Pt-Co 3 O 4 Superstructures by One-Pot Reduction / Precipitation in Bicontinuous Microemulsion for Electrocatalytic Oxygen Evolution Reaction

: Bicontinuous microemulsions (BCME) were used to synthesize hierarchical superstructures (HSs) of Pt-Co 3 O 4 by reduction / precipitation. PtAq-CoAq, PtAq-CoOi, PtOi-CoAq and PtOi-CoOi were prepared (where Aq and Oi stand for the precursor present in aqueous or oily phase, respectively). The characterization of the Pt-Co 3 O 4 -HS conﬁrmed the formation of metallic Pt and Co 3 O 4 whose composition and morphology are controlled by the initial pH and precursor combination, determining the presence of the reducing / precipitant species in the reaction media. The electrocatalytic activity of the Pt-Co 3 O 4 -HSs for oxygen evolution reaction (OER) was investigated using linear sweep voltammetry in 0.1 M KOH and compared with Pt-HS. The lowest onset overpotentials for Pt-Co 3 O 4 -Hs were achieved with PtOi-CoOi (1.46 V vs. RHE), while the lowest overpotential at a current density of 10 mA cm − 2 ( η 10 ) was obtained for the PtAq-CoAq (381 mV). Tafel slopes were 102, 89, 157 and 92 mV dec − 1 , for PtAq-CoAq, PtAq-CoOi, PtOi-CoAq and PtOi-CoOi, respectively. The Pt-Co 3 O 4 -HSs showed a better performance than Pt-HS. Our work shows that the properties and performance of metal–metal oxide HSs obtained in BCMEs depend on the phases in which the precursors are present. spectra were recorded with a Labram HR Evolution from Horiba, equipped with a 532 nm Nd: Yag laser (50 mW); the measurements were run with 10 s acquisition time, 50 accumulations. The metal content was measured using Inductively coupled plasma atomic emission spectroscopy (ICP-AES) with a Thermo Jarrell Ash iCAP 6000 equipment; appropriate sample treatment and dilution was carried out prior to ICP-AES measurement by open-beaker digestion with aqua regia.


Introduction
Electrocatalytic reactions involving HER (hydrogen evolution reaction) and OER (oxygen evolution reaction) are required in the production of hydrogen and oxygen by water electrolysis. These reactions, relevant to renewable energy technologies, require the lowering of onset potential and increasing the reaction kinetics to make the energy balance efficient [1][2][3].
Hydrogen evolution reaction (1) 4OH − → 2H 2 O + 4e − + O 2 Oxygen evolution reaction (2) respectively. The OER kinetic measurements showed Tafel slopes of 57.55, 60.70 and 68.01 mV dec −1 for Pt-Co 3 O 4 , Pd-Co 3 O 4 and Au-Co 3 O 4 electrocatalyst, respectively. This showed that the introduction of metals into cobalt oxides promotes higher OER activity than that observed on either the metal or cobalt oxide nanomaterial alone. Furthermore, it has been shown that oxygen vacancies in Co 3 O 4 -based nanomaterials would result in a higher Co 2+ /Co 3+ ratio, but conversely provide more active sites and lead to improved performance for OER. Xiao et al. 2020 studied the real active sites during OER process using operando characterization techniques such as EIS (electrochemical impedance spectroscopy), CV (cyclic voltammetry), XPS (X-ray photoelectron spectroscopy) and showed that oxygen vacancies promote the pre-oxidation of low-valence Co (Co 2+ ) as well as the formation of cobalt oxyhydroxide (Co III -OOH) intermediate species for OER. The authors also showed that a high ratio of Co 2+ /Co 3+ plays a role in the rapid OH ion adsorption and deprotonation of CoOOH [18].
Recently, we reported for the first time the synthesis of hierarchical metallic superstructures using bicontinuous microemulsions (BCME) [28]. A nano-meso-macrostructure of Pt electrocatalyst with a nanocoral morphology was obtained from a soft and facile synthesis approach. Since BCMEs comprise both water and oil interconnected nanochannels, it was possible to obtain platinum hierarchical superstructures (Pt-HSs) using both water-soluble and oil-soluble precursors; it was observed that the resultant morphology depends on the type of Pt precursor. The resultant materials showed an efficient and stable electrocatalytic performance for hydrogen evolution reaction. In the present work, we extend this strategy for the synthesis of a hybrid metal-metal oxide HSs material. To the best of our knowledge, synthesis of Pt-Co 3 O 4 superstructures by bicontinuous microemulsion (BCME) has not been previously investigated. The synthesis of Pt-Co 3 O 4 and Pt hierarchical superstructures with various characteristics and performance is achieved by choosing different combinations of Co and Pt metallic precursors, based on their solubility, and employing bicontinuous microemulsion as a soft template, for achieving unique hierarchical morphologies.

