CERAMIC ENERGY: Thin Electrolyte Tapes for SOFCs

Very thin electrolyte tapes are being consistently manufactured for use in high-performance, low-temperature SOFCs.

Materials for solid oxide fuel cells (SOFCs) are being developed in laboratories all over the world. As new formulations are evaluated, it is important to know that high-quality materials can be reproduced. In a manufacturing environment where ISO 9001 certification is in place, certain procedures for formulas and processes must be followed. This starts with the initial development of any new product or process, and includes careful tracking and documentation of any changes that are made as the development progresses.

The process begins with characterization of the raw materials. Intermediates and final product properties are monitored as part of the process, and control charts are maintained for each measurement. Measurement techniques are documented, and gauge repeatability and reproducibility (gauge R&R or GRR) studies are performed to determine the within-system and between-system measurement variation for a given measurement technique. Gauge R&R studies enable measurement variation to be minimized and ensure that measurements are meaningful.

The tapes used for SOFCs can be manufactured economically under these conditions. Since formulations and processes are reproducible, waste is correspondingly minimized. Scale-up issues are also minimized,¹ and economies of scale can apply to larger lots of materials required for prototyping cells and stacks.

The tapes used for the electrolyte function present an unusual challenge. Cell efficiency depends on the properties of this tape, and even a small tape defect can harm or destroy the overall cell. The internal resistance of a cell determines the ionic conductance and, consequently, the efficiency of electron production.² Since resistance is a product of ionic resistivity and thickness, these properties are the focus of many development programs.³ A variety of chemical formulations have been evaluated in efforts to increase conductivity at lower stack temperatures, which would lower the cost of materials. Although a number of different electrolyte compositions have been cast at ESL, the focus of the company's work is on tape thickness. The tapes described in this article are cast in the range of 10 to 20 μm green (unfired), and fired electrolyte thicknesses as low as 6 to 8 μm have been attained. More importantly, tapes in this thickness range have been routinely manufactured, and ISO 9001 manufacturing documentation has been established to ensure control consistency from lot to lot and with scale-up.

The Need for Thin Tape

In a SOFC, hydrogen and carbon monoxide are reduced at the anode (H₂

Making Thin Tape Reliably

Methods of casting ceramic tape depend on whether the tape is to be thick, intermediate or thin-generally defined as <50 μm, 50 to 1000 μm and >1000 μm, respectively.⁵ Each method has similarities with the others but also involves specific techniques, formulations and equipment. Tapes in the middle range include multi-layer ceramic packages, electrode-supported SOFCs and low-temperature co-fired ceramic (LTCC), while thick tapes are more commonly used for armor, substrates or other structural applications.

The manufacture of thin tapes (~<50 μm) has, until recent years, been the realm of passive components alone-multi-layer ceramic capacitors (MLCC), multi-layer ceramic inductors (MLCI), etc. Producing such thin tapes entails manufacturing and handling challenges, and casting variables such as dispersion and thickness variation become magnified in importance.

Thin tapes require greater dispersion of inorganic particles. While this is true of most wet ceramic processes, the need becomes more critical as tape thickness decreases in order to minimize structural defects stemming from even low levels of agglomeration. Achieving excellent dispersion becomes more difficult due to the need for extremely fine particles, with a correspondingly high surface area.⁶ Submicron or nanoscale powders are known for their tendency to agglomerate. However, the particle size must decrease to ensure a packed particle bed, which is necessary to increase structure uniformity in both green and fired states, minimize or negate through-layer defects, homogenize densification shrinkage and increase the strength of the layer after sintering.⁷

Thin tape requires enough green strength per unit of cross-sectional area to maintain integrity through device assembly. Building this strength into a high-surface-area particle bed requires a delicate balance between dispersion, strength, green density and viscosity. Controls must be established and maintained for consistency.

Many equipment-related demands also accompany the manufacture of micron-scale tapes. Simple tasks such as tape thickness measurement as an in-process control check must be performed on more sensitive and accurate equipment than is typically found in a production area. For green tape thickness in the range of 12.5 μm, production floor measuring equipment must have accuracies in the single micron range. Commercially available precision-milled or surface-ground casting heads (doctor blades) often have flatness tolerances no better than 2.5 µm and up to 5 µm after use. Solid granite precision casting surfaces may only be certified flat to 8 µm. Polymer carrier films claimed to have been processed in clean rooms have been seen to include static-related debris larger than 8 μm. These mechanical factors-along with additional factors such as web speed uniformity, head pressure repeatability, slip homogeneity, lot-to-lot repeatability of casting slips and in-depth operator training-become key process variables as the layer thickness decreases to <50 μm.⁸

Procedure and Results

The procedures and controls described in the previous section were put to use to cast very thin (10-20 μm) electrolyte tape. YSZ powders were obtained from several sources to determine the parameters required to cast consistent tapes. Incoming powders were characterized by measuring surface area, particle size distribution and other parameters. Powder lot records were maintained so that cast and fired tape properties could be traced back to the original lots of powders and their properties. The submicron YSZ powders used had surface areas in the range of 8-10 m²/gm.

