Electrodeposited PdNi on a Ni rotating disk electrode highly active for glycerol electrooxidation in alkaline conditions

The development of alcohol-based electrolysis to enable the concurrent production of hydrogen with low electricity consumption still faces major challenges in terms of the maximum anodic current density achievable. Whilst noble metals enable a low electrode potential to facilitate alcohol oxidation, the deactivation of the catalyst at higher potentials makes it diﬃcult for the obtained anodic current density to compete with water electrolysis. In this work the effect of signiﬁcant parameters such as mass transport, glycerol and OH − concentration and electrolyte temperature on the glycerol electrooxidation reaction (GEOR) in alkaline conditions on a bimetallic catalyst PdNi/Ni RDE (Pd 0.9 Ni 0.1 ) has been studied to discern experimental conditions which maximise achievable anodic current density before deactivation occurs. The ratio of NaOH:glycerol in the electrolyte highly affects the rate of the GEOR. A maximum current density of 793 mA cm − 2 at -0.125 V vs. Hg/HgO through steady state polarisation curves was achieved at a moderate and intermediate rotation rate of 500 RPM in a 2 M NaOH and 1 M glycerol (ratio of 2) electrolyte at 80 ° C. Shown here is a method of catalyst reactivation for enabling the long-term use of the PdNi/Ni RDE for electrolysis at optimal conditions for extended periods of time (3 h at 300 mA cm − 2 and 10 h at 100 mA cm − 2 ). Through scanning electron microscopy (SEM), X-ray photon electron spectroscopy (XPS) and X-ray diffraction (XRD) it is shown that the electrodeposition of Pd and Ni forms an alloy and that after 10 h of electrolysis the catalyst has chemical and structural stability. This study provides details on parameters signiﬁcant to the maximising of the GEOR current density and the minimising of the debilitating effect that deactivation has on noble metal based electrocatalysts for the GEOR.


Introduction
According to the 2018 Intergovernmental Panel on Climate Change special report, [1] in order to maintain global temperatures to within a 1.5 °C increase of the pre-industrial baseline and to have net-zero emissions by 2050, drastic changes need to occur, before 2030, in the methods by which society sources its energy supply.The most recent report by the World Meteorological Organization on the state of the global climate, [2] indicates that the temperature increase up until 2020 was a global mean of around 1.2 °C so, time is running out.Therefore, the changes in the global energy infrastructure that take place in the next decade will determine the future of the globally increasing population.[3] A way to support the sustainability and decarbonisation of the economy is through biodiesel utilisation from vegetable and seed oils, i.e. triglycerides, and through hydrogen from electrolysis.Biodiesel production, consumption and technological development has therefore seen a significant increase in the last decade [ 4 , 5 ] as it can be implemented as a fuel replacement for vehicles based on fossil fuel diesel, reducing between 20 and 80% of greenhouse gas emissions.[6] A usual 10 wt% by-product of bio-diesel production is glycerol.[6] With global biodiesel production expected to continue at an approximately constant rate in the coming decade, [7] the need for glycerol utilisation to increase the positive impact of biodiesel and complete the energy cycle becomes increasingly important.[ 8 , 9 ] Glycerol oxidation to produce organic acids, ketones and aldehydes for various important industries [10] has been studied through non-electrochemical, homogeneous [11][12][13] and heterogeneous [14][15][16] catalysis at both low and high temperatures.Furthermore, due to glycerol being a non-toxic tri-hydroxy functionalised tri-carbon molecule it is particularly suitable for electrocatalytic oxidation.[17] Both the primary and the secondary carbon can adsorb to electrocatalysts for oxidation, depending on the crystal facets, [ 18 , 19 ] and 10 electrons can potentially be passed before any energy intensive C -C bond splitting is required.[20] As a result, the electrocatalysis of glycerol oxidation has developed as a hot topic of research over the last decade, especially due to the prevalence of glycerol at low prices on the global market.[21] Hydrogen is currently produced predominantly using natural gas and coal with less than 0.1% of H 2 produced from direct water electrolysis and between 2 and 5% through the chlor-alkali process.[ 22 , 23 ] With large scale industries like fertiliser synthesis and low carbon steel production, as well as applications such as fuel cells, it is imperative that the global increase in hydrogen production and utilisation be uncoupled from fossil fuels.[24] This can be achieved by renewable electricity powered water electrolysis.In water electrolysis oxygen gas is produced on the anode and hydrogen gas on the cathode of an electrochemical cell.By electro-oxidising glycerol on the anode instead of water, one can potentially produce H 2 using less than half of the electrical energy needed for water electrolysis.[21] The importance of catalyst selection has been the key research question as the glycerol electrooxidation reaction (GEOR) mechanistic pathway and maximal current density that can be achieved depends highly on the chemistry and structure of the anode material.There has been significant work conducted on a variety of catalysts [25] for the electro-oxidation of glycerol ranging from noble [ 26 , 27 ] and non-noble [28][29][30] metals to carbon [31] electrodes, in acidic and alkaline media.Electrocatalysts in alkaline media, as opposed to acidic media, have generally demonstrated higher activity.[ 10 , 27 ] Amongst the highest current densities reported (see table S1 for a summary) was shown in two glycerol fuel cell publications using carbon supported Pt and Au nanoparticles, [ 32 , 33 ] a pure Au disk, [34] and Au nanoparticles deposited on a carbon ceramic electrode.[20] For Pd-based catalysts, the highest current densities (normalised by geometric area) were demonstrated in alkaline electrolytes.[ 20 , 35 , 36 ] Composite Pd and Ni based catalysts have attracted much attention for the electrooxidation of alcohols and carbon based molecules such as methanol, [ 37 , 38 ] ethanol, [39][40][41] ethylene glycol, [ 42 , 43 ] formic acid [ 44 , 45 ] and carbohydrazide.[46] For the GEOR on PdNi there has been several studies with these catalysts synthesised by various chemical reduction methods.[ 35 , 47-54 ] Several of these studies have reported the GEOR pathway in alkaline conditions to result in the formation of glycerate, tartronate, oxalate and glycolate with higher potentials leading to increased formate production due to C -C bond cleavage.[ 35 , 47 , 48 , 53 ] These studies report the first oxidation step (2e − transfer) in the mechanistic pathway to be glyceraldehyde before being further oxidised to glycerate.Though product formation is a significant factor when considering GEOR, this work will focus on maximising the anodic current density in order to compare the applicability of the GEOR as a means to contribute to the future hydrogen economy.For comparison, the current densities of industrial scale alkaline water electrolysers operate at around 400 mA cm −2 .[ 55 , 56 ] In contrast to the aforementioned studies on PdNi for GEOR, here electrodeposition is used in the fabrication of the catalysts, a comparably simpler preparation process occurring at room temperature and, after the electrodeposition solution is made, requiring only 60 s.To the best of our knowledge there are no reports on electrodeposited PdNi catalysts for GEOR, though there are some for ethanol oxidation.[ 39 , 41 , 57 ] Additionally, on PdNi catalysts and generally, the GEOR has mainly been studied at ambient temperatures and low glycerol concentrations, where long-term stability of the catalysts, elevated temperatures and glycerol concentrations > 0.5 M are not so frequently explored.[ 20 , 32 , 47 , 58-61 ] Furthermore, the hydrodynamic conditions for optimising the GEOR have been reported to only a small extent using a rotating disk electrode (RDE).[62][63][64] By contrast, all of the above will be reported here.Since glycerol is particularly viscous, [ 65 , 66 ] knowledge of how convection, temperature and glycerol concentration affect activity is of high importance to aid in determining fluid dynamics for appropriate operating conditions for large scale application.[67] Through cyclic voltammetry, IR-corrected polarisation curves (ICPC) and chronopotentiometry, conditions for the facilitation of a high current density for the GEOR through variations in mass transport, electrolyte concentration and temperature are reported here for a PdNi electrocatalyst electrodeposited on a Ni RDE (PdNi/Ni RDE ).The nanoflower structure of the catalyst, developed similarly in a previous study by our group, [41] characterised here before and after electrolysis through scanning electron microscopy (SEM), X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS), has been seen to provide very high GEOR current densities.Additionally, the PdNi/Ni RDE electrocatalyst retains its mechanical and chemical structure well, in addition to its electrocatalytic activity over a relatively long period of electrolysis at elevated temperatures.Through variations in electrolyte concentration, a stoichiometric optimal was obtained which shows the importance of electrolyte composition when optimising the GEOR rate.Finally, a moderate and intermediate mass transport rate resulted in the highest current densities, providing significant insight into the GEOR on an electrodeposited bimetallic catalyst.
The electrodeposition solution was composed of 0.10 M Ni(NO 3 ) 2 and 1.7 mM PdCl 2 in 0.144 M HCl and 0.48 M NaCl.For the monometallic Pd catalyst (Pd/Ni RDE ), the coating was electrodeposited using the same solution composition but without Ni(NO 3 ) 2 .All the experiments of GEOR were performed in alkaline solutions made up with NaOH.

