The develop process of traditional Novolac I-line resists with 2.38% TMAH developers was optimized at 60 second
puddle time. This process was embedded in the lithographers' tradition, and so it was transferred to newer resist
technologies i.e. CA resists.
The dissolution rate of exposed I-line resists is in the order of 1000A/sec, while KrF resists have dissolution rate of
~12,000A/sec, and ArF resists at ~60,000A/Sec. The resist thickness of KrF and ArF resists is usually less than 0.5
micron, which means that the exposed film is dissolved within a fraction of the first second. The extended exposure of
the resist features to developer was found to be a major reason for defects described as flying lines in center.
There are two routes to eliminate the extended development of the resist: one is to dilute the developer as a method for
improving the contrast and reducing the dissolution rate. This approach is very demanding and involves renewing the
OPE data package, adding a new line to the tracks with a new material that is not a standard developer, but a special
The other option is reduce develop time in order to minimize the attack on the resist-substrate interface, and resist edge.
Experiments were run at 10, 20, 25 and 40-second puddle time, showing only a minor increase in exposure energy is
necessary to reach CD target.
OPE data was checked by simulation and SEM measurements, showing no significant difference between 60 and 25
second puddle time.
Reducing puddle time improved defect density and eliminated the phenomenon of flying lines in the center.
This paper presents our experiments towards improved CD control over 4000A' poly steps using a single layer resist. The techniques used for reduction of CD variation were a special develop process to enhance surface inhibition effect, the use of dye additive and increase of resist thickness. As resist thickness varies over the poly steps it goes through various minima and maxima of the swing curve. This phenomenon causes the line width to vary and forms a challenge to modern lithography as shrinking design rules are more and more demanding, and CD variation due to topography can cause the device to fail. We have looked at the effect of dye additive, a special develop process and choice of thickness on CD control, focus and exposure latitude. The recommended process is chosen based on those parameters and presents the best achievable results with single layer resist. The geometries we investigated were 0.5 and 0.6 Micron showing that an acceptable process window for those geometries is achievable under a production environment.
This paper studies e effects of process conditions and dye additives on sidewall profiles using RS—1 designed experiments. We have explored the mechanism of forming negative sidewall profiles (above 90) in a positive photoresist, the dependence of depth of focus (DOF) and exposure latitude upon sidewall angle The resist studied is a positive, dual purpose (g—, i—line) resist, ULTRAMAC EL2015. The study presented in this paper was done in i—line but similar phenomena were observed in g—line as well. The effect of two different actirtic dyes in various concentrations on wall angle control is demonstrated. The two dyes investigated are both soluble in resist solvent but their solubility in developer is very different and their effect on required exposure energy is different. The contribution of the dyes to the absorption of the resist, their effect on swing curve and surface inhibition is evaluated and compared. The surface inhibition effect of undyed resist has been determined to be the most critical parameter influencing profiles, as increased post exposure bake temperature increases sidewall profile angle.
The accelerated drive toward 0.5 and 0.3 urn optical imaging is putting severe
pressure on both the equipment and photoresist vendors. Materials must give
both high resolution and high process latitude. These needs led to the develop-
ment of EL2015 positive photoresist series. This is a high contrast, high
resolution photoresist series based on the same chemical concepts used in our
PR1O24MB resists. This new resist contains a 2,1,5-naphthoquirionediazidesulfonate
ester of a special novolac as its photoactive component. This is combined with
an improved novolac resin for better film integrity. EL2015 is a versatile
positive resist optimized for g-, h-, i-line, and broadband exposure tools.
Resolution of better than 0.50 urn has been routinely obtained with current
0.40 NA i-line steppers and 0.40 urn resolution at best focus. A focus latitude
of greater than 1.5 urn for 0.5 urn images with an exposure latitude of 30-50%
has been demonstrated on i-line equipment. Finally, the thermal stability
and plasma etch resistance of EL2015 has been found to be comparable to current
high resolution g-line photoresists.