Results and Discussion
A one-step chemical reduction/precipitation process was used to synthesize the Pt-Co 3 O 4 -HSs. The structures of the nanomaterials obtained reflect the confinement imposed by the bicontinuous microemulsion which acted as an interconnected nanocage network. The SEM images reveal that the Pt-Co 3 O 4 -HSs, as well as the reference Pt-HS solid, are made up of hierarchical nanostructures with a unique tailored arrangement. Figure 1A,B show the morphology of the Pt-HS. An interconnected-needle-like structure resembling a nanocoral is observed for Pt-HS. The SEM images for PtAq-CoAq revealed a morphological arrangement similar to that of Pt-HS ( Figure 1C,D). This seems to follow from the fact that the precursors for PtAq-CoAq and Pt-HS are dissolved in the same medium in which the reducing agent NaBH 4 was dissolved. Therefore, the resulting framework can be attributed to the soft template, morphology-directing nature of the water channels in the BCME and the presence of NaBH 4 in the same aqueous phase. Similar observation has been reported by Zhuang et al., where the authors attributed the formation of hierarchical microstructure to the use of CTAB micelles and a strong reducing agent dissolved in the same aqueous solution as the metal precursors [6]. In the reacting systems containing either or both metallic precursors dissolved in the organic phase; (PtAq-CoOi, PtOi-CoAq, and PtOi-CoOi), the SEM images ( Figure 1E-J), show a different morphology from that of Pt-HS. In the case of PtAq-CoOi, nanoplatelet-like morphology was observed, and PtOi-CoOi, exhibited a fiber-like morphology, which is attributed to the nanocage property of the oily channels in our BCME and the restricted access of reducing agent through the interface. Also, it is expected that the oil channels to be wider [28] than the water channels, since there is a larger fraction of the oil phase in the BCME, which permits the formation of different shapes when one or both precursors are present in the oil phase. Figure 2A,C present the STEM images of PtAq-CoAq and PtOi-CoAq. Agglomeration of interconnected needle-like nanomaterial was obtained for PtAq-CoAq, while PtOi-CoAq showed small nanoparticle agglomerates without a defined shape. Figure 2B,D show that PtAq-CoOi and PtOi-CoOi are made up of nanoplatelet-like and fiber-like nanomaterials respectively, which can also observed from SEM.
The Energy-dispersive X-ray spectroscopy (EDS) analysis showed the presence of Pt, Co, and O in all the samples ( Figure S1). The ICP-AES analysis results in Table 1 reveal that Pt-Co 3 O 4 -HSs synthesized using Pt aqueous precursor yielded higher amounts of Pt than the nominal Pt/Co mole ratios. This observation suggests that a higher Pt reduction was achieved for the aqueous Pt precursor than when Pt organic precursor was used. In our synthesis, two scenarios should be considered: one in which both metallic precursors are present in the same nanophase and another in which they are individually dissolved in each liquid (water or isooctane). NaBH 4 is added to the BCME as a freshly prepared solution in water, and it is expected to incorporate to the aqueous nanochannels initially. It is well known that NaBH 4 hydrolysis is a pH-dependent reaction producing basic metaborate ions (which increase the solution pH) and hydrogen gas as products [29]. When NaBH 4 is in acidic media, H 2 is produced quickly. In contrast, at neutral or alkaline pH, NaBH 4 is much more stable, leading to slow H 2 release [30]. Therefore, the initial pH in the BCME greatly affects the presence of either the hydrolysis products or reducing BH 4 − ions as majoritarian species. On the other hand, the catalytic effect of several noble metal nanoparticles and their salts on the NaBH 4 hydrolysis reaction has been widely reported, which increases H 2 production [31][32][33]. On formation of H 2 gas, microbubbles that evolve in situ in the water phase may reach the oil phase due to the low surface tension inside the BCME conferred by the nonionic surfactant [34]. These microbubbles are surrounded by a surfactant monolayer. The hydrocarbon tails orient towards the inside H 2 bubble and EO chains towards the aqueous phase. It has been reported that EO chains are capable of preferentially solvate anions [35]; thus, we hypothesize that it is feasible that these H 2 microbubbles may be dragging some reducing species (BH 4 − ) adsorbed on its surface, similarly to ion flotation phenomena [36], see Figure S2. Furthermore, H 2 itself is a slow reducing agent for platinum salts [37][38][39]. Thus, few nuclei of reduced metal in the oil phase are sufficient to act as seeds for superstructure growth when one of both reagents are present in this phase. For the preparation of the PtAq-CoAq solid, the presence of H 2 PtCl 6 .6H 