Electrochemical instruments
A Princeton Applied Research PAR273A potentiostat/galvanostat from Ametek (Minneapolis, MN, USA) was used for all electrochemical measurements and methods performed in this study.For ICPCs, a National Instruments (NI) cDAQ-9178 chassis with a NI9223 voltage digitaliser (Austin, TX, USA) was used in addition as seen similarly in a previous study.[68] See Figure S1 for a schematic of the current interrupt method.The RDE was controlled using a RDE710 Rotating Electrode setup from Gamry Instruments (Warminster, PA, USA) and a Hg/HgO reference electrode (RE-A6P, Bio-Logic, 1.0 M NaOH) was used for all measurements.

Catalyst electrodeposition
A two-electrode cell was setup in a 25 mL beaker, a schematic of which is seen in Figure S2, with a graphite rod as the counter (CE) and reference electrode (RE).A 1 cm 2 Ni RDE (Ni RDE ) working electrode (WE) was used as the substrate for electrodeposition.The WE, prior to immersion in the electrodeposition solution, was lightly sanded with a 1200 grit sandpaper to refresh the electrode surface, then polished with 1 μm MicroAlumina polish until a uniform reflective surface was observed.The RDE was rotated at 500 rotations per minute (RPM) at a depth of 0.5 cm into the electrodeposition solution for 60 s at room temperature before a cathodic current of −50 mA cm −2 was applied to the WE for 60 s.The electrodeposited catalyst was then washed in Milli-Q water under rotation before being placed in the electrochemical cell.

Electrochemical surface area calculations
To calculate the electrochemical surface area (ECSA) of the Pd and PdNi catalysts, a three-electrode cell was used seen in Figure S3.A N 2 -purged electrolyte of 1 M NaOH at 25 °C was used to determine the ECSA of Pd in the bimetallic catalyst for the GEOR active sites.A Luggin capillary was used to limit the IR-drop effect on the calculation of the ECSA.The RDE WE was kept static, a Pt grid was the CE and the RE was Hg/HgO (1 M NaOH).Cyclic voltammetry was used in a potential window of −0.600 to 0.258 V vs. Hg/HgO at a scan rate of 50 mV s −1 for two cycles, utilising the second cycle for ECSA calculation to compare the maximum achievable current density of the Pd-based catalysts by normalising by the active Pd electrodeposited onto the Ni disk.Since it was determined that in 1 M NaOH at approximately 1.25 V vs. RHE, a monolayer of PdO is formed.[69] This anodic vertex potential value was closely corroborated in acidic conditions.[ 70 , 71 ] Figure S4 demonstrates a typical cyclic voltammogram (CV) and the corresponding PdO reduction peak from which the calculations were made.
Eq. ( 1) was used to determine the ECSA of active Pd in the deposited catalyst: where Q is the charge passed during the reduction of the PdO to Pd and S is the characteristic charge density (405 μC cm −2 ) [71] of the reduction of a mono-oxide layer of PdO to Pd.

GEOR electrochemical measurements
GEOR was undertaken in a three-electrode cell seen in Figure S3.A PdNi catalyst was the RDE and WE, a Pt grid was the CE and the RE was Hg/HgO (1.0 M NaOH).The RE was kept at ambient temperature using a Luggin capillary.The N 2 -purged electrolytes studied were combinations of different concentrations of NaOH and glycerol.Electrolysis was conducted at several temperatures with the RDE at several rotation rates.Activity of the PdNi catalyst for GEOR was analysed via CVs, galvanostatic techniques including chronopotentiometry and ICPCs (conducted from low to high anodic current densities).

Catalytic material Characterisation
The electrodeposited PdNi/Ni RDE catalyst was analysed through SEM, XPS and XRD.A Zeiss LEO 1550 with an Oxford Aztec EDS microscope was used for SEM imaging with an acceleration voltage of 5 kV and a working distance of 7.7 mm.A Physical Electronics Quantera II Scanning XPS Microprobe instrument using a monochromatic Al K α operated at 15 kV with a total power of 50 W was used for the XPS measurements.The spot size was 100 μm.The base pressure in the measurement chamber was maintained at about 7 × 10 -10 bar.Four different regions in each sample were selected for survey scan, examined and the results showed good reproducibility.Surveys (Figure S5) were obtained in quintuplicate in the region 0-1040 eV, using a pass energy of 224 eV and a step size of 0.1 eV.High resolution spectra were acquired using spectra of a 26 eV pass energy and a 0.05 eV resolution.A typical survey scan and elemental scans lasted approximately 15 min and 4 hrs, respectively.XPS analysis and deconvolution of peaks were carried out using CASA XPS software.The Pd and Ni peak area were determined by the peak integration with Shirley type background function.Grazing incidence X-ray diffraction (GI-XRD) analysis was carried out with a Bruker D50 0 0 θ -2 θ parallel beam diffractometer, with a Cu microfocus X-ray source (1.54 Å) and a CCD detector with an incident angle of 1 °and a step size of 0.01.

