Resist contrast curves

Figure 3 Representative contrast curves for a) positive resists and b)...

The response of crosslinkers 1-4 to pTSA catalysis as a function of acid concentration is shown in Figure 2 for a 75°C/1 minute softbake and 105°C/1 minute hardbake cycle. The concentration of acid required to crosslink these films to a given LP is an indication of the relative resist sensitivities, while the steepness of the LP curves reflects resist contrast. Crosslinkers 3 and 4 are about twice as sensitive to pTSA catalysis as 1, while 2 requires a higher concentration of pTSA for crosslinking. Furthermore, the steepness of the curve for 3 suggests that it would show higher contrast in a resist formulation. 

Figure 2 Typical lithographic response (contrast) curves for (a) positive and (b) negative resists.

Contrast curves were obtained for each resist by measuring the thickness after development of a series of 1 mm by 5 mm exposed areas the exposure dose typically varied from approximately 1 mJ/ cm2 to several J/cm2 for the slowest resists. The majority of the resists were developed in ethyl acetate for 30 to 60 sec followed by a 20-sec rinse in 2-propanol. Initially, THF or a THF/2-propanol mixture was used as the developer they were replaced by ethyl acetate because it provided superior contrast. Resist sensitivity was taken to be the incident dose which resulted in 50% exposed thickness remaining after development, Dg 5. This is the standard convention for a negative resist. 

As expected, the incorporation of pendant unsaturation in the resists greatly enhances sensitivity as demonstrated by a comparison of the contrast curves for poly(N-aiiyl maleimide-VBC) and the structurally similar poiy(N-ethyl maleimide-VBC) (Figure 4). Both polymers have similar molecular weights and nearly identical mass absorption coefficients but the allyl-containing copolymer is 5X faster. 

The effect of increased x-ray absorption on sensitivity was explored by conducting monochromatic exposures of a bromine-containing resist, poly(N-allyl maleimide-vinyl benzyl bromide), at photon energies which bracket the bromine absorption edges between 1.6 and 1.8 keV contrast curves obtained for these monochromatic exposures are shown in Figure 7. The results are also plotted as l/D "5 vs absorption coefficient in Figure 8 the data accurately follow the predicted inverse relationship defined by Equation 1. 

For uniform large area exposure such as one encounters in the determination of the contrast curve, the incident electron energy dissipation density depends only on the depth z into the resist. Explicitly, it can be written in the form (47,48)... 

For a positive resist, the film thickness of the irradiated region after development decreases until eventually a critical dose Dp is reached which results in complete removal of the film 8,9). The sensitivity and contrast (7p) are evaluated in a manner similar to that for a negative resist. After they have been spin-coated and prebaked, a series of pads of known area are exposed to varying doses. The substrate is developed in a solvent that does not attack the unexposed film and the thickness of the film remaining in the exposed areas measured. The film thickness is normalized to the original thickness, and this value is plotted as function of log dose, as shown in Figure 5 where Dp represents the sensitivity of the positive resist. Contrast (7p) is determined from the extrapolated slope of the linear portion of the response curve as... 

The resist formulation was spin-coated onto a silicone wafer on which a bottom antireflective coating had been previously applied and then soft-baked for 60 seconds at 90°C on a hot plate to obtain a film thickness of 1000 nm. The resist film was then exposed to i-line radiation of 365 nm through a narrowband interference filter using a high-pressure mercury lamp and a mask aligner. Experimental samples were then baked for 60 seconds at 90°C on a hot plate and developed. The dose to clear, E0, which is the dose just sufficient to completely remove the resist film after 60 seconds immersion development in 2.38% aqueous tetramethyl ammonium hydroxide, was then determined from the measured contrast curve. Testing results are provided in Table 1. 

A variety of techniques have been used in the present work to establish the relative sensitivity of positive electron-beam resists made from copolymers of maleic anhydride (Table I). The term sensitivity is used rather loosely at times. In the most practical sense, sensitivity is a comparative measure of the speed with which an exposure can be made. Thus, the exposure conditions, film thickness, developing solvent and temperature may be involved. Most often, the contrast curve is invoked as a more-or-less objective measure of sensitivity. The dose needed to allow removal of exposed film without removing more than about 70% of the unexposed film can be a measure of sensitivity. The initial film thickness and the developing conditions still must be specified so that this measure is not, strictly speaking, an intrinsic property of the polymeric material. 

