.S., Benzing, W.C. and Baldo, J., A production reactor for low2.3.4.5.6.7.temperature plasma enhanced silicon nitride deposition. Solid StateTechnol. 19(6) :45 (1976).Hollahan, J. R., Wauk, M.T. and Rosier, R.S., Plasma-enhanced chemicalvapor deposition of thin films and some of their etching characteristics,in Proceedings of the Sixth International Conference on ChemicalVapor Deposition, eds.
Donaghey, L.F., Rai-Choudhury and R.N. Tauber,Electrochem. Soc., Pennington, NJ, 1977, p. 224. This table was originally presented at the Fall 1977 Meeting of The Electrochemical Society, Inc. held in Atlanta, Georgia.Claasen, W.A.P., Valkenburg, W.G.J.N., Willemsen, M.F.C., and v.d. Wijgert,W.M., Influence of deposition temperature, gas pressure, gas phasecomposition, and RF frequency on composition and mechanical stressof plasma silicon nitride layer. J. Electrochem. Soc.
132:893 (1985).Claasen, W.A.P., Valkenburg, W.G.J.N., Habraken, F.H.P.M., and Tamminga, Y., Characterization of plasma silicon nitride layers. J. Electrochem. Soc. 130:2419 (1983).Chow, R., Lanford, W.A., Ke-Ming, W. and Rosier, R.S., Hydrogen contentof a variety of plasma-deposited silicon nitrides. J. Appl. Phys. 53:5630(1982).Fang, Y.K., Huang, C.F., Chang, C.W. and Lee, R.H., Preparation andcharacterization of boron- and phosphorus-doped hydrogenated amorphous silicon nitride films. J. Electrochem.
Soc. 132:1222 (1985).Allaert, K., Van Calster, A., Loos, H. and Lequesne, A., A comparison between silicon nitride films made by PECVD of N 2 -SiH 4 /Ar and N 2 SiH 4 /He. J. Electrochem. Soc. 132:1763 (1985).Plasma-Enhanced CVD1498. van de Yen, E.P.G.T., Plasma deposition of silicon dioxide and siliconnitride films. Solid State Technol.
24(4) :167 (1981 ).9. Adams, A.C., Alexander, F.B., Capio, C.D. and Smith, T.E., Characterization of plasma-deposited sil icon dioxide. J. Electrochem. Soc. 128:1545 (1981 ).10. Yokoyama, S., Dong, D.W., DiMaria, D.J. and Lai, S.K., Characterizationof plasma-enhanced chemically-vapor-deposited silicon rich silicon dioxide/thermal silicon dioxide dual dielectric systems.
J. Appl. Phys.54:7058 (1983).11. Avigal, I., Inter-metal dielectric and passivation-related properties of plasmaBPSG. Solid State Technol. 26(10) :217 (1983).12. Tong, J.E., Schertenleib, K. and Carpio, R.A., Process and film characterization of PECVD borophosphosilicate films for VLSI applications.Solid State Technol. 27 (1 ) :181 (1984).13. Nguyen, V.S., Burton, S. and Pan, P., The variation of physical propertiesof plasma-deposited silicon nitride and oxynitride with their compositions. J. Electrochem. Soc.
131 :2348 (1984).14. Reif, R., Plasma enhanced chemical vapor deposition of thin crystallinesenliconductor and conductor films. J. Vac. Sci. Techno I. A2(2) :429(1984).15. Donohue, T.J., Burger, W. R., and Reif, R., Low-temperature silicon epitaxy using low pressure chemical vapor deposition with and withoutplasma enhancement. Appl. Phys. Lett. 44 :346 (1984).16. Hess, D.W., Plasma-enhanced chemical vapor deposition of metal andmetal silicide films, in Proceedings of the MatI.
Res. Society Symposium, 1985, Vol. 38.17. Tang, C.C., and Hess, D.W., Plasma Enhanced chemical vapor depositionof ~-tungsten, a metastable phase. Appl. Phys. Lett. 45(6) :633 (1984).18. Akitomoto, K., and Watanabe, K., Formation of WxSi 1 - x by plasma chemical vapor deposition. Appl. Phys. Lett. 39 :445 (1981 ).19. Tabuchi, A., Inoue, S., Maeda, M. and Takagi, M., Formation of Mo andMo-sil icide film by plasma assisted chemical vapor deposition, in Proceedings of 23rd Symposium on Semiconductor and IC Technologyof Japan, 1982, p.
60.20. Hieber, K., Stolz, M. and Wieczorek, C., Plasma enhanced chemical vapordeposition of TaSi 2 , in Proceedings of Ninth International Conferenceon Chemical Vapor Deposition, eds. Robinson, McD., Cullen, G.W.(The Electrochemical Society, Pennington, NJ, 1984), Vol. 84-6, p. 205.This figure was originally presented at the Spring 1984 Meeting of TheElectrochemical Society, Inc. held in Cincinnati, Ohio.21.
Rosier, R.S. and Engle, G.M., Plasma enhanced CVD of titanium silicide.J. Vac. Sci. and Technol. B2(4) :733 (1984).22. Morgan, A.E., Stacy, W.T., DeBlasi, J.M., and Chen, T-V. J., Materialcharacterization of plasma-enhanced chemical vapor deposited titaniumsilicide.J. Vac. Sci. and Technol. B4(3):723 (1986).23. Ito, T., Sugii, T. and Nakamura, T., Aluminum plasma-CVD for VLSIcircuit interconnections, in Digest of Papers of 1982 Symposium onVLSI Technology (IEEE, NY, 1982).6Production CVD Reactor Systems6.1 INTRODUCTIONIn Chapters 1 and 2, we not only covered the basics of thern1al and plasmaenhanced CVD, but we described the general reactor configurations that researchers have explored over the years. From these concepts have come a fewproduction CVD reactors that satisfy the commercial needs of the integratedcircuit manufacturing process.When considering a production reactor, we first assume that the requisitequality film can be made at least one at a time.
The challenge then is to develop a reactor that is capable of acceptable wafer throughput with each waferhaving film thickness within an acceptable tolerance. For example, we maywant a reactor that can process 30 wafers per hour with thickness uniformityon a single wafer, and from wafer to wafer, of ± 5%. In addition, we mayimpose other conditions such as permissible number of particles per cm 2 , orfor epi silicon films, the allowable number of defects per cm 2 • When we speakof wafer throughput, we are concerned with the actual cost per wafer for thisprocess step.The cost per -wafer will depend on many factors. First, the reactor can bequite expensive, so it is a capital item and must be amortized.
Also, if thereactor has to be cleaned very frequently or is unreliable and experiences a lotof down time, then this will also add to the capital cost. If the reactants areexpensive and not utilized efficiently, then this is another expense item. Energy requirements can be high for heating either the chamber or the susceptor.So, a system with high wafer throughput leads in the direction of lower costper wafer, provided film quality is acceptable.One way to achieve high wafer throughput is to pack many wafers intoeach reactor load. This is what is referred to as a "batch" system.
![](https://s.studizba.com/z.php?f=/templates/si/i/noavatar.png&w=100&h=100&t=1"){kind=avatar}
