Transmutation Sciences Materials (TRP) Transmutation Research Program Projects

7-2004

Spectroscopic and Microscopic Investigation of theCorrosion of 316/316L Stainless Steel by Lead-Bismuth Eutectic (LBE) at Elevated Temperatures:Importance of Surface PreparationAllen L. JohnsonUniversity of Nevada, Las Vegas, aljohnson@ccmail.nevada.edu

Denise ParsonsUniversity of Nevada, Las Vegas

Julia ManzerovaUniversity of Nevada, Las Vegas

Dale L. PerryLawrence Berkeley National Laboratory

Daniel KouryUniversity of Nevada, Las Vegas, dan@physics.unlv.edu

See next page for additional authorsFollow this and additional works at: http://digitalscholarship.unlv.edu/hrc_trp_sciences_materials

Part of the Biological and Chemical Physics Commons, Chemistry Commons, MetallurgyCommons, and the Oil, Gas, and Energy Commons

This Postprint is brought to you for free and open access by the Transmutation Research Program Projects at Digital Scholarship@UNLV. It has beenaccepted for inclusion in Transmutation Sciences Materials (TRP) by an authorized administrator of Digital Scholarship@UNLV. For moreinformation, please contact digitalscholarship@unlv.edu.

Repository CitationJohnson, A. L., Parsons, D., Manzerova, J., Perry, D. L., Koury, D., Hosterman, B. D., Farley, J. (2004). Spectroscopic and MicroscopicInvestigation of the Corrosion of 316/316L Stainless Steel by Lead-Bismuth Eutectic (LBE) at Elevated Temperatures: Importance ofSurface Preparation. Journal of Nuclear Materials, 328(2-3), 88-96.Available at: http://digitalscholarship.unlv.edu/hrc_trp_sciences_materials/160

AuthorsAllen L. Johnson, Denise Parsons, Julia Manzerova, Dale L. Perry, Daniel Koury, Brian D. Hosterman, andJohn Farley

This postprint is available at Digital Scholarship@UNLV: http://digitalscholarship.unlv.edu/hrc_trp_sciences_materials/160

1

Spectroscopic and Microscopic Investigation of the

Corrosion of 316/316L Stainless Steel by Lead-Bismuth Eutectic (LBE) at Elevated Temperatures: Importance of Surface Preparation

Allen L. Johnson,* Denise Parsons, Julia Manzerova

Department of Chemistry University of Nevada, Las Vegas

4505 S. Maryland Parkway, Box 4003 Las Vegas NV 89154-4003 USA

Dale L. Perry

Mail Stop 70A-1150 Lawrence Berkeley National Laboratory

Berkeley CA 94720 USA

Dan Koury, Brian Hosterman, John W. Farley Department of Physics

University of Nevada, Las Vegas 4505 S. Maryland Parkway, Box 4002

Las Vegas NV 89154-4002 USA

*Corresponding author: (Tel.: 702-895-0881; fax: 702-895-4072. E-mail address: aljohnson@ccmail.nevada.edu

Key Words: lead, bismuth, lead-bismuth eutectic, LBE, 316/316L stainless steel, transmuter, transmutation, coolant, spallation, corrosion, X-ray photoelectron spectroscopy, XPS, scanning electron microscopy, SEM, energy-dispersive X-ray analysis, EDAX

2

Abstract

The corrosion of steel by lead-bismuth eutectic (LBE) is an important issue in proposed nuclear

transmutation schemes. Steel samples were exposed to oxygen-controlled LBE at high

temperature and exposure times that simulate actual reactor systems. Scanning electron

microscopy (SEM), combined with energy-dispersive X-ray analysis (EDAX) and X-ray

photoelectron spectroscopy (XPS), has been used to study post-exposure steel samples and

unexposed controls. Our investigation of the reacted samples revealed preferential segregation of

metals contained in the initial alloy formulations. Sputter depth profiling of the exposed

annealed sample and cold-rolled sample showed a marked difference in oxide layer composition,

depending on surface preparation. The annealed sample showed a complex oxide structure (iron

oxide over chromium/iron oxide mixtures) of tens of microns thickness. The cold-rolled sample

was covered with a rather simple, primarily chromium oxide layer of ~1 micron thickness. The

cold-rolled sample had an order of magnitude less corrosion (i.e., both lower oxidation and less

weight change) than the annealed sample.