Slurries were formulated using phosphate-free dispersants and solvents that could be used safely in manufacturing. This precluded the use of toluene. Polyvinyl butyral (PVB) binder systems were chosen because they provide good strength for handling thin green tapes, and the formulations included enough binder to ensure adequate laminating characteristics. The YSZ powders were dispersed for 24-48 hours before binder was added, followed by dissolution milling for an additional 24-48 hours. The viscosity of the dispersed powder slurry was monitored to ensure that agglomerates of the high surface area powders were adequately broken up. After complete dissolution of the binder, viscosity was adjusted to a range that is consistent with practical casting parameters.

The casting slurry was then de-aired to remove dissolved bubbles and was finally filtered before being fed into the doctor blade reservoir. An initial set-up of the doctor blade height was adjusted after the dry green tape came off the caster and was measured to ensure that the thickness was within the control range. The leading part of the cast was discarded after this adjustment was made. The drying parameters and casting rate were optimized for the slurry composition to maximize throughput. The tape was periodically monitored for thickness along the length and across the width of the cast.

Figure 1 shows thickness data taken on a 250-m (~800 ft) cast of electrolyte tape. Each point represents the average of four points taken across the 8-in. width of the cast. This lot (Number 8 in Figure 2) is typical of the data obtained on all the thin tape casts shown in Figures 2 and 3. The dried tape was then taken onto rolls at the end of the caster. Each cast was re-rolled on a light table for visual inspection and additional thickness measurements.

Figures 2 and 3 show the results obtained for two thicknesses of the green tape. Figure 2 shows data obtained from nine separate lots of tape cast to a target green thickness of 14.5 μm. The target thickness for the five lots of tape shown in Figure 3 was 12 µm.

Figure 4 shows green density data obtained for the five lots of electrolyte tape shown in Figure 3. Green density is an indication of solids loading and powder dispersion in the casting slurry. High green density led to lower shrinkage and a more consistent fired tape. All measurements were performed using procedures that had been certified to pass gauge R&R. This ensured acceptable measurement variation among operators.

These electrolyte tapes attained >97% of theoretical density when fired at 1450øC for two hours. For anode-supported cells, the electrolyte tape was typically laminated to the anode support structure and co-fired with the anode tape. SOFC developers reported that the resulting fired electrolyte tape was free of leaks when subjected to helium leak detection.

Consistent Tapes

As SOFC developers work to improve the performance of their cell designs, the need for consistent materials on a large scale becomes increasingly important. Fortunately, the ability to consistently cast very thin electrolyte tape in a documented prototype and production environment already exists. While the data presented in this article is for YSZ tapes, other electrolyte materials have been cast with similar controls and specifications. ISO 9001 compliance ensures consistency from the incoming inspection of raw materials to the final tape and cell properties.

For more information about ceramic tapes, contact ESL Electro-Science, 416 East Church Rd., King of Prussia, PA 19406-2625; (610) 272-8000; fax (610) 272-6759; e-mail technicalservice@electroscience.com; or visit http://www.electroscience.com.

References

1. Webb, R.S.; Feingold, A.H.; Topka, Z.J.; and Twiname, E., Fuel Cell Science, Engineering and Technology-2004, Rochester, N.Y., June 14-16, 2004, pp. 477-481.

2. Ishihara, T.I.; Sammes, N.M.; and Yamamoto, O., High Temperature Solid Oxide Fuel Cells, S.C. Singhal and K. Kendall, Eds., Elsevier, 2003, p. 103.

3. Ibid., p. 96.

4. Singhal, S.C., "Solid Oxide Fuel Cells VII," H. Yokokawa and S.C. Singhal, Eds, The Electrochemical Society Proceedings, Pennington, N.J., PV2001-16, 2001, p. 166.

5. Mistler, R.E. and Twiname, E., Tape Casting Theory and Practice, American Ceramic Society, Ohio, 2000, p. 96.

6. Ibid., p. 13.

7. Ibid., p. 214.

8. Ibid., pp. 83-151.