Comparative study of porosity in 3Y-TZP superplastic ceramics by USANS and SEM image analysis

V. Ryukhtin^1,2, J. Šaroun¹, S. Harjo³, Y. Motohashi³, M. Baron^4,5, R. Loidl^4,5

1- Nuclear Physics Institute, 25068 Řež near Prague, Czech Republic.

2- Charles University, Faculty of Mathematics and Physics, Ke Karlovu 3, 121 16 Praha 2, Czech Republic.

3- Ibaraki University, Faculty of Engineering, The Research Center for Superplasticity, Hitachi, Ibaraki, 316-8511 Japan.

4- Institute Laue-Langevin, BP 156, F-38042 Grenoble Cedex 9, France.

5- Atominstitut der Österreichischen Universitäten, A-1020 Wien, Austria.

Evolution of cavities in superplastic ceramics is a significant indicator of mechanism of the superplastic deformation process. Moreover the cavities have large influences on mechanical and functional properties [1]. The cavities and pores in fine-grained 3Y-TZP have sizes typically between 0.1 mm and 1 mm and can therefore be studied by both scanning electronic microscopy (SEM) image analysis and ultra small-angle neutron scattering (USANS). However, quantitative comparison of results obtained by the two techniques is complicated by surface artefacts, uncertainty in pore identification, assumptions of the model used to fit the USANS data and other factors. We show that scattering functions measured at double-crystal USANS instruments can be calculated directly from sufficiently large SEM images of the sample sections parallel to the scattering plane. The results of both techniques can therefore be compared directly. The porosity in 3Y-TZP ceramics samples deformed from 0% to 200% were measured in broad range of Q (2·10^-5 – 8·10^-3 Å^-1) using Bonse-Hart SANS (S-18 at ILL, Grenoble) and double crystal (DC) SANS (DN-2 at NPI, Řež near Prague) instruments. The measured scattering curves were compared with the scattering curves, calculated directly from the SEM images. Remarkable agreement of both methods was observed (Fig. 1), except of small Q-range, where the contribution of very large pores prevails. For such big pores, numerical and statistical errors of calculation of the scattering function from SEM images become large [2].

USANS experiments enabled to measure alternatively the bulk porosity with the results consistent to those of other techniques (SEM, Archimedes) (Fig. 2). The presented correlation between measured data and calculated from SEM images indicates, that the SEM images represent well the bulk microstructure of the material and the pore identification procedure doesn't introduce significant bias to the SEM results for pores with radii <1 mm. However, at strains near 200%, large cavities (R~5mm) have been observed on some SEM images. Comparison with SANS indicates, that these cavities are either surface artefacts or are not distributed in the whole bulk investigated by SANS.

Fig. 1. Comparison of measured data with scattering function calculated from SEM images.

Fig. 2. Comparison between the volume fraction of cavities measured by SANS, those by image processing and those by Archimedes method.

References

[1] Y. Motohashi, N. Sugeno, S. Koyama and T. Sakuma, Mater. Sci. Forum, 243-245, 1997, pp. 399-404.

[2] V. Ryukhtin, J. Šaroun. Direct comparison of SANS data with SEM image analysis. Proc. ICNS-2001. J. Appl. Phys. A. (in print).