3

1. Introduction

Lead Bismuth Eutectic (LBE) is of interest as both a coolant and a spallation target in

proposed transmutation schemes for radioactive waste. The Russians have decades of experience

with LBE as a coolant in their Alpha-class submarines, generating a body of engineering

knowledge on the material. Unfortunately, hot LBE corrodes most engineering materials.

Several studies addressing the corrosion issue in these reactor operating systems have

been published in the research literature, addressing the corrosion of the steel in the presence of

either lead or LBE. Early work focused on ferritic steels, guided by the high solubility of

chromium in LBE. Barbier and Rusanov [1] reported the corrosion behavior of several steels,

including Optifer IV, T91, and EP823 with LBE. Ghetta et al. [2] designed a detailed and precise

description for procedures for corrosion studies of steel samples in molten lead. Glasbrenner et

al. [3] conducted corrosion studies of steels in flowing lead at temperatures of 673 K and 823 K.

Mueller et al. [4] tested corrosion of various steels after 2000 h of exposure to flowing LBE at

693 K and 873K. Nicaise et al. [5] studied the embrittlement of Grade 91 martensitic steel in

liquid lead. Park et al. [6] suggested various other alloys and combinations.

In their Alpha-class nuclear submarine power plants, the Russians found that the addition

of controlled amounts of oxygen to LBE greatly suppressed the corrosion of steel by LBE [10].

The suppression of corrosion is believed to arise from the formation of a protective layer of iron

oxide [11, 12]. The corrosion of steel by oxygen controlled LBE is a complex phenomena, with

both oxidation and dissolution of the steel occurring at different rates at different times. The

growth of the oxide layer follows a parabolic law in time, whereas the dissolution rate is more

constant. The use of controlled amounts of oxygen in the LBE to form protective oxide layers

led to the renewed interest in the use of austenitic (high-chromium) steels for this purpose.

4

Several investigators [2, 10-14] report detailed studies of the effects of various levels of oxygen

on the corrosion processes involving steels with liquid lead and/or LBE. Chemical compatibility

tests were conducted of martensitic and austenitic steels [8, 12, 14-18], and mechanical tests [19]

were conducted in both liquid lead and LBE.

2. Varieties of Steel

One of the most common varieties of alloy steels is stainless steel, which has enhanced

corrosion resistance due to the formation of a protective passivation layer. Steel is classified as

stainless if the chromium content is at least 11.5 % by weight. The main components of stainless

steel are iron, chromium, and nickel [7]. Stainless steels are classified as ferritic, martensitic, or

austenitic, based on the amounts of alloying components and their lattice structures. Austenitic

alloys have chromium content of 16-18 % by weight. Of the 300 series, 316 is the alloy most

resistant to corrosion attack. 316 stainless steel and its variants (316L, 316LN, etc.) are

commonly used in highly corrosive environments and in nuclear construction. Austenitic steels

have higher oxidation resistance in LBE due to their relatively high chromium content [8].

The purpose of the present study was to obtain data that would be helpful in detailing a

mechanism for oxygen-controlled LBE/stainless steel corrosion. Scanning electron microscopy

(SEM), combined with energy-dispersive X-ray analysis (EDAX) and X-ray photoelectron

spectrometry (XPS), has been used to study the products formed in the reaction between oxygen-

controlled lead-bismuth eutectic (LBE) and 316/316L nuclear-grade stainless steel at elevated

temperatures. The steel samples were reacted with LBE under temperature and exposure times

that simulate actual reactor systems and were spectroscopically and microscopically compared to

unreacted blank samples. XPS spectral lines for the various elements were studied for both the

5

reacted and un-reacted samples, and the elemental compositions were extracted for reacted and

un-reacted samples. The data are discussed in the context of known corrosion mechanisms for

the alloys and the results of previous studies of reactions of the alloys with LBE at elevated

temperature.

3. Experimental

The 316/316L nuclear grade stainless steel samples studied were among a batch of steels

corrosion-tested by scientists at the Institute of Physics and Power Engineering (IPPE) in

Obninsk, Russia, under contract to Los Alamos National Laboratory (LANL). The samples were

inserted in IPPE's CU-1M non-isothermal LBE loop for time intervals of 1000, 2000, and 3000

hrs at temperatures of 733 K and 823 K. The oxygen level in the LBE was maintained at 30-50

ppb. The samples were an 8 mm diameter tube (6 mm ID) of annealed material and 8 mm

diameter rod of cold-rolled material. The samples were cut to convenient sizes before analysis

using abrasive cutoff wheels and cleaned in methanol.

