Glycerol Electrooxidation at Industrially Relevant Current Densities Using Electrodeposited PdNi/Nifoam Catalysts in Aerated Alkaline Media

Through glycerol electrooxidation, we demonstrate the viability of using a PdNi catalyst electrodeposited on Ni foam to facilitate industrially relevant rates of hydrogen generation while concurrently providing valuable organic chemicals as glycerol oxidation products. This electrocatalyst, in a solution of 2 M NaOH and 1 M glycerol at 80 °C, enabled current densities above 2000 mA cm − 2 (in a voltammetric sweep) to be obtained in atmospheres of both air and N 2 . Repeated potential cycling under an aerated atmosphere to these exceptional current densities indicated a high stability of the catalyst. Through steady state polarisation curves, 1000 mA cm − 2 was reached below an anodic potential of 0.8 V vs RHE. Chronoamperometry showed glycerate and lactate being the major oxidation products, with increased selectivity for lactate at the expense of glycerate in aerated systems. Aerated atmospheres were demonstrated to consistently increase the apparent Faradaic ef ﬁ ciency to > 100%, as determined by the concentration of oxidation products in solution. The excellent performance of PdNi/Ni in aerated solutions suggests that O 2 removal from the electrolyte is not needed for an industrial glycerol electrooxidation process, and

The world is shifting to renewable electricity and electrical energy storage as its main means of powering homes, businesses and transport.A chemical means of electrical energy storage is to produce hydrogen gas (H 2 ) through renewable-powered water electrolysis for use in applications such as fuel cells for electricity generation in long-range transport vehicles. 1H 2 can also be utilised in the low-carbon manufacturing of steel 2 and ammonia. 3However, the current status of H 2 production through renewables-powered conventional water electrolysis, from a global perspective, is negligible, with electrolytic processes in general only making up around 5%, and the overwhelming bulk of that being formed through the chlor-alkali process. 4,5The remaining 95% of H 2 production is mainly achieved through steam reforming of natural gas and coal gasification.With the climate crisis expected to be the challenge of the century, it is vital that H 2 production be uncoupled from fossil fuels if it is to be a green energy carrier and fuel of a low-carbon economy. 6onventional water electrolysis for H 2 generation requires the oxygen evolution reaction (OER) to occur on the anode.This reaction is energy intensive as it proceeds through a four-electron transfer with a high equilibrium potential.As a result, overpotentials of around 200 mV are required, even on some of the best electrocatalysts reported, 7,8 to enable the OER on the anode.Since the HER proceeds readily on a variety of electrocatalysts at low overpotentials, it is imperative that the problem of the high energy demand on the anodic reaction from the OER is solved.An alternative to having the OER on the anode is to replace it with oxidation processes with lower equilibrium potentials.A promising option is the electrooxidation of alcohols, specifically, glycerol. 9lthough ethanol and methanol have also been studied for this purpose, glycerol has three hydroxyl functional groups and can therefore undergo oxidation on both the primary and secondary alcohol sites, making it an important platform chemical.Furthermore, glycerol is an unavoidable by-product of biodiesel production through a transesterification reaction and is therefore readily available.
There are several advantages to the glycerol electrooxidation reaction (GEOR) over the OER.Not only does the GEOR occur at a lower equilibrium potential than that of the OER, it also results in organic products such as glyceric, tartronic, glycolic, lactic and oxalic acids.These glycerol oxidation products (GOPs) are valuable for a variety of applications in e.g. the pharmaceutical, cosmetic and food industries. 10Hence, replacing the OER with the GEOR leads to an overall lower cell potential, requiring less energy input for the generation of H 2 while simultaneously producing industrially relevant organic chemicals for further downstream utilisation.Therefore, the GEOR results in the production of two value streams as opposed to only one as, the oxygen generated by the OER is commonly vented to the air.Some inextricably linked challenges for the development of an industrially relevant GEOR process are; systems that enable low anodic potentials, high current density, low degrees of electrocatalyst deactivation/poisoning and high GOP selectivity.With regard to the high current density, for the GEOR to be viable the current densities need to be comparable to those of conventional water electrolysers which can typically operate at around 400 mA cm −2 for alkaline electrolytes. 11,12These high current densities can be difficult to achieve for the GEOR and can also be problematic if, when running galvanostatically at high current densities, the electrocatalyst potential increases as this increase in potential can cause a change in GOP selectivity.Some of the highest current densities for the GEOR in the literature are reported in our previous study. 13To achieve low anode potentials for the GEOR, noble metal catalysts can be employed; however, noble metals have a difficulty with deactivation during alcohol oxidation if either C-C bond cleavage of the glycerol molecule occurs (leading to CO poisoning) or oxidation of the metal occurs, depending on the type of metal.For Pt, deactivation is believed to be from CO-poisoning, 14 whereas for Au and Pd it is believed to be oxidation of the catalytic sites. 15,16On PdNi, this deactivation was shown to be potential dependent for the GEOR, where, at a variety of different temperatures and electrolyte compositions under an atmosphere of nitrogen, the deactivation occurred at similar potentials. 138][19] When using alkaline electrolytes, for Pd-based electrocatalysts, the GOPs detected are typically glycerate, tartronate, glycolate and formate. 18However, experimental conditions such as electrode potential, temperature and the electrolyte composition can play a large role in the observed selectivity of a catalyst.The addition of Ni to Pd-based electrocatalysts enhances the activity for the oxidation of not only glycerol, discussed in our previous study, 13 but also for ethanol 20 and methanol. 21PdNi on Ni foam has only recently been studied, 22 where the synthesis of PdNi involved a hydrothermal method and the GEOR was conducted at low temperature, relatively low concentrations of glycerol and supporting electrolyte and only in deaerated solutions.