2 O as a water-soluble Pt precursor lowers the pH of the BCME prior to NaBH 4 addition (pH = 1.57). Under this condition, NaBH 4 hydrolysis is promoted; thus, H 2 and OH − are produced, as evidenced by the increase of the pH after the recovery of the solids (i.e., 6.36). Both reducing species, as well as OH − ions, become available in the same nanophase, resulting in a one-step reduction (Pt)/precipitation (Co 3 O 4 ) inside the water nanochannels. In this case, the relatively high standard reduction potentials of the platinum precursor (E 0 obtain PtOi-CoAq, the microemulsion had an initial pH of 4.10, closer to BCME without any precursors. In this case, the presence of the reducing agent in the water nanochannels and the platinum precursor in the oil nanochannels suggests that the reduction reaction can only occur at the interface or by contact with the H 2 microbubbles dragging BH 4 − through the oil channels. Consequently, a lower production of metallic Pt is observed, as revealed in the lower Pt/Co ratio compared to the PtAq-CoAq solid. On the other hand, the use of cobalt 2-ethylhexanoate as Co oil-precursor to obtain either PtAq-CoOi or PtOi-CoOi increases the initial pH up to neutral values due to the presence of 2-ethylhexanoate species at the interface [40]. For the PtAq-CoOi solid, the H + provided by the acidic platinum precursor neutralizes the unreacted hydroxyl species (produced from the NaBH 4 hydrolysis) causing a lower final pH in this system. It is worth noticing that the pH of the BCME barely changes after the synthesis of Pt-HS (initial pH = 0.57; final pH = 0.69). In the case of PtOi-CoOi, the experimental Pt/Co ratio is the lowest of all the samples, since platinum is in the oil phase and the BCME showed the highest initial pH value (Table 1); thus, the production of H 2 is lessened and the presence of reducing species passing through the oil phase is not favored.
Regarding the formation of Co species, metallic Co was not obtained. For all the compositions tested, the conditions were not appropriate for the reduction of Co 2+ to metallic cobalt; instead, Co 3 O 4 was formed. The formation of this oxide depends on the pH of the BCME; its formation is favored as the pH becomes more alkaline. As observed in Table 1, the final pH of the BCME for the PtAq-CoAq sample was 6.36, leading to a low Co 3 O 4 content and the highest Pt/Co ratio obtained. The relatively low pH did not allow the formation of a large amount of Co 3 O 4 , leaving a considerable amount of precursor unreacted. In contrast, PtOi-CoOi, which started at pH = 7.34 and reached pH = 10.30 after NaBH 4 addition, allowed the formation of a larger amount of Co 3 O 4 . According to Carlo et al., 2015, Co particles were obtained with NaBH 4 . It seems that in this work the simultaneous presence of Pt and Co precursors implied a competition, in which Pt 4+ is reduced to Pt 0 and cobalt precipitates due to a consequent increase in pH as shown in Table 1. Actually, Co 3 O 4 is the product obtained when NaOH or Oxalic acid are used [25]. Overall, we observed a significant dependence of initial and final pH values on the combination of precursors used. Figure 3 shows the X-ray diffraction pattern of the PtAq-CoAq-HS compared to Pt-HS. Reference peaks for fcc Pt and cubic Co 3 O 4 are included. Figure 3a shows the diffraction peaks located at 40.01 • , 46.39 • and 67.69 • , these are assigned to the (111), (200) and (220) planes, respectively and were indexed to face-centered cubic (fcc) Pt (JCPDS file 98-002-1963). In Figure 3b, the broad diffraction peak at Bragg angle of 40.38 • corresponds to Pt (111) and a possible contribution from Co 3 O 4 (222) reflection, which is expected in PtAq-CoAq. Likewise, the diffraction peak at around 47.17 • is ascribed to fcc crystalline Pt (200), and it closely coincides with cubic crystalline Co 3 O 4 (400), while the peak at 68.39 • is ascribed to fcc crystalline Pt (220), coincident with cubic crystalline Co 3 O 4 (440). [Co 3 O 4 JCPDS file number is 04-022-7366 (space group Fd-3m)]. The Pt (111) peak is broadened and shifted to a higher angle in comparison to the reference Pt and Pt-HS peaks. Following the Debye-Scherrer equation, the crystallite size was estimated for PtAq-CoAq to be 3.21 ± 0.45 nm, while the size of Pt-HS was estimated to be 4.81 ± 0.24 nm. The Pt and Co 3 O 4 peaks superimposition, as well as the decrease in the crystallite size could be responsible for the broadening of the XRD pattern of PtAq-CoAq compared to Pt-HS. However, a possible microstrain in the Pt lattice should not be ruled out, due to replacement of some Pt by Co atoms, which have a lower atomic radius (r Co = 1.52 Å vs. r Pt = 1.77 Å) [41][42][43][44]. PtAq-CoAq exhibit a diffractogram coinciding to a greater extent with the reference Pt pattern, indicating that crystalline Pt is the main component in this sample, also in agreement with the ICP-AES results which revealed more Pt in the Pt-Co 3 O 4 -HSs synthesized with Pt aqueous soluble precursor. The diffractograms for the other Pt-Co 3 O 4 -HSs (PtAq-CoOi, PtOi-CoAq, and PtOi-CoOi) present broad, noisy, and undefined peaks ( Figure S3).  