JOURNAL

OF MATERIALS

SCIENCE

LETTERS

1 (1982)

288--291

Transmission electron microscopy of ceria gels J. A. DESPORT, P. T. M O S E L E Y , D. E. W I L L I A M S

Materials Development Division, AERE Harwell, Didcot, Oxen, UK

Refractory ceramic coatings derived from aqueous oxide sols by painting them onto a surface, allowflag them to dry to a gel and then firing them are of general interest for the control of corrosion and oxidation of metals. Major advantages are low cost and ease of application, particularly over large areas [1, 2]. Such coatings are being examined for their efficacy in controlling the oxidation of stainless steels and in controlling or preventing oxide spallation [3]. In order to optimize coating quality it is useful to know the microstructure and composition of ceramic coatings obtained by calcining gels applied to the surface of stainless steels. A simple technique for obtaining this information has been devised and is presented in this paper. Thin films of ceria gels were obtained by dipping into a ceria sol substrates of stainless steel (20Cr25NilNb) through which variously shaped holes had been etched (Fig. 1). The substrate (3ram in diameter) was masked using photolithographic techniques and the holes etched out with ferric chloride solution. The different shapes and sizes of holes were chosen in an attempt to display how well the coating covered gaps and to display shrinkage effects during drying and firing. As the sol drained and dried to a gel, thin Films were formed which partially covered the holes and which were transparent to 100keV electrons at the extremes. The air-dried gels were examined before heat treatment and after firing for 1 h at various temperatures (up to 1200 ° C), raising the temperature slowly from room temperature. Platinum substrates with Free drilled holes were employed for material calcined at 1000 ° C or over. Transmission electron micrographs were recorded using a Philips EM400 instrument. Elemental composition was established by analysis of the characteristic X-rays emitted under electron irradiation. Estimates of average particle size were obtained by

measurement of X-ray powder diffraction patterns recorded on a diffractometer. The cerium oxide sol was prepared by peptization in nitric acid [1]. Examination of the coated grids in an optical microscope were allowed empirical establishment of the sol composition necessary to obtain electron transparent films. Typically, dilution with distilled water to an oxide concentration of 50g1-1 was sufficient. In order to observe the substrate-coating interface, coated grids were placed in an ion-beam mill, and the edges of the holes etched to a 100 keV-electron-transparent wedge. Comparison was made with tray-dried and fared gel flakes, ion beam thinned to electron transparency, and with a gel coating fired onto Fecralloy* steel, the steel substrate subsequently being dissolved away in dilute HC1 to leave an oxide wafer transparent to 1 MeV electrons [4]. Fig. 2 shows a general view of a substrate coated with an unfired gel film. Fig. 3 shows an

Figure 1 Etch pattern through metal substrate.

*Fecralloy is a registered trademark of the UK Atomic Energy Authority. Its nominal composition is Fel6Cr5A10. 36YO.23Si. 288

0261--8028/82/070288--04502.98]0

© 1982 Chapman and Hall Ltd.

Figure 2 General view of substrate coated with gel film.

electron-transparent area at the edge of one of the teeth-like regions that protruded from the edges of the holes in the grid. The film had curled over at its extreme edge and comprised extremely small particles around 2 to 5 nm diameter (comparable with the crystallite size indicated by X-ray line broadening). There was a certain amount of porosity visible on the very fine scale. At a lower magnification, some aggregation was apparent and furthermore as the liquid film drained off, a

preferential segregation of unaggregated material to the edges of the film occurred. Fig. 4 shows a whisker spanning a hole, and illustrates in this case aggregation, non-uniform distribution of the aggregates and fluctuations in the film thickness presumably caused by the flow of the liquid layer as it dried and became progressively more viscous. One might hypothesize that the aggregation occurred during the drying of the Film. As well as the sort of aggregation illustrated in

Figure 3 Electron-transparent extremes of the tooth-like proj~eetions. 289

Figure 4 Whisker of unfired gel showing aggregation of gel particles. The dark patches are assumed to comprise aggregates of finer particles which make up the semi-transparent areas and which are revealed in Fig. 2.

