Glycerol Electro-Oxidation in Alkaline Medium with Pt-Fe/C Electrocatalysts Synthesized by the Polyol Method: Increased Selectivity and Activity Provided by Less Expensive Catalysts

We have investigated platinum catalysts containing iron as a modifier to obtain catalysts with superior electrocatalytic activity toward glycerol electro-oxidation in an alkaline medium. The electrocatalysts, supported on carbon Vulcan, were synthesized by the polyol method. The physicochemical characterization data showed that the metals were well distributed on the carbon support and had small particle size (2 nm). The Pt:Fe metal ratio differed from the nominal composition, indicating that reducing iron with platinum was difficult, even though some parameters of the synthesis process were changed. Electrochemical analyses revealed that PtFe/C was more active and stable than commercial Pt/C was, and analysis of the electrolysis by-products showed that iron addition to Pt/C boosted the glycerol conversion and selectivity for glyceric acid formation.


Introduction
Glycerol, a renewable, low-cost, and widely available compound, is a waste product from the transesterification of vegetable oils for biodiesel production-10% m/m of the products is glycerol [1][2][3][4]. Given that the rate of world biodiesel production is high, new applications for glycerol as a platform molecule for the chemical industry must be found [2][3][4][5]. The industrial demand for value-added products, including C3 compounds such as glyceric acid, tartronic acid, glyceraldehyde, dihydroxyacetone, hydroxypyruvic acid, and mesoxalic acid derived from glycerol oxidation, is high because these compounds have numerous commercial applications. They are mainly used in the manufacture of biodegradable polymers and as emulsifiers or raw materials for organic synthesis, so they have high market values [4,5]. In addition, glyceric acid can be used to prepare medicines to treat skin diseases, and it can be employed as an intermediate in serine amino acid synthesis and as an anionic monomer for packaging material, not to mention other medicinal applications [4][5][6].
Direct glycerol electro-oxidation to C3 compounds could become a clean electrosynthesis route. The coupling of this reaction in a fuel cell device allows energy to be partially recovered from the waste fuel, while value-added products are produced [4][5][6][7][8]. Recently, electroreduction processes that produce hydrogen, ammonia, or carbon products have also been investigated as promising ways to add economic value to the glycerol electro-oxidation process [5,7].
On the laboratory scale, platinum-based nanocatalysts are still the most promising anode material for glycerol electro-oxidation [9,10]. Conducting glycerol electro-oxidation

Electrocatalyst Synthesis
Pt/C, Fe/C, and Pt 50 Fe 50 /C containing 20% metal loading were prepared by the polyol method [47][48][49]. Briefly, to prepare 200 mg of each electrocatalyst, 100.0 mL of ethylene glycol (J. T. Baker, Mexico City, Mexico) was added to a three-necked flask (250 mL), the pH was adjusted to 13 with NaOH (400 mg), and the flask was subjected to ultrasound for 30 min. After NaOH was completely dissolved, 106.3 mg of H 2 PtCl 6 .6H 2 O (Sigma-Aldrich, Saint Louis, MO, USA) was quantitatively transferred to the flask and subjected to ultrasound stirring for over 30 min. To allow electrocatalyst nanoparticles to form, the mixture was maintained under N 2 flux and refluxed (130 • C) for 4.5 h. The heating was turned off, the solution was allowed to reach room temperature, and 160.0 mg of thermally activated carbon Vulcan (400 • C under N 2 atmosphere) was added [50]. The resulting carbon dispersion was placed in the ultrasound bath for 30 min. Finally, the system was magnetically stirred for 24 h to enhance nanoparticle adsorption on Carbon Vulcan. Fe/C and Pt 50 Fe 50 /C were synthesized in a similar way by using 145.0 mg of FeCl 2 .4H 2 O as the iron precursor and 31.9 mg of FeCl 2 .4H 2 O (Sigma-Aldrich, Saint Louis, MO, USA) and 80.0 mg of H 2 PtCl 6 .6H 2 O (Sigma-Aldrich, Saint Louis, MO, USA) as the iron and platinum precursors, respectively. In addition, 200 mg of Pt 50 Fe 50 /C-hydrazine was also synthesized as described above. Briefly, after the iron and platinum salts were dissolved, a stoichiometric amount of hydrazine (N 2 H 4 ) was added to the reaction flask to reduce all the metal ions.
The suspensions were filtered through a hydrophobic membrane (0.2 µm PTFE membrane, Millipore, Molsheim, France) and washed with 1.0 L of water (Milli-Q-Molsheim, France) and two 10 mL aliquots of ethanol. The membranes containing the electrocatalysts were dried in an oven at 80 • C for 4 h. After that, the electrocatalysts were removed from the membrane and properly stored.