Since only one of the Pt-Co 3 O 4 -HSs presented a defined XRD pattern (PtAq-CoAq), these samples were analyzed by Raman Spectroscopy in order to confirm the nature of the cobalt phase. Figure 4 shows the Raman spectra of Pt-Co 3 O 4 -HSs; all samples presented the characteristic spectra for Co 3 O 4 with a cubic spinel structure. Five Raman bands were observed at about 186, 463, 504, 601, and 662-668 cm −1 for all the samples, which could be ascribed to the F 2g , and A 1g symmetry of crystalline cubic Co 3 O 4 phase [45,46]. In particular, the Raman band at 186 cm −1 can be assigned to F (1) 2g symmetry in tetrahedral sites (CoO 4 ), which can be attributed to vibration of Co 2+ -O 2− ; while the band at 662-668 cm −1 can be assigned to A 1g species in octahedral sites (CoO 6 ), which can be attributed to vibration of Co 3+ -O 2− [18].     Furthermore, as observed in the inset of Figure 4, the position of the A 1g band varied consistently from 668 to 662 cm −1 depending on the sample. This shift towards lower frequencies has been associated with the increase in oxygen vacancies [46,47], although other studies attribute this shift to the amorphous state of the samples [48]. In our case, XRD spectra of PtAq-CoOi, PtOi-CoAq and PtOi-CoOi indicate that the samples are rather amorphous ( Figure S3), although given the reducing conditions of the reaction, oxygen vacancies cannot be ruled out. Figure 4 shows the Raman spectra of Pt-Co3O4-HSs; all samples presented the characteristic spectra for Co3O4 with a cubic spinel structure. Five Raman bands were observed at about 186, 463, 504, 601, and 662-668 cm −1 for all the samples, which could be ascribed to the 2 (1) , 2 , 2 (2) , 2 (3) , and 1 symmetry of crystalline cubic Co3O4 phase [45,46]. In particular, the Raman band at 186 cm −1 can be assigned to 2 (1) symmetry in tetrahedral sites (CoO4), which can be attributed to vibration of Co 2+ -O 2-; while the band at 662-668 cm −1 can be assigned to 1 species in octahedral sites (CoO6), which can be attributed to vibration of Co 3+ -O 2- [18]. Furthermore, as observed in the inset of Figure 4, the position of the 1 band varied consistently from 668 to 662 cm −1 depending on the sample. This shift towards lower frequencies has been associated with the increase in oxygen vacancies [46,47], although other studies attribute this shift to the amorphous state of the samples [48]. In our case, XRD spectra of PtAq-CoOi, PtOi-CoAq and PtOi-CoOi indicate that the samples are rather amorphous ( Figure S3), although given the reducing conditions of the reaction, oxygen vacancies cannot be ruled out. The surface composition and chemical oxidation states of Pt-Co3O4-HSs and Pt-HS was determined by X-ray photoelectron spectroscopy. Figure 5A shows the XPS survey spectra which revealed the expected photoelectron peaks of Pt, C, O and Co for Pt-Co3O4-HSs, while for Pt-HS, as expected, the peak due to Co was not found. Boron or chloride species from the precursors were also disregarded. The highest intensity for the Pt 4f photoelectron peak was found in Pt-HS, while the intensity of the O 1s peak in this sample was much lower as compared to the Pt-Co3O4-HSs solids.
The Pt 4f high-resolution spectra in Figure 5B show the corresponding spin-orbit coupling (4f7/2 and 4f5/2) centered around 71 and 74 eV, respectively. For the Pt-HS, the deconvolution of the 4f7/2 peak showed the contribution of zerovalent platinum (Pt 0 ), as well as Pt 2+ forming Pt-O and Pt-OH type-bonds (71.47, 72.39, and 73.17.4 eV) [49]. The binding energy for metallic Pt 0 (71.47 eV) is slightly higher than the value expected for bulk Pt (71.2 eV) [50], as well as the values obtained in the Pt-Co3O4-HS(s) samples. This is explained to be as a result of small cluster-size effect [16,51]. Regarding the platinum species existing in the Pt-Co3O4-HSs nanomaterials, the deconvoluted 4f7/2 peak revealed The surface composition and chemical oxidation states of Pt-Co 3 O 4 -HSs and Pt-HS was determined by X-ray photoelectron spectroscopy. Figure 5A shows the XPS survey spectra which revealed the expected photoelectron peaks of Pt, C, O and Co for Pt-Co 3 O 4 -HSs, while for Pt-HS, as expected, the peak due to Co was not found. Boron or chloride species from the precursors were also disregarded. The highest intensity for the Pt 4f photoelectron peak was found in Pt-HS, while the intensity of the O 1s peak in this sample was much lower as compared to the Pt-Co 3 O 4 -HSs solids.