Results and discussion
The PdNi/Ni RDE catalyst is analysed using SEM, XPS and XRD ( ex-situ ) to evaluate the atomic composition and morphology before and after electrolysis, as well as determining the portion and stability of the PdNi alloy and surface structure ( Section 3.1 ).Subsequently, the electrochemical activity of the Ni RDE substrate, Pd/Ni RDE and PdNi/Ni RDE is evaluated through CVs to illustrate the oxidative behaviour of the different catalysts with and without glycerol present in the electrolyte ( Section 3.2.1 ).Further, the enhancement of catalytic activity due to the addition of Ni to Pd is analysed via CVs and ICPCs ( Section 3.2.2 ).To observe the effects of different experimental operating conditions on the PdNi/Ni RDE catalyst the following parameters are observed through CVs and ICPCs; RDE rotation rate ( Section 3.2.3 ), temperature ( Section 3.2.4 ) and electrolyte composition ( Section 3.2.5 ).The stability of PdNi/Ni RDE under optimal electrolysis conditions is then examined through chronopotentiometry for a period of 3 and 10 h ( Section 3.2.6 ).

Scanning electron microscopy
SEM was used to analyse the morphology and particle size distribution of the PdNi/Ni RDE electrocatalyst before (pristine) and after ten-hour electrolysis in 2 M NaOH and 1 M glycerol at 100 mA cm −2 (PTHE) ( vide infra ) the images of which can be seen in Fig. 1 .
The as synthesized pristine PdNi/Ni RDE nanoparticles in Figs. 1 (a) and (b) show a uniform deposition with flower-like morphology and the particle size is predominantly between 150 -200 nm with a distribution of 100 -250 nm.Figs. 1 (c) and (d) describe the images of PTHE PdNi/Ni RDE nanoflowers, where an increment in the particle size distribution ranging from 150 -350 nm after electrolysis with a predominant particle size of 225 -275 nm indicates an increase in particle size compared to the pristine catalyst.Though, when comparing Figs. 1 (b) and (d) it is clear that there is very little change in the morphology of the catalyst and that there are no significant structural defects that appear as a result of the extended electrolysis.To determine the chemical stability, and since electrocatalysis is predominantly a surface phenomenon, XPS was used to analyse the surface atomic compositions of the pristine and PTHE catalysts.

X-Ray photoelectron spectroscopy and X-Ray diffraction
XPS was used to check the surface elemental composition, analysing a depth profile up to 10 nm, [72] and oxidation state of both the pristine and PTHE catalysts.The resultant spectra can be seen in Figs. 2 (a) -(e).The surface analysis of the pristine catalyst shows an elemental composition of Pd 89.4 ± 0.17% and Ni 10.6 ± 0.10%, whereas the PTHE catalyst shows an atomic composition of Pd 84 ± 1.7% and Ni 16 ± 2.2%, thus a slight increase in the Ni concentration can be seen on the surface of the catalyst after electrolysis.In order to determine the oxidation state of the pristine and PTHE PdNi/Ni RDE , high-resolution XPS was carried out for core levels of Pd 3d and Ni 2p.The Pd 3d core level spectrum of the pristine catalyst in Fig. 2 (a) shows two intense well resolved asymmetric peaks at energies ∼ 335.3 eV and ∼ 341 eV associated with Pd(0) 3d 5/2 and 3d 3/2 peaks, with traces of Pd-O at higher binding energies.[ 73 , 74 ] In contrast, in Fig. 2 (d) the PTHE catalyst shows ∼ 335.6 eV and ∼ 341.3 eV for Pd(0) 3d 5/2 and 3d 3/2 peaks, respectively, that aligns well with the XPS of pure Pd. [ 15 , 75 ] A negative shift of 0.3 eV to lower binding energies of the pristine catalyst compared to the PTHE catalyst shown in Fig. 2 (c) reveals a slight electron transfer from Ni to Pd during PdNi alloy formation.[76][77][78] This shift was consistent with different sample spots and thus reduced the probability of any experimental artefact or errors.This is indicative of a slight change occurring in the PdNi alloy where Ni is segregating which correlates with the increase in Ni in the elemental composition after 10 h of electrolysis.This was further confirmed by GI-XRD ( vide infra ).Moreover, no significant oxidation of Pd could be observed in the catalysts indicating that Pd is chemically stable during the ten hours of electrolysis.that Ni was completely oxidized to Ni(OH) 2 .Though exposure of Ni to air can result in a thin film of Ni(OH) 2 formation, it was reported that at elevated oxygen pressures over 15 h a film of only 0.34 nm of Ni(OH) 2 formed.[82] It can be seen from the pristine catalyst ( Fig. 2 (b)) that there is still Ni(0) present even though the measurements have been conducted ex-situ , where both catalysts were exposed to air for approximately the same period.For the PTHE catalyst ( Fig. 2 (e)) there is no Ni(0) present indicating the oxidative effect of the electrolysis rather than just exposure to air.Therefore the oxidation of metallic Ni to Ni(OH) 2 upon extensive electrolysis is likely the reason behind the de-alloying of the sample and the Ni enrichment observed through XPS.
The pristine and PTHE catalyst crystal structure was investigated using GI-XRD, shown in Fig. 2 (f), to analyse the bulk composition of the catalysts with a potential depth profile of up to 1 μm depending on the incident angle of the x-rays.[72] The pristine sample shows 2 θ values at 40.3, 46.8, 68.5, 82.5 and 86.9 °, whereas the PTHE sample shows 2 θ values at 39.4, 45.8, 66.9, 80.3, 84.5 and 87.0 °, with both corresponding to the fcc unit cell of Pd with (1 1 1), (2 0 0), (2 2 0), (3 1 1) and (2 2 2) crystal planes, respectively.[ 75 , 83 ] Lattice constants for the pristine and PTHE catalyst were calculated using Braggs law and found to be 0.387 and 0.396 nm, respectively, compared to 0.389 nm for pure Pd and 0.352 nm for pure Ni. [84][85][86] In pristine PdNi/Ni RDE , the peaks are slightly shifted to higher angles, representing a contraction of the unit cell relative to pure Pd, which reveals the formation of PdNi alloy in the bulk.[ 77 , 86 ] As Pd and Ni are known to form solid solutions, Vegard's law is used (Figure S6) for compositional analysis and calculated to be a Pd:Ni ratio of 95:5.Compared to pure Pd, the PTHE PdNi/Ni RDE observed a shift to lower angles in the 2 θ values and thereby an increment in lattice parameter corresponding to an expansion of the Pd unit cell, likely due to the presence of interstitial oxygen in the Pd unit cell.[87] However, note that there is no presence of PdO or Pd(OH) 2 peaks (which occur at a similar binding energy around 337 eV [88] ) in Figs. 2 (a) and (d) indicating the stable oxidation state of Pd after being exposed to air before the ex-situ XPS measurements.Hence, the interstitial oxygen is likely a result of Ni oxidation during electrolysis as mentioned previously.Additionally, the peaks corresponding to the Ni substrate at the 2 θ values 44.6, 51.9, and 76.5 corresponding to Ni (1 1 1), (2 0 0), and (2 2 0), respectively, remains unchanged for both the pristine and the PTHE catalyst.

