Any attractive system for efficient light-driven water splitting needs to involve water-oxidation catalysts, which ideally are based on earth-abundant chemical elements only. Prime candidates are oxides of first-row transition metals, in particular the elements manganese to copper. In 2008, Kanan and Nocera described electrodeposition of an amorphous cobalt-based oxide film (CoCat) in phosphate buffer, which promotes water oxidation in the neutral pH regime and excels by self-healing properties. In this thesis, stability and self-healing of a cobalt catalyst (CoCat), a nickel catalyst (NiCat) and a nickel-iron catalyst (NiFeCat) were investigated at various electrochemical potentials. Cyclic voltammograms were measured every 30 min, to diagnose the catalytic activity and loss in redox-active cobalt ions. Furthermore, X-ray absorption spectroscopy is used for assessment of structural changes at the atomic level. The optimal electric potential window for having a comparably stable (> 80 %) CoCat sample is 1.1 < V vs. NHE < 1.35, where reasonable activity (> 65 %) is maintained after 24 hours operation in Co-free phosphate buffer (0.1 M KPi, pH 7). On the other hand, operation in Co-containing buffer results in an almost stable performance (in terms of current density per geometrical surface area), but the film thickness is growing significantly. Thus, self-assembly of CoCat is not well described as self-healing. Structural analysis of CoCat operated in 20 μmol L-1 Co-containing KPi buffer for 24 hours at catalytic potential (1.3 V vs. NHE) exhibit small, but significant changes in EXAFS spectra because long time operation results in increasingly ordered structure or enhanced CoCat fragment size, which cause a drop in the activity of the CoCat. XAS In-situ measurements show a slowdown of redox kinetics of the CoCat after being operated for 24 hours. This finding can be attributed to charge transport limitations in the thicker film, operated for 24 hours, and a different kinetic response of the Co-oxide fragment. One-week operation of NiCat in Ni-free buffer at non-catalytic and catalytic potentials in glass container reveals that the redox charge drastically decreases within the first day of operation and then reaches a stable condition. However, the catalytic current has a different behavior, means that first it drops and reach a minimum amount and then continuously increased. The steady increase of catalytic current is related to Fe contamination of glassware because operation in a plastic container or Fe-containing buffer confirmed that. TXRF result also confirmed that the ratio Fe:Ni increases after the operation. The Tafel slope of NiCat after one-week operation is lower than as-deposited sample, meaning that lower over-potential is needed for operated sample to increase the reaction rate. XAS measurements show that the minimum catalytic activity of NiCat operated at non-catalytic potentials (0.8 and 0.9 V vs. NHE, 11 hours) is related to a rather ordered amorphous nickel-oxide. However, after 48 hours of operation at these potentials, the catalyst is more disordered. On the other hand, operation at catalytic potential (1.4 V vs. NHE) induces a structure similar to γ-NiOOH, which is majorly stable over electrolysis time. Furthermore, long-time operation of a NiCat results in a shift of the edge position towards higher energies, which corresponds to higher oxidation state. XAS In-situ measurements show that the derivative of the CVs fluorescence data matches very closely to the reductive wave of the CV, meaning that mostly related to redox and structural changes of the Ni metal centers and not to changes in other species (no formation of peroxides or superoxides for example).