The Pt 4f high-resolution spectra in Figure 5B show the corresponding spin-orbit coupling (4f 7/2 and 4f 5/2 ) centered around 71 and 74 eV, respectively. For the Pt-HS, the deconvolution of the 4f 7/2 peak showed the contribution of zerovalent platinum (Pt 0 ), as well as Pt 2+ forming Pt-O and Pt-OH type-bonds (71.47, 72.39, and 73.17.4 eV) [49]. The binding energy for metallic Pt 0 (71.47 eV) is slightly higher than the value expected for bulk Pt (71.2 eV) [50], as well as the values obtained in the Pt-Co 3 O 4 -HS(s) samples. This is explained to be as a result of small cluster-size effect [16,51]. Regarding the platinum species existing in the Pt-Co 3 O 4 -HSs nanomaterials, the deconvoluted 4f 7/2 peak revealed only two contributions assigned to Pt 0 and Pt 2+ -O centered at 71. 16 On the other hand, the presence of oxidized platinum species, suggests that reduction is not fully completed under these conditions. In the case of the synthesis using chloroplatinic acid the reduction must be done from a Pt 4+ oxidation state; while in the PtCOD precursor the Pt exists as Pt 2+ . The Pt 0 /Pt 2+ ratio was dependent on the combination of precursors used; for Pt-HS, which was prepared using aqueous Pt precursor H 2 PtCl 6 • 6H 2 O, the Pt 0 /Pt 2+ ratio was 1.36; in comparison, for samples PtAq-CoAq and PtAq-CoOi, which were also prepared with the same precursor, the ratios were 0.96 and 0.7, respectively, indicating that the presence of Co precursor (and the formation of Co 3 O 4 ) inhibited the reduction of Pt 4+ to Pt 0 . On the contrary, for PtOi-CoAq and PtOi-CoOi that were synthesized using the oil-soluble platinum precursor, the Pt 0 /Pt 2+ ratio is higher. In particular, the sample synthesized with both Pt and Co oil-soluble precursors, presented the highest Pt 0 /Pt 2+ ratio with a value of 8.12. According to these results, it seems easier to reduce the oil-soluble Pt precursor to Pt 0 , despite the reducing species are less available in the oil phase. However, it should also be considered that in the case of synthesis using the platinum water-soluble precursor, the aqueous environment may favor stabilization of higher Pt oxidation state, in comparison with the oil environment.
The high-resolution spectra of Co 2p are shown in Figure 5C. The Co 2p 3/2 and Co 2p 1/2 doublet resulted from a contribution of four signals assigned to Co 3+ , Co 2+ species and satellite peaks [11,52]. The full peaks list with the corresponding assignation is shown in  [10]. Likewise, the binding energy at 778.00 eV, which indicates the presence of metallic Co 0 , was not present in the XPS spectra of the synthesized nanomaterials [4,53]. Lower Co 2+ /Co 3+ ratio obtained in the case of the nanostructures prepared with both precursors dissolved in the same phase ( Figure 5C) suggests that in these materials the oxidation of Co 3+ to Co 4+ through an electron-withdrawing inductive effect by Pt should be more favored.
In Figure 5D, The peak labelled as O3 appears at 532.65 eV for PtAq-CoAq, 532.52 eV for PtAq-CoOi, 532.72 eV for Pt-HS is due to the oxygen atoms in -CO bonds, probably arising from the surface contamination with adventitious carbon. However, the contribution of -OH species should not be ruled out [54]. In the case of the peaks at 533.53 eV in PtOi-CoAq and 533.11 eV in PtOi-CoOi solids, it could be ascribed to adsorbed molecular water [6,19,55,56]. A higher O2:O1 ratio, as estimated from the ratio between the area of the peaks, may indicate a larger amount of surface oxygen vacancies [6,18,57,58]. The higher O2:O1 ratio were achieved in those samples prepared incorporating the precursors in the same phase: for PtOi-CoOi this value is 1.51 in agreement with Raman results (higher shift in the A 1g band), while for PtAq-CoAq this value is 1.41. This result also correlates with the lower Co 2+ /Co 3+ ratio found in these solids, and it is independent of the Co 3 O 4 content, as the Pt/Co ratio is very different in both samples. The surface oxygen vacancies are useful in OER, and thus, these results would surely reflect in the different performance of the materials prepared. Therefore, having both precursors in the same phase competing for the reducing/precipitating species leads to higher surface defects in Co 3 O 4 and a higher Co 3+ population. In this report, the water splitting capability of Pt-Co3O4-HSs and Pt-HS electrocatalysts was evaluated by investigating the OER activity through linear sweep voltammetry (LSV) in 0.1 M KOH with a sweep rate of 1 mV/s. The total catalysts loading of Pt-Co3O4-HSs and Pt-HS of 200 µ g was maintained and drop-casted onto a glassy carbon electrode. As shown in Figure 6A, Pt-HS and Pt-Co3O4-HSs showed different OER performances. In Figure 6A inset, the LSV curve of the Pt-Co3O4-HSs shows anodic peaks at 1.15 V, 1.32 V, 1.19 V and 1.16 V vs. RHE for PtAq-CoAq, PtAq-CoOi, PtOi-CoAq and PtOi-CoOi, respectively, which correspond to Co 2+ oxidation to Co 3+ (Co 