Physical Characterization
Metal loading in the electrocatalysts was analyzed using simultaneous DTA-TGA equipment, TA Instruments (New Castle, USA), the model SDT 2960 at a constant heating rate of 10 • C min −1 from 30 to 900 • C under air atmosphere flow (100 mL min −1 ). Around 5.0 mg of the sample was placed in a Pt-crucible.
The electrocatalyst composition was analyzed by energy dispersive X-ray spectroscopy (EDX) from IXRF System Inc., model 500 Digital Process (Houston, TX, USA) using a scanning electron microscope (SEM), model EVO 50 (Cambridge, UK).
Diffraction patterns were obtained on an X-ray diffractometer (D5005 from Siemens-Munich, German) operating with Cu Kα radiation (λ = 1.5406 Å) generated at 40 kV and 40 mA. The following parameters were kept constant during the analyses: 2θ range = 30 • -90 • , step = 0.01 • s −1 , and total analysis time = 100 min. XRD data were corrected by the background and refined by fitting the experimental angular range of interest to the pseudo-Voigt1 function per crystalline peak with a computer refinement program (Profile Plus Executable, Siemens AG, Munich, German). The crystallite size was estimated by using Scherrer's equation [51]. The electrocatalyst morphology was also investigated by Transmission Electron Microscopy (TEM) using an FEI TECNAI G 2 F20 HRTEM (200 kV) microscope (Fei Company, Hillsboro, Oregon).

Electrochemical Measurements
The electrode was prepared by the deposition of 5 µL of catalyst ink solution on a previously polished glassy carbon electrode (0.07 cm 2 ). The amount of catalyst (m cat ) that was used to prepare each ink is summarized in the Supplementary Information (Table S1). The catalyst ink was obtained by suspending an ideal amount of catalyst (m cat ) in 100 µL of diluted Nafion ® solution (95 µL of isopropanol/5 µL of 5% Nafion ® ) and placing the ink suspension in an ultrasonic bath for 30 min to achieve good dispersion and homogeneity.
The electrochemical profile of the electrocatalysts was obtained by cyclic voltammetry (CV) in an N 2 -purged 0.1 mol L −1 NaOH (Mallinckrodt) solution. The measurements were conducted in a conventional electrochemical cell that included Hg/HgO/KOH (0.1 mol L −1 ) and a platinum wire as the reference and counter electrode, respectively. The activity tests were carried out by conducting CV at a scan rate of 10 mV s −1 , and chronoamperometry (CA) was accomplished in the presence of 0.5 mol L −1 glycerol and 0.1 mol L −1 NaOH at +0.7 V vs. RHE for 1.5 h. A stability test was performed by recording cyclic voltammograms at 10 mVs −1 before and after 1000 CV cycles at 50 mV s −1 in support electrolyte ( Figure S1).
The electrochemically active surface area (ECSA) of the electrocatalysts was calculated by the CO stripping technique in 0.1 mol L −1 NaOH solution; 0.30 V vs. RHE was applied for 40 min. In the first 15 min, CO was bubbled into the solution and adsorbed onto the electrocatalyst surface. Afterward, the system was purged with N 2 to remove free CO from the electrolyte solution. The cyclic voltammograms were recorded at a scan rate of 10 mV s −1 , and ECSA was obtained by integrating the CO voltammetric charge and comparing it with the standard value of 420 µC cm −2 , which corresponds to the charge required to oxidize a CO monolayer on the platinum catalyst [52,53].