Fig. 4, areas where the crystallite size was much greater than that shown in Fig. 3 were observed. Areas like this were quite rare - there was certainly too little of the coarser crystalline material to show up in X-ray line broadening measurements. The larger grains could have been formed by a crystallization of polynuclear oxocerium species from the solution during drying of the fdm. Electron diffraction patterns of both the very fine material and the coarser material indexed as CeO~. A substantial level of impurities derived from the substrate (Fe, Cr and Ni) was observed within the ceramic film; observation of the drying t'rims using an optical microscope revealed gas evolution in some cases as the film dried and it is presumed that, since the sol itself was fairly acidic (pH ~ 2), the impurities were derived from acid attack on the substrate. It appeared that Cr was relatively enhanced and Ni relatively depleted in the ceramic film in comparison with their level in the metal. Firing the gel to 400 ° C had no effect upon the crystallite size; /his remained in general around 2.5 nrn diameter. Firing at progressively higher 290

temperatures caused the crystallites to grow. Grain size was around 30nm after firing at 760°C (Fig. 5), 60 to 200nm after firing at 1000 ° C and > 5 0 0 n m after faring at 1200 ° C. These results were in broad agreement with those of X-ray line broadening measurements. The thin electrontransparent sheets observed with the unfired gel and with the material fired at up to 400 ° C had shrunk into whiskers after firing at 800 ° C and disappeared entirely after Faring at 1200 ° C. Some of the grains after faring at intermediate temperature s (Fig. 5) showed internal striations reminiscent of the contrast produced by parallel twins in small growing particles of cobalt metal [7]. This contrast was not apparent in grains of materials fired at higher temperatures. The results described above pertain to essentiaUy free Films of ceria gel, dried and fired out of contact with the substrate. Qualitatively similar results, although micrographs were not nearly so clear, were obtained from tray-dried and Ftred ceria gel flakes, ion-beam thinned to electron transparency. It was of interest to consider whether interaction with a substrate might modify the

Figure 5 Oxide derived from gel after firing at 760° C.

Figure 6 Substrate-coating interface on a Pt grid revealed at the edge of an ion-beam thinned wedge; oxide fired at 1200 ° C.

Acknowledgements grain size in a gel coating fired onto it. The results obtained so far suggest that any effect might vary with the nature o f the substrate. Thus Fig. 6 shows the substrate-coating interface on a platinum grid, revealed at the edge o f an ion-beam milled wedge; the grain size revealed is little different to that ~ o w n in the free films. In contrast, the gel coating fired onto Fecralloy ® steel (at 850 ° C for 10min) showed a smaller grain size, around 10 nm. These preliminary results would therefore seem to show that with some substrates, grain growth in the ceria coating is constrained, whereas with other substrates it is not. A simple technique for the transmission electron microscopic study o f the microstructure and composition of ceramic coating materials has been demonstrated and used to study the drying and sintering characteristics of ceria sols. A variety of information concerning the composition, drying and sintering of thin films o f such materials, not otherwise easily obtainable, can thereby be gained.

The ceria sols and X-ray line broadening results were provided by J. L. Woodhead, with whom we have had many stimulating discussions. We would also like to thank Mr K. T. Harrison, Drs P. McGeehin and B. C. Tofield for discussing various aspects of this work with us, and Mr G. Bishop for photolithography.

References 1. 2. 3. 4. 5.

R.A. NELSON, J.D.F. RAMSAY, J.L. WOODHEAD, J.A. CAIRNS and J.A.A. CROSSLEY, Thin Solid Films 81 (1981) 329. United Kingdom Patent 2,023,453a. M.J. BENNETT, M. R. HOULTON and J. B. PRICE, UK Atomic Energy Research Establishment Report R10169 (UK Atomic Energy Authority, 1981). P.T. MOSELEY, J. S. SEARS and G. TAPPIN, Thin Solid Films 78 (1981) 349. A.R. THOLEN, in "Sintering Processes", edited by G. C. Kuczynski, Material Science Research, Vol. 13 (Plenum Press, New York, 1979) p. 539.

Received 10 February and accepted 5 March 1982

291