2+ /Co 3+ ) [6,17,59,60]. This anodic peak precedes the onset of the OER and the sudden increase in the current density. However, Pt-HS did not show a distinct anodic peak for Pt oxidation, this process might be occurring in the same range at which OER happens on the surface of Pt-HS catalyst. The overpotential corresponding to a current density of 10 mA cm −2 (η10) is of high relevance and a primary figure of merit for catalytic activity of oxygen evolution. It is a measure of the current density expected for an integrated solar water-splitting device operating at 10% solar-to-fuels efficiency. This is a common and widely used criterion for assessing OER activities [61,62]. Table 2 shows the overpotential corresponding to the current density at 10 mA cm −2 as well as the onset of oxygen evolution and Tafel slopes for Pt-Co3O4-HSs and Pt-HS. η10 for Pt-HS was calculated by extrapolation of the polarization curve. In this report, the water splitting capability of Pt-Co 3 O 4 -HSs and Pt-HS electrocatalysts was evaluated by investigating the OER activity through linear sweep voltammetry (LSV) in 0.1 M KOH with a sweep rate of 1 mV/s. The total catalysts loading of Pt-Co 3 O 4 -HSs and Pt-HS of 200 µg was maintained and drop-casted onto a glassy carbon electrode. As shown in Figure 6A, Pt-HS and Pt-Co 3 O 4 -HSs showed different OER performances. In Figure 6A inset, the LSV curve of the Pt-Co 3 O 4 -HSs shows anodic peaks at 1.15 V, 1.32 V, 1.19 V and 1.16 V vs. RHE for PtAq-CoAq, PtAq-CoOi, PtOi-CoAq and PtOi-CoOi, respectively, which correspond to Co 2+ oxidation to Co 3+ (Co 2+ /Co 3+ ) [6,17,59,60]. This anodic peak precedes the onset of the OER and the sudden increase in the current density. However, Pt-HS did not show a distinct anodic peak for Pt oxidation, this process might be occurring in the same range at which OER happens on the surface of Pt-HS catalyst. The overpotential corresponding to a current density of 10 mA cm −2 (η 10 ) is of high relevance and a primary figure of merit for catalytic activity of oxygen evolution. It is a measure of the current density expected for an integrated solar water-splitting device operating at 10% solar-to-fuels efficiency. This is a common and widely used criterion for assessing OER activities [61,62]. Table 2 shows the overpotential corresponding to the current density at 10 mA cm −2 as well as the onset of oxygen evolution and Tafel slopes for Pt-Co 3 O 4 -HSs and Pt-HS. η 10 for Pt-HS was calculated by extrapolation of the polarization curve. The overpotential was calculated by subtracting water oxidation potential (1.229 V vs. RHE) from the applied potential [63]. In comparison, the Pt-Co 3 O 4 -HSs electrocatalysts synthesized in this work showed that PtAq-CoAq had η 10 of 381 mV, while PtOi-CoAq had the highest η 10 of 513 mV. PtAq-CoOi and PtOi-CoOi had overpotential value of 406 mV and 411 mV respectively. Pt-HS in comparison to the Pt-Co 3 O 4 -HSs, exhibited the highest overpotential of 627 mV to achieve a current density of 10 mA cm −2 . A good OER electrocatalyst is expected to produce a current density of 10 mA cm −2 at low overpotential. The high surface population of Co 3+ species and the interconnected hierarchical structure in PtAq-CoAq could be favoring the Co 3+ to Co 4+ oxidation driven by the electron-withdrawing inductive effect of Pt, favoring the OER [14]. The potential at which oxygen evolution commences is called the onset potential and it varies between the electrocatalysts. According to the values in Table 2, PtOi-CoOi exhibited the lowest onset potential value of 1.46 V, while Pt-HS showed the onset at 1.56 V. On the other hand, the Tafel slope provides information associated with the rate determining steps and hence, it allows to compare the kinetics of the electrocatalytic reactions occurring on the surface of each material. The potential range at which the Tafel plot was calculated is labeled as the Tafel region in Figure 6A, while The Tafel slope of Pt-Co 3 O 4 -HSs and Pt-HS are shown in Figure 6B, these values are also listed in Table 2. PtAq-CoOi exhibited the lowest Tafel slope (89 mV dec −1 ), which is synonymous with the catalyst having a faster kinetics to evolve oxygen from water as compared to Pt-HS which had a Tafel slope of 117 mV dec −1 . Tafel slopes near 120 mV dec −1 are observed when the surface species formed just before the rate-determining step are predominant; under alkaline conditions and assuming a single site mechanism it corresponds to the following equilibrium: M + OH − ↔ MOH + e − . However, when the surface adsorbed species produced in the early stage of the OER remains predominant the Tafel slope decreases [64]. Additionally, the presence and high ratio of Co 2+ /Co 3+ ions in PtAq-CoOi indicate an increase in the pre-oxidation of Co 2+ , which enhances O 2 -related electrocatalysis [18,56]. These factors could account for the lower Tafel slope of PtAq-CoOi as compared to Pt-HS.
