The blur caused by the nonzero mean free path of electrons in photoresists exposed by extreme ultraviolet lithography has detrimental consequences on patterning resolution, but its effect is difficult to quantify experimentally. So far, most mean free path calculations use the dielectric formalism, which is an approximation valid in the optical limit and fails at low kinetic energy. In this work, we used a modified substrate-overlayer technique that exploited the attenuation of the Si 2p core level originating specifically from the native silicon dioxide to evaluate the attenuation of electrons traveling through 2 and 4 nm of photoresist overlayers to provide a close estimation of the inelastic mean free path relevant for photoresist lithography patterning and for electron microscopy. The photoemission spectra were collected in the proximity of the Si 2p edge (binding energy ∼101 eV) using synchrotron light of energy ℏω ranging between 120 and 550 eV. The photoresist films were prototypical chemically amplified resists based on organic copolymer of poly hydroxystyrene and tertbutyl methacrylate with and without triphenyl sulfonium perfluoro-1-butanesufonate photoacid generator and trioctylamine quencher. The inelastic mean free path of electrons, in the range that is relevant for photoresist exposure in extreme ultraviolet lithography (20-92 eV), was found to be between 1 and 2 nm. At higher kinetic energy, the mean free path increased, consistently with the well-known behavior. The presence of the photoacid generator and quencher did not change the mean free path, within experimental error. Our results are discussed and compared with the existing literature on organic molecules measured via dielectric formalism and electron transmission experiments.
Read full abstract