Since glycerol oxidation can also be achieved chemically by using heterogeneous catalysts with the reaction typically involving an atmosphere of oxygen, with oxygen acting as the oxidant, there is an opportunity to combine electrochemical and chemical oxidation of glycerol.The chemical oxidation steps are not thought to occur directly through the dissociation of the oxygen molecule but by regenerating hydroxide ions formed via catalytic decomposition of a peroxide intermediate which requires the reduction of the oxygen molecule by adjacent water molecules. 23The authors note that this decomposition of the peroxide intermediate is said to occur most readily on Pd when comparing to other noble metals such as Pt and Au. 23Therefore, for Pd-based electrocatalysts, greater amounts of GOP formation could be achieved by providing oxygen to the solution without additional electrical energy input.This implies that it could be possible to achieve an apparent Faradaic efficiency (FE) of >100% for the GEOR when considering the charge required for the amount of GOPs generated from a combination of electrochemical and chemical processes.
Generally, the literature concerning GEOR denotes experimental work undertaken in atmospheres of inert gas.However, there was recently one study that detailed the effect of having oxygen present in solution for GEOR undertaken in acidic media. 24The authors concluded that the presence of oxygen using a Pt/C and Pt@Au core-shell catalyst for the GEOR had the following effects: a significant reduction in the electrochemical surface area (ECSA) of the catalysts with the worst case resulting from a O 2 -saturated solution; a net cathodic current in a O 2 -saturated solution during potentiostatic measurements (0.776 and 0.876 V vs RHE) at low glycerol concentrations; the same GOPs formed but in different proportions.
In this work we provide an insight into the effect of an aerated atmosphere during the GEOR on a PdNi catalyst electrodeposited on Ni foam (PdNi/Ni) at a high concentration of glycerol and NaOH and at an elevated temperature.The characterisation of the PdNi/Ni catalyst, by scanning electron microscopy (SEM), energy dispersive X-ray spectroscopy (EDS), grazing incidence X-ray diffraction (GI-XRD), and X-ray photoelectron spectroscopy (XPS), shows that the catalyst varies in composition across the substrate due to the nature of the Ni foam substrate during electrodeposition.Through cyclic voltammetry (CV), linear sweep voltammetry (LSV) and iR-corrected polarisation curves (ICPCs), we show an electrocatalyst, fabricated through a scalable method, that is capable of reaching current densities comparable to that of alkaline water electrolysers.By altering the atmosphere of the electrochemical studies, we evaluate the effect this has on the performance of the PdNi/Ni catalyst and the selectivity for GOPs, determined by high performance liquid chromatography (HPLC), from which we discuss possible reaction mechanisms for the GEOR and how these can be altered by providing an aerated atmosphere.Similar to the previously mentioned study, 24 the introduction of O 2 to the anolyte results in the same GOPs but with different selectivity for the organic product fractions (PFs).Finally, we show that having an atmosphere of air instead of N 2 can increase the apparent FE in a process combining electrochemical and chemical oxidation.

Experimental
Materials and reagents.-TheNi foam (>99.99%) was purchased from MTI Corporation (Richmond, CA, USA) and cut into 1 cm 2 pieces before further processing.Palladium chloride (>59.0%Pd; >99.9%, metal basis) PdCl₂ (Alfa Aesar), nickel nitrate hexahydrate Ni(NO 3 ) 2 .6H 2 O (Alfa Aesar), sodium chloride NaCl (VWR), sodium hydroxide NaOH (Merck), hydrochloric acid HCl 37% (Sigma Aldrich) and glycerol (Merck) were all purchased from VWR (Radnor, PA, USA).All reagents were of analytical grade.The ultrapure water used to make up all solutions in this study was obtained with a Millipore DirectQ3 purification system from Millipore (Burlington, MA, USA).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), the coating was electrodeposited using the same solution composition but without the presence of Ni(NO 3 ) 2 .All the experiments of GEOR were performed in alkaline solutions made up with NaOH.