Electrolysis and By-Product Determination
Four-hour-long electrolysis experiments at +0.7 V vs. RHE (−0.4 V vs. Hg/HgO/KOH, 1.0 mol L −1 ) were performed in an electrochemical cell with separate compartments for the cathode and the anode, separated by an anion exchange membrane Fumatech (35 µm thickness). The anode was fabricated by depositing 50 µL of catalyst ink (prepared as described above) over 2.0 cm 2 (both sides of 1.0 cm × 1.0 cm) Toray ® carbon paper.
The reaction products were analyzed by High-Performance Liquid Chromatography (HPLC) on a Shimadzu apparatus equipped with a refractive index (RID) detector. The products were separated on an Aminex HPX-87H column (Bio-Rad Laboratories, Hercules, USA) under isocratic conditions by using 3.33 mmol L −1 H 2 SO 4 at a flow rate of 0.6 mL min −1 at 45 • C. Every hour, a 50 µL aliquot of electrolytic solution was injected into the equipment to monitor the formation of by-products. Figure 1 presents the thermogravimetric (TG) analysis of the electrocatalysts. Only Pt/C achieved an experimental metal loading value that was equal to the nominal one (20 % wt.). In Fe/C, the amount of iron was 25% higher than that in the nominal composition. The iron content increased probably because part of the iron was converted to its oxide during the TG experiment carried out in an air atmosphere. We also observed that co-reduced iron and platinum in the metal loading decreased significantly compared to the nominal value, indicating that incomplete reduction occurred during the synthesis, or that no metal nanoparticles adsorbed over the carbon support. To clarify this point, we conducted atomic absorption analysis of the solution filtered during the synthesis that employed hydrazine. This analysis indicated that about 70% of the iron remained in the solution. This analysis corroborated the statement above; nevertheless, given that the analytical equipment does not differentiate between Fe 2+ and Fe 0 , the iron precursor may not have been reduced, or the iron nanoparticles may not have been adsorbed. Considering that hydrazine addition during the synthesis increased the amount of iron in the PtFe electrocatalyst, we can infer that non-reduction might have been the main cause of the lower iron content. carbon support to only 6%, as shown by TG analysis. To overcome this low metal and iron content, we repeated the synthesis in the presence of hydrazine by chang parameters such as the pH, amount of hydrazine, reflux time, and temperature, results did not improve. Our findings agreed with the literature reports about th culty in achieving the nominal composition when dealing with platinum and iro lysts [28][29][30][31].   Figure 2 illustrates the XRD diffraction patterns and structural profiles of the catalysts. Fe/C did not display any characteristic peak of the cubic structure (spac Im-3m), which should have appeared at 43.50°, 63.20°, and 79.85° (01-071-4650). Th the as-prepared nanoparticles were in an amorphous state. Pt/C exhibited the ch istic peaks of the crystalline structure of face-centered cubic (fcc) platinum. Th were related to the reflection planes  Table 1 summarizes the physical characterization of the electrocatalyst nanoparticles. The EDX analysis of Pt x Fe y /C evidenced that it was difficult to co-reduce iron with platinum by using the polyol method. The amount of iron obtained at the end of the syntheses was only 5%, rather than the targeted 50%. When we added hydrazine as a reducing agent (Pt x Fe y /C-hydrazine), the iron content in the composition increased to 25% (half of the desired percentage). However, the use of hydrazine reduced the metal loading over the carbon support to only 6%, as shown by TG analysis. To overcome this low metal loading and iron content, we repeated the synthesis in the presence of hydrazine by changing the parameters such as the pH, amount of hydrazine, reflux time, and temperature, but the results did not improve. Our findings agreed with the literature reports about the difficulty in achieving the nominal composition when dealing with platinum and iron catalysts [28][29][30][31].  Figure 2 illustrates the XRD diffraction patterns and structural profiles of the electrocatalysts. Fe/C did not display any characteristic peak of the cubic structure (space group Im-3m), which should have appeared at 43.50 • , 63.20 • , and 79.85 • (01-071-4650). Therefore, the as-prepared nanoparticles were in an amorphous state. Pt/C exhibited the characteristic peaks of the crystalline structure of face-centered cubic (fcc) platinum. The peaks were related to the reflection planes (111) Pt95Fe05/C and Pt75Fe25/C-hydrazine did not present the characteristic peaks of platinum or the iron structure because low metal loading over the carbon support did provide enough material for defined diffraction data to be obtained, even at a slow rate (0.03° s −1 ). The other reasons for the absence of peaks could be the small crysta size or the attainment of an amorphous material [54]. For Pt95Fe05/C and Pt75Fe25/C-hy zine, we were able to identify a not well-defined peak at around 40°, which overlap with the peak at around 46°, both of which are related to the platinum structure. Howe the resolution was poor, which prevented the crystallite size or lattice parameter f being calculated. Figure 3 shows the TEM images and histograms of Pt/C, Pt95Fe05/C, and Pt75Fe hydrazine. The nanostructures were randomly dispersed on the carbon support, and l clusters did not emerge in the materials. The high-resolution TEM images evidenced g crystallinity and a sphere-like shape for the platinum-based electrocatalysts. The par size remained at around 1.8 nm, and the EDX analysis from TEM characterization ag well with the results reported in Table 1. Pt 95 Fe 05 /C and Pt 75 Fe 25 /C-hydrazine did not present the characteristic peaks of the platinum or the iron structure because low metal loading over the carbon support did not provide enough material for defined diffraction data to be obtained, even at a slow scan rate (0.03 • s −1 ). The other reasons for the absence of peaks could be the small crystallite size or the attainment of an amorphous material [54]. For Pt 95 Fe 05 /C and Pt 75 Fe 25 /C-hydrazine, we were able to identify a not well-defined peak at around 40 • , which overlapped with the peak at around 46 • , both of which are related to the platinum structure. However, the resolution was poor, which prevented the crystallite size or lattice parameter from being calculated. Figure 3 shows the TEM images and histograms of Pt/C, Pt 95 Fe 05 /C, and Pt 75 Fe 25 /Chydrazine. The nanostructures were randomly dispersed on the carbon support, and large clusters did not emerge in the materials. The high-resolution TEM images evidenced good crystallinity and a sphere-like shape for the platinum-based electrocatalysts. The particle size remained at around 1.8 nm, and the EDX analysis from TEM characterization agreed well with the results reported in Table 1.

