The MLs chemical composition was measured using EDS. Color-coded composition maps of the metal elements of the ML deposited at 300°C are shown in Fig. 1, those of the MLs deposited at 500°C look alike and can be found in Supplementary Fig. 1. The color-code ranges from blue (8 at.%) over green and yellow to red (37 at.%) and is the same for all composition maps. The composition gradients appear linear: the Ti-content increases from 11.6 at the upper right to 29 at.% at the lower left, Co from 14 at the lower right to 38 at.% at the upper left, Mo from 12 to 34 at.% from upper left to lower right, Ta slightly lower from about 8 at the lower left to 28 at.% at the upper right und W ranges from 13 at the upper right to 28 at.% at the lower left. At the center of the ML the metal composition is near equiatomic with 20 at.% Ti, 24 at.% Co, 21 at.% Mo, 15 at.% Ta and 20 at.% W.
EDS measurements indicate a rather uniform N-content across one ML and a decreasing N-content with increasing deposition temperatures: 30 to 40 at.% N vs. metal elements in the ML deposited at 300°C and 26 to 35 at.% in the ML deposited at 500°C, see Fig. 2. The composition maps also indicate an increasing N-content for increasing Mo- and decreasing Co-contents.
The crystallinity of the MLs was investigated using XRD. For both MLs five exemplary diffractograms from Ta-, Mo-, (Ti-W)- and Co-rich and the center MAs are displayed in Fig. 3a) and 3d). The XRD diffractograms from both MLs show peaks at very similar locations: around 36.7, 42.5, 62 and 74.5°. These peaks can be attributed to the (111), (200), (220) and (222) planes of the NaCl-type structure and are marked in the diffractograms. Figure 3b), c), e), f) show color-coded peak intensity maps for the (111) and (200) NaCl-type structure peaks. They visualize the relative peak size compared to the rest of the diffractogram and can indicate a preferential orientation of the thin films. A red-colored MA indicates a comparatively large peak and blue is for a very low or no peak at that position. The crystal structure is roughly the same for all deposition temperatures, but the orientation changes. The strongly (111) oriented area shifts from Mo-rich MAs on the ML deposited at 300°C to a small and weak Ta-rich area for the ML deposited at 500°C. The peak intensity maps shows that large areas of the MLs are (200) oriented. For deposition at 300°C this is a large (Ti-W)- and Co-rich area covering more than 50% of the ML. On the ML deposited at 500°C this (200) oriented area shifts more to the Ti- and W-rich corner and spares out two small Ta- and Mo-rich areas.
Focusing on the more prominent (111) and (200) NaCl-structure peaks, it is evident for both MLs that the (111) peak shifts slightly with increasing Ti- and W-content and thus decreasing Ta-content from lower diffraction angles, as 36.5° to higher diffraction angles, 36.9°. The (200) NaCl-structure peak shifts accordingly, from 42.4° on the ML’s Ta-rich side to 42.7° on the Ti- and W-rich side. This is consistent with Vegard’s law: The atomic radius of Ta is larger than those of Ti and W (0.2 nm vs. 0.176 and 0.193 nm, respectively), which shifts the peaks to lower diffraction angles. 35 On the other diagonal, from high Co- to high Mo-contents, the (111) peak appears constantly at a diffraction angle of approximately 37° and the (200) peak appears at approximately 42.7°. According to Vegard’s law, this peak should shift to higher diffraction angles with increasing Co-content (atomic radii of Co and Mo are 0.152 and 0.19 nm, respectively). Looking on the N-content EDS maps, the N-content increases with increasing Mo-content. Consequently, the lattice expands with higher nitrogen content and shifts the diffraction peaks to lower angles. This trend was inter alia observed and described by Lai et al. 36
The deposition temperature affects the nitrogen content of the MLs (see above) and consequently the peak position. Furthermore, the preferential orientation of the thin films is influenced by the deposition temperature, while the basic crystal structure remains the same for increasing deposition temperatures.
SEM images of the ML deposited at 300 and 500°C are shown in Fig. 4, alongside photos of the as-deposited MLs. The approximate MAs are marked for each ML. The surface structure appears quite uniform across the whole ML deposited at 300°C and consists of a mixture of triangular and spheric shaped grains with an average grain size ranging from 52 nm at Ta- and Co-rich MAs to 77 nm at a Mo-rich MA. The average roughness values Ra of these areas range approximately from 3.8 to 6.7 nm, measured by AFM roughly at the same positions. The morphology changes for deposition at 500°C: the center, Ta- and Ti- and W-rich area appear very smooth, with only a few small grains attached to the surface, while the Co-rich area consists of cauliflower-shaped grain clusters. The Mo-rich area appears very similar compared to the ML deposited at 300°C. Ra values range from 0.5 nm at the Mo-rich area over 5.4 nm at the Mo-rich area and up to 6.8 nm at the Co-rich MA.
Both, the deposition temperature, and the chemical composition, influence the surface morphology of this HEN system. Increasing Co-content promotes a very rough surface consisting of cauliflower-shaped grain-clusters, while areas with a higher Mo-content tend to have a rather smooth surface. Increasing deposition temperatures changes the surface morphology completely in the step from 300 to 500°C.
