Method to Optimize Polymer Film Spin Coating for Polymer LED Displays
M. Cecchi1, H. Smith2, D. Braun3, *
Engineering Department, 2 Department of Statistics, 3
Electrical Engineering Department,
Cal Poly State University, San Luis Obispo, California, U.S.A.
Research and development of displays and image sensors based on semiconducting polymers require design of new polymer materials and evaluation of film properties. Application of statistical methods can expedite process development. Optimizing device performance entails determining the effect of several process parameters, necessitates numerous samples, and may consume more scarce new material than desired. Uniform film thickness, with no voids, pinholes, inhomogeneities, or particulate contamination across the substrate can improve display efficiency and uniformity. After performing and analyzing a two-cubed full-factorial experiment with three replicated center points, this paper concludes that spin velocity and polymer concentration have statistically significant effects on film thickness, film uniformity, and device efficiency. Ramp acceleration does not. Analysis of variance methodology determines the effect of the three spin coating factors and their two-way interactions.
Keywords: Spin coating, Electroluminescence, Poly(phenylene vinylene) and derivatives.
Semiconducting polymers blend several attractive material properties that enable inexpensive and unique applications, including the next generation of flat and flexible image sensors and multimedia displays [1-8]. The application of a homogeneous, uniform thin film with no imperfections is fundamental to the manufacture of flat-panel-displays . To meet these stringent requirements, researchers and scientists may employ statistically designed experiments. "By the statistical design of experiment, we refer to the process of planning the experiment so that appropriate data will be collected, which may be analyzed by statistical methods resulting in valid and objective conclusion.". Experimental design assists engineers and researchers in understanding the relationships between input factors and responses in the most efficient way; by building mathematical models to relate the response to the input factors.
2. Experimental Setup
This experiment considered the factors of polymer concentration, spin speed and ramp acceleration. The responses were coefficient of variation (CV)  for thickness also called film non-uniformity, film thickness, and device efficiency. A three factor, two-level, full factorial statistically designed experiment was run in order to estimate the desired factor effects. Figure 1 shows the design matrix.
Figure 1a. Full Factorial Design with three factors, two levels plus center points. 2-level design prove useful to screen factors and to estimate first order models with interactions. The number of run for a k-factor design is n = 2k. 8 runs can analyze a total of 7 effects.
Figure 1b. Design Matrix showing each combination of the factors: polymer concentration, spin speed, and ramp acceleration. For example: At the center point (4, 3000, 5500) means (4 mg/ml, 3000 rpm, 5500 rpm/sec.)
Single layer devices were prepared on 30 mm x 30 mm glass substrates covered with patterned indium-tin oxide (ITO) electrodes using a previously reported technique . Solutions of OC1C10-PPV were made in three concentrations of 7:3 mixtures of Toluene and THF [4, 5]. The OC1C10-PPV polymer films were spin-coated from toluene/THF solutions with concentrations between 2 and 6 mg/ml. Spin speed between 1000 and 5000 rpm, and ramp acceleration between 1000 and 10000 rpm / sec. were used. Calcium contacts deposited on top of the polymer films by vacuum evaporation at pressures below 2x10-6 Torr yield active areas of 54 mm2.
2. Results and Discussion
The following statistical model was developed to analyze this experiment [10, 11]:
This model allows the determination of the effect of each of the three factors (a , b , g ) plus the three two-way interactions (a b , a g , b g ) on the response (y). Finally, all experiments include both experimental error, which is quantified by e , and a constant m .
Analysis of variance methods show that both polymer concentration and spin speed significantly influence CV (film non-uniformity). The corresponding p-values  are 0.019 and 0.003 respectively. Ramp acceleration does not effect CV, with a p-value >0.05. See Figure 2. There is no significant interaction between these factors, with p-value >0.05.
Figure 2. Mintab summary of the main effects for film non-uniformity. For example, at low polymer concentration, the average CV is ~ 10%, while at high polymer concentration, the average CV is ~ 20%. A steeper slope indicates a stronger factor effect on the CV. To minimize film non-uniformity, one may choose 2mg/ml polymer concentration and a 5000 rpm spin angular velocity.
Polymer concentration and spin speed significantly influence film thickness. Figure 3 shows a significant interaction between these two factors (p-value of 0.024). No statistically significant factor effects are found to influence device efficiency (all p-values exceeded 0.05).
Figure 3. Minitab summary of the factor effects for the film thickness. There is a spin angular velocity effect, because the average thickness over both polymer concentration levels decreases as spin angular velocity increases from 1000 rpm to 5000 rpm. However the amount of decrease depends on the level of polymer concentration.
Figure 4. The process window: The white region bound by the red (solid) and blue (dashed) lines on the contour plot is a likely process window, where film non-uniformity (CV) is below 5%, and the film thickness is at least 600 Å.
This work applies the Design of Experiment (DOE) methodology to optimize polymer LED film spin coating. The study analyzed the effect of polymer concentration, spin speed, and ramp acceleration on polymer film non-uniformity, thickness, and device efficiency. Rather than the 26 samples required by a One Variable at a Time (OVT) experimental approach, DOE provides an exhaustive study of the same design space with only 10 samples. Our experimental findings suggest that polymer film thickness and film non-uniformity depend on polymer concentration and spin speed. Ramp acceleration did not seem to be important.
This material is based upon work supported by the National Science Foundation under Grants No. 9702320 and 9820781. "Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation." The project has also received support from the Cal Poly Plan Project and used polymer from Covion Corporation.
 C. W. Tang and S. A. VanSlyke, Appl. Phys. Lett. 51 (1987) 913.
 J. H. Burroughes, D. D. C. Bradley, A. R. Brown, R. N. Marks, K. Mackay, R. H. Friend, P. L. Burns, and A. B. Holmes, Nature, 347, 539-541 (1990); R. H. Friend, R. W. Gymer, A. B. Holmes, J. H. Burroughes, R. N. Marks, C. Taliani, D. D. C. Bradley, D. A. Dos Santos, J. L. Brédas, M. Lögdlund & W. R. Salaneck, Nature 397 (1999) 121.
 D. Braun and A. J. Heeger, Appl. Phys. Lett., 58 (1991)
G. Gustafsson, Y. Cao, G. M. Treacy, F. Klavetter, N. Colaneri, and A. J. Heeger, Nature 357 (1992) 477.
 Links contained in http://www.ee.calpoly.edu/~dbraun/polyelec/moreinfo.html and http://www.chipcenter.com/eexpert/dbraun/main.html, accessed June 27, 2000.
 M.D. McGehee, E.K. Miller, D. Moses and A.J. Heeger, in P. Bernier, S. Lefrant and G. Bidan (eds.), Advances in Synthetic Metals: Twenty Years of Progress in Science and Technology, Elsevier, Amsterdam 1999, p. 98.
 G. Yu, J. Wang, J. McElvain, and A. J. Heeger, Adv. Mater.
10 (1998) 1431;
N.S. Sariciftci, L. Smilowitz, A.J. Heeger and F. Wudl, Science 258 (1992) 1474
 M.L.Parodi, W.T.Batchelder, J.McKibben, and P.D.Haaland, SID 94 Digest (1994) 933-935.
 Douglas C. Montgomery, Design and Analysis of Experiments, John Wiley and Sons, New York, 1976, p.2, 223
 Jay Devore, Nicholas Farnum, Applied Statistics for Engineers and Scientists, Duxbury Press, Pacific Grove 1999, p. 76, 331, 411-451.
A MS Word 97 file containing the manuscript to be published in Synthetic Metals is also available.