XPS has been reviewed by Perry [20], while SEM has been reviewed by Goldstein [21].

Our XPS instrument is a Surface Science Instruments SSX-100 which has been fitted with a

Nonsequitur Technologies Model 1401 ion gun for sputtering. The XPS binding analyses were

referenced to the adventitious carbon 1s line at 284.6 eV. The X-ray source for the XPS was a

monochromatized and refocused aluminum Kα (1486 eV) source with a 0.5-1.0 mm diameter X-

ray spot size. The SEM was a JEOL JSM-5600 with an Oxford EDS detector.

4. Theory: basic mechanism of corrosion

One of the most predictive of the current theories [11] follows from the observation that

for a (metal) substrate to dissolve into an overlying liquid metal, it must be in the reduced (i.e.,

metallic) state. Thus, for a substrate oxide to dissolve, it must first be reduced to the metal and

6

dissolved oxygen. Using the thermodynamic stability of the oxides leads to a new equilibrium

concentration of the (metal) substrate in the liquid LBE, which can be lowered by raising the

dissolved oxygen concentration. With the addition of the diffusion equation and the effective

thickness of the stagnant layer between the substrate wall and the fully mixed liquid, the

corrosion rate can be found [11]. Thus, oxygen control in LBE forces a lower corrosion rate by

the formation of a more thermodynamically stable passivating oxide layer.

The gas phase oxidation behavior of stainless steels, including 316 family steels, has been

extensively reported in the literature [7]. In oxidation in air at elevated temperatures (0.l3 atm.

O2, T=1273 K) the oxide layer depends on the chromium content. At a chromium content of less

than 2 %, the oxide layer is pure iron oxide, while at chromium contents near 9 %, an iron oxide

layer is found above mixed iron and chromium oxides. At chromium contents of 16 %, pitting

corrosion is found with iron oxides over mixed iron/chromium oxide. At high (e.g., 28 %)

chromium content, a pure chromium oxide scale is found. The question naturally arises whether

the insights resulting from gas phase oxidation of steels can be applied to the oxidation of steel in

non-isothermal LBE loops. While the surface is definitely being oxidized, similar to the gas

phase case, in liquid metal the oxide is also being dissolved. Thus the results of this and similar

studies are necessary for full understanding of the O/LBE/stainless steel system.

Many of the studies (e.g., 4) of stainless steel in LBE found that the oxide at the surface

of the steel was composed of iron oxide, with an iron/chromium oxide layer underneath. Li [11]

uses the commonly observed iron oxide surface layer to apply the theory to practical systems.

7

5. Results

The Russian collaborators measured weight change and oxide thickness of these samples

after LBE exposure. They found that all of the steels had similar properties, except for the cold-

rolled sample, which had an order of magnitude less oxidation and weight loss than the other

steels (in particular, annealed steel). Our initial investigations focus then on the differences

between the annealed and cold-rolled samples.

5.1 SEM studies

Our SEM investigations of the annealed and the cold-rolled samples showed that the

316L sample had an evident ~10 micron grain structure (Figure 1). The sample was not etched to

enhance the grain structure. Annealed 316 generally shows an ~40 micron grain on etching. The

composition of the unexposed annealed and cold-rolled 316/316L by EDAX was essentially

similar, based on the average composition of the top few microns of the surface.

5.2 XPS studies

The surface composition examined by XPS (Figure 2) reveals some differences between

the samples. In comparison to EDAX, the region of analysis by XPS is only the top few atomic

layers. After an initial short (85 s) sputter to remove most of the carbonaceous surface

contamination (typical of samples that have been exposed to the atmosphere for extended times),

we find that the cold-rolled sample surface is enhanced in chromium (Fe/Cr = 1.8) with respect

to the annealed sample (Fe/Cr = 3.3). The large oxygen and carbon component in the top few

atomic layers make it necessary to look at ratios rather than absolute values of the surface atomic

percentage of the components of interest (Table 1).

After exposure to oxygen-controlled LBE at 823K for 3000 hrs, and 85 s of sputtering,

the surface of the annealed sample is primarily iron oxide (Fe/Cr = 6.5) , whereas the cold-rolled

8

sample had large chromium content (Fe/Cr = 2.7). While there is still large carbon

contamination, the residual lead and bismuth falls below 1 atom % in both cases, indicating that

there is little to no penetration of the LBE into the forming oxide layers. The exposed samples

showed no nickel in the XPS, in agreement with previous work and the high solubility of nickel

in LBE. Similar enhancement of chromium was observed by EDAX in the near-surface region of

the cold-rolled sample as compared to the annealed sample.

