This paper describes the use of statistical design experimentation to improve the photoresist performance properties of Dynachem's Nova 2070. A full factorial design was employed to investigate the effects of changes in the weight percent of both the minor resin and sensitizer in the total solids and of changes in the major resin's molecular weight on the after-hardbake wall profiles. The effect of the formulation changes on lithographic properties such as process latitude and resolution has also been measured. Scanning electron micrographs (SEMs) were generated to measure wall profile, thermal, and lithographic properties. A SEM measurement technique was then developed to quantify resist thermal stability. From these measurements models were generated to show the effects of the various formulation changes and to make predictions with respect to optimum formulations. Graphs of profile tendencies as a function of formulation changes and hardbake temperature and response surfaces generated from the various models are presented to help illustrate the optimization trends. With respect to lithographic performance, the experimental and model data indicate that the optimum resist formulation within the tested experimental matrix has the following make-up: high major resin molecular weight, low minor resin content, and high sensitizer content. With respect to thermal stability, the data suggests that the optimum resist formulation is the following: high major resin molecular weight, high minor resin content, and low to medium sensitizer content. The lithographic property optimum formula was retested to optimize its performance as a function of process changes according to a quadratic statistical design. Comparative process latitude graphs contrasting the optimum formula to alternative formulas under their respective optimized process conditions are also presented. These studies are collectively analyzed to indicate the direction that future resist formulation changes could be made to further optimize resist performance.
A series of experiments was run to determine the scope, magnitude and possible solution to the frequently observed proximity effect on CD uniformity, which has become an acute problem in submicron processing. The matrix of experiments compared linewidth control for lines that were in an isolated location versus lines in high-density locations. The effects of g- line versus i-line exposure, substrate type, softbake temperature, post-exposure bake temperature, developer strength, and develop time on linewidth control were measured for Dynachem's NovaTM 2050 and EL 2015. These two resists represent examples of conventional diazoquinone/novolac g-line and g/i-line resists, respectively. Additionally, the effects of actinic dyes on proximity effects were measured by evaluating the dyed versions of the aforementioned resists, NovaTM 2050 AR and EL 2015 0.5 AN. Actinic dyes were initially introduced to control antireflective notching on metal substrates, but have now been shown with this work to have an effect on the isolated versus dense line phenomenon. The matrix of experiments used was a full factorial statistical design. When the results were collated and analyzed by the software package RS/1, the relative effects of each parameter on CD control was assessed. The statistical data, graphs and predicted best processes for various resist/dye/substrate combinations are presented. This data provides general guidelines for the control of the proximity effect for 0.8 micrometers g-line and i-line processes.
A series of experiments was run to determine the chemical and
process effects on CD uniformity for Dynachem's new dyed
photoresist, NOVA 2050 AR, when developed with metal ion free and
metal ion type developers. The matrix of experiments included
variations in spray time and puddle time, but with the constraint
that all experiments had the same total time for spray and puddle
times. In order to do this type of matrix it was necessary to make
an unusual use of a mixture model to devise the experimental
design.
The design was devised to look at the effects of one-puddle, twopuddle,
three-puddle and all-spray processes. Also, the effect of
using either metal ion free or metal type developers was examined
by running the matrix with Dynalith EPD-85 and DE-3 (0.5%) as
respective examples of the two developer types.
Before beginning the above matrix the baseline parameters for
nozzle position, puddle time, spray time, spray speed, pre-wet
time, spray pressure and spray time were determined via a Plackett-
Burinan design of experiments.
These three studies when compiled and analyzed with the statistical
software package, RS-l, served to separate the effects of process
parameters and developer type according to their effect on CD
uniformity. These results suggest that three-puddle metal ion free
develop processes were superior.
Graphs will be shown that illustrate which parameters are the most
influential concerning CD uniformity. A mathematical model will be
presented that will allow the calculation of the standard deviation
for CD uniformity with 95% confidence for any given set of spray
and puddle times.
These data and models will thereby give guidance to the process
engineer how to best use developer type and process in order to
obtain the best CD uniformity for a 1.0 micron process.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.