Electrochemical Characterization
Figure 4a depicts the cyclic voltammogram of the platinum-based electrocatalysts in the supporting electrolyte (NaOH 0.5 mol L −1 ), which were recorded from 0.05 to 1.0 V vs. RHE. Peaks in the region between 0.0 and +0.4 V are related to hydrogen adsorption/desorption on the platinum surface [55]. At around +0.80 V (positive sweep) and +0.70 V (negative sweep), the peaks correspond to platinum oxide formation and reduction, respectively [36]. The addition of even small amounts of iron to platinum, as in the case of Pt95Fe05/C, shifted the hydrogen adsorption/desorption and platinum oxide formation/reduction peaks to less-positive values compared to that of bare platinum. Furthermore, the presence of iron is evidenced in pseudo-capacitive double layer region between 0.4 and 0.6 V vs. RHE by the presence of the Fe +2 /Fe +3 redox couple. Moreover, reversible peaks appeared near 0.6 V vs. RHE (forward scan) and 0.45 V vs. RHE (backward scan), which are associated with iron ion oxidation and reduction and iron oxide formation. Hydrazine addition enhanced the amount of iron in the electrocatalyst (Pt75Fe25/C-hydrazine), so the redox peaks became more evident. The characteristic peaks of hydrogen adsorption/desorption on platinum sites became poorly defined and the pseudo-capacitive area became wider as an effect of a greater amount of conductive support (Vulcan Carbon) since this composition presented the smallest metal loading value (only 6% wt.), as evidenced by TGA analysis.
The formation of iron oxides at lower potentials is interesting for electrocatalysis: the availability of oxygenated groups facilitates the electro-oxidation of carbonic species on platinum, a consequence of the bifunctional mechanism, during which iron/iron oxide  Figure 4a depicts the cyclic voltammogram of the platinum-based electrocatalysts in the supporting electrolyte (NaOH 0.5 mol L −1 ), which were recorded from 0.05 to 1.0 V vs. RHE. Peaks in the region between 0.0 and +0.4 V are related to hydrogen adsorption/desorption on the platinum surface [55]. At around +0.80 V (positive sweep) and +0.70 V (negative sweep), the peaks correspond to platinum oxide formation and reduction, respectively [36]. The addition of even small amounts of iron to platinum, as in the case of Pt 95 Fe 05 /C, shifted the hydrogen adsorption/desorption and platinum oxide formation/reduction peaks to less-positive values compared to that of bare platinum. Furthermore, the presence of iron is evidenced in pseudo-capacitive double layer region between 0.4 and 0.6 V vs. RHE by the presence of the Fe +2 /Fe +3 redox couple. Moreover, reversible peaks appeared near 0.6 V vs. RHE (forward scan) and 0.45 V vs. RHE (backward scan), which are associated with iron ion oxidation and reduction and iron oxide formation. Hydrazine addition enhanced the amount of iron in the electrocatalyst (Pt 75 Fe 25 /C-hydrazine), so the redox peaks became more evident. The characteristic peaks of hydrogen adsorption/desorption on platinum sites became poorly defined and the pseudo-capacitive area became wider as an effect of a greater amount of conductive support (Vulcan Carbon) since this composition presented the smallest metal loading value (only 6% wt.), as evidenced by TGA analysis. the electrocatalysts containing iron provided greater activity for glycerol electro-oxidation in an alkaline medium, especially in the case of Pt75Fe25/C-hydrazine. In the presence of 0.5 mol L −1 glycerol (Figure 4b), alcohol adsorption on the platinum catalytic sites suppressed the hydrogen adsorption/desorption peaks (+0.05 to +0.4 V vs. RHE). The onset potential for glycerol electro-oxidation was +0.4 vs. RHE for all the investigated electrocatalysts, with very slight variations. An explanation for this is that at high pH values, the supporting electrolyte furnishes OH-species for the alcohol molecule electro-oxidation [10,11,58]. This is a very different situation from acid solutions, where hydroxyl species (-OH) are obtained by water activation on the electrocatalyst surface, and the catalyst site plays a key role in the adsorption and bond cleavage of the water molecules. In a basic solution, the catalytic activity does not depend entirely on this process, so alcohol electro-oxidation starts nearly at the same potential, irrespective of the electrocatalyst composition [59,60].