Catalysts 2020, 10, x FOR PEER REVIEW 12 of 21 The overpotential was calculated by subtracting water oxidation potential (1.229 V vs. RHE) from the applied potential [63]. In comparison, the Pt-Co3O4-HSs electrocatalysts synthesized in this work showed that PtAq-CoAq had η10 of 381 mV, while PtOi-CoAq had the highest η10 of 513 mV. PtAq-CoOi and PtOi-CoOi had overpotential value of 406 mV and 411 mV respectively. Pt-HS in comparison to the Pt-Co3O4-HSs, exhibited the highest overpotential of 627 mV to achieve a current density of 10 mA cm −2 . A good OER electrocatalyst is expected to produce a current density of 10 mA cm −2 at low overpotential. The high surface population of Co 3+ species and the interconnected hierarchical structure in PtAq-CoAq could be favoring the Co 3+ to Co 4+ oxidation driven by the electron-withdrawing inductive effect of Pt, favoring the OER [14]. The potential at which oxygen evolution commences is called the onset potential and it varies between the electrocatalysts. According to the values in Table 2, PtOi-CoOi exhibited the lowest onset potential value of 1.46 V, while Pt-HS showed the onset at 1.56 V. On the other hand, the Tafel slope provides information associated with the rate determining steps and hence, it allows to compare the kinetics of the electrocatalytic reactions occurring on the surface of each material. The potential range at which the Tafel plot was calculated is labeled as the Tafel region in Figure 6A, while The Tafel slope of Pt-Co3O4-HSs and Pt-HS are shown in Figure 6B, these values are also listed in Table 2. PtAq-CoOi exhibited the lowest Tafel slope (89 mV dec −1 ), which is synonymous with the catalyst having a faster kinetics to evolve oxygen from water as compared to Pt-HS which had a Tafel slope of 117 mV dec −1 . Tafel slopes near 120 mV dec −1 are observed when the surface species formed just before the ratedetermining step are predominant; under alkaline conditions and assuming a single site mechanism it corresponds to the following equilibrium: M + OH -↔ MOH + e -. However, when the surface adsorbed species produced in the early stage of the OER remains predominant the Tafel slope decreases [64]. Additionally, the presence and high ratio of Co 2+ /Co 3+ ions in PtAq-CoOi indicate an increase in the pre-oxidation of Co 2+ , which enhances O2-related electrocatalysis [18,56]. These factors could account for the lower Tafel slope of PtAq-CoOi as compared to Pt-HS.  However, in PtAq-CoAq the lowest Co 2+ /Co 3+ ratio was estimated which suggests a smaller amount of Co 2+ ( Figure 5C) is available to be effectively bounded by surface oxygen vacancies and oxidized to a higher valence [65]. This is suggested to be responsible for the higher Tafel slope of PtAq-CoAq as compared to PtAq-CoOi. Based on the morphological differences between interconnected needle-like PtAq-CoAq and nanoplatelet-like PtAq-CoOi, it is worthy to note that the electrocatalytic species in the nanoplatelet PtAq-CoOi ( Figure 1E,F and Figure 2B) are sufficiently exposed to the electrolyte, thereby improving its surface reaction rate.
On the other hand, the atomic ratio of Co 2+ /Co 3+ and O2/O1 for PtOi-CoOi is 0.62 and 1.51, with Tafel slopes of 92 mV dec −1 ; whereas PtOi-CoAq had a Co 2+ /Co 3+ ratio of 1.08 and O2/O1 of 0.75, with a Tafel slope of 157 mV dec −1 . In these two HSs, we identified that their morphology and arrangement ( Figure 1G-J and Figure 2C,D), higher Co/Pt ratio (Table 1 ICP values), high surface oxygen vacancies (Raman) as well as atomic Co 2+ /Co 3+ ratio from XPS, had significant effect on the electrocatalytic activity. These favored the early OER onset potential, comparable overpotential for 10 mA cm −2 ( Table 2) and lower Tafel slope for PtOi-CoOi when compared with PtOi-CoAq ( Figure 6B). The fibrous morphological structure of PtOi-CoOi permits rapid electron transfer through the OER catalytic active sites [6,63].