Electrochemical characterisation 3.2.1. GEOR on Ni, Pd and PdNi
To clarify the oxidative behaviour of the bimetallic PdNi/Ni RDE catalyst, the Ni RDE substrate and Pd/Ni RDE , the monometallic Pd component of the catalyst, CVs were undertaken at ambient temperature with and without glycerol in alkaline electrolytes.In Figs. 3 and S7, illustrations of overlaid CVs from the three different catalysts in two electrolytes and the same CVs separated for particular emphasis in different regions of the potential scans, respectively, are shown.
When studying alcohol oxidation on noble metal based catalysts, it is common to observe a deactivation of the electrode which, during voltammetry, results in a peak current density appearing as a sudden drop after the initial climb.For Pd-based catalysts, the mechanism of deactivation is thought to be a result of the oxidation of Pd, [89] which is also what will be proposed in this study.The effect of having a noble metal present in the catalyst for the GEOR can be seen clearly in Figs. 3    The deactivation of the PdNi/Ni RDE catalyst in the presence of glycerol is described by the sharp peak at ∼ 0.0 V vs. Hg/HgO.This prominent oxidation peak at low potentials is not present in the CV using only Ni, seen more clearly in Figures S7 (a which is what we also observe.[93][94][95] Consequently, this higher oxidation state of Ni (NiOOH) is generally thought to make it active for OER [93][94][95] and GEOR [92] which can be seen in Figure S7 (a).From the brown and orange curves in Fig. 3 the positive effect of Pd when the GEOR takes place on Ni can be observed, where the onset potential of GEOR is approximately 100 mV lower on Ni in the presence of Pd, which is consistent with a previous report.[47] From Fig. 3 , there is a strong correlation between the potentials at which the PdNi/Ni RDE catalyst is active for GEOR and the reduced state of Pd defined by the low-potential anodic peak observed between −0.1 and 0 V vs. Hg/HgO which is not seen for solutions without glycerol.Furthermore, the difference of 0.9 V between the onset potential of the GEOR and OER on the PdNi/Ni RDE electrode ( ∼ −0.3 V vs Hg/HgO, and 0.6 V vs Hg/HgO, respectively), with still over a 0.8 V difference at 10 mA cm −2 , shows a significantly lower required potential for GEOR over OER.For the GEOR at ∼ −0.2 V vs. Hg/HgO, this corresponds to an approximately 0.5 V difference when comparing to some of the lowest overpotentials at 10 mA cm −2 for OER reported in the literature.[96][97][98] This shows the benefit of GEOR over OER in the production of hydrogen at lower electrode potentials.

Enhancement of Ni in the bimetallic catalyst
To illustrate the positive effect of having Ni electrodeposited with Pd to form a bimetallic catalyst, the performance of PdNi/Ni RDE and monometallic Pd/Ni RDE for the GEOR in 2 M NaOH 1 M glycerol at 80 °C and at an electrode rotation rate of 500 RPM (optimal conditions, vide infra ), is examined in Fig. 4 .Here, the current is normalised by ECSA and the IR-drop measurement for each catalyst in these conditions can be seen in Figure S8 for reference as the CVs in Fig. 4 (a) are not IR-corrected and because the IR-drop can vary slightly between experiments.
The introduction of Ni to the Pd catalyst has a positive effect on the achievable peak current density when normalised by ECSA as shown in Fig. 4 (a), where the PdNi/Ni RDE catalyst has a peak current density of approximately 75 mA cm −2 compared to the Pd/Ni RDE catalyst at 59 mA cm −2 .It is also clear from figure S9, utilising the derivative of the forward scan of the CVs [99] from Fig. 4 (a), that the GEOR onset potential is lower for the PdNi/Ni RDE catalyst indicating a catalytic effect.
For defining the regions in the ICPCs in Fig. 4 (b) and herein, the terms critical current density (i cr ) and critical potential (E cr ) will be used.Above a certain current density, i cr , a large increase in electrode potential appears and the corresponding potential at i cr is here defined as the E cr .This jump in potential is attributed to the oxidation of the Pd sites on the catalyst to Pd(OH) x and PdO x and consequently the deactivation of the electrode for GEOR men-tioned previously.In Fig. 4 (b), the ICPCs support the conclusions from Fig. 4 (a) as the PdNi/Ni RDE catalyst is at a lower potential for the entirety of the polarisation curve.Furthermore, PdNi/Ni RDE has a higher i cr of 52.7 mA cm −2  ECSA than that of Pd/Ni RDE at an i cr of 38.4 mA cm −2  ECSA .This indicates that the PdNi/Ni RDE is indeed the more desirable GEOR electrocatalyst and that the PdNi alloy aids in a higher activity.This is likely a result of the oxidation of Ni to Ni(OH) 2 seen in the XPS from Fig. 2 (e) and is indicative towards an affinity for Ni to OH groups which likely, when placed adjacently to glycerol molecules adsorbed to Pd, facilitate the GEOR.This agrees with previous studies on the role of Ni in PdNi catalysts in the electrooxidation of other alcohols.[ 37 , 40 ]