Electrochemical Characterization
Shifting toward more positive values, we observed that the oxidation peak was reactivated in the negative scan. This behavior originates from platinum surface regeneration due to the removal/oxidation at higher potentials of adsorbed carbonaceous species produced by direct glycerol electro-oxidation. In the reverse scan, glycerol was electrooxidized on the renewed surface in a typical alcohol electro-oxidation process [14,55,61,62].
The presence of iron boosted the platinum catalytic activity, as shown in Figure 4b. This behavior depended on the amount of iron. As observed, just 5% wt. of iron (Pt95Fe05/C) increased the peak current (at +0.87 V vs. RHE) by 35% compared to that of pure platinum. When the iron content was higher, as in the case of Pt75Fe25/C-hydrazine, the effect was even more significant: the peak current was three times higher compared to that of Pt/C ( Table 2).  The formation of iron oxides at lower potentials is interesting for electrocatalysis: the availability of oxygenated groups facilitates the electro-oxidation of carbonic species on platinum, a consequence of the bifunctional mechanism, during which iron/iron oxide provides oxygenated species for the oxidation process [36,56,57]. As shown in Figure 4b, the electrocatalysts containing iron provided greater activity for glycerol electro-oxidation in an alkaline medium, especially in the case of Pt 75 Fe 25 /C-hydrazine.
In the presence of 0.5 mol L −1 glycerol (Figure 4b), alcohol adsorption on the platinum catalytic sites suppressed the hydrogen adsorption/desorption peaks (+0.05 to +0.4 V vs. RHE). The onset potential for glycerol electro-oxidation was +0.4 vs. RHE for all the investigated electrocatalysts, with very slight variations. An explanation for this is that at high pH values, the supporting electrolyte furnishes OH-species for the alcohol molecule electro-oxidation [10,11,58]. This is a very different situation from acid solutions, where hydroxyl species (-OH) are obtained by water activation on the electrocatalyst surface, and the catalyst site plays a key role in the adsorption and bond cleavage of the water molecules. In a basic solution, the catalytic activity does not depend entirely on this process, so alcohol electro-oxidation starts nearly at the same potential, irrespective of the electrocatalyst composition [59,60].
Shifting toward more positive values, we observed that the oxidation peak was reactivated in the negative scan. This behavior originates from platinum surface regeneration due to the removal/oxidation at higher potentials of adsorbed carbonaceous species produced by direct glycerol electro-oxidation. In the reverse scan, glycerol was electro-oxidized on the renewed surface in a typical alcohol electro-oxidation process [14,55,61,62].
The presence of iron boosted the platinum catalytic activity, as shown in Figure 4b. This behavior depended on the amount of iron. As observed, just 5% wt. of iron (Pt 95 Fe 05 /C) increased the peak current (at +0.87 V vs. RHE) by 35% compared to that of pure platinum. When the iron content was higher, as in the case of Pt 75 Fe 25 /C-hydrazine, the effect was even more significant: the peak current was three times higher compared to that of Pt/C ( Table 2).  Table 3 compares the peak current (I f ) obtained in the oxidation of glycerol by different materials investigated in the literature. Even considering different experimental conditions, such as concentration of glycerol, support electrolyte and scanning speed, we observed that the PtFe catalyst presents results that are compatible with other compositions that have been investigated. In these cases, the activity is better compared to those ones that use only noble metals (Pt, Pd, and Au) as catalysts.  We carried out chronoamperometry (CA) to investigate the electrochemical performance of electrocatalysts during steady-state glycerol electro-oxidation. We chose the applied potential (+0.7 V vs. ERH) on the basis of the glycerol cyclic voltammograms and because it corresponds to a potential where all the electrocatalysts exhibited electrocatalytic activity toward glycerol electro-oxidation (Figure 4b). Figure 5 shows that the j-t curves were consistent with the cyclic voltammetry curves. The activity of the platinum-based electrocatalysts decreases due to the poisoning of electrocatalytic sites by CO-like species and other glycerol electro-oxidation intermediates, a typical behavior of the electro-oxidation of small molecules on platinum-based catalysts [73]. We carried out chronoamperometry (CA) to investigate the electrochemical perfor mance of electrocatalysts during steady-state glycerol electro-oxidation. We chose the ap plied potential (+0.7 V vs. ERH) on the basis of the glycerol cyclic voltammograms and because it corresponds to a potential where all the electrocatalysts exhibited electrocatalytic activity toward glycerol electro-oxidation (Figure 4b). Figure 5 shows that the jcurves were consistent with the cyclic voltammetry curves. The activity of the platinum based electrocatalysts decreases due to the poisoning of electrocatalytic sites by CO-like species and other glycerol electro-oxidation intermediates, a typical behavior of the elec tro-oxidation of small molecules on platinum-based catalysts [73]. By analyzing iron addition to platinum in the steady-state studies, we observed tha small amounts (5% wt.) of this metal improved the catalytic activity compared to that o bare platinum. This indicates that even a small amount of iron was enough to enhance the reaction kinetics given that its presence helped to renew the platinum catalytic sites. When the amount of iron was 25% wt., the platinum catalytic activity of increased significantly and remained stable. As pointed out earlier, in an alkaline medium, platinum catalytic activity depends on the amount of iron. Proper amounts of iron can act directly on glycero By analyzing iron addition to platinum in the steady-state studies, we observed that small amounts (5% wt.) of this metal improved the catalytic activity compared to that of bare platinum. This indicates that even a small amount of iron was enough to enhance the reaction kinetics given that its presence helped to renew the platinum catalytic sites. When the amount of iron was 25% wt., the platinum catalytic activity of increased significantly and remained stable. As pointed out earlier, in an alkaline medium, platinum catalytic activity depends on the amount of iron. Proper amounts of iron can act directly on glycerol electro-oxidation through a bifunctional and electronic effect [63]. Figure 6 presents the electrochemical surface area (ECSA) of the electrocatalysts obtained by the CO stripping technique in 0.1 mol L −1 NaOH at 10 mV s −1 after the CO monolayer was adsorbed. The onset of CO oxidation was at around +0.45 V vs. RHE for Pt/C, and iron addition shifted the CO oxidation to less-positive potentials, 0.3 and 0.40 V vs. RHE for Pt 75 Fe 25 /C and Pt 95 Fe 05 /C, respectively. Moreover, the CO oxidation peak was broader for large potential ranges after iron was added in the electrocatalyst composition. The greater the amount of iron there was, the broader the peak was. This behavior has been reported before and is due to the presence of transition metals on the platinum surface. These metals can modify the platinum electronic structure and contribute to the oxidation of small molecules [16][17][18]74,75]. The greater the amount of iron there was, the broader the peak was. This behavior has been reported before and is due to the presence of transition metals on the platinum surface. These metals can modify the platinum electronic structure and contribute to the oxidation of small molecules [16][17][18]74,75]. Figure 6b shows the ESCA measured from the CO stripping experiments and the H2 adsorption-desorption region. The data obtained from both methods agree, indicating that both approaches can be used for estimating the electrochemical area. As shown in Table 2, the ECSA of the PtFe-based electrocatalysts is much higher and remained at around 400 m 2 gPt −1 compared to 240 m 2 gPt −1 for pure platinum. Although the active area is not the only effect that must be evaluated regarding catalyst efficiency, it can significantly contribute to catalytic activity. As seen from the results above, the proper amount of iron and a higher ECSA provide an electrocatalyst with significant activity for glycerol electro-oxidation. Figure 7 shows the glycerol electro-oxidation by-products and their respective percentages obtained at 4 h of electrolysis (+0.7 V vs. ERH) in the presence of one of the prepared electrocatalysts. In this analysis, we only compared the best iron-containing composition (Pt75Fe25/C-hydrazine) with Pt/C and the commercial compositions PtRu/C and Pt/C both from BASF. All the investigated compositions electro-oxidized between 5% and 10% of the glycerol present in the solution. In all the cases, the major product was glyceric acid, and the presence of iron or ruthenium in the electrocatalyst improved the selectivity for this by-product, as presented in Figure 7.  Figure 6b shows the ESCA measured from the CO stripping experiments and the H 2 adsorption-desorption region. The data obtained from both methods agree, indicating that both approaches can be used for estimating the electrochemical area.

