Research Highlights of Polymers Division

Introduction

Due to successes in pharmaceuticals research, combinatorial and high-throughput methods for searching composition space have received increasing attention for the synthesis and discovery of new inorganic materials, catalysts, and organic polymers. Combinatorial methods can also allow rapid scanning of parameter space to make fundamental measurements and develop physical models for polymers. One limitation is the difficulty of preparing parallel libraries and performing high-throughput screening with conventional instrumentation and sample preparation techniques.

We present combinatorial methods for measuring two important fundamental properties of polymer thin films: dewetting and phase behavior of blends. In each case library creation, high-throughput measurements, and informatics are used to generate combinatorial maps of wettability and phase behavior. The temperature and film thickness dependence of the dewetting behavior of polystyrene on silicon is evaluated in combinatorial libraries in which thickness and temperature are varied systematically. Automated scanning optical microscopy is used to determine the time-temperature-thickness superposition of dewetting kinetics and to observe transitions between film stability regimes. By a similar methodology, the phase boundary for a polystyrene / poly(vinylmethyl ether) blend film is observed with composition-temperature libraries. The combinatorial method is validated by comparison to previous results. In this procedure, the wafers are cleaned in a Piranha solution to grow a thin (˜10 Å) layer consisting primarily of SiOx. The results show that high-throughput experimentation is useful not only for the discovery of new materials, but also for observation of fundamental materials properties.

Experimental

Library Creation. To investigate polymer thin film dewetting phenomena, we used sample libraries with orthogonal gradients in thickness (h) and temperature (T) to create a large database of dewetting information in a few hours of experiments. Thickness gradients were prepared on “Piranha-etched” silicon wafers (Polishing Corporation of America) with a velocity-gradient, knife-edge coating apparatus. A drop of polymer solution was spread over the substrate under an angled steel blade (5 degrees relative to substrate) at constant acceleration. The solvent dried within seconds of spreading, and resulted in a polymer film with a gradient in thickness. The initial and final thicknesses and the slope of the gradient were controlled with solution concentration, blade-substrate gap width, and spreading velocity and acceleration. The thickness profiles, measured with 0.5 mm diameter spot-ellipsometry, had power law or polynomial dependence on position on the wafer, depending on the flow conditions. Solutions with mass fractions of 2% and 5% polystyrene (Goodyear, Mw = 1900 g/mol, Mw/Mn = 1.19, where Mw and Mn are the mass and number average relative molecular masses) in toluene were used to prepare 25 mm wide sample wafers.

Our high-throughput method for studying polymer blend phase separation involves the creation of libraries with orthogonal gradients in blend composition and temperature. Three steps are involved in preparing composition gradient films: gradient mixing, deposition, and film spreading. Two syringe pumps introduce and withdraw polymer solutions to and from a mixing vial at rates I and W, respectively, where I = W = 1.7 mL/min. Pump I contained mass fraction PS = 0.080 of polystyrene (PS, Mw = 96.4 kg/mol, Mw / Mn = 1.01, Tosoh) in toluene. The vial was loaded with an initial 2.0 mL of mass fraction 0.080 of poly(vinylmethyl ether) (PVME, Mw = 119 kg/mol, Mw/Mn = 2.5) in toluene from pump W. The infusion and withdrawal syringe pumps were started simultaneously while vigorously stirring the vial solution, and a third syringe, S, was used to manually extract solution from the vial. The rates I, W and S, the initial volume in the vial, and the sampling time control the end points and slope of the composition gradient, which has been verified in situ with FTIR spectroscopy.

Next the gradient solution from the sample syringe is deposited as a thin 31 mm long stripe on the silicon substrate. The gradient stripe was quickly placed under a stationary knife edge of equal length. The gradient stripe was spread as a film, orthogonal to the composition gradient direction, for a distance of 40 mm with the flow coating procedure described above. After a few seconds most of the solvent evaporated, leaving behind a thin film with a gradient of polymer composition.

High Throughput Screening

Both the thickness (dewetting) and composition (phase behavior) gradient samples were annealed on a linear T gradient heating stage, over a large range of T values, 90 °C < T < 160 °C. A CCD camera (Kodak ES1.0) coupled to an optical microscope (Nikon Optiphot 2) sent (1024 x 1024) pixel, 8 bit digitized images to a computer that also synchronized sample stage movement over a grid of T and h conditions using a robotic x,y translation stage. For quantitative analysis of images, we averaged the library T and h values over the image area. At the beginning of each time cycle, the translation stage returned to a home position to within ± 0.5 µm. In a typical experiment, the T-h-t dependence of dewetting structures was followed by collecting a 5 by 5 array of images every 5 min over a period of 2 h, for a total of 600 images. Optical microscope images and photographs capture the phase separation process as a function of T and ö PS with average standard uncertainties of T = ± 1.5 °C and ö PS = ± 0.006.

Results and Discussion

Figure 1 presents a photograph of a typical temperature-composition library after 2 h of annealing, in which the LCST phase boundary can be seen with the unaided eye as a diffuse curve. Cloud points measured with conventional light scattering are shown as discrete data points and agree well with the phase boundary observed on the library. The diffuse nature of the phase boundary reflects the natural dependence of the microstructure evolution rate on temperature and composition. Near the LCST boundary the microstructure size gradually approaches optical resolution limits (1 mm), giving the curve its diffuse appearance.

micrographs (one square image is 6 mm on a side) of a temperature-thickness library used to study thin film dewetting. Bright regions are more dewet than dark regions. A wide range of structures (holes [D] and polygons [A]) and both metastable and unstable (capillary nucleated [B]) film regions are observed in this image. By using a combinatorial technique to prepare libraries in which the temperature and thickness are varied systematically and continuously, we have also measured the temperature-thickness-time dependence of dewetting

By using a combinatorial technique to prepare libraries in which the temperature and thickness are varied systematically and continuously, we have also measured the temperature-thickness-time dependence of dewetting structures, stability, and kinetics in a small number of experiments. Figure 2 presents a composite optical image of a typical T-h library that indicates different dewetted structures and growth rates as a function of T and h. The libraries are able to reproduce a wide range of dewetted structures observed with our own uniform control samples and in previous work of others for similar systems, validating our methodology. The method presented here is unique in its ability to rapidly identify fundamental trends in film stability, nucleation mechanisms, and dewetting kinetics over a wide range of parameter space. Delineating these effects early in a scientific study can lead to observations of novel regions of phase space for thin film phenomena and can be extremely useful in guiding detailed analysis with conventional one-sample for one-measurement approaches.

For more information on this topic:

J.C. Meredith, A. Karim, E.J. Amis, Macromolecules, 33, 5760 (2000).