In S i tu Studies of PassiveFilm Chemistry using X-ray Absorption Spe c troscopy

The passive film on sputter-deposited thin film Fe and Fe-26Cr electrodes has been examined using in situ X-ray Absorption Near Edge Spectroscopy (XANES). During the x-ray spectroscopic measurements, samples were maintained under potentiostatic control in p H 8.4 borate buffer. The pa s sive film on both Fe and Fe-26Cr is formed from the metal without any detectable losses due to dissolution. Cathodic reduction leads to a loss of Fe from the film in both case s , but substantially less Fe is lost in the case of Fe-26Cr. Transpassive dissolution of Cr from Fe-26Cr only occurred on the second potential cycle after the film had been enriched i n Cr due to the reductive dissolution of Fe. S i multaneous collection of both the Fe and Cr x-ray absorption edges provides quantitative information on the relative amounts of material lost during passivation and reduction.


Introduction
The chemistry of passive films formed on pure metals and alloys in aqueous electrolytes is the subject : of much experimental investigation.Surface analytical work has been carried out using e.g.Auger, Xray Photoelectron Spectroscopy (XPS) and Secondary Ion Mass Spectroscopy (SIMS), leading to a greater understanding Ofthe composition of passive films.However, these measurements must be made ex situ in UHV and it has been demonstrated I that the passive film on Fe-Cr alloys is not stable when removed from an aqueous environment for ex situ examination.Thus, an in situ technique is to be preferred, particularly if it can yield detailed information about the composition and valence states of elements in the film.These requirements are met by in situ X-ray Absorption Spectroscopy 24.In the present work, X-ray Absorption Near Edge Structure (XANES) has been used to study the passive films on Fe and Fe-26Cr.

Experimental
The experimental setup for in situ XANES measurements has been described previously 3 and is shown in are quoted with respect to a saturated mercurous sulfate reference electrode (-0.4V(SCE)).The electrolyte was a pH 8.4 borate buffer.The solution was de,aerated throughout the experiments by bubbling with nitrogen which also ensured that ali dissolution products were removed from the vicinity of the electrode.The samples consisted of a 6_m Mylar film onto which -100A of Nb or Ta were sputtered to provide electrical contact.20 or 40A of the alloy under investigation (Fe or Fe-26Cr) was then deposited on top of the conducting layer.Measurements were made at Beamline X19A at the National Synchrotron Light Source.Data were collected in fluorescence geometry using a 13-element solid state detector (Canberra).
A set of absorption edges of standard compounds of known valency were measured in transmission geometry.The compounds were ground finely and sealed in adhesive Kapton tape.Absorption edges of metallic chromium and iron were collected from thin metal foils, also measured in transmission

