Discussion

A major requirement for the interpretation of high-resolution images acquired by STM is a precise knowledge of the chemistry of the object investigated. For example, the interpretation of the data recorded in the study of the passive films formed on Ni and Cr [4-6] would have been difficult without the knowledge of the distribution of the different phases within the passive films. Such chemical information can only be obtained in ex-situ conditions using surface analytical techniques (e.g., XPS). To correlate this chemical information with the structural information deduced from STM is a major objective of ex-situ STM investigations of passive films on various substrates. This approach must be considered as the first step of a more complete structural investigation which should include both ex-situ and in-situ measurements. Besides, one advantage of ex-situ investigations is to stop dynamic processes taking place in-situ and this favors high-resolution imaging. However, the drawback is that the electrode must be removed from solution, which may cause modifications of the surface species. The extent of the modifications due to removal of the electrode depends on the stability of the anodic layer upon exposure to atmospheric pressure of air or inert gas, and/or exposure to reduced pressure (from vacuum to UHV) and therefore this stability should be tested in the conditions of the ex-situ investigation. A critical evaluation of the differences related to the conditions of the STM measurements is necessary. The similarities between ex-situ and in-situ studies reported both on Ni [4, 5, 14] and Cr [6, 15] show the relevance of this combined approach.

In addition to the precise knowledge of the chemistry of the passive film, another requirement is necessary if the crystallinity of the passive films is investigated: it is to work with well-defined substrate surfaces. The criterion of a well-defined surface is the observation of atomic terraces. This is obtained with single-crystal surfaces. The use of such surfaces allows us to study passive films formed on metal substrate planes of the desired crystallographic orientation and to minimize and control the possible influence of the substrate defects on the defects of the passive film.

On both Ni and Cr substrates, the passive film is made of an inner oxide layer at the metal interface and of a hydroxide layer in the outer part of the film. On Ni [10-13], the thickness of the oxide inner part is about 0.4 - 1.2 nm and that of the hydroxide outer part is at most that of one monolayer (about 0.6 nm). No distinct evidence of the hydroxide layer has been found in the STM images of the passive film on Ni. This can be explained if one considers two possible tunneling mechanisms. The first possible mechanism involves tunneling from the oxide layer to the tip (or vice versa) through the hydroxide layer. In this case the oxide layer is imaged but the hydroxide layer is not. The integrity of the hydroxide layer during STM measurements depends on the width of the tunneling gap. If it is smaller than 0.6 nm, the hydroxide layer may be damaged by the scanning tip. The second possible mechanism involves direct electron transfer from the metal substrate to the surface hydroxyl groups and tunneling from these groups to the tip (or vice versa). In this case, the hydroxide layer can be imaged. The recorded images suggest then that the monolayer-thick hydroxide layer is in epitaxy on the crystalline NiO host lattice with a (lxl) relationship, and thus the structure of the hydroxide monolayer duplicates the structure of the NiO host lattice. The point defects visible in Fig. 1 could then correspond to OH vacancies (-OH on Ni2+ or -H on O2"). On Cr [6], the nonordered areas measured by STM have been assigned to this hydroxide layer on the basis of data recorded on passive films containing only dispersed islands of oxide in the inner part whereas the ordered areas were assigned to oxide nanociystals. It cannot be excluded that OH groups are present in (lxl) epitaxy above the ordered domains assigned to a-Cr203, hence suggesting that a mechanism involving tunneling from (or at) OH groups could also exist for oxide areas on Cr. A similar situation can be postulated for Fe-Cr alloys. The presence of these OH groups could possibly explain the measured deviations (see Fig. 4) from the perfect hexagonal arrangement of the basal plane of a-Cr203.

STM offers the possibility of performing local spectroscopic measurements (/ vs. V curves). These measurements can be performed in-situ and ex-situ. Ex-situ UHV conditions are however more appropriate to ensure the nonconductivity of the tunneling barrier between surface and tip. Such measurements on passive films formed on Ni and Cr should provide valuable information on the conductivity of the films. This is a promising perspective for the local characterization with high resolution of the electronic properties of passive films. On the subject of the relation between chemistry at the atomic scale and atomic structure, the STM results on the passive film formed on Ni also show promising perspectives for further characterization: accurate bias-dependent measurements of the terraces of the NiO oxide should provide information on their chemical termination (cation or anion). Also, such measurements should allow us to characterize the nature of point defects such as those shown in Fig. 1 (i.e., cation or anion vacancy).

0 0

Post a comment