Electrochemical instruments.-Thevoltammetry experiments were performed with an Ivium XP20 potentiostat from Ivium (Eindhoven, The Netherlands) using 85% iR-compensation for which the electrolyte resistance was determined prior to the voltammetry using electrochemical impedance spectroscopy (EIS).For ICPCs, a current interrupt method was used through a Princeton Applied Research PAR273A potentiostat/galvanostat from Ametek (Minneapolis, MN, USA) in connection with a National Instruments (NI) cDAQ-9178 chassis and a NI9223 voltage digitaliser (Austin, TX, USA) as seen similarly in a previous study. 25A Hg/HgO reference electrode (RE) (RE-A6P, Bio-Logic, 1.0 M NaOH) with a Luggin capillary used for all measurements.
Catalyst electrodeposition.-Atwo-electrode cell was setup in a 25 ml beaker, a schematic of which is seen in Fig. S1, with a graphite rod either side of the Ni foam connected in parallel as the counter electrode (CE).A 1 cm 2 Ni foam working electrode (WE) with a thickness of 1.6 mm was used as the substrate for electrodeposition.The WE, prior to immersion in the electrodeposition solution, was ultrasonicated in 37% HCl, MQ-water and ethanol, respectively, for 5 min each and then patted dry with fibreless tissue paper.A stirring bar was placed in the electrodeposition solution at 600 rotations per minute (RPM).The Ni foam WE was then placed into the electrodeposition solution for 60 s at room temperature before a cathodic current of −50 mA cm −2 was applied for 120 s.The electrodeposited catalyst was then washed in Milli-Q water before being placed in the electrochemical cell.
Electrochemical surface area calculations.-Tocalculate the ECSA of the Pd/Ni and PdNi/Ni electrocatalysts, a three-electrode cell was used (Fig. S2).A N 2 -purged electrolyte of 1 M NaOH at 25 °C was used as the anolyte to determine the ECSA of Pd on both the Pd/Ni and PdNi/Ni catalysts for the GEOR active sites.A Luggin capillary was used to limit the iR-drop effect on the calculation of the ECSA.The electrolyte was kept stagnant, a Pt grid was the CE in 1 M NaOH and the RE was Hg/HgO (1 M NaOH).CV was used in a potential window of 0.265 to 1.250 V vs RHE 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 normalised by the active Pd electrodeposited onto the Ni foam.It was previously determined in 1 M NaOH that at approximately 1.25 V vs RHE a monolayer of PdO is formed. 26This anodic vertex potential value was closely corroborated in acidic conditions. 27,28Fig. S3 demonstrates a typical cyclic voltammogram and the corresponding PdO reduction peak from which the calculations of the ECSA were made using Eq. 1.
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 ) of the reduction of a mono-oxide layer of PdO to Pd. 28 GEOR electrochemical measurements.-AllGEOR experiments were undertaken in a divided three-electrode cell seen in Fig. S2.The cell was separated using a Nafion 212 membrane (VWR).The Pd/Ni and PdNi/Ni electrocatalysts were the WEs, a Pt grid was the CE and the RE was Hg/HgO (1 M NaOH).The WE potential was converted from Hg/HgO (1 M NaOH) to the reversible hydrogen electrode (RHE) using Eq. 2. is the measured or controlled potential during the experiments when using the Hg/HgO (1 M NaOH) and the pH is 14.3 as the supporting electrolyte was 2 M NaOH.

