Cobalt-Doped ZnO Nanorods Coated with Nanoscale Metal–Organic Framework Shells for Water-Splitting Photoanodes

Developing highly efficient and stable photoelectrochemical (PEC) water-splitting electrodes via inexpensive, liquid phase processing is one of the key challenges for the conversion of solar energy into hydrogen for sustainable energy production. ZnO represents one the most suitable semiconductor metal oxide alternatives because of its high electron mobility, abundance, and low cost, although its performance is limited by its lack of absorption in the visible spectrum and reduced charge separation and charge transfer efficiency. Here, we present a solution-processed water-splitting photoanode based on Co-doped ZnO nanorods (NRs) coated with a transparent functionalizing metal–organic framework (MOF). The light absorption of the ZnO NRs is engineered toward the visible region by Co-doping, while the MOF significantly improves the stability and charge separation and transfer properties of the NRs. This synergetic combination of doping and nanoscale surface functionalization boosts the current density and functional lifetime of the photoanodes while achieving an unprecedented incident photon to current efficiency (IPCE) of 75% at 350 nm, which is over 2 times that of pristine ZnO. A theoretical model and band structure for the core–shell nanostructure is provided, highlighting how this nanomaterial combination provides an attractive pathway for the design of robust and highly efficient semiconductor-based photoanodes that can be translated to other semiconducting oxide systems.

. Literature comparison.              Table S1 shows a comparison of different semiconducting metal oxides and their performance as water splitting photoanodes, be it in their native state or engineered with some modifications to further improve their properties. Information about the electrolyte used for the measurement is provided in Table S1 as these may influence the performance of the photoanodes. All the applied potentials values given are referenced to the reversible hydrogen electrode (RHE), with the exception of the N-doped ZnO NRs which is reference to a saturated calomel electrode (SCE). The potential values presented in the table are those at which the IPCE values were extracted.

Literature comparison
Regarding the other semiconducting metal oxides, it is clear that TiO2 shows a much higher IPCE than any other base material. Hematite and ZnO nanoparticles (NPs) yields a poor IPCE of 4%, four times lower than that of the ZnO NPs. Nonetheless, adequate engineering of the hematite by coupling it with titanium and a CoFe outer layer results in a vast improvement in the IPCE, increasing it by more than four times, thus resulting in a competitive photoanode.
With respect to ZnO, it is shown that thin films yield a poor performance both in terms of IPCE.
However, as soon as a third dimension is introduced in the material in form of NPs, the IPCE significantly increases. Nitrogen doping of ZnO nanorods (NRs), another nanotechnological approach, proves to increase the efficiency up to 25%, far above from the 16% obtained for the NPs. Furthermore, a core-shell structure based on ZnO NRs with a Ni(OH)2 decoration proves to soar the IPCE to 40%. However, comparing these results with those of the core-shell engineered NRs of this work, it can be seen that the efficiency at 350 nm is five and two times higher, respectively, rising up to 75% at 350 nm. This clearly sets this work apart its competitors in the ZnO field in terms of efficiency.      The ZnO:Co@ZIF-8/electrolyte system is described by the equivalent circuit shown in Figure   S10. ZnO nanorods and the electrolyte form a Schottky junction, which fits the Mott-Schottky equation, i.e. the potential drop across the interface is dominated by the formation of a space charge layer within ZnO 9 :

S4
where C is the capacitance obtained by electrochemical impedance spectroscopy, A the contact area, e the elemental charge and 0 is the permittivity. ND is the donor carrier concentration, V the applied potential and Vfb is the flatband potential. Thus, the flatband potential can be calculated from the intercept of the resulting fitted line with the voltage-axis. Assuming that the contact area A between electrode and electrolyte does not vary, we can analyse the relative carrier concentrations of the photoanodes by attending the slopes of in the corresponding representation (Fig. 2d). However, it is not possible to accurately determine ND with this method unless the effective contact area between the NRs and the electrolyte is precisely known, which is beyond the scope of this work. For any MOF concentration, the maximum current density is increased compared to the sample without a MOF shell. When the shell is added to the NRs the cathodic current in the -0.51 V to 0.3 V applied potential range disappears completely, which results in an almost flat current density.
Nonetheless, increasing the shell thickness more than 1 M resulted in a decrease of the current density obtained, illustrated by the shell deposited at a 2 M concentration. Similarly, the IPCE reflects the same trend by which any shell vastly improves the performance compared to the pristine NRs, with 1 M being the best performing out of the different studied samples. S17 Figure S12. Linear Sweeping Voltammetry under AM1.5 G illumination and in dark condition of a ZIF-8 thin film. It is shown that ZIF-8 has no activity in dark conditions. Under illumination, the ZIF-8 film provides a small photogenerated current density, evidencing that ZIF-8 actively generates electron-hole pairs when illuminated. This measurement was carried out in a ZIF-8 thin film on an ITO substrate to ensure that any photocurrent obtained came from ZIF-8.