Effect of rotation rate
In order to better understand the effect of convection on the GEOR for the PdNi/Ni RDE , CVs and ICPCs were undertaken under various rotation rates using an RDE.The optimal rotation rate in terms of maximal peak current density achieved was 500 RPM, and not at the expected highest rotation rate, 20 0 0 RPM.The same optimal 500 RPM for two high concentration solutions at 80 °C can be seen in Figs. 5 (a) and (b) and the IR-drop measurement for each electrolyte in these conditions can be seen in Figure S10.
It can be seen that 500 RPM provided the highest peak current density for both 2 M and 1 M NaOH with 1 M glycerol, with the magnitude by which mass transport affects the peak current density differing only slightly depending on the concentration of the electrolyte, highlighted in Fig. 5 (c).For both solutions there was an increase in current density for increasing rotation rates up until 500 RPM and a decrease in current density for rotation rates above 500 RPM.
In a study from our group, [41] this phenomenon was shown for ethanol oxidation, on a PdNi catalyst electrodeposited on a Ni RDE, though at a greater rotation rate than reported here, 800 RPM.Martín-Yerga et al. also showed that rotation of the electrode resulted in a different product distribution, where a stagnant solution showed a slightly higher selectivity towards acetic acid which requires more oxidation steps than its precursor acetaldehyde.This was thought to be due to the longer residence time of ethanol and acetaldehyde in the porous catalyst without convection.This type of phenomena is key to understanding the most appropriate conditions for the optimisation of mass transport conditions towards the goal of operating industrial electrolysers for GEOR.
Here, the difference between deactivation and the decrease in current density with rotation rates above 500 RPM must be emphasised.Here, deactivation refers to the fact that in potential sweeps there is a peak current density, after which, at higher potentials, the catalyst becomes inactive for further GEOR and the current density drops to almost zero.This is most likely due to the oxidation of Pd during electrolysis, as previously stated, and is different to the decrease in peak current density above 500 RPM, which we assume is not a result of deactivation, but more related to mass transport effects and the oxidation of GEOR intermediates.However, it must be stated, and will be discussed later, that it cannot be concluded here that GEOR intermediates and deactivation are independent of each other.[64] Studying the GEOR through the use of the RDE has only been seen in a handful of reports.[62][63][64] A focused discussion on the intermediate rotation rate resulting in the highest GEOR current density was reported, to the best of our knowledge, in only one study using a Pt RDE [63] in acidic and ambient conditions.There the explanation of the phenomena referred to a separate study on Pt thin film and porous electrodes for the oxidation of various alcohols but not glycerol.[100] It was concluded that catalytic poisoning by alcohol oxidation products and the retention time of partial oxidation intermediates in the structure of the catalyst were the determining factors in the effect of RDE rotation rate on the an- odic peak current density.It was also observed that an increase in retention time in the more corrugated Pt/C catalyst for ethanol and n-butanol allowed for further oxidation of intermediate species leading to a higher current density with increased rotation rate, as opposed to a decrease on the Pt film.However, for methanol a decrease in current density with increased rotation rate occurred on both catalyst types.Both of these studies indicated that CO adsorption on the Pt catalysts was responsible for the decrease in current density at higher rotation rates altering the reaction mechanism to favour a route towards more CO ads .
This explanation, however, does not align with some studies regarding the oxidation of alcohols such as ethanol and ethylene glycol on Pd-incorporated catalysts in alkaline conditions.[101] It was observed for ethanol oxidation [102] at 30 °C in a CO saturated solution of 1 M KOH 1 M ethanol, that the peak current density was not diminished for a PdNiO/C catalyst during CV cycling, whereas it was for a PtNiO/C catalyst.In a study on ethylene glycol oxidation [103] at 20 °C in alkaline conditions it was concluded that Pd depresses the cleavage of the C -C bond and reduces the effect of poisoning by intermediates such as CO on the electrocatalyst performance.Furthermore, it has been shown using CVs that CO oxidation on PdNi catalysts occurs between −0.2 and −0.1 V vs. Hg/HgO (1 M OH − ), [ 37 , 102 , 104 ] it is thus likely that any CO ads will be oxidised in our experiments as these potentials coincide with the potential window observed for GEOR in Fig. 5 (a) and (b).Therefore, the decrease in peak current density at rotation rates greater than 500 RPM observed in Fig. 5 (a) is unlikely a result of poisoning by intermediates and more likely an effect of diffusion processes in the porous PdNi/ RDE catalyst.
From Fig. 5 it is seen that the rotation rate of 100 and 2000 RPM for both concentrations exhibited the biggest difference from 500 RPM in regards to the peak current density before deactivation of the electrode.Therefore, ICPCs were conducted to verify this in 2 M NaOH 1 M Glycerol at 80 o C and can be seen in Fig. 6 (see Table 1 for a summary of all ICPC results presented herein).
The ICPCs conducted from low to high current densities in Fig. 6 can be divided into two regions.The first (between −0.5 and −0.1 V vs. Hg/HgO) is the GEOR region with no significant effect of mass transport observed for the various rotation rates.The similar slopes of the curves indicate that the rate determining step for GEOR does not seem to depend on mass transport.However, it can be seen that 100 and 20 0 0 RPM deactivate at the same current density, 630 mA cm −2 , whereas 500 RPM reaches a higher i cr of 793 mA cm −2 .This demonstrates that 500 RPM is the most appropriate rotation rate to maximise current density output.The second region, after i cr , at the highest potentials between 0.4 and 0.6 V vs. Hg/HgO the reactions are believed to be glycerol oxida-  tion on Ni, OER and further oxidation of Pd sites which correlates with the results in Fig. 3 .Significant to observe is that 500 RPM reaches an approximately 25% higher i cr than 100 and 2000 RPM in the steady state ICPCs, yet the E cr for deactivation is the same, which strongly emphasises that 500 RPM provides the optimal mass transport conditions for GEOR at these operating conditions.These similar E cr values support the idea that the deactivation of the electrode is due to the oxidation of Pd since this a potential dependant reaction.