E
The RE was kept at ambient temperature using a Luggin capillary.The electrolyte consisted of 2 M NaOH and 1 M glycerol and was studied under an atmosphere of nitrogen and air, respectively.Therefore, this study utilises the purging of air, not an openair experiment, studied previously, 24 as it is conducted at 80 °C and would therefore be susceptible to extensive evaporation if it were not a closed-lid system with only a small outlet to allow gas to escape.Extensive evaporation would alter the concentration of the electrolyte and consequently the concentration of GOPs.Hence, there was careful observation of the volume of the electrolyte during experiments to ensure that the volume was not significantly changing.The electrochemical measurements were conducted at 80 °C with the stirring bar at 200 RPM.The activity of the PdNi/Ni catalyst for GEOR was analysed via CV (85% iR-compensated), LSV (85% iRcompensated) and ICPCs (conducted from low to high anodic current densities).For the GOP analysis, chronoamperometry was used.Journal of The Electrochemical Society, 2023 170 086504 High performance liquid chromatography.-HPLCanalyses were performed with an Agilent 1260 Infinity II isocratic pump, multisampler and multicolumn thermostat with a 1290 Infinity II refractive index detector.The analytical columns, in series, included a Bio-Rad guard column with a standard cartridge holder and a Micro-Guard cation H + cartridge (4.6 × 30 mm), a Bio-Rad Aminex HPX-87H column (7.8 × 300 mm), and Shodex Sugar SH1011 column (8 × 300 mm).All columns in series were kept at 30 °C.To analyse the glycerol oxidation products with the best possible peak separation to get the most accurate results, two mobile phases were utilised.The two mobile phases were 1 mM H 2 SO 4 and 8 mM H 2 SO 4 at a flow rate of 0.25 ml min −1 .All standards used to detect the GOPs were calibrated from 0.1 mM to 10 mM with linear regression R 2 values greater than 0.99.
Catalytic material characterisation.-Theelectrodeposited PdNi/Ni catalyst was analysed through SEM, EDS, XPS and GI-XRD.SEM imaging was performed using a JEOL JSM-7000F microscope equipped with an energy dispersive spectrometer operating at an acceleration voltage of 20 kV.The EDS analysis was carried out on three different sample regions by recording nine spectra on each of the characteristic sites depicted in Fig. 1.A Physical Electronics Quantera II Scanning XPS Microprobe instrument using a monochromatic Al Kα operated at 15 kV with a total power of 50 W 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 scans and the results showed good reproducibility.Surveys (Fig. S4) 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 scan lasted approximately 15 min and 4 h, respectively.XPS analysis and deconvolution of peaks were carried out using CasaXPS software.The Pd and Ni peak area were determined by peak integration with Shirley type background function.Grazing incidence X-ray diffraction (GI-XRD) analysis was carried out with a Bruker D5000 θ-2θ parallel beam diffractometer, with a Cu microfocus X-ray source (1.54 Å) and a charge coupled device (CCD) detector with an incident angle of 1°and a step size of 0.01.

Results and Discussion
Physical characterisation.-ThePdNi/Ni electrocatalyst was analysed using SEM, EDS, XPS and GI-XRD, all of which were done ex situ, to evaluate the atomic composition and morphology.The Ni foam substrate has many areas of irregularities which can be seen in the SEM images in Fig. S5.As a result, uniform electrodeposition of a precise composition was not achieved and there was a difference in particle size depending on which section of the Ni foam the catalyst was deposited on.The three distinct areas of the Ni foam substrate that provide different types of surfaces are referred to below as the "arm" (Figs.1a, 1d and 1g), "arm edge" (Figs.1b, 1e  and 1h) and "arm joint" (Figs.1c, 1f and 1i), where the arm is a flatter surface between the arm edges which are curved in nature and the joint between arms is the most pointed and with the most visible defects.The non-uniform electrodeposition of the catalyst illustrates that there is a variation in current distribution depending on the geometry of the Ni foam substrate influencing the Pd:Ni ratio of the deposited catalyst.
For Figs. 1c, 1f and 1i, the arm joint, there is the highest composition of Pd, averaging around 70 at%, this is an indication that at these sites, with likely the most defects and mechanical stress, Pd appears to be depositing more readily.For Figs. 1a, 1d and 1g, the arm of the substrate, the particles are smaller in size and have a higher Ni composition than that of the Pd, which is around 30 at%.Figures 1b, 1e and 1h shows that the arm edge section of the Ni foam provides a surface that allows for a more even co-deposition of Pd and Ni with the ratio of Pd to Ni being 60:40 at%.At these sites the surface is not flat but appears to have enough augmentation and structural defects to provide suitable conditions for co-nucleation of the Pd and Ni.
