Hostname: page-component-586b7cd67f-t7czq Total loading time: 0 Render date: 2024-11-25T17:32:53.175Z Has data issue: false hasContentIssue false
  • English
  • Français

Microscopic Estimates for Electromigration Velocities of Intragranular Voids in Thin Aluminum Lines

Published online by Cambridge University Press:  15 February 2011

L. K. Wickham  and
J. P. Sethna
L. K. Wickham
Affiliation:
Laboratory of Atomic and Solid-State Physics, Cornell University, Ithaca, NY 14853-2501
J. P. Sethna
Affiliation:
Laboratory of Atomic and Solid-State Physics, Cornell University, Ithaca, NY 14853-2501
Get access

Abstract

We explore the effect of faceting on possible mechanisms for mass transport around electromigration voids in aluminum interconnects. Motivated by linear response estimates which suggest that particle flux would be much higher along steps than across terraces on a clean aluminum surface, we study step nucleation in the presence of a small driving force along a surface. We find that step nucleation, even on a nearly defect-free void surface, would be slow if the step energy is equal to that calculated for a clean aluminum surface. In the presence of a uniform electromigration force, the creation of new steps between existing ones should not occur unless the free energy cost of a step is much less than thermal energies. We conclude that voids cannot move intragranularly at μm/hr rates without help from other factors such as local heating and impurities.

Type
Research Article
Information
MRS Online Proceedings Library (OPL) , Volume 428: Symposium L – Materials Reliability in Microelectronics VI , 1996 , 171
Copyright
Copyright © Materials Research Society 1996

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

[1] Several observations of void motion appear in the references below. Marieb et al performed SEM in situ studies of passivated lines and Arzt et al have studied unpassivated ones. In addition, TEM in situ studies are described in Riege, S.P., Hunt, A. W., and Prybyla, J. A., MRS Symp. Proc 391, p. 249.Google Scholar
[2] Marieb, T.N., Abratowski, E., and Bravman, J. C. in Stress-induced Phenomena in Metallization, ed. Ho, P.S., Li, C.-Y., and Totta, P., (A.I.P. Proc. 305, NY, 1994) pp. 1–14.Google Scholar
[3] Rose, J. H., Appl. Phys. Lett. 61, 2170 (1992).Google Scholar
[4] Arzt, E., Kraft, O., Nix, W.D., and Sanchez, J.E., J. Appl. Phys. 76, 1563 (1994).Google Scholar
[5] Marieb, T., Flinn, P., Bravman, J. C., Gardner, D., and Madden, M., J. Appl. Phys. 78, 1026 (1995).Google Scholar
[6] Suo, Z., MRS Symp. Proc. 338, p. 379.Google Scholar
[7] Li, C.-Y., Borgesen, P., and Sullivan, T. D., Appl. Phys. Lett. 59, 1464 (1991);Google Scholar
Ho, P. S., J. Appl. Phys. 41, 64 (1970);Google Scholar
Nix, W.D. and Arzt, E., Met. Trans. A 23A, 2007 (1992).Google Scholar
[8] Shingubara, S., Kaneko, H., Saitoh, M., J. Appl. Phys. 69 207 (1991). Y.-C. Joo and C.V. Thompson, MRS Symp. Proc. 309, p. 319.Google Scholar
[9] Burton, W. K., Cabrera, N., and Frank, F. C., Phil. Trans. Roy. Soc. A 243, 299 (1951).Google Scholar
[10] Stumpf, R., and Scheffler, M., Phys. Rev. B 53, 4958 (1996).Google Scholar
[11] In the results of Stumpf and Scheffler, the sum of diffusion barrier plus cost to produce a mobile carrier is even higher for vacancies than adatoms on the surface.Google Scholar
[12] Since the roughening transition of the Al(111) surface is near its melting temperature, we use zero temperature free energies. In the square lattice Ising model, the zerotemperature step free energy is accurate to within about one percent even at T = .5T c ,: Avron, J.E., van Beijeren, H., Schulman, L.S., and Zia, R.K.P., J. Phys. A 15, L81 (1982).Google Scholar
[13] Screw dislocations can, of course, produce steps, but dislocation densities in interconnects have been estimated at 10 – 50/μm 2: Rose, J. H., Loyd, J. R., Shepela, A., and Riel, N.,Proc. 49th Mtng. of Elec. Microsc. Soc. Am., ed. Bailey, G. W., and Hall, E. L. (S. F. Press, San Fransisco, 1991) pp. 820–821.Google Scholar
[14] The following step flow calculation is in a similar spirit to that of Natori, A., Fujimura, H., and Fukuda, M., Appl. Surf. Sci 60/61, 85 (1992).Google Scholar
[15] See, for example, Wynblatt, P., and Gjostein, N. A., in Progress in Solid State Chemistry, ed. McCaldin, J. O., and Somorjai, F. (Pergamon, Oxford, 1975) vol. 9, p. 21. Tiny voids or high adatom density may require some correction to Gibbs-Thomson theory: see B. Krishnamachari, J. McLean, B. Cooper, and J. Sethna, sub. to Phys. Rev. B.Google Scholar
[16] Even if there is no flux entering a step, setting a local linear response flux leaving the step equal to D step ∇2 n step , the divergence at the step, gives a step adatom profile whose variation from the equilibrium density goes as exp(-y/A), where y is the distance to the nearest kink and , roughly a thousand lattice sites for clean aluminum parameters.Google Scholar
[17] Kondo, S., and Hinode, K., Appl. Phys. Lett. 67, 1606 (1995).Google Scholar
[18] Sorbello, R.S., to appear in Adv. Metal. for Future ULSI, (MRS Symp. Proc., 1996).Google Scholar
[19] Oates, A. S., J. Appl. Phys. 79, 163 (1996).Google Scholar
[20] LeGall, M., Lesage, B., Phil. Mag. A 70, 761 (1994).Google Scholar
[21] Augur, R.A., Wolters, R.A.M., Schmidt, W., and Kordic, S., MRS Symp. Proc. 391, p. 259.Google Scholar
[22] Hart, R. K., and Maurin, J. K., Surf. Sci. 20, 285 (1970).Google Scholar
[23] Brune, H., Wintterlin, J., Trost, J., and Ertl, G., J. Chem. Phys. 99, 2128 (1993).Google Scholar