Effect of temperature
To discern the effect of increasing the temperature on the GEOR in the 2 M NaOH 1 M glycerol electrolyte, four temperatures were studied, 25, 40, 60 and 80 °C, see Figs. 7 (a) and (b).All measurements herein have been conducted at 500 RPM.
Figs. 7 (a) and (b) show that increases in temperature had a significant increase in the peak current density that could be reached before the deactivation of the catalyst, effectively doubling with each increment of temperature increase.From 25 to 80 °C the peak current density increases almost tenfold from 127 to 1100 mA cm −2 in Fig. 7 (a).The deactivation potential also appears to increase; however, this is a result of the IR drop (values for which can be seen in figure S11) and not the kinetics of the reaction at those higher potentials.From figure S11, it can be seen that as the temperature increases, the IR-drop decreases.This is to be expected since increasing the temperature in aqueous glycerol solutions results in a decrease in the viscosity and an increase in the diffusion coefficient and conductivity.[ 65 , 66 , 105-107 ] This is likely to play a role in the increased the current densities seen at higher temperatures.By collating data, it can be seen that with every 25 °C increase starting at 25 °C up to 100 °C, the diffusion coefficient is almost doubled in glycerol-water binary mixtures with the same mole fraction.[ 105 , 106 ] Note that the changing IR-drop is not a result of any temperature effects on the reference electrode as it was maintained at ambient temperature using a Luggin capillary.Also note that although the experimental conditions in Fig. 5 (b) (500 RPM, 2 M NaOH 1 M glycerol and 80 °C) are the same as the CV at 80 °C in Fig. 7 (a), the differences between the two figures can be explained by the IR-drops and the use of several Ni RDE substrates for electrodeposition over the series of experiments.ICPCs were further recorded to correlate the effect of temperature on the deactivation of the catalyst without the effect of IR-drop.Fig. 7 (b) shows that with increasing temperature, the onset potential for GEOR decreases.It can also be seen that from 25 to 80 °C the i cr increased approximately by a factor of 8 from ∼ 100 to 793 mA cm −2 , whereas the E cr occurs at approximately the same value, not significantly influenced by temperature.This also supports the result from Fig. 6 that the deactivation of the electrode is mostly dependant on the potential at which the electrode is reached indicating that is most likely due to the oxidation of the Pd active sites in the catalyst.
The apparent activation energy (E a ) was determined for two regions seen in Fig. 8 and was done so according to a previous study from our group [108] using the Arrhenius equation for electrochemical reactions.[109] One from the lower current density region nearer the onset for GEOR and one from the highest current density region closest to the i cr .Arrhenius plots for the various electrode potentials in which linearity was observed can be seen in Figure S12 where the R 2 value was greater than 0.95 for all fitted data.
The E a values from Fig. 8 (a) near the onset of GEOR range from 38.1 to 27.8 kJ mol −1 and linearly decrease with increasing electrode potential.This trend has been reported for electrochemical catalysis of GEOR on various materials [ 61 , 110 ] and was modelled for reversible and irreversible reactions.[111] Studies conducted on electrocatalytic [61] and heterogenous [ 112 , 113 ] Pd-based catalysts have determined the activation energy for glycerol oxidation to be within the range 51 to 10 kJ mol −1 with the lowest values (between 22 and 10 kJ mol −1 ) being found from the electrochemical study.
Fig. 8 (b) shows a linearly increasing E a from 22.2 to 34.6 kJ mol −1 with an increasing electrode potential towards the i cr .This phenomena was described by Danilov and Protsenko [111] where they stated that it results from the reaction process becoming diffusion limited.The study notes that a minimal E a occurs during the increase in overpotential at which point there is a reversal towards higher E a during further polarisation as there is a transition from linear Tafel dependence to the limiting current.This trend of an increased overpotential resulting in higher E a after a minimum had been reached was seen by Yahya et al. [61] and Habibi and Razmi.[20] Though the studies did not mention diffusion limitations, it was hypothesised that glycerol oxidation products were covering active sites and inhibiting further GEOR on Pd-based electrocatalysts by referencing a study by Simões et al. [35] It was stated that the breaking of a C -C bond at low potentials ( < 0.7 V vs. RHE) through a dissociative adsorption step resulted in a bridge bonded CO on the Pd surface however, there was limited direct evidence given to establish poisoning by other intermediates at higher potentials.Note, this does not contradict what was discussed in Section 3.2.3, as that was in regards to the effect of poisoning on the peak current density being lower for rotation rates above 500 RPM for the CVs seen in Fig. 5 , not deactivation, where it can be seen from the ICPCs in Fig. 6 that RPM had little effect on the rate determining step for glycerol oxidation, as previously mentioned.
Therefore, in this study we consider the deactivation of the PdNi/Ni RDE catalyst to be a combinatorial effect of net diffusion limitations of glycerol to the electrode surface at such high reaction rates (793 mA cm −2 in the case of 80 °C) and there being an inhibition of GEOR by oxidation products not desorbing or diffusing away from the surface due to characteristics; physical (as the catalyst is porous) and electrochemical (as at high enough overpotentials further oxidation of the products can occur).As a result, an increase in electrode potential leads to the oxidation of the Pd sites, thereby completely deactivating the catalyst.This type of combinatorial effect was discussed by Pérez-Martínez et al. [64] in a study of GEOR using a Au RDE and it was stated that the actual deactivation was caused by the oxidation of Au not the adsorbed intermediates, which is what is also proposed here for Pd.

Effect of concentration
Variations in concentration were studied at 80 °C to minimise any effect of viscosity on mass transport and to approach industrial conditions.Several ratios of NaOH to glycerol (NaOH:glycerol [mol:mol]) concentration were studied; 20, 10, 4, 2 and 1, as well as NaOH solutions without glycerol.The influence of the electrolyte ratio can be seen in Fig. 9 where the experiments were conducted at 500 RPM and 80 °C.
From Fig. 9 (a) through CVs, a distinct visual representation of the effect of the NaOH:glycerol ratio can be observed.For the electrolytes containing 0.1 M glycerol where the two ratios studied were 20 and 10, the effect of doubling the concentration of NaOH is an approximately 30% increase in the peak current density from 200 mA cm −2 to approximately 300 mA cm −2 .However, for solutions containing 0.5 M glycerol in Fig. 9 (a), the studied NaOH:glycerol ratios were 2 and 4 and the peak current density was around 550 mA cm −2 in both cases, i.e. doubling the NaOH concentration in this case had little effect.
Increasing the glycerol concentration to 1 M in 1 M NaOH (ratio 1), seen in Fig. 9 (c), did not increase the peak current density, thus it may be that the NaOH concentration is limiting the GEOR.In a solution containing 2 M NaOH, doubling the glycerol concen- tration from 0.5 M to 1 M for a ratio of 2, approximately doubled the peak current density from 550 mA cm −2 to 1100 mA cm −2 .However, there was no significant increase in peak current density when further doubling the glycerol concentration for a ratio of 1 from 1 M to 2 M glycerol.This agrees with the results seen when using 0.5 M glycerol and indicates that a ratio of 2 achieves the best performance.To our knowledge this phenomenon for more than one NaOH concentration has not been reported for electrochemical studies for the GEOR on Pd-based catalysts.However, an optimal ratio of 2 for glycerol oxidation in solutions of NaOH on Au heterogeneous catalysts has been reported previously [114] and its desirable effect on maintaining selectivity of oxidation products was also observed.[115] From the stoichiometry of the first oxidation step for the GEOR seen in Eq. ( 2) , there is a requirement of 2 OH − for every glycerol molecule to complete the reaction.Therefore, it can be that either glycerol or OH − becoming rate limiting at the anode is dependant on the electrolyte composition.
The optimal ratio of 2 was clarified by the ICPCs in Fig. 9 (b) and is highlighted in Fig. 9 (d).From Fig. 9 (b) it can be seen that the solutions containing 2 M NaOH with ratios 1 and 2 achieved the highest i cr over all the other electrolyte compositions.The comparison between only ratios of 1 and 2 for solutions containing 1 M NaOH in Fig. 9 (d) show clearly that a ratio of 2 with a i cr of 397 mA cm −2 outperforms that of ratio 1 with a higher glycerol concentration having a i cr of 250 mA cm −2 .This large difference in i cr was not present in the solutions of 2 M NaOH with ratios of 1 and 2. However, in the applied current density region of 1 mA cm −2 to 300 mA cm −2 in Fig. 9 (d), the curve for the ratio of 2 for the 2 M NaOH solution shows lower potential values compared to those for the ratio of 1 solution.It can be concluded that these electrolyte conditions are optimal for GEOR at 500 RPM and 80 °C as no improvement in the i cr can be reached by doubling the glycerol concentration.
For the set of concentrations studied, applying a current density above 100 mA cm −2 results in different catalytic behaviour for the different concentrations at 80 °C and 500 RPM affects largely the i cr but less so the E cr , values of which are summarised in Table 1 .This is most likely because the complete deactivation of the electrode is due to the oxidation of the surface.However, below the E cr , close to the i cr, there is the competing reactions of GEOR, gradual oxidation of the Pd surface (Pd(OH) x , [69]   fully utilise the glycerol, the stoichiometric ratio is ideal as we have shown that increasing only the glycerol concentration does not always result in achieving a higher i cr .It must be stressed that although the ratio of NaOH to glycerol is important, it is also the foundational concentration of the NaOH and glycerol in solution, i.e. beginning with 1 M NaOH and not having a ratio of 2 will be better than having 0.1 M NaOH and having a ratio of 2 for example.Though, increasing both and keeping the ratio of 2 seems to improve the maximum achievable current density.