Since the penetration depth of EDS is in the order of microns and therefore may be affected by the Ni foam substrate for sections with smaller sized particles, XPS was used to identify the surface composition and chemical state of the as synthesised PdNi/Ni catalyst (Figs.2a and 2b).It was found that the electrodeposited Pd portion of the bimetallic catalyst is mainly in the elemental form (Fig. 2a) whilst the Ni is present in a divalent state which likely indicates the presence of Ni(OH) 2 (Fig. 2b). 29igures 2a and 2b, represent the high resolution XPS for core levels of Pd 3d and Ni 2p of the electrodeposited PdNi/Ni catalyst.2][33] In Fig. 2b, the sharp peaks at binding energies of ∼856 eV and ∼873.7 eV describe Ni (II) 2p 3/2 and 2p 1/2 of Ni(OH) 2 . 34,35The broad peaks at binding energies of ∼861.8 eV and 879.8 eV ("sat.")represent the satellite peaks of Ni(OH) 2 . 34ence, the electrodeposited Ni in the PdNi/Ni catalyst is in the form of Ni(OH) 2 .The variation in the Pd:Ni atomic composition observed from the SEM EDS can also be seen through XPS. Figure S4 details samples of the raw XPS spectra with the atomic composition of Pd and Ni on the PdNi/Ni catalyst.There are two compositional ratios of Pd:Ni observed, with the average compositional ratios being 50.1:49.9± 6.7 at% and 85.7:14.3± 1.3 at%.These results indicate that the electrocatalytic surface is a composite and that the electrodeposition of PdNi onto the Ni foam substrate did not occur uniformly across the entire structure.
Figure 2c shows the GI-XRD spectrum of the pristine PdNi/Ni catalyst.The GI-XRD spectrum without the logarithm taken can be seen in Fig. S6. Figure 2c  Journal of The Electrochemical Society, 2023 170 086504 Ni (1 1 1), (2 0 0) and (2 2 0) crystal planes, respectively.[39][40][41][42] Electrochemical characterisation.-Voltammetry.-Voltammetry was undertaken in a divided electrochemical cell with an electrolyte consisting of 2 M NaOH and 1 M glycerol, as in our previous work it was concluded that, for a PdNi/Ni RDE electrocatalyst, a ratio of 2:1 NaOH:glycerol provided the highest current density. 13For reference, the electrochemical activity of the Pd/Ni and PdNi/Ni catalysts with and without glycerol present in the electrolyte can be seen in Fig. S7.As a means to screen the performance of the catalysts, the activity of the PdNi/Ni and Pd/Ni was compared through CV undertaken at 25 °C, see Fig. 3a.The resultant current from the GEOR was normalised by the ECSA of Pd on the catalysts as within the potential range studied, where the peak current is observed, the GEOR does not occur on Ni but solely on the Pd sites (see previous work 13 ).For PdNi/Ni, the ECSA was 66 ± 9 cm 2 (rounded to the nearest integer) over eight different samples.The ECSA for the PdNi/Ni and Pd/Ni samples used for the measurement in Fig. 3 was 66 cm 2 and 59 cm 2 , respectively.
From Fig. 3a it can be seen that the PdNi/Ni electrocatalyst reached a higher current density than that of Pd/Ni with a current density of around 4.7 mA cm −2 ECSA compared to around 3.7 mA cm −2 ECSA , i.e. the PdNi/Ni has higher peak current density of around 21%. Furthermore, the deactivation of the catalysts occurs at different potentials with the Pd/Ni catalyst deactivating at a lower potential of approximately 1.03 V vs RHE, around 60 mV less than that of PdNi/Ni at approximately 1.09 V vs RHE.Hence, for the study conducted at 25 °C, the higher current density and the higher tolerance to deactivation as a function of potential, makes the PdNi/ Ni catalyst superior to the Pd/Ni catalyst.In order to achieve a higher current density and compare the two catalysts at an industrially relevant temperature, the temperature was increased to 80 °C, see Fig. 3b.At 80 °C it can be seen that the current densities are approximately the same for PdNi/Ni and Pd/Ni but that the onset of deactivation occurs for Pd/Ni at a lower potential than that of PdNi/ Ni.Therefore, the PdNi/Ni catalyst was chosen for further study.To investigate the effect of atmosphere on the GEOR for PdNi/Ni at elevated temperatures, LSV measurements were conducted under both a N 2 atmosphere and an atmosphere of air, respectively, with purging before and during the experiments.In Fig. 4a, the comparison of the performance of the PdNi/Ni catalyst can be seen for both atmospheres at 80 °C.
From Fig. 4a it can be seen that there is little difference in the overall performance and shape of the curve when comparing the PdNi/Ni catalyst in different atmospheres, with increased current densities in the presence of air.This is a positive result since it would be an advantage going forward to an industrial application to not have the added cost of purging the electrolyte with N 2 to ensure the high performance of the catalyst, i.e. the electrolyte could be left aerated.The onset potential for both atmospheres is ∼0.6 V vs RHE.Under both atmospheres, a current density of above 2000 mA cm −2 was achieved at 1.3 V vs RHE.This significantly high current density shows that with the right catalyst and substrate, GEOR can achieve current densities relevant to industry.4][45][46][47][48] For Pd/Ni, under these conditions, the results were similar although the difference in current density between the atmospheres was not as significant, see Fig. S8.
