Normalized residual resist thickness vs exposure dose for large features in (a) normal PMMA, and PMMA coated with (b) aluminum, (c) chromium, or (d) copper at primary beam energies from . In all cases the presence of a metal overlayer increases the required clearing dose, but has little effect on the contrast.
SEM images of Cr lines patterned using liftoff with Cu-coated and normal PMMA samples. Exposures were conducted at (a) and (b) with the same dose of .
Linewidth variation as function of exposure dose for (a) Cu coated and (b) normal PMMA at beam energy from . In both cases linewidth increased at higher exposure dose. In both experiments the measured lines are Cr patterned via lift-off.
Plot of change in linewidth (, measured after Cr lift-off) between bare and Cu coated PMMA at primary beam energies ranging from . The predicted change in linewidth based on the Monte Carlo simulations is also shown. The Monte Carlo curve represents a best fit to the experimental data using a binary resist development model with critical dose as the only free parameter. The experimentally measured change in linewidths follow the trend suggested by the Monte Carlo simulation and exhibit a distinct peak in the linewidth difference at low doses.
(a)Monte Carlo simulations of the energy density deposited in PMMA by a -diameter, electron beam. Energy profiles are shown for uncoated PMMA and PMMA coated with of Al, Cr, or Cu. The plots depict the energy density at the midpoint of the PMMA layer. As expected, as the atomic number and density of the coating increases the energy in the central region decreases and the energy profile broadens. (b) Monte Carlo simulation results (open symbols) with double-Gaussian plus exponential fits (solid lines). The model provides an excellent fit for the data over all radii and energy levels. Only uncoated and Cu-coated PMMA are shown for clarity.
Comparison of double-Guassian plus exponential model parameters for different metal coatings on PMMA. The parameters were obtained by fitting to Monte Carlo simulations. In each case the primary beam was in diameter, the metal was thick, and the PMMA was thick. The parameters were determined at the midpoint of the PMMA layer. Clearly the forward scattering and fast-secondary contribution, given by , increase with increasing atomic number, but remains satisfactory for many processes.
Article metrics loading...
Full text loading...