Electrolysis Study
As shown in Table 2, the ECSA of the PtFe-based electrocatalysts is much higher and remained at around 400 m 2 g Pt −1 compared to 240 m 2 g Pt −1 for pure platinum. Although the active area is not the only effect that must be evaluated regarding catalyst efficiency, it can significantly contribute to catalytic activity. As seen from the results above, the proper amount of iron and a higher ECSA provide an electrocatalyst with significant activity for glycerol electro-oxidation. Figure 7 shows the glycerol electro-oxidation by-products and their respective percentages obtained at 4 h of electrolysis (+0.7 V vs. ERH) in the presence of one of the prepared electrocatalysts. In this analysis, we only compared the best iron-containing composition (Pt 75 Fe 25 /C-hydrazine) with Pt/C and the commercial compositions PtRu/C and Pt/C both from BASF. All the investigated compositions electro-oxidized between 5% and 10% of the glycerol present in the solution. In all the cases, the major product was glyceric acid, and the presence of iron or ruthenium in the electrocatalyst improved the selectivity for this by-product, as presented in Figure 7. Pt75Fe25/C provided the best glycerol electro-oxidation as compared to those of the other compositions, including the commercial ones, Pt/C and PtRu/C (BASF) (Figure 7b). Another important feature concerned the effective cost of the electrocatalysts: for instance, Pt75Fe25/C is much less expensive than Pt/C with Ru as a co-catalyst is. Moreover, iron addition to platinum improved the glycerol electro-oxidation and selectivity for glyceric acid formation. Figure 8 illustrates the proposed mechanism for glycerol electro-oxidation based on the literature [21,25,62,68]. Considering the products observed by liquid chromatography analysis, the mechanism for PtFe/C follows the solid arrows represented by route II; under the investigated experimental conditions, oxidation occurs via glyceraldehyde formation and not through dihydroxyacetone formation (arrows with dashed lines). In fact, oxidation occurs with the primary carbon to form glyceric acid as the major product. On the other hand, DHA formation (route I) is only observed for Pt/C and PtRu/C. Electrolysis with platinum and ruthenium as electrocatalysts applied to glycerol electro-oxidation also gave glyceric acid as the major product [62] and a mixture of DHA, glyceric acid, and tartronic acid, [68], indicating that both routes I and II may occur simultaneously on Pt and PtRu electrocatalysts. Pt 75 Fe 25 /C provided the best glycerol electro-oxidation as compared to those of the other compositions, including the commercial ones, Pt/C and PtRu/C (BASF) (Figure 7b). Another important feature concerned the effective cost of the electrocatalysts: for instance, Pt 75 Fe 25 /C is much less expensive than Pt/C with Ru as a co-catalyst is. Moreover, iron addition to platinum improved the glycerol electro-oxidation and selectivity for glyceric acid formation. Figure 8 illustrates the proposed mechanism for glycerol electro-oxidation based on the literature [21,25,62,68]. Considering the products observed by liquid chromatography analysis, the mechanism for PtFe/C follows the solid arrows represented by route II; under the investigated experimental conditions, oxidation occurs via glyceraldehyde formation and not through dihydroxyacetone formation (arrows with dashed lines). In fact, oxidation occurs with the primary carbon to form glyceric acid as the major product. On the other hand, DHA formation (route I) is only observed for Pt/C and PtRu/C. Electrolysis with platinum and ruthenium as electrocatalysts applied to glycerol electro-oxidation also gave glyceric acid as the major product [62] and a mixture of DHA, glyceric acid, and tartronic acid [68], indicating that both routes I and II may occur simultaneously on Pt and PtRu electrocatalysts.