To investigate the stability of the PdNi/Ni catalyst to reach such high current densities consistently, a test using 800 cycles was conducted under an atmosphere of air.The catalyst was swept between 0.265 V and 1.465 V vs RHE for 800 cycles.Here we show, in Fig. 4b, the sweep in the anodic direction up until the peak current density of the 1st and 800th cycle.The 1st cycle reaches a current density of approximately 2600 mA cm −2 and after 800 cycles a current density of 2300 mA cm −2 was still achieved.Therefore, the PdNi/Ni retained 88% of its activity over 800 cycles in the presence of air.This demonstrates that not only is the PdNi/catalyst relatively stable at high current densities, high temperatures and high concentrations of glycerol and NaOH, it is also able to retain its performance in the presence of air.
To analyse the activity of the catalysts under a more steady-state environment, ICPCs were conducted.The ICPCs were undertaken from 1 mA cm −2 to 1000 mA cm −2 , see Fig. 4c.We see that below 0.8 V vs RHE, 1000 mA cm −2 can be reached for both atmospheres of N 2 and air.Interestingly, the performance of the PdNi/Ni catalyst appears to be better under an atmosphere of air as the potential is lower, however the Tafel slopes are very similar at 122 mV dec −1 for nitrogen and 124 mV dec −1 for air.This further implies that the GEOR is not being inhibited by the presence of dissolved oxygen in solution using the PdNi/Ni catalyst.
Chronoamperometry and HPLC analysis.-Tounderstand the effect of the different atmospheres on the GOP composition, chronoamperometric tests were conducted followed by HPLC analysis of the products formed.Two potentials were used for these studies, 0.765 V vs RHE and 0.865 V vs RHE, see Fig. S9.The two potentials were chosen 100 mV apart to increase the possibility for variability in the product distribution as the anodic potential of the electrocatalyst can have an effect on the GOPs.For both potentials the current densities obtained in the two different atmospheres converge and the resultant charge passed is similar; with 1275 C and 1252 C for N 2 and air, respectively, at 0.765 V vs RHE and with 2110 C and 1829 C at 0.865 V vs RHE for N 2 and air, respectively.
Samples of the bulk solutions after 2 h of electrolysis were analysed for GOPs using HPLC.Examples of the chromatograms for the calibrated standards and the experimental samples can be seen in Fig. S10.The dominant GOPs determined from the GEOR on PdNi/ Ni were glyceric and lactic acid which are carboxylic acids but in alkaline electrolytes are carboxylates, i.e. deprotonated and therefore have a negative charge.The presence of tartronate, formate, glycolate, oxalate and pyruvate were also observed.The molecular structure of the GOPs and the organic PFs plotted as a percentage of the total concentration of products generated can be seen in Fig. 5.
From Fig. 5, it can be seen that lactate (light blue) and glycerate (grey), are the two most abundant products.A collection of the overall PFs (%) and yield (mM) of each product can be seen in Tables S1 and S2, respectively.In addition to the two main products, tartronate, formate, glycolate, oxalate and pyruvate were observed to form in minor quantities in all cases, indicating that further electrochemical oxidation steps of glycerate and lactate takes place at both potentials assessed, with tartronate and formate having the largest proportion of these minor quantities.The product analysis from the chronoamperometric experiment at 0.865 V vs RHE (Fig. 5a), in an atmosphere of N 2 , resulted in glycerate as the main product with a PF of 52% and lactate as the second most abundant product at 34%.Whilst in an aerated atmosphere, the PF increased for lactate to 39% at the expense of glycerate which decreased to 46%.This corresponds to a decrease in the selectivity of glycerate to lactate from 1.53 to 1.18.A similar trend was observed for the experiments carried out at 0.765 V vs RHE.With N 2 , the PFs for glycerate and lactate were determined to be 46% and 42%, respectively, whereas for electrolysis carried out with air, the PF for glycerate decreased to 41% but increased to 45% for lactate.This corresponds to a decrease in the selectivity of glycerate to lactate from 1.10 to 0.91 and therefore, a switch in product selectivity from glycerate to lactate as the main product.Although there are differences in the PFs between atmospheres for both potentials, the presence or absence of O 2 in solution does not affect the PFs very substantially, which supports the electrochemical data in Fig. 4c where similar Tafel slopes were observed for both atmospheres indicating there was no change observed in the ratedetermining step of the electrochemical reaction mechanism.Although the effect is not so substantial, the switch from inert atmosphere to air seems to predominantly affect the first two oxidation steps, i.e. the oxidation of glycerol to glyceraldehyde/ dihydroxyacetone and subsequent oxidation of glyceraldehyde to glycerate, thus affecting the selectivity for glycerate and lactate which are expected to form by electrochemical and chemical routes, respectively.A proposed reaction network for the GOPs is found in Scheme 1 with the products identified by HPLC outlined in coloured boxes, as seen in Fig. 8.For simplicity, only the number of electrons for each reaction step is included in Scheme 1, excluding the processes of water formation and the number of oxygen and hydrogen passed during each reaction.