Galvanostatic electrolysis
To determine the stability of the PdNi/Ni RDE electrocatalyst, several galvanostatic electrolysis experiments were carried out.To establish the effect of the gradual oxidation of the electrocatalyst (seen in the ECSA measurement in Figure S4 between potentials −0.3 and −0.1 V vs. Hg/HgO [69] ) on the operating potential at a specific current density, two experiments were conducted and can be seen in Fig. 10 .One with the reactivation of the electrode during electrolysis every 1 hour over 3 h and one without reactivation of the electrode with an electrolysis duration of also 3 h.
For Fig. 10 the reactivation of the electrode involved a 60 s potentiostatic hold at −0.6 V vs. Hg/HgO at the end of the 1st and 2nd hour, whereas for the red curve there was no reactivation of the electrode.Fig. 10 shows that the reactivation of the electrode results in the electrode potential remaining within 80 mV of the initial electrolysis potential at 300 mA cm −2 of around −0.15 V vs. Hg/HgO which by the end of the 3rd hour was under −0.07 V vs. Hg/HgO.From the red curve it can be seen that without the hourly reactivation of the electrode there is a much higher increase in electrode potential and results in the deactivation of the electrode before the 3 h of galvanostatic electrolysis at 300 mA cm −2 has completed.These results indicate that it is important to reactivate the electrode periodically.
In these batch experiments the electrolyte composition due to GEOR may also result in the increase in electrode potential.From the blue, orange and green curves it can be seen that the potential does not return to the original starting potential after reactivation.To rule out that this could be a result of the reduction in glycerol concentration, calculations of the charge passed (described in the supporting information) show that the decrease in concentration, if all the charge was associated with the first 2e − oxidation step seen in Eq. ( 2) , is only around 11%, thus making it unlikely.Therefore, a Fig. 11.Galvanostatic electrolysis at 100 mA cm −2 g eom (IR-corrected) over 10 h of GEOR on PdNi at 500 RPM and 80 °C in 2 M NaOH and 1 M glycerol with reactivation at −0.6 V vs. Hg/HgO every 0.5 h with electrolyte refresh every 2 h.better reactivation process is required to ensure a stable resultant potential under chronopotentiometry.
To further examine this and in an effort to observe the stability of the electrocatalyst under less intensive conditions for a longer period of time, a 100 mA cm −2 chronopotentiometric experiment was undertaken for 10 h with reactivation every half hour.Additionally, the electrolyte was refreshed every two hours, see Fig. 11 .
The initial resultant potential with an applied current density of 100 mA cm −2 in the first half hour, was approximately −0.156 V vs. Hg/HgO (taken from the point where the curve becomes more linear at around 250 s).The final electrode potential was −0.126 V vs. Hg/HgO and so over a period of half an hour there was a 30 mV increase.This increase in potential from start to finish was the largest that was observed for each separate half hour chronopotentiometric measurement.
Over a period of ten hours the electrode potential remains within a 100 mV window.The electrode potential at the beginning and end of the final half hour curve only differs from the first half hour experiment, by 20 and 10 mV, respectively.This result indicates that reactivation of the electrode at a reduction potential more frequently than illustrated in Fig. 10 enables the continuous application of the electrocatalyst for a longer period without a significant increase in operating voltage.Note, the oxidation of Pd cannot be seen in the XPS data from Fig. 2 (d) as the reactivation process was completed after the tenth hour of electrolysis.
Though the reactivation of the electrode every half hour helps to maintain the operating potential within 100 mV in the first 2 hr period, the combination of the electrolyte refresh and the reactivation have a much a larger effect.This can particularly be seen in the first six hours of electrolysis, as every two hours there is a larger drop in potential than can be seen for the prior half hour reactivations in the same two-hour period.This reinforces our previous assertions about the increase in operating potential being a combination of catalyst oxidation and GEOR inhibition by oxidation products where diffusion limitations are not applicable at 100 mA cm −2 at these experimental conditions.After six hours, the potential remains within a 60 mV window for the remaining four hours.The results from Figs. 10 and 11 show that the PdNi/Ni RDE catalyst can operate at high current densities for extended periods of time at an elevated temperature and high glycerol concentration by carefully considering ways to alleviate the mechanisms by which GEOR is inhibited.

Conclusions
Our study has shown the benefit of having a bimetallic catalyst of Pd and Ni to that of only Pd for the GEOR in alkaline electrolytes.It was seen that the addition of approximately 10% into the surface of the Pd catalyst significantly improved the achievable anodic current density over pure Pd electrodeposited on a Ni substrate.This was a result of the increased OH − adsorption to the Ni shown as Ni(OH) 2 from XPS.For the PdNi/Ni RDE electrocatalyst there was an intermediate and moderate rotation rate of 500 RPM that achieved the highest current densities both in CVs and in steady state polarisation curves, where current densities close to 800 mA cm −2 were obtained.
A critical anode potential, E cr , was identified, above which the Pd is oxidized and becomes inactive.The corresponding critical current density, i cr , depended on the glycerol concentration as well as on the NaOH:glycerol concentration ratio, where it was discerned that a ratio of 2 for both 1 M NaOH and 2 M NaOH solutions was optimal and an electrolyte containing 2 M NaOH and 1 M glycerol showed the highest values of i cr .Variations in the temperature for the 2 M NaOH 1 M glycerol solution studied through ICPCs showed that the apparent activation energy, E a , with an increasing electrode potential, decreases (38.1 to 27.8 kJ mol −1 ) near the GEOR onset potential region due the increased rate of reaction but increases (22.2 to 34.6 kJ mol −1 ) nearing the i cr due to likely diffusion limitations and the inhibition of the GEOR by surface adsorbed oxidation products.
Long term chronopotentiometric electrolysis in batch experiments with 2 M NaOH 1 M glycerol showed the PdNi/Ni RDE to be relatively stable at 300 mA cm −2 for three hours and at 100 mA cm −2 for ten hours.It was discerned that to maintain low operating potentials in galvanostatic conditions, the catalyst must undergo a reactivation (reduction potential) periodically to ensure the oxidation and thereby deactivation of the catalyst does not occur.In addition to the reactivation, it was shown that refreshing the electrolyte also aids in lowering the operating potentials at constant current densities.
Thus, to maximise the GEOR current density, an intermediate mass transport rate at elevated temperatures with an electrolyte composition of the stoichiometric ratio of 2:1, NaOH:glycerol is required.However, it was also seen that these three variables had little effect on the E cr (the potential at which i cr occurs) indicating that the complete deactivation of the PdNi/Ni RDE catalyst was likely due to oxidation of the Pd sites.Finally, the inhibition of the GEOR on PdNi/Ni RDE resulting in increasing operating potentials in galvanostatic conditions was surmised to be the combination of the gradual oxidation of the Pd sites and adsorbed glycerol oxidation products.
To continue the development of alcohol-based electrolysis at low electrode potentials, noble metals must be an integral component.In further research there will need to be a significant devotion to alleviating the deactivation of noble metal based catalysts at higher potentials to enable industrially relevant rates of hydrogen production.Operating conditions, in addition to the design of catalysts, should be optimized to balance between the GEOR, the further oxidation of products, the subsequent removal of said products for value added chemical production and the poisoning and complete oxidation of the catalyst.For further understanding of PdNi catalysts in the optimised operating conditions established here, product selectivity and the amount glycerol conversion will be addressed in future work.

