Defect-Free Ceramic Capacitors
Achieving CPVCAt the CPVC, the ceramic particles in the pigment-binder system form the densest packing. The voids between the ceramic particles are filled with binder.1 This is the most desirable condition.
Below CPVC, excess binder is present and ceramic particles lose contact with each other. When the ceramic slurry (sometimes called "paint") formulated below CPVC dries, "Swiss cheese" or void structures are formed in the film. Above CPVC, insufficient binder is present, and air voids are trapped in the coating. A system of ceramic powder, binder and air exists, which often shows itself in the development of pinholes and surface craters.
In lab tests, values of CPVC determined through a repetitious binder volume dilution procedure were repeatable with an accuracy of I0.8%. The films prepared at critical pigment (ceramic powder) volume concentrations showed smooth surfaces without pinholes, craters or Swiss cheese structures. The "paint system" tested was a slurry of ceramic powder used to produce a fired monolithic body that could have no physical defects, or else it would have electrical deficiencies.
Incoming batches of ceramic raw material varied somewhat in physical properties. Therefore, a means of characterizing each of those batches for CPVC was needed to avoid large yield losses in production from physical defects in the films formed. In addition, the production process required buildups of 30 to 200 or more layers of thin films without complete drying, still free of physical defects.
Binder Volume Dilution ProcedureIn the lab tests, multiple paint samples with a range of pigment (ceramic powder) volume concentration (PVC) from 20% to 85% were made according to the following routine:
1. Determine the ratio of ceramic powder volume (Vceramic) to binder volume (Vbinder) desired, depending on end-use criteria. For example, the Vceramic/Vbinder ratio in the manufacture of ceramic capacitors is chosen to promote the development of optimum mechanical and electrical properties after high temperature treatment (firing) of the parts. Select a range of PVCs around the estimated Vceramic/Vbinder ratio. (In this set of experiments, a range of 55% to 85% was chosen.) Prepare a starting mill base with high ceramic powder and solvent concentrations. Binder concentration should be kept low but sufficient to give some stability to the resulting slurry.
2. Pour a known amount of mill base into a 1-liter mixing jar. Mixing can be accomplished by means of a rubber mill or a high-speed shaker.
3. Prepare a stock binder solution, with the binder concentration high, but still in solution.
4. Incrementally add calculated amounts of binder solution into the mill jar. Record the exact weight of the binder solution added.
5. Put the paint sample on a fast speed mill for 20 to 30 minutes to allow sufficient mixing.
6. After adequate mixing, make a 1-gram paint drawdown on a glass plate, using a knife-coater set at a 6 mil wet gap setting.
7. Allow the film to dry (30 to 60 minutes). Then take a gloss measurement at a high value of incidence angle (e.g., 85°), which seems to produce the best results.
8. Check percent solids of the paint sample after drying 1 gram of the sample in a flowing air oven at 150°C for 30 minutes. This determines the actual solids content of the wet paint sample.
9. Repeat steps 4 through 8 for the remaining five binder dilution paint samples.
10. Plot gloss vs. PVC. Select the lowest point of the curve of gloss vs. PVC as the CPVC (see the sidebar for PVC calculations).
Paint Preparation Using CPVCOnce the CPVC has been determined, the paint can be prepared as follows:
Pre-mix the solvent and dispersant thoroughly in a beaker. Pour the solvent/dispersant solution into a dry mill jar one-half filled with dry ceramic grinding media. Roll the mill several turns to wet the grinding media. Load approximately half of the ceramic powder and roll the mill until the ceramic powder looks thoroughly wet (approximately 20 minutes for a 1-liter mill jar). Then add approximately 50% of the remaining ceramic powder and roll again until the ceramic powder looks thoroughly wet. Put in the remainder of the ceramic powder and roll until all of the ceramic powder looks thoroughly wet. Roll the mill for several hours (approximately 16 hours in these lab tests). The content in the mill jar is called the mill base. A small amount of binder, which is accounted for in the letdown stage, is added to the mill base for stability.
After grinding is complete, add the binder solution with its known CPVC value (determined by the binder dilution procedure, minus the small amount of binder added to the mill base as mentioned in the section on "grinding"), plasticizer and extra solvent to adjust for the desired viscosity. Roll the mill for two hours. At the end of milling, pull the samples for percent solids, viscosity and density check.
Results and DiscussionFigure 1 shows a minimum PVC value of 60%, which is the CPVC value determined in the lab tests. In a large-scale manufacturing run, 59.2% was the actual CPVC value used. This shows an accuracy or deviation of 0.8%. At PVC = 85%, massive pinholes can be observed on the layer of the paint film drawn down on the glass plate. At PVC = 75%, Benard cells mixed with some pinholes are formed on the surface of the drawdown. From PVC = 65% to PVC = 55%, we find very smooth paint surfaces without any pinholes or defects. At PVC = 60%, the glossmeter reaches its lowest reading. For PVC <55%, Benard cells and holes larger than pinholes are shown on the glass plates.
Since the slurry rheology is the result of a complex relationship in the electrochemistry of the paint system, zeta potential values of the paint system are measured. Figure 2 shows a minimum zeta potential at PVC = 60.8%, consistent with good particle dispersion and stability of the paint system.
Particle size distribution determines packing efficiency, the interstitial pore volume and the surface area of the ceramic powders. Dinger and Funk have presented a method that combines particle-size distribution and specific surface area to allow the prediction of the rheological performance of a slip or slurry, or the minimum deflocculated viscosity.2 Bierwagen introduced a model for CPVC calculation using particle size distribution data.3 In the lab tests described in this article, particle size distributions of ceramic powders of each batch and standard deviations of the particle size distributions-as well as surface areas, mean diameters and standard deviations of mean diameters-were measured using a particle size analyzer. This serves not only as a quality control tool for accepting and rejecting the incoming ceramic powders, but also for determining packing efficiency and predicting the minimum deflocculated viscosity of the paint system.
The effect of grinding on particle size can readily be determined. By comparing the particle size distribution and mean diameter of the ceramic powder before and after grinding (2 to 16 hours), the optimum grinding time can be determined. If the particle size distribution and mean diameter of the ceramic remain the same after five hours of grinding, there is no reason to grind six hours before letdown. Although ball milling was used in this study, a high-speed disperser would also be an appropriate and faster mode of mixing.