Surface Inhibition

Our research group has studied the dissolution of photoresists over the last 8 years in the effort to determine a fundamental mechanism for the aqueous base dissolution of phenolic polymers.  This research has led to the critical ionization model for dissolution, and the incorporation of this model into a molecular level simulation algorithm.  (See Critical Ionization Model.) However, one aspect of dissolution that has been difficult to explain is surface inhibition (or induction). Surface inhibition describes a slow dissolution rate at the beginning of dissolution (near the surface), and then an acceleration to a faster, bulk dissolution rate (Figure 1).  In general, this behavior is observed in nonchemically amplified resists (novolac polymer materials), but not in chemically amplified materials (polyhydroxystyrene).  The goal of this project is to determine the fundamental mechanism of surface inhibition.  With a fundamental explanation for this phenomenon, dissolution models can be improved, rational design and processing of resists may be influenced, and we will gain a more complete understanding of the dissolution process.

Although a nonlinear dissolution rate might seem detrimental, it actually can improve the shape of resist features. A Prolith simulation is shown in Figure 2. in which a resist feature is simulated with and without the surface inhibition effect. With surface inhibition, the sidewall angle is improved, and unwanted dissolution in unexposed areas of the resist (dark loss) is also reduced. Thus, the phenomenon is actually a beneficial effect.

Several theories exist for the origins of surface inhibition. Many theories involve a maldistribution, or concentration gradient, of resist components throughout the thickness of the film. For example, it has been suggested that a concentration gradient of residual casting solvent is responsible for surface inhibition.1 Our group has used radio-labeling techniques to measure the bulk concentration of residual casting solvent in thin films.2 A natural extension of those studies is to determine concentration gradients of residual solvent, but this requires a method of separating individual layers of a resist.  We have developed a method known as the halt development technique (Figure 3) to separate and analyze individual layers of a resist.  We have used this method to measure the distribution of residual casting solvent, low MW components, photoactive compound and density throughout ~1 mm resist films. For more detailed reading, please consult references (3-5).

Figure 1. Dissolution profile of a film with and without Surface inhibition

Figure 2. Simulated resist line with and without surface inhibition

Figure 3. Diagram of Halt Development Technique. 1) A coated wafer is loaded into the cell. 2) Dilute developer is poured into the cell. 3) The resist is allowed to partially dissolve. 4) The dissolved resist is drained from the cell. 5) The intermediate thickness is determined. 6) The top layer is analyzed. The process is repeated for the bottom layer of resist.

Figure 4. Typical results of the halt development technique.  Shown is the PGMEA profile in a 1.3 mm novolac film at 70 and 110°C bake temperature.

An interesting theory for surface inhibition is derived from the lattice model for dissolution. In the lattice model, thousands of polymer chains are equilibrated onto a lattice, with the boundary condition that the top and bottom are impenetrable. This creates smooth surfaces at the top of the film. As dissolution proceeds, the surface roughness increases, the surface area available to the developer increases, and the dissolution rate increases accordingly.  This is shown in Figure 5. A pictorial of the lattice model is shown, as well as the average roughness and the dissolution rate versus thickness.  In this example, the degree of polymerization is 30, the critical ionization fraction (fcrit) is 0.6, and the fraction of ionized sites (a) is 0.85. The void fraction in the lattice is 0.2. A more detailed discussion can be found in references 3 and 6.

Figure 5. Increase in surface roughness and dissolution rate predicted by the critical ionization lattice model for dissolution.

References

B.T. Beauchemin Jr. and C.E. Ebersole, Proc. SPIE, 2438, 261, 1995

A.B. Gardiner, A. Qin, C.L. Henderson, S. Pancholi, W.J. Koros, R.R. Dammel, C. Mack, W.D. Hinsberg, C.G. Willson, Proc. SPIE, 3049, 850, (1997)

S.D.Burns, A. B. Gardiner, V.J. Krukonis, P.M. Wetmore, G.M. Schmid, J.Lutkenhaus, L.W. Flanagin and C. Grant Willson, Proc SPIE (2001)

A.B. Gardiner, S.D. Burns, A. Qin, and C. G. Willson, J. Vac. Sci. Technol. B, 19(1), 136, (2001)

S.D.Burns, A.B. Gardiner, V.J. Krukonis, P.M. Wetmore,A. Qin, and C.G. Willson, Proc. Amer. Chem. Soc., PMSE, 81, 81-84 (1999)

L.W. Flanagin, V.K.Singh, C.Grant Willson, J. Vac. Sci. Technol. B, 17(4), 1371-1379 (1999)

© 2006 Willson Research Group, University of Texas at Austin
Last updated
 Site design by Arrion Smith and Obi
WEL 5.240, 512.471.3975