From Scheme 1, it can be seen that the first two electrochemical oxidation steps for the GEOR is to either dihydroxyacetone or glyceraldehyde.Glyceraldehyde and dihydroxyacetone can undergo chemical interconversion [49][50][51] via a reaction with OH which involves an enediol intermediate in what is known as a hydride shift.Since glycerate is the dominant product in three of the four cases studied, it is likely that a majority of glyceraldehyde is rapidly oxidised to glycerate electrochemically and the remaining glyceraldehyde can participate in further chemical reactions to form the minor side products.
As discussed above, the selectivity for the formation of glycerate over lactate is dominant at 0.765 and 0.865 V vs RHE in inert atmosphere as well as at 0.865 V in an atmosphere of air, whereas lactate was most abundant at 0.765 V vs RHE in air (Fig. 5).To understand the possible origins of this switch in selectivity, the potential at which the oxygen reduction reaction (ORR) can take place on Pd-based catalysts should be considered.][54] Analogously, it is likely that more ORR occurs at 0.765 V vs RHE resulting in a higher concentration of OH − on the surface of the electrode, either occurring electrochemically 54 or chemically 23 through hydrogen peroxide decomposition as discussed previously.In turn, these species can promote subsequent chemical reactions of GOPs formed at the electrode surface and thus lead to a higher selectivity for base-catalysed formation of lactate (Scheme 1).That is, an increase in alkalinity can result in an increased conversion of glyceraldehyde to dihydroxyacetone 49 and consequently in higher lactate formation as has been seen previously for the GEOR 55 and for the homogeneous oxidation of dihydroxyacetone at a higher pH. 56However, both dihydroxyacetone and glyceraldehyde can both undergo dehydration to form pyruvaldehyde 57,58 which can then react with OH − and undergo an intramolecular Cannizzaro rearrangement 59,60 to form lactate. Lactate can then undergo electrooxidation to form pyruvate in alkaline condition as seen previously on a PdNi catalyst with an average Pd:Ni composition of 87:13; 61 only a minor amount of pyruvate was detected in this study.Similarly, only minor amounts of glycolate and formate were formed in the present study.][64][65] However, the minor amount of tartronate formed is a result of the electrochemical oxidation of glycerate, likely through a glyceric aldehyde intermediate. 66 significant result from the change of atmosphere is the increase in apparent FE.The apparent FE is defined here as the ratio of the charge required for the formation of each product, given in Scheme 1, to the charge passed during the chronoamperometric experiments conducted at 0.765 and 0.865 V vs RHE.In Table I we consider the charge passed through the pathway where glyceraldehyde has undergone C-C bond cleavage through the retro-aldol reaction and keto-enol tautomerisation to form glycoaldehyde and formaldehyde.That is, in Table I, there are 8 electrons and 4 electrons passed for the formation of oxalate and formate, respectively.With oxalate having shown previously to be a result of glycolate being oxidised electrochemically 67,68 and chemically using Pd nanoparticles 63 and the likely intermediate being oxoacetate, also known as glyoxylate. 69This pathway is considered here due to the more significant proportion of tartronate when compared to glycolate and oxalate indicating that glycerate is more likely being electrochemically oxidised to tartronate rather than C-C bond cleavage through the retro-aldol reaction to form glycolate.However,  S3, the apparent FEs for the chemical pathway of C-C bond cleavage for glycerate has been included for clarity.The possibility of the decarboxylation of tartronate to form glycolate and CO 2 , shown in Scheme 1, is not considered in either Table I or Table S3.It can be seen in Table I and Table S3 that, by having an atmosphere of air, the apparent FE can be increased when compared to having an inert atmosphere.
From Table I, the apparent FE for all experiments can be seen to be close to or greater than 100%.This is likely due to chemical reactions involving dissolved oxygen as discussed above and the possibility that not all of the dissolved oxygen was removed during the purging of N 2 .Important to observe is that when there has been an introduction of air to the GEOR there has been an increase in the apparent FE for both electrode potentials studied.For the 0.865 V vs RHE study this corresponded to an increase of 6% and for the 0.765 V vs RHE study there was an increase of 7%.Therefore, having an aerated atmosphere during the GEOR can provide a further advantage over conventional water electrolysis in that it can generate a more than 100% return on the energy input in the form of electricity.If this were to be considered on an industrial scale it would correspond to an extra 14% of products formed through electrolysis in the case of 0.765 V vs RHE whilst using the same amount of energy to provide electricity which is a significant factor.