Electrolysis Study
The selectivity for the by-products of glycerol electro-oxidation in the basic medium depends on the metal associated with platinum or palladium, the applied potential, and the OH concentration, as shown in Table 3. Jin et al. [42] also investigated different Pt x Fe y /CeO 2 compositions. They found that glyceric acid was the main product (75%) of glycerol electro-oxidation, which is the same finding as that observed in our study. Fashedemi et al. [34] evaluated FeCo@Fe@Pd/MWCNT-COOH core-shell catalysts in a passive glycerol cell operated until it became inactive. These authors reported that the presence of iron and cobalt contributes to complete glycerol electro-oxidation, with mostly carbonate ions being formed. Zhou et al. [25] studied the PtAg composition in electrolysis experiments at different potentials for 2 h and obtained dihydroxyacetone as the main product. Furthermore, the authors showed that the KOH concentration, the glycerol concentration, and the electrolysis time significantly affected the distribution of the electro-oxidation by-products and the mechanism through which they were formed.  The selectivity for the by-products of glycerol electro-oxidation in the basic medium depends on the metal associated with platinum or palladium, the applied potential, and the OH concentration, as shown in Table 3. Jin et al. [42] also investigated differen PtxFey/CeO2 compositions. They found that glyceric acid was the main product (75%) o glycerol electro-oxidation, which is the same finding as that observed in our study Fashedemi et al. [34] evaluated FeCo@Fe@Pd/MWCNT-COOH core-shell catalysts in a passive glycerol cell operated until it became inactive. These authors reported that the presence of iron and cobalt contributes to complete glycerol electro-oxidation, with mostly carbonate ions being formed. Zhou et al. [25] studied the PtAg composition in electrolysis experiments at different potentials for 2 h and obtained dihydroxyacetone as the main product. Furthermore, the authors showed that the KOH concentration, the glycerol con centration, and the electrolysis time significantly affected the distribution of the electro oxidation by-products and the mechanism through which they were formed.
Zhou et al. [26] evaluated how nickel or ruthenium addition affects the platinum cat alytic activity and found that both metals favor C3-products and that selectivity depends on the applied potential. González-Cobos et al. [19] evaluated how bismuth addition af fects platinum and palladium catalysts and obtained glyceraldehyde as the main produc after electrolysis for 4 h.

