1548
Tuning Electrocatalytic Activity of Ultrathin Pd Shells: CO2 to Fuels

Wednesday, 31 May 2017: 08:40
Grand Salon A - Section 6 (Hilton New Orleans Riverside)
D. Fermin (School of Chemistry, University of Bristol), J. Humphrey (UK National Physics Laboratory), and V. Celorrio (University of Bristol)
It is well established that thin metal layers on a foreign support can exhibit very distinctive electrocatalytic activity with respect to bulk electrodes. In the case of -metals, changes in reactivity can be rationalised in terms of the so-called ligand or strain (also known as geometric) effects. Although is commoly argued that ligand effects are rather short range (sub-monolayer to few atomic layers), it is notoriusly difficult to experimentally differentiate the contributions from these two phenomena. In this contribution, we shall explore ligand and strain effects on the electrocatalytic properties of Au-Pd core-shell nanoparticles towards CO2 reduction to fuels. The electrochemical properties and product generation are probed by off-line 1H NMR, differential electrochemical mass spectrometry (DEMS), on-line Mass spectrometry (OLEMS) and in-situ FTIR.

Thin Pd shells were grown onto 20 nm Au nanoparticles employing a colloidal synthesis method at low temperatures, generating continuous shells with tuneable mean thickness between 1.3±0.1 (CS1) and 9.9±1.1 nm (CS10). Selected area electron diffraction patterns revealed that the effective Pd strain decreases from 3.5 to less than 1% in this thickness range, which is consistent with strain relaxtion mechanisms at Au {111} surfaces. DEMS was used to calculate the faradaic efficiency associated with the CO2 reduction vs hydrogen evolution at the core-shell nanostructures supported on a porous carbon layer. The data show that over 90% of the faradaic charge in the range between 0.0 to -0.5 V vs RHE is consumed in CO2 conversion at CS1 nanoparticles, as opposed to approximately 40% in the case of CS10. The analysis of the electrolyte solution employing 1H NMR revealed the formation of HCOO- only in the case of CS10 over a narrow potential range centred at -0.4 V, with faradaic efficiencies close to 30%. Furthermore, CH4 and C2H6 were also detected using OLEMS at potentials as negative as -0.7 V at the CS10 nanoparticles. No products in solution were detected In the case of CS1, indicating that CO is the main product generated over the entire potential range.

The contrast in reaction pathway and electrocatalytic performance are rationalised in terms of the CO affinity to the Pd nanoshells. To probe this, in-situ FTIR studies were performed revealing that the potential dependence of the adsorbed C-O stretching band (stark slope) is significantly weaker on CS1 in comparison to CS10. Furthermore, the effective CO coverage increases with increasing Pd thickness, reaching values reported for polycrystalline bulk Pd electrodes in the case of CS10. These results show that CO has a significantly weaker binding to strain Pd shells (CS1). Finally DFT calculations are used to estimate the contributions of the so-called ligand and strain effects on the local density of states of the Pd d-band. The calculations strongly suggest that the key parameters contributing to the change in mechanism is the effective lattice strain.