310h Coupling Gas Phase and Surface Reaction Kinetics In C4F8 and SF6 Plasmas Used for Si and SiO2 Etching

George Kokkoris1, Evangelos Gogolides1, Andy Goodyear2, and Mike Cooke2. (1) Institute of Microelectronics, NCSR Demokritos, Terma (End) of Patriarhou Gregoriou St., Aghia Paraskevi, 15310, Greece, (2) Plasma Technology, Oxford Instruments, North End, Yatton, Bristol, United Kingdom

C4F8 plasma has been widely used for dielectric etching in microelectronics fabrication.1 It is also met in the area of micro-electro-mechanical systems (MEMS) fabrication, where in combination with SF6 plasma is used for deep Si etching during the Bosch2 process. In particular, SF6 plasma is used for the etching of Si and C4F8 plasma is used to deposit a fluorocarbon (fc) film which protects the sidewalls of the etched Si structure and sustains the anisotropy of the etching process. C4F8 has been also used for plasma enhanced chemical vapor deposition of fc films,3 which are exploited either as low-k films or in bio-fluidic applications.4

Several models (0D, 1D, and 2D) for C4F8 plasmas5-8 have been reported, while there is a lack of models for SF6 plasmas9 in low pressure conditions. None of the models has focused on the interaction of the gas phase with the reactor surfaces. The importance of the interactions with the reactor walls increases as the constraints for manufacturing become stricter; the interaction with the walls can affect the reproducibility of the process.10

In this work, a 0D or global type11 model for C4F8 and SF6 plasmas is combined with a surface reaction model. It consists of a) mass balances for all species (neutral or charged) except electrons, b) charge neutrality equation, and c) energy balance for the electrons. The reaction set in the gas phase comprises several types of reactions such as electron impact reactions (e.g. ionization, dissociation, attachment), ion – ion recombination, and neutral – neutral recombination; it is assembled based on literature sources for cross sections (electron impact reactions) and rate coefficients. The surface model describes the interaction of the species produced in the gas phase with the reactor walls (e.g. sticking, recombination, deposition); it is phenomenological, it is assembled so as to take into account pertinent experimental measurements, and its formulation is the surface site balances on the reactor surfaces. The rate coefficients of the surface model are fitted to the available experimental measurements.

The combined model, not only takes into account the effect of surface reactions in the densities of species in the gas phase, but allows the calculation of derived outputs which extend the potential experimental measurements for the validation of global models. In particular, it allows the calculation of the pressure rise after the ignition of the discharge (compared to that before) which links to the degree of dissociation of the parent gas. The combined model also allows the calculation of the effective sticking coefficients of the species which signify the net consumption of the species on the reactor surfaces and are the values measured in the experiments. Finally, it allows the calculation of the deposition rate and the ratio of F/C of the fc film deposited on the reactor surfaces (C4F8 plasma).

The results of the combined model compare well with experimental measurements concerning pressure rise as well as densities of F atoms, CF2 and CF radicals, and ion flux in an inductively couple plasma reactor. Summarizing the results for C4F8 plasma, the pressure rise and the densities of CF and CF2 radicals are compared with experimental measurements in Figures 1a and 1b respectively. In addition, the parent gas (C4F8) is vastly dissociated, CF4 dominates after 1000 W, CF3 radical is predicted to be produced on the reactor walls, and the fc film deposited on the reactor walls is mainly due to deposition of light fc radicals; Concerning the results for SF6 plasma, the loading phenomenon12 observed during Si etching by SF6 plasma is predicted.

The combined model for C4F8 and SF6 plasmas is also coupled with a previously developed simulation framework13,14 for feature scale etching, which links the output of a global model to the shape of the etched feature (e.g. trench or hole), in order to investigate the effect or the operational parameters of the reactor on the shape of the features during Si and SiO2 etching.

Figure 1. Comparison of model results with experimental measurements for C4F8 plasma. Pressure before plasma ignition is 1.29 Pa and gas temperature is 300 K. a) Pressure rise. b) Densities of CF2 and CF radicals.

References

  1. H. H. Doh, J. H. Kim, S. H. Lee, and K. W. Whang, J. Vac. Sci. Technol. A 14, 2827 2834 (1996).
  2. F. Laermer and A. Schilp, German Patent DE 4241045 (25 May 1994).
  3. K. Endo, K. Shinoda, and T. Tatsumi, J. Appl. Phys. 86, 2739-2745 (1999).
  4. P. Bayiati, A. Tserepi, P. S. Petrou, K. Misiakos, S. E. Kakabakos, E. Gogolides, and C. Cardinaud, Microelectron. Eng. 84, 1677-1680 (2007).
  5. S. Rauf and P. L. G. Ventzek, J. Vac. Sci. Technol. A 20, 14 (2002).
  6. A. V. Vasenkov, X. Li, G. S. Oehrlein, and M. J. Kushner, J. Vac. Sci. Technol. A 22, 511-530 (2004).
  7. D. Bose, S. Rauf, D. B. Hash, T. R. Govindan, and M. Meyyappan, J. Vac. Sci. Technol. A 22, 2290-2298 (2004).
  8. G. I. Font, W. L. Morgan, and G. Mennenga, J. Appl. Phys. 91, 3530-3538 (2002).
  9. C. Riccardi, R. Barni, F. De Colle, and M. Fontanesi, IEEE Trans. Plasma Sci. 28, 278-287 (2000).
  10. G. Cunge, B. Pelissier, O. Joubert, R. Ramos, and C. Maurice, Plasma Sources Sci. Technol. 14, 599-609 (2005).
  11. C. Lee, D. B. Graves, M. A. Lieberman, and D. W. Hess, J. Electrochem. Soc. 141, 1546-1555 (1994).
  12. R. A. Gottscho, C. W. Jurgensen, and D. J. Vitkavage, J. Vac. Sci. Technol. B 10, 2133-2147 (1992).
  13. G. Kokkoris, A. Tserepi, A. G. Boudouvis, and E. Gogolides, J. Vac. Sci. Technol. A 22, 1896-1902 (2004).
  14. G. Kokkoris, A. G. Boudouvis, and E. Gogolides, J. Vac. Sci. Technol. A 24, 2008 (2006).