Conclusions
In this study it has been shown that a PdNi catalyst electrodeposited on Ni foam can generate current densities above 1000 mA cm −2 and 2000 mA cm −2 at an anodic potential of 1.0 and 1.3 V vs RHE (with 85% iR compensation), respectively, in atmospheres of both air and N 2 .The PdNi/Ni electrocatalyst was seen to maintain these very high current densities for 800 voltammetric cycles in an aerated atmosphere, thus demonstrating the ability of the catalyst to remain highly active for the GEOR in the presence of O 2 in solution.These very high current densities were replicated through steady state ICPCs where it was shown that 1000 mA cm −2 can be reached below an anodic potential of 0.8 V vs RHE in both atmospheres with Tafel slopes of 122 mV dec −1 and 124 mV dec −1 when in N 2 and air, respectively.The maximum current density of the PdNi/Ni electrocatalyst was similar in both an aerated and inert atmosphere, suggesting that the necessity to remove all of the O 2 from a glycerol solution can be obviated.As such, these results suggest that a scale-up of the present GEOR setup would enable a more practical and cost-efficient industrial process.
Additionally, the presence of air influenced the product selectivity in the GEOR.The dominant GOPs in this study were glycerate and lactate, with glycerate being the overall dominant product in chronoamperometric measurements at 0.765 and 0.865 V vs RHE under inert atmosphere.Introduction of air increased the selectivity for lactate in all cases and afforded a switch in product selectivity from glycerate to lactate as the main product for 0.765 V vs RHE.Furthermore, it was found that apparent Faradaic efficiencies of more than 100% were achieved at both potentials upon saturation of the glycerol solution with air.This impressive increase in GOPs production with respect to the energy input will present significant advantages if implemented on industrial scale.Thus, the present study demonstrates that the robust PdNi/Ni catalyst enables highly energy efficient electrooxidation of glycerol in the presence of air with the production of value-added chemicals at industrially relevant current densities for concurrent hydrogen generation and indicates that GEOR may serve as a practical alternative to conventional water electrolysis.
Table I.Values of the apparent FE determined by the ratio of charge from the concentration of GOPs determined by HPLC to the charge passed during the 2 h chronoamperometric experiments of 0.765 and 0.865 V vs RHE with the consideration of the pathway of glyceraldehyde undergoing electrochemical oxidation to glycerate and chemical oxidation through the retro-aldol reaction and keto-enol tautomerisation to form glycoaldehyde and formaldehyde in the process of forming glycolate and formate.

Figure 3 .
Figure 3. Voltammetry of PdNi/Ni and Pd/Ni in a 2 M NaOH and 1 M glycerol electrolyte under an atmosphere of nitrogen at a scan rate of 10 mV s −1 and normalised by the electrochemical surface area of Pd.(a) Cyclic voltammograms at 25 °C and (b) linear sweep voltammograms at 80 °C.

Figure 4 .
Figure 4. (a) Linear sweep voltammograms of PdNi/Ni in a 2 M NaOH and 1 M glycerol electrolyte at 80 °C in atmospheres of air and N 2 , respectively, at a scan rate of 10 mV s −1 .(b) Stability test.Linear sweep voltammograms of PdNi/Ni in a 2 M NaOH and 1 M glycerol electrolyte at 80 °C under an atmosphere of air at a scan rate of 100 mV s −1 .(c) iR-corrected polarisation curves of PdNi/Ni in a 2 M NaOH and 1 M glycerol electrolyte at 80 °C in atmospheres of air and N 2 .All current densities depicted are normalised by geometric area.

Figure 5 .
Figure 5. Above: GOPs determined through HPLC analysis of the GEOR on PdNi/Ni in a 2 M NaOH and 1 M glycerol electrolyte at 80 °C in atmospheres of air and N 2 at potentials of 0.865 V vs RHE and 0.765 V vs RHE.Below: Product fractions (PFs) of GOPs at potentials of (a) 0.865 V vs RHE and (b) 0.765 V vs RHE after 2 h of electrolysis.All standards used to detect the GOPs were calibrated from 0.1 mM to 10 mM with linear regression R 2 values greater than 0.99.

Scheme 1 .
Scheme 1. Reaction pathways of the GEOR on PdNi/Ni in a 2 M NaOH and 1 M glycerol electrolyte at 80 °C in atmospheres of air and N 2 at potentials.