Results and Discussion
A series edges of Cr and Fe standard compounds is shown in Figures 2 and 3 respectively.In general the absorption edge moves to higher energy as the oxidation state of the metal increases.The spectra of Fe and Cr both show relatively flat features after the edge, where the spectra for Fe304, Fe203 and Cr203 show a sharp peak.The 6-valent Cr shows a distinctive pre-edge peak.These characteristics can be used to easily visualize the changes in oxidation state in a sample containing a particular element in more than one valence state.
Figure 4 shows a series of spectra collected at the iron K edge from a 40A film of iron polarized cyclically in a borate buffer, lt should be noted that the edges are not normalized so changes in the edge height represent quantitative changes in the amount of iron present.The initial curve was collected after cathodic reduction of the air-formed oxide film at -I.SV.The edge position and flat top to the edge is characteristic of metallic iron.On stepping the potential to -0.4V and then to +0.4V, the edge position shifts to higher energies and the top of the edge begins to sharpen into a peak (both are characteristic of oxides).The height of the edge at the right hand side of the figure does not change indicating that in the solution used (a borate buffer), oxidation of iron proceeds by a solid state reaction without any detectable dissolution.Reduction of the passive film on stepping the potential back to -1.5V results in a spectrum identical in shape to the first one, but the drop in the edge height indicates the dissolution of a significant amount of iron.Spectra from the next cycle show that this process is repeatable a second time.
ubsequent ¢,ycles do not show such clear behavior, probably due to a loss of integrity of the very thin film.The loss of material during reduction suggests that the 3-valent oxide is reduced to the soluble 2valent state which dissolved away leaving a bare metal surface.If this reductive dissolution process were 100% efficient, then it should be possible to reconstruct the XANES signal originating from the oxide alone by subtracting curve (d) from curve (c).The result of this subtraction is shown as (1).Its shape is characteristic of 3-valent iron oxides (2) but the negative deviation in the curve around 7120 eV indicates that Fe in the metallic state has been gained upon cathodic reduction.Approximate fitting of this curve indicates that the ratio of reductive dissolution to solid state reduction is in the region of 2:1.
The total amount of metal lost in each potential cycle is listed in Table 1.lt is calculated on the following basis: it is assumed that the height of the Fe K edge for the as-deposit_ sample (which has a very thin air-formed oxide film), not shown, corresponds to 40A of iron.If, for example, the edge height drops by 20%, this corresponds to a loss of iron equivalent to 8A of metal.From Table 1, it appears that each potential cycle for pure iron involves the loss of approximately 7A.
A similar set of experiments was carried out for thin films of Fe-26Cr.One of the advantages of using a multi-element solid state detector is that different detector elements can be used to collect tluorescence data from different chemical elements.Thus iroP, and chromium K edges can be collected by scanning over both edges in a single spectrum whilst different detector elements collect Fe Kt, and Cr K_ tluorescent photons.Figure 5(a) shows details of the Fe edge iq a polarization experiment on Fe-26Cr.The results are qualitatively similar to those tbund for pure iro_l: metallic Fe present at -1.5V undergoes oxidation at -0.4V and +0.4V without any loss of material, th,;n undergoes significant dissolution as the sample is reduced back to -l.5V.This behavior is repeated on a second potential cycle (not shown).However, the amount of iron lost in each cycle is -2A, considerably less th',m the -7A lost in the case of pure iron.
The chromium K edge, measured during the same two potential cycles, is shown in Figure 5rb).Whilst the data are noisy due to the low concentration of chromium and the diminished number of detector elements used, it is clear that the data fall into two bands.Ali the spectra are roughly similar in shape, but there is a significant loss of chromium indicated by a drop in the edge height during the second potential cycle when the potential is raised into the transpassive region (increased from -0.4V to +0.4V).Concurrent monitoring of the iron and chromium edges demonstrates the interplay between the two elements in the film' on the first potential cycle, both elements are oxidized into the film.However, chromium is not present in the film in a sufficient concentration for significant transpassive dissolution to take place.During reduction of the oxide at the end of the first potential cycle, iron is lost by reductive dissolution leading to enrichment of chromium in the passive film.On the second potential :ycle, the chromium is now no longer protected by the iron and is present in a sufficient concentration or transpassive dissolution to occur.This model is consistent with similar work carried out on AICr ,dioys5.In the presence of a protective alumina layer, chromium is trapped in the film but if the alumina dissolves, then transpassive dissolution of chromium can take piace.
Cyclic voltammograms of the thin film Fe and Fe-26Cr electrodes were similar to those for bulk electrodes (after the second cycle in the case of Fe-26Cr) indicating that the electrochemistry is not affected by the thinness of the sample.
In situ x-ray absorption spectroscopy is thus a powerful technique for studying alloy passivation.In addition to providing in situ valency information, the approach allows sensitive measurement of the quantity of material lost from the alloy through dissolution or passivation.In principle, it is possible to determine which components of an alloy undergo selective dissolution, which are enriched in the passive film, and which are enriched under the film in the metallic state.

Conclusions -
I. Formation of the passive film on both iron and Fe-26Cr proceeds without any metal dissolution.
2. Cathodic reduction of the passive film on iron leads to the loss of the equivalent of -7/_ of metal in dissolution.The ratio of reductive dissolution to solid state reduction is approximately in the region of 2'1. 3. Cathodic reduction of the passive film on Fe-26Cr leads to the loss of only --2_ of iron.4. Transpassive dissolution of chromium from Fe-26Cr can only take place on the second potential cycle once iron in the film has undergone reductive dissolution leading to enrichment of chromium in the passive film. 5.In situ x-ray absorption spectroscopy of passive films on thin film electrodes is an excellent technique for obtaining quantitative data on valency and dissolution during alloy passivation.

Figure 3 .Figure 4 .
Figure 3. Fe K edges of standard compounds measured in transmission.

Figure 5 .
Figure 5. 40A of Fe-26Cr stepped through two oxidation/reductioncycles with the potential sequence indicated: (a) Fe K edge (only the first potential cycle is shown); Co)Cr K edge.

Table 1 .
Amount of metal lost during potential cycling measured from the edge height and converted into an equivalent thickness of pure metal.
Figure 2. Cr K edges of standard compounds measured in transmission.