Examination of the peak shape and location of the carbon and chromium core levels gives

further information. Peak positions are referenced to the common aliphatic carbon peak (284.6

eV) found on as-received samples. The different binding energies of chromium (both Cr III and

Cr VI) and their oxidation state-dependent spin-orbit splittings are described by Perry et al. [22].

The binding energies of other elements are found in the standard NIST database [23].

All samples were lightly (5 sec) argon ion sputtered to remove the majority of the

overlying carbon. The chromium in the unexposed annealed sample was found at peak location

(Cr 2p3/2 574.7 eV) and spin orbit splitting (9.2 eV) consistent with metallic and/or carbidic

chromium. The unexposed cold-rolled sample, as well as the exposed cold-rolled and annealed

samples, show chromium at a binding energy and splitting consistent with the formation of

Cr2O3. (Cr 2p3/2 576.6-576.8 eV, splitting 9.7-9.9 eV). The carbon in the unexposed samples

exhibited primarily a single peak (taken to be at 284.6 eV) except for the case of annealed

unexposed sample, which showed a small carbide peak at 282.9 eV, consistent with chromium

carbide (Figure 3). The finding of metallic or carbidic chromium in the annealed sample,

coupled with the evidence of carbidic carbon, indicates a lower stability of chromium oxide and

the formation of chromium carbide. On extended sputtering (85 s), a carbide peak was found for

9

both annealed and cold-rolled unexposed samples, presumably due to ion beam induced reaction

between the chromium and the residual carbon.

5.3 Sputter depth profiling studies

These studies were complicated by the presence of carbon (which can lead to loss of

chromium in the passivation layer by sequestration of the chromium in the grain boundaries as

the carbide) and the possibility of precipitation of iron oxide onto the surface during exposure

and extraction. We undertook a sputter depth profiling study to determine the overall oxide

compositional morphology. The very long sputtering times (up to ten hr) alters the chemical

composition of the surface, but elemental analysis of the surface is still useful.

The sputter depth profiles of annealed and cold-rolled samples show marked differences

(Figures 4, 5). The oxide on the annealed exposed sample has a top layer of iron oxide over a

layer of mixed chromium and iron oxides (either a mixed oxide or a spinel). We observed that

the total oxide thickness on the annealed sample was ~30 microns, based on the sputter rate

through SiO2 and other metal samples whose thickness was determined microscopically. The

cold-rolled sample had an oxide thickness of ~1 micron, based on similar information. The cold-

rolled sample showed a surface chromium oxide layer, with a possible mixed chromium/iron

oxide underneath, whereas the annealed sample showed iron oxide at the surface, with a clear

mixed oxide underneath.

One complicating factor is the variation in the thickness of the oxide layer from one point

to another in the sample. This affects the measurement of the elemental composition of bottom of

the oxide layer, because experimental measurements are combinations of the signals from the

oxide layer (in regions where the oxide layer is thicker) and from the bulk material (in regions

10

where the oxide layer is thinner). This complicating factor does not affect the measurement of

the elemental composition of the top of the oxide layer.

The results for the samples exposed to LBE at 823K for 3000 hrs show the same

morphology as the 1000 hr samples, with the observation of an increase in the iron percentage in

the mixed layer in the annealed sample and the chromium layer in the cold-rolled sample. These

shifts may be due to a slow dissolution of the underlying iron into the passivating oxide layer.

6. Discussion

Other workers [15-18] examining 316 and 316L samples have found both behaviors: the

formation of thin protective layers and thicker heterogeneous layers with iron oxide on the

surface. Since LBE is not included in the oxide layer, we feel that a comparison with gas phase

oxidation studies is appropriate. Thus, the dependence of passivating oxide scale composition

and structure with chromium content discussed above [7] might lead to insights as to how to

optimize the oxygen/LBE/steel system to minimize corrosion.

Seemingly small amounts of carbon can play a role in corrosion [24]. In 316 stainless

steels chromium carbides or chromium/molybdenum carbides can precipitate out at grain

boundaries, producing a near-surface layer depleted in free chromium that cannot create a

passivating chromium oxide layer. Consequently, corrosion occurs easily, especially at grain

boundaries. The 773-1073 K temperature range is particularly prone to carbide enhanced

corrosion: at lower temperatures, neither carbon nor chromium migrates very rapidly, while at

high temperatures, both carbon and chromium can migrate, and thus chromium can be

replenished at the metal surface [3].

