The aim of the paper is to give examples of XPS analysis, showing how 0 1s bands can be affected by surface carbonates. A proper peak assignment (e.g. to 0-OH-species) needs the preliminary determination of a carbonate contribution. Although the samples examined represent Co and La compounds only, the principal conclusions can be extended also to other oxide systems.
INTRODUCTIONDetailed spectroscopic investigations of metal oxides with the use of XPS, AES or SIMS, for example, very often show a presence of different surface contaminants. Apart from the hydrocarbon deposit, hydroxide and carbonate species also are present generally in samples exposed previously to air.The ammount of carbonate contamination is often significant and can affect the results of analysis. It is especially important in the case of oxides of alkali metals or some rare earth or transition metals. Under convenient conditions, carbonation occurs very quickly, eventually even with total conversion of the relatively thick surface layer.When the content of adventitious carbon is high, its C 1s peak at a binding energy (BE) of 285.0 eV is dominant and the carbonate peak at 289.5 f 0.5 eV easily can be overlooked. Nevertheless, since the 0 1s peak corresponding to carbonate species is -10 times more intense (cross-section ratio tso/a, = 2.93 and atomic ratio O/C = 3), it can contribute considerably to the 0 Is band. 0 1s bands in oxides usually are complex. In addition to the peak of lattice oxygen 02-, other forms corresponding to OH-, 0-and C0:-are present. All these peaks are placed very near to each other so it can be difficult sometimes to distinguish and assign them correctly.In the case of metal oxides, the position of the lattice oxygen peak can be within the 521-535 eV range;'-, for oxides with tetrahedral coordination, it is >531.5 eV; in the remaining cases it is between 531.0 and 529.0 eV?Other oxygen species of great significance in catalysis are the 0-ions.' According to early works from Roberts' they can be distinguished by the use of XPS since their characteristic peak has a chemical shift of -2 eV with respect to that of lattice oxygen. Unfortunately, the shift due to OH -groups is almost the ~ame.'-~.'O Also, carbonates have a peak in the same region (531.5 & 0.3 eV).'s3 * Author to whom correspondence should be addressed The presence of surface 0-ions was suggested also in recent works on lanthanide-Me oxides."-13 Unfortunately, the authors do not consider the possibility that the second (higher BE side) peak in the 0 1s band may be due partly, or even totally, to carbonates.Our experience concerning the analysis of surface oxygen on rare earth or transition metal oxides indicates that a careful analysis of the C 1s band is indispensable. In this paper we give a few examples of XPS spectral analysis, showing how the 0 1s band structure can be affected by the carbonate contribution. Although the oxides studied are selected from compounds of the perovskite group or their substrates, as in some of the mentioned works, the main co...