Christopher L. Soles, Polymers Division, Materials Science and Engineering Laboratory
The deprotection reaction front profile between exposed and unexposed regions of a chemically amplified photoresist can limit the ultimate feature resolution (i.e. image blur and CD control). Factors that affect the reaction front include the diffusion of the photogenerated acid, the catalytic length of the regenerated acid, composition of the photoresist polymer, additives, temperature, and time. In the past, the spatial evolution of the reaction front propagation was indirectly determined from the final developed structure, through a combination of theoretical modeling and experimental measurement, or by inducing contrast to electrons with additional chemical modification. Here, we use neutron and x-ray reflectometry to provide the first direct, in-situ measurements of the deprotection reaction front profile (chemical composition and density), with nanometer spatial resolution. This detailed knowledge of the deprotection and development processes will be invaluable for CD and LER control in sub-100 nm lithography, where error budgets become increasingly stringent.
In these experiments, an acid-sensitive protected polymer specially synthesized with a deuterated volatile protection group (a tert-butoxycarbonyl derivative) is spin-coated onto a silicon substrate. The deuterated group provides strong neutron scattering contrast without significantly changing the deprotection chemistry. The deprotected polymer, poly(hydroxystyrene), loaded with a photoacid generator (PAG) is spin-coated onto the protected polymer layer forming a bilayer structure with a sharply defined interface. Upon exposure to UV radiation, acid generated in the upper layer migrates across the interface and initiates the deprotection reaction in the lower layer. As the polymer undergoes the deprotection reaction, the deuterated protection group volatilizes, leaving a clear demarcation of the location, percent deprotection (composition), and shape of the reaction front in the neutron reflectivity profile. Further, the subsequent development of the upper layer and the deprotected parts of the lower layer provides detailed information about the relationship between the reaction front and the final image. We find that the in-situ interfacial profile has the form of a simple error function, but the developed compositional profile does not have this functional form. In addition, we observe additional deprotection at the surface of the lower layer after development. This methodology is currently being generalized to study the effect of PAG loading, exposure dose, and post-exposure bake temperatures and times on the spatial extent of the deprotection reaction in photoresist polymers; critically needed information to fabricate sub-100 nm features with chemically amplified photoresists.