(Color online) Schematic of the 6-inch Si substrate pre-patterned with a 19 mm wideband of 72.6 nm pitch nano-grooves. We measured the spin-coated resist thickness both inside (tpattern ) and outside (tblank ) the patterned area. Since the resist thickness is defined as its volume divided by the area, tpattern is always smaller than the filling height hf.
(Color online) SE-OCD models for an over-filled (a) and an under-filled (b) structure. The trapezoidal profile Si grating is defined by four parameters: the pitch P (72.6 nm), the etch depth H, the top width WT , and the bottom width WB. The HSQ is assumed to have a flat surface that is at a distance hf from the bottom of the Si trenches.
(Color online) SEM images of the Si gratings filled with the spin-coated HSQ and the (Ψ, Δ) spectra for samples without and with the HSQ. The SE-OCD best fit models are superimposed on the SEM images for comparison. Both the measured (lines) and the calculated (dots) spectra are presented, which show a close fit.
(Color online) The HSQ thickness in the patterned area for samples with different etch depths. Within the margin of measurement error, the spin-coated HSQ thickness tpattern shows no observable dependence on the pattern depth.
(Color online) A simplified model of the spin-coating process: (a) initially, the resist thinning is dominated by a radial outflow driven by the centrifugal force; vr (z) is the relative fluid flow velocity with z = 0 at the crest of the grating and the no-slip boundary condition requires that vr (0) = 0; (b) the transition point where the resist ceases to flow due to the increased viscosity and the reduced film thickness; (c) the final stage where a solid film is formed and the residual solvent evaporation produces a non-planar film over the Si topography.
Summary of the fitting results.
Article metrics loading...
Full text loading...