One standard picture has corrosion enhanced at grain boundaries. Alternatively, previous

workers have found both grain size and steel composition to be important parameters in the

11

formation of passivating oxide coatings on austenitic stainless steels. Higgins et al. [25] found

that for NiCr alloys, abrasive blasting of the surface reduced grain size and led to stabilization of

the chromium oxide layer. Baer et al. [26, 27] found for 304 stainless steel that small grain size

led to protective oxide layers, but that large (>40 micron) grains had a propensity to form surface

layers that were chromium oxide enriched at grain boundaries, with an iron oxide surface region

at the center of the grain [26]. This model is appealing, because it explains how the iron oxide

layer can form on top of the chromium layer - e.g., how the iron penetrates the passivating

chromium oxide layer. In subsequent work Baer found that grain size, surface preparation,

surface carbon, oxygen partial pressure during oxidation, and other factors were all important in

determining the degree of passivation.

7. Conclusions

We have observed that a particularly stable and protective chromium oxide layer can be

formed on 316/316L nuclear grade stainless steels when exposed to oxygen controlled LBE at

823 K. This layer is distinct from the iron oxide on top of iron/chromium oxide layers observed

by other workers and is similar to behavior seen by some workers at lower temperatures for

316L. Further, similar behavior is seen for the gas phase oxidation of stainless steels. We

propose a mechanism similar to that seen for nickel/chromium alloys and other 300 series

stainless steels where grain sizes, and perhaps other factors, are important in the formation of the

passivation layer. Our laboratory is examining other 316 steels with varying surface preparation

and composition to further determine optimal operational parameters for 316 class steels in LBE.

If a chromium oxide layer can be obtained and stabilized, then new operational regions of

temperature and oxygen concentration might be utilized [11] with acceptable corrosion rates.

12

8. Acknowledgments

This work is funded through the University of Nevada, Las Vegas Transmutation

Research Program administered through the Harry Reid Center for Environmental Studies (U. S.

Department of Energy Advanced Fuel Cycle Program Grant No. DE-FG04-2001AL67358). This

work was supported by the Office of Nuclear Energy, Science, and Technology, U. S.

Department of Energy, under the auspices of the Transmutation Research Program at the

University of Nevada, Las Vegas. One of the authors (DLP) wishes to acknowledge support

from the U. S. Department of Energy under Contract Number DE-AC03-76SF00098. We

gratefully acknowledge the assistance of Lindsay Wylie with the figures.

13

References

[1]. F. Barbier and A. Rusanov, J. Nucl. Mater. 296 (2001) 231.

[2]. V. Ghetta, F. Gamaoun, J. Fouletier, M. Henault, and A. Lemoulec, J. Nucl. Mater. 296

(2001) 295.

[3]. H. Glasbrenner, J. Konys, G. Mueller and A. Rusanov, J. Nucl. Mater. 296 (2001) 237.

[4] G. Muller, A. Heinzel, J. Konys, G. Schumacher, A. Weisenburger, F. Zimmermann, V.

Engelko, A. Rusanov, V. Markov J. Nucl. Mater. 301 (2002) 40.

[5]. G. Nicaise, A. Legris, J. B. Vogt, and J. Foct, J. Nucl. Mater. 296 (2001) 256.

[6]. J. J. Park, D. P. Butt, and C.A. Beard, Nuc. Eng. Design 196 (2000) 315.

[7]. J. R. Davis, Carbon and Alloy Steels, ASM International Materials Park, OH (l996) p.

510 Table 1.

[8]. C. Fazio, G. Benamati, C. Matini and G. Paolmbarini, J. Nucl. Mater. 296 (2001) 243.

[9]. M. M. Gonzales, Biotechnology/Pharmaceutical Facilities Design (2001) 2.

[10]. B. F. Gromov, Yu. I. Orlov, P. N. Martynov, K. D. Ivanov, V. A. Gulevski, in: H. U.

Borgstedt, G. Frees (Eds.). Liquid Metal Systems, Plenum, New York, 1995, p. 339.

[11]. N. Li, J. Nucl. Mater. 300 (2002) 73.

[12]. J. U. Knebel, G. Muller, and G. Schumacher, Jahrestagung Kerntechnik (1999) 645.

[13] C. F. Lefhalm, J. U. Knebel, K. J. Mack, J. Nucl. Matl. 296 (2001) 301.

