Fluorine atom subsurface diffusion and reaction in photoresist

Kinetic studies of fluorine and deuterium atoms interacting with an OiR 897 10i i-line photoresist (PR) are reported. All experiments were conducted at room temperature. Films of this PR were coated on quartz-crystal microbalance (QCM) substrates and exposed to alternating fluxes of these atoms in a high vacuum apparatus. Mass changes of the PR were observed in situ and in real time during the atom beam exposures using the QCM. A molecular-beam sampled differentially pumped quadrupole mass spectrometer (QMS) was used to measure the species desorbing from the PR surface during the F and D atom exposures. During the D atom exposures, hydrogen abstraction and etching of the PR was observed, but no DF formation was detected. However, during the F atom exposures, the major species observed to desorb from the surface was DF, formed from fluorine abstraction of deuterium from the photoresist. No evidence of film etching or fluorine self-abstraction was observed. The film mass increased during F atom exposure, evidently due to the replacement of D by F in the film. The rate of DF formation and mass uptake were both characterized by the same kinetics: An initially rapid step declining exponentially with time (e−t/τ), followed by a much slower step following inverse square root of time (t−1/2) kinetics. The initially rapid step was interpreted as surface abstraction of D by F to form DF, which desorbs, with subsequent F impacting the surface inserted into surface C dangling bonds. The slower step was interpreted as F atoms diffusing into the fluorinated photoresist, forming DF at the boundary of the fluorinated carbon layer. The t−1/2 kinetics of this step are interpreted to indicate that F diffusion through the fluorinated carbon layer is much slower than the rate of F abstraction of D to form DF, or the rate of F insertion into the carbon dangling bonds left behind after DF formation. A diffusion-limited growth model was formulated, and the model parameters are consistent with values reported in related systems. The implications of these results for plasma etch applications with respect to radical diffusion through surface-passivating films is discussed.