CRediT authorship contribution statement
Jai White -Conceptualisation, methodology, validation, investigation, writing original draft Athira Anil -Investigation, validation, writing parts of original, review and editing Daniel Martín-Yerga -Conceptualisation, methodology, review and editing Germán Salazar-Alvarez -Supervision, visualisation, formal analysis, review and editing Gunnar Henriksson -Supervision, review and editing Ann Cornell -Supervision, project administration, funding acquisition, review and editing

Declaration of Competing Interest
The authors declare no conflicts of interest including employment, consultancies, stock ownership, honoraria, paid expert testimony, patent application registrations, and grants or other funding.

Figs. 2
(b) and (e) show the XPS spectra of Ni in the pristine catalyst and PTHE catalyst, respectively.The peaks at ∼ 855.9 eV and ∼ 873.8 eV correspond to the 3p 3/2 and 3p 1/2 of the Ni( + 2) state of Ni(OH) 2 .[ 79 , 80 ] The other broad peak around ∼ 860 eV represents the satellite peaks of Ni(OH) 2 [79] labelled 'sat.'.Additionally, a peak present at ∼ 852.7 eV in the pristine PdNi/Ni RDE sample is attributed to the metallic Ni(0) phase, [81] where the Ni(2 + ):Ni(0) ratio of the sample is ∼4.3.Zero concentration of the metallic Ni(0) phase was observed in the PTHE catalyst, indicating and S7 (a) where the onset potential for the oxidation of glycerol on the Pd-based catalysts, Pd/Ni RDE and PdNi/Ni RDE , occurs around −0.3 V vs. Hg/HgO compared to Ni RDE with an onset potential around 0.5 V vs. Hg/HgO.

Fig. 3 .
Fig. 3. CVs of Ni RDE , Pd/Ni RDE and PdNi/Ni RDE catalysts in 1 M NaOH with and without 0.1 M glycerol.All CVs were undertaken at 25 °C with a scan rate of 10 mV s −1 except for Pd/NiRDE at 50 mV s −1 .
) and (h).It is also not observed at potentials > 0.6 V vs. Hg/HgO for the Pd-based catalysts either, seen in FigureS7(b), as in this region the oxygen evolution reaction (OER) is the dominant reaction and requires higher Pd oxidation states to occur.[90]This indicates that the deactivation is specific to the Pd-glycerol interaction on the electrocatalyst.On the reverse scan the reactivation of Pd/Ni RDE and PdNi/Ni RDE occurs in correlation with the reduction of Pd(OH) x and/or PdO x (depending on the anodic potential reached) back to Pd after having been deactivated on the forward scan.This is evident by the positions of the oxidation peaks for Pd/Ni RDE and PdNi/Ni RDE in 1 M NaOH and 0.1 M glycerol compared to the reduction peaks of the same catalysts in 1 M NaOH without glycerol.A closer view of this region can be seen in FigureS7 (c).In regards to the attributes of Fig.3that relate to Ni, in the potential range ( −0.1 to 0.1 V vs. Hg/HgO) in which the GEOR on Pd occurs, the oxidation of Ni to Ni(OH) 2 dominates, not GEOR.The low current densities (visible in FiguresS7 (g) and (h)) as reported previously,[ 91 , 92 ]  and evidenced in the XPS results from Fig.2 (e), describes the full conversion of the surface Ni to Ni(OH) 2 .The visible anodic peak for Ni in 1.0 M NaOH in Figs. 3 , S7 (b), (d), (e) and (g) is characteristic of the oxidation of the Ni 2 + to Ni 3 + and is widely reported to be around 0.45 V vs. Hg/HgO in 1 M NaOH (1.45 V vs. RHE)

Fig. 4 .
Fig. 4. Effect of Ni introduction to Pd catalyst electrodeposited on Ni RDE on GEOR at 500 RPM and 80 °C in 2 M NaOH and 1 M glycerol (a) CVs at a scan rate of 10 mV s −1 (not IR-corrected) (b) ICPCs.

Fig. 5 .
Fig. 5. CVs of glycerol oxidation at 80 °C on PdNi catalysts on Ni RDE at 0, 100, 300, 500, 1000 and 2000 RPM where 500 RPM was the optimal rotation rate showing the highest peak current density.(a) 1 M NaOH and 1 M Glycerol (b) 2 M NaOH and 1 M Glycerol (c) RPM versus peak current density for both electrolytes.CVs undertaken at 10 mV s −1 (not IR-corrected).

Fig. 6 .
Fig. 6.ICPCs of glycerol oxidation in 2 M NaOH and 1 M Glycerol on PdNi catalysts on Ni RDE at 10 0, 50 0 and 20 0 0 RPM and 80 °C between 1 and 10 0 0 mA cm −2 .500 RPM was the optimal rotation rate showing the highest current density before the potential sharply increased.

Fig. 7 .
Fig. 7. Temperature effect on GEOR on on Ni RDE at 500 RPM in 2 M NaOH and 1 M Glycerol (a) CVs at a scan rate of 10 mV s −1 (not IR-corrected) (b) ICPCs between 1 and 10 0 0 mA cm −2 .

Fig. 8 .
Fig. 8. Determination of apparent activation energy, E a , of GEOR on PdNi on Ni RDE at 500 RPM in 2 M NaOH and 1 M glycerol (a) Potentials closest to onset potential (b) potentials closest to i cr .

Fig. 9 .
Fig. 9. Effect of concentration on GEOR on PdNi on Ni RDE at 500 RPM and 80 °C (a) CVs for all concentrations (not IR-corrected) (b) ICPCs for all concentrations (c) CVs ratio comparison of 2 and 1 for 1 M NaOH and 2 M NaOH (not IR-corrected) (d) ICPCs ratio comparison of 2 and 1 for 1 M NaOH and 2 M NaOH.CVs undertaken at 10 mV s −1 and ICPCs between 1 and 10 0 0 mA cm −2 .
see Figure S4) and possible GEOR inhibition by intermediates, resulting in the rise in electrode potential until E cr is reached.All of which are shown here to be affected by the composition of the electrolyte with the stoichiometric ratio of NaOH:glycerol of 2 resulting in the most favourable conditions for GEOR.It is likely that in order to

Table 1
Summary of ICPC i cr and E cr results for rotation rate, temperature and electrolyte ratio.