The exposure characteristic curves were obtained by plotting normalized film thickness versus the logarithm of exposure time. Ihe improvement in resist contrast was evaluated by comparing the y-value (slope of the exposure characteristic curve) for... 

Ihe bleaching curves of the CEL layers are shown in Fig.7. Two parameters which express the optical efficiency of the CEL layers can be derived from the bleaching curves. One is the Tco/To ratio where t represents the initial transmittace of the CEL layer and T represents the transmittance of the completely bleached CEL layer. High contrast enhancement is provided by a high Wfc ratio(l). Since the three diazoniun salts have a large T /1b ratio, a good improvement in resist contrast can be expected. 

A CEL curve was presented which can express the relationship between resist contrast and optical properties such as molar absorption coefficient and quantun yield. This curve is thought to be a useful tool for designing CEL materials. 

CEL curve relationship between resist contrast ratios ptical characteristic parameters of CEL materials(P and a). 

Resist contrasts are defined as the slopes of the linear portion of the sensitivity curves (Fig. 3) and depend on process conditions. Thus, the sensitivity (contrast) curves are constructed to semi-optimize process conditions for a given formulation. However, final optimization of a resist formulation and process conditions requires lithographic imaging of target features. A plot of a dissolution rate as a function of exposure dose (cf. Fig. 172) is very useful in assessing the developer selectivity (development contrast) as mentioned earlier. 

Another useful contrast values are related to the resist chemistry in the film, which will subsequently affect the lithographic contrast. Sensitivity (contrast) curves similar to Fig. 3 can be generated by following the degree of reaction (deprotection, for example) with IR or by measuring thinning (in deprotection, for example) as the function of exposure dose. Comparison of a chemical contrast curve with a development contrast curve provides useful information on resist behavior, such as a degree of deprotection at E0. 

The sensitivity (or contrast) curves for positive and negative resists are shown schematically in Fig. 12.4. To generate this curve, the thickness of the exposed resist film after development (normalized to the original thickness... 

Figure 12.4 Contrast curves for (a) positive-tone and (b) negative-tone resists. The intercept of the curve and abscissa in positive-tone resists is called the dose to clear" and is designated as Do, while in negative resists, it marks the onset of cross-linking, and is designated as Dq. This should not be confused with the lithographic dose to print, which tends to be approximately 1.6-2.2 times higher. The absolute value of the slope of the tangent to the contrast curve at its intercept with the abscissa is defined as the resist contrast. It is usually defined in terms of an auxiliary dose value Di, which is obtained by continuing the above tangent line to the full resist film thickness (normalized to 1.0).

The theoretical resist contrast is different from the characteristic resist contrast, which is defined as a slope on a characteristic curve generated from normalized resist thickness versus an exposed log dose plot. 

For a negative resist, a contrast curve also can be consuiicted (Fig. 2). In a crosslinked system, the time of development should not alter the curve since extraction of sol from the thin film network is voy rtq>id. Contrast is defined now as (1/slope). A problem with all polymer resists, but especially acute with negative ones, is that of distortion of the remaining pattern by solvent swelling during development... 

The idealized contrast curve consisting of a linear thickness dependency on logarithm of dose is not likely to be realized in practice. Indeed, such a curve would contradict the clear predictions of theory for both positive and negative resists. 

Diffusion vs. Reaction Controlled Kinetics Contrast curves obtained from e-beam experiments at the lithographically usefiil doses at 90 °C and 110 °C for different PER times show a very characteristic behavior for this epoxy chemistry based chemically amplified resist. The gel dose does not change in this PER temperature and time range whereas the contrast increases in the higher temperature and time regime (Figures 7 and 8). 

Figure 9 Contrast curves of epoxy novolac resists exposed to DUV and postexposure-baked at 40 °C for the times shown.

---