surface, interface and nano-structures 490 # 2001 International Union of Crystallography  Printed in Great Britain ± all rights reserved J. Synchrotron Rad. (2001). 8, 490±492 XAFS analysis of corroded metal surfaces with molten salts by conversion-electron- yield method Etsuya Yanase, a* Kazuo Arai, a Iwao Watanabe, b Masao Takahashi c and Yoshinori Dake d a The New Industry Research Organization, 1-5-2, Minatojima-minamimachi, Chuo-ku, Kobe, 650-0047, JAPAN, b Department of Chemistry, Graduate School of Science, Osaka University, 1-1 Machikaneyama-cho, Toyonaka, Osaka, 560-0043, JAPAN, c Institute of Scientific and Industrial Research, Osaka University, 8-1 Mihogaoka, Ibaraki, Osaka, 567-0047, JAPAN, d Kanto Technical Institute, Kawasaki Heavy Industries, Ltd, 118 Futatsuzuka, Noda, Chiba, 278-8585, JAPAN. Email:yanase@ri.niro.or.jp We have measured XAFS spectra of metal surfaces corroded with melting salt (NaCl, KCl, and Na 2 SO 4 ). Steel samples used were S45C, SCM435, SUS310S, and SUS304. We measured the Fe K- edge XAFS spectra for all samples and the Ni K-edge for SUS310S and SUS304 samples before and after the corrosion. The XANES spectra of samples before the corrosion show metallic structure because surface oxide thickness is thinner than probing depth with a conversion yield XAFS method. Each result of these XAFS spectra gives good agreements with the FEFF calculation in the assumption of bcc and/or fcc structure. The Fe K-edge spectra of steel samples except SUS310S after corroded treatment show existence bonding between Fe and another light element although the spectra of SUS310S samples before and after corroded treatment are much the same. Keywords: metal, molten salts, corrosion. Introduction It is recently expected to operate at a higher temperature for suppression of dioxins in operating refuse incinerating power plants. However, the accelerated corrosion of structural steel is a major problem in the development of the plants. This is because passivation oxides dissolve in melting eutectic mixture salt, which is included in fly ash. In order to develop new material for such plants, it is essential to collect basic data of corroded steel surface. Despite the large number of investigations (Shardakov et. al, 1999; Okuyama et. al, 1988), there has not been a systematic microscopic study of surface structure corroded with molten salt. Therefore, we study steel surface structure corroded with molten salt by using XAFS technique. XAFS clarifies not only chemical states but also local structures of corroded surface. In our study, we used conversion electron yield method as XAFS measurement technique. This method is especially valuable for analyses of chemical system where conventional XAFS methods are not applicable. We have measured XAFS spectra of steel surfaces corroded with NaCl, KCl and Na 2 SO 4 mixture that is adopted as a model of salt in fly ash. We report and discuss the results of XAFS measurements and the corrosion mechanism with molten salt. Experimental The chemical composition of the steel samples investigated in this study is given in Table 1. All steel sample surfaces were polishing with emery paper. To prepare the corroded samples for our measurements, we immersed the specimens in melting salt for a few seconds at a temperature 973K under an atmosphere. Equal molarity of NaCl, KCl and Na 2 SO 4 powder were uniformly mixed to blended salt and melted in alumina crucibles in an electric furnace at 973K. Blended salt on the specimen was removed with filter paper for XAFS measurements after it got cool. A conversion electron yield method was used for XAFS measurements. When a sample placed in atmospheric He gas is irradiated with X-ray photons, Auger electrons are ejected from the sample surface in proportion to the photon absorption, and there ionize He atoms. The electrons or He ions current so generated can be used for XAFS measurement. The probing depth of this method depends on Auger electron mean free path. We can estimate the probing depth at about a few 10nm. The details of our conversion electron method has been described elsewhere (Yanase et. al, 1999). All XAFS measurements were performed at the beamline BL01B1 SPring-8 in Hyogo Prefecture, Japan. The X-ray beam was monochromatized by using Si(111) double crystal monochromators. The higher harmonics were rejected by using a rhodium coated mirror system. The sample size of 10mm x 10mm x 1mm was placed on the sample holder with conductive carbon tape. The X-ray incidence angle to the sample was fixed at 25 degree. The beam intensity was monitored with a nitrogen gas ionization chamber of 5.5 cm in path length. We measured the Fe K-edge XAFS spectra for all samples and the Ni K-edge for SUS310S and SUS304 samples before and after the corrosion treatment. The EXAFS analysis procedure used in this work has been described elsewhere (Sakane et. al, 1993). Results and Discussion S45C and SCM435, which are ferrite steels, have bcc structure of Fe atoms. On the other hand, SUS304 and SUS310S have fcc structure because these steels are austenite stainless steel. Therefore, we calculated k3-weighted EXAFS spectra, k 3 (k), by using FEFF7.02 code in the assumption of bcc or fcc structure. The k 3 (k) spectra obtained by our calculated and experimental results at Fe K-edge are shown in Fig. 1. The k 3 (k) spectrum of SUS310S gives good agreement with the FEFF calculation. The spectrum at Ni K-edge as shown in Figure 2 also gives good agreement. This result shows that the thickness of oxide layer on the surface is sufficiently thinner than probing depth with a conversion yield XAFS method. The result of SUS304 is different although it is the same austenite stainless. Table 1 Chemical composition of steel used (wt.%) Steel Composition SUS304 SUS310S SCM435 S45C Fe Bal. Bal. Bal. Bal. C M 0.08 M 0.08 0.42-0.48 0.33-0.38 Si M 1.00 M 1.50 0.15-0.35 015-0.35 Mn M 2.00 M 2.00 0.60-0.90 0.60-0.85 P M 0.045 M 0.045 M 0.030 M 0.030 S M 0.030 M 0.030 M 0.035 M 0.030 Ni 8.00-10.50 19.00-22.00 M 0.20 M 0.25 Cr 18.00-20.00 24.00-26.00 M 0.20 0.90-1.20 Cu M 0.30 Cu M 0.30 Mo 0.15-0.30