|
This paper uses simulation to examine the effect of four specific process contrast modifiers on lithographic performance of a chemically amplified process. With exposure latitude as a process contrast indicator, it is shown that all four modifiers (1) dissolution curve steepness, (2) minimum dissolution rate, (3) bake kinetic contrast, and (4) diffusion, have similar effects on some important responses such as dense line resolution and isolated:dense proximity bias. For other simulation outputs, some of the modifiers behave differently from the others. In particular increasing diffusion length gives a rapid decrease in exposure latitude and resolution, but has relatively little effect on CD-related depth of focus. Also the lithographic predictor exposure margin (EM equalsV E1:1/E0), shown to be a good indicator of process contrast in the past, fails if diffusion length varies. The question of how the contrast modifiers affect performance of dense lines, isolated lines, and both dense and isolated together is addressed. The results show dense line performance is driven by process contrast: all important responses tested (resolution, exposure latitude, and DOF) are optimized by maximizing the process contrast. This condition is met by maximizing dissolution curve steepness and bake kinetic contrast, and minimizing Rmin and diffusion length. Isolated lines show different behavior. While isolated line exposure latitude is similarly maximized with highest process contrast, maximum DOF requires a low process contrast. Particularly effective for maximizing isolated line DOF with the least detrimental impact on exposure latitude is use of a low contrast dissolution curve and/or a high minimum dissolution rate. When dense and isolated lines must be printed simultaneously, highest contrast maximizes exposure latitude for both but gives poor isolated line DOF. There is also a distinct trade-off in dense line and isolated line DOF; simultaneous optimization is apparently not possible, under the conditions varied in this work.
|