The electrical resistance values of all MAs of both MLs were measured using 4PP at room temperature and are visualized in Fig. 5 as color-coded maps. All maps use the same color code with blue as the lowest measured resistance of 0.6 Ω to the highest resistance of 2.2 Ω in red. The resistance increases on the ML deposited at 300°C from left to right, i.e., starting from 1 Ω for high Co-, Ti- and W-contents towards 2.2 Ω for high Ta- and Mo-contents. The ML deposited at 500°C shows a similar trend with increasing resistance from left to right, but with lower overall resistance in the range of 0.8 to 1.2 Ω. The thin film thickness was measured on the ML deposited at 300°C at exemplary positions via cross-sectional SEM images. With this information, the resistivity of the thin film was calculated to be in the range from 1,814 to 5,255 µΩ*cm. On the right side of the Fig. 5 the resistivity values for five MAs per ML are listed in a table. The corresponding MAs are marked on the resistance maps on the left. Overall, the resistance values are composition-dependent and increase with increasing Ta- and Mo-contents for both MLs. The resistivity follows the same trend. The decreasing resistivity for the higher deposition temperature could arise from the lower N-contents and the consequently more metallic composition of the thin films.
The electrochemical activity for OER was evaluated using the current value at 1700 mV vs. RHE, which is plotted in color-coded maps for the MLs deposited at 300 and 500°C in Fig. 6a) and b), respectively. The color ranges from white (almost no current measured) to blue (highest measured current). A comparison of the linear sweep voltammograms (LSVs) obtained in the areas with different electrocatalytic activities is as well presented in Fig. 6a and b). All recorded LSVs can be found in Supplementary Fig. 3. The maximum activity, with a current of 13.1 µA, of the ML deposited at 300°C is found in an area with more than 35 at.% Co or more than 27 at.% Ta, normalized to the metal elements in the thin film. This corresponds to a current density of 1.78 \(\frac{\text{m}\text{A}}{{\text{c}\text{m}}^{2}}\). The activity is decreasing with increasing Mo-content and decreasing Co-content. The ML deposited at 500°C shows slightly lower OER activity with a maximum of 11.9 µA at 1700 mV vs. RHE, which is a current density of 1.62 \(\frac{\text{m}\text{A}}{{\text{c}\text{m}}^{2}}\), in an area with more than 30 at.% Co and more than 19 at.% Ta. Overall, the Co- and Ta-rich areas show higher OER activity.
There are several randomly distributed “bad points” (no current measured) apparent on the activity maps. These are most likely from incorrect measurements, e.g., bubbles in the SDC. These areas were not remeasured, because the surface was altered during OER measurements, which indicates insufficient stability for the applied electrochemical condition. Furthermore, it would not change the general trend which was already evident. The MAs are already visible with bare eye, as the photos in Supplementary Fig. 1show. The optical appearance of the OER MAs can be related to two different things. First: the thin film surface could be oxidized during measurement, which is described in more detail below, and second: some electrolyte will remain and dry on the surface after measurement. The overall higher electrocatalytic activity of samples prepared at 300°C, compared to those prepared at 500°C, seems to be correlated with the nitrogen content in the films (lower for the 500°C ML) and the electrical resistivity of the films. As can be seen in LSVs in Supplementary Fig. 3, in most of the MAs there is some oxidation of the film appearing before the start of the OER. Especially for the ML prepared at 300°C it seems like the more pronounced is the oxidation peak the lower is the OER activity. Positions of oxidation signals (most probably oxidation of the film) before start of the OER were observed at 1.5 and 1.2 V vs. RHE. Plotting a color-coded map of the difference values of current density at 1.5 V and at 1.2 V vs. RHE revealed that for the ML prepared at 300°C the oxidation peak is strongly correlated with the amount of Mo and Co in the film, see Fig. 1, (the more Mo in the film the more pronounced is the oxidation peak and the lower the OER activity). For the ML prepared at 500°C the trend is disrupted in the area with more than 30 at.% Co and more than 19 at.% Ta, where the oxidation initially increases and decreases again before increasing significantly for the Mo-rich compositions. The area of high activity corresponds to the area with the highest surface roughness and consequently also the area with the largest surface area. This is logical, since more surface area also means more active sites for OER.
Next, for each ML, 4 MAs located on the diagonal of the ML, marked with the arrow on the left side of Fig. 7 and in Supplementary Fig. 4, were chosen for conducting additional measurements. In each of those MAs, 2 cyclic voltammograms in the potential range 1-1.85 V vs. RHE were conducted separated by the 10 cycles in the potential range of 1-1.6 V vs. RHE, see Supplementary Fig. 4. For all the spots pre-oxidation is visible only in the first CV cycle in the full potential range and not in the CV after 10 cycles in narrower potential range (1-1.6 V). During cycling between 1-1.6 V vs. RHE pre-oxidation signal is getting smaller suggesting either irreversible formation of the surface oxide or dissolution of the material.