[14]. D. G. Briceno, F. J. M. Munoz, L. S. Crespo, F. Eseban and C. Torres, J. Nucl. Mater.

296 (2001) 265.

[15]. M. Takahashi, H. Sekimoto, Null. Res. Lab. Nucl. Reactor, 25 (200l) 42.

14

[16]. J. U. Knebel, G. Muller, J. Konys, Proc. ICONE 10: 10th International Conference on

Nuclear Engineering, Arlington, VA, April 14-18, 2002.

[17]. F. Barbier, G. Benamati, C. Fazio and A. Rusanov, J. Nucl. Mater 295 (200l) 149.

[18]. G. Benamati, C. Fazio, H. Piankova, A. Rusanov, J. Nucl. Mater. 301 (2002) 23.

[19]. B. Schmidt, S. Guerin, J. L Pastol, P. Plaindoux, J. P Dallas, C. Leroux and D. Gorse, J.

Nucl. Mater. 296 (200l) 249.

[20] D. L. Perry (editor), Applications of Analytical Techniques to the Characterization of

Materials, Plenum Press, New York and London (1991).

[21] Joseph Goldstein (editor), Scanning Electron Microscopy and X-Ray Microanalysis,

Kluwer Academic Publishers/Plenum Press, New York and London, 3rd Ed., (2003).

[22] D. L. Perry, L. Tsao, and J. A. Taylor, Inorg. Chim. Acta, 85 (1984) L57 and references

therein.

[23] NIST on-line database ( http://srdata.nist.gov/xps/).

[24] D. Talbot and J. Talbot, Corrosion Science and Technology, CRC Press, 1998, p. 275.

[25]. C. S. Higgins and F. S. Pettit, Trans. Metall. Soc. AIME, (1969).

[26]. D. R Baer, Appl. Surf. Sci. 7 (1981) 69.

[27]. D. R. Baer, and M.D. Merz, 1980. Metall. Trans.A 11A (1980).

15

SEM of uncorroded 316 stainless steel SEM of corroded 316 stainless steel

SEM of uncorroded 316L stainless steel SEM of corroded 316L stainless steel

Fig. 1. SEM of uncorroded 316/316L and corroded 316/316L samples, annealed and cold rolled.

16

Fig. 2 (a). Characteristic XPS survey spectra of exposed annealed 316/316L sample before and after 85 s of argon ion sputtering. The source is monochromatized aluminum K� X-rays at 1486 eV.

Fig. 2 (b). Same as 2(a) but cold-rolled sample.

17

Fig. 3 (a) Comparison of high-resolution XPS spectra of unexposed annealed and cold-rolled samples in the carbon region. Note the evidence of carbidic carbon in the annealed sample.

Cold-rolled

Annealed

Cold-rolled

Annealed

Fig. 3 (b). Same as 3(a) but the chromium region. Note the evidence of chromium oxide in the cold-rolled sample.

18

Cr 2p3/2

O 1s

Fe 2p3/2

Cr 2p3/2

Fe 2p3/2 O 1s

Ni 2p3/2

Ni 2p3/2

Fig. 4. Depth profile of annealed 316/316L after exposure to LBE for 1000 hr at 823 K. Total sputter depth is approximately 30 microns (SiO2 equivalent) and correlates well with independent SEM determinations. Note enhancement of iron and oxygen at the surface and the depletion of chromium. The composition of the 3000 hr 823 K sample (not pictured) was similar, with more iron to be found in the underlying Fe/Cr mixed oxide layer.

Fig. 5. Depth profile of cold-rolled 316/316L. Total sputter depth is approximately one micron (SiO2equivalent) and correlates well with independent SEM determinations. Note that initial chromium oxide surface is preserved over metal surface. Only a small amount of surface iron oxide is present. The composition of the 3000 hr, 823 sample (not pictured)was similar, with slightly more iron to be found in the primarily chromium oxide layer.

19

Table 1. Composition of steel samples (atom percent), before and after exposure to LBE,

as measured by XPS analysis after 85 s of argon ion sputtering.

Sample exposure Fe Ni Cr Co O C Pb Bi Other

Annealed unexposed 36 14 11 7 7 25 - -

Cold-

rolled

unexposed 16 24 9 7 17 23 - - Mo 2

Annealed 823 K

3000 hrs

13 - 2 4 34 45 <1 <1

Cold-

rolled

823 K

3000 hrs

16 - 6 2 38 31 <1 <1 Mn 5