Conclusions
The results obtained herein show that addition of iron, an inexpensive metal, is a good alternative for improving the platinum catalytic activity without affecting the pro cess' efficiency. The main findings of this investigation are:

•
The polyol method provides good metal distribution and small nanoparticles (<2 nm in PtFe electrocatalysts. • Iron addition to platinum increases the active area and the catalytic activity for glyc erol and CO electro-oxidation.

•
Initially, glycerol electro-oxidation depends on the type of co-catalyst. PtFe favors the Zhou et al. [26] evaluated how nickel or ruthenium addition affects the platinum catalytic activity and found that both metals favor C3-products and that selectivity depends on the applied potential. González-Cobos et al. [19] evaluated how bismuth addition affects platinum and palladium catalysts and obtained glyceraldehyde as the main product after electrolysis for 4 h.

Conclusions
The results obtained herein show that addition of iron, an inexpensive metal, is a good alternative for improving the platinum catalytic activity without affecting the process' efficiency. The main findings of this investigation are:

•
The polyol method provides good metal distribution and small nanoparticles (<2 nm) in PtFe electrocatalysts. • Iron addition to platinum increases the active area and the catalytic activity for glycerol and CO electro-oxidation. • Initially, glycerol electro-oxidation depends on the type of co-catalyst. PtFe favors the glyceraldehyde oxidation route, and glyceric acid is the main product. • PtFe presents high selectivity for high-value-added C3-products.
Data Availability Statement: All data included in this study are available upon request upon contacting the corresponding author.