From the LSV plots for the OER process, (Figure 6A,B) Pt-HS had the highest onset potential value, a Tafel slope higher than all the Pt-Co 3 O 4 (s) evaluated except PtOi-CoAq, whereas the potential corresponding to a current density of 10 mA cm −2 for Pt-HS was obtained by extrapolation. This shows that the catalytic performance of Pt-HS is significantly low without the presence of the mixed metal oxide. This highlights that the formation of interconnected structures in which Pt atom and mixed valence Co 3 O 4 exist together enhances the performance of nanomaterials as electrocatalyst.
In general, the presence of Co 3 O 4 in the nanomaterial using either oil soluble or aqueous soluble cobalt precursor showed an enhancement in the oxygen evolution reaction electrocatalytic activity of platinum. Therefore, for practical applications, the use of PtAq-CoAq is preferred not only because it has a unique interconnected needle-like morphology but because the overpotential to achieve the primary figure of merit η 10 is low, earlier onset potential and a comparable Tafel slope (Table 3). In addition, the use of bicontinuous microemulsion as confined reaction media resulted in a good control of the morphology, as well as platinum and cobalt oxide incorporation into the nanomaterial, which are key for electrocatalytic applications.
The catalytic performance of the BCME synthesized Pt-Co 3 O 4 -HSs is comparable to reported literature for metal-metal oxides, and performed better than most metal oxides catalysts. This suggests that the Pt-Co 3 O 4 -HSs can serve as efficient electrocatalysts.
Chronoamperometry is a useful technique to evaluate the stability and durability of electrocatalysts [16]. Figure 7 shows the stability test of Pt-Co 3 O 4 -HSs and Pt-HS conducted at a potential of 1.50 V vs. RHE in 0.1 M KOH using chronoamperometry method for 4000 s. This potential was used because it is close to onset of OER as observed from the LSV ( Figure 6A). The reaction is performed at constant potential with oxygen bubbles steadily generated on the surface of the electrocatalyst. The generated bubbles attach to the surface of the electrode, which resulted in decreased current density. The steady oxygen bubbles grow bigger with time and leave the catalyst surface. This increases current density immediately after the bubbles leave the active sites (Figure 7 inset). This process repeats throughout time under which chronoamperometry is performed. This gives a fluctuating current density-time curve. The decrease in the current density was 25%, 10.7%, 2.8%, 11.6% for PtAq-CoAq, PtAq-CoOi, PtOi-CoAq, PtOi-CoOi and 20.5% for Pt-HS after a long time of polarization. However, this change shows that the electrocatalyst are stable and can be used for long duration without substantial changes. The mechanism of the OER using noble metal-cobalt oxide as electrocatalyst has been investigated [4,6,16,49]. It is reported that in metal-cobalt oxide catalysts, highly electronegative metals such as Pt facilitate the oxidation of Co 3+ to produce Co 4+ cations. This is majorly the rate limiting step, which highlights the importance of the noble metal. The electronegative metal allows for the easier formation of Co 4+ , which is the active center for OER. Higher number of active centers on the electrode correspond to increase in the OER performance. Co 4+ is more electrophilic, and will form intermediates by nucleophilic reactions with O species and OH − at the reaction interface [16,70].
The OER activity of Pt-Co 3 O 4 -HSs is better when compared to Pt-HS because multielectron transfer steps involved are promoted by the synergistic effect of the electrophilic Pt, the oxidation of Co 3+ to Co 4+ cation and surface oxygen vacancies. Additionally, the hierarchical structure of these materials made the active sites accessible for a higher electrocatalytic activity. Co 3+ (s) → Co 4+ (s) + e - Co 4+ (s) + OH -→ Co 4+ (OH)ad + e - Co 4+ (OH) ad + OH -→ Co 4+ (s) + 1/2O2 + H2O + e - The OER activity of Pt-Co3O4-HSs is better when compared to Pt-HS because multielectron transfer steps involved are promoted by the synergistic effect of the electrophilic Pt, the oxidation of Co 3+ to Co 4+ cation and surface oxygen vacancies. Additionally, the hierarchical structure of these materials made the active sites accessible for a higher electrocatalytic activity.

Synthesis of Pt-Co 3 O 4 -HSs and Pt-HS
BCME was prepared using 52% of Isooctane, 27.36% of Synperonic TM 91/5 and 20.64% of Milli-Q water. More details on this microemulsion system and its characterization can be found elsewhere [28]. Pt-Co 3 O 4 -HSs were produced by the reduction/precipitation of the aqueous or organic soluble precursors present in the BCME as follows: Pt and Co aqueous precursors both present in the aqueous phase is referred as PtAq-CoAq II. Pt and Co organic precursors both present in the organic phase is referred to as PtOi-CoOi III. Pt aqueous precursor and Co organic precursor is referred to as PtAq-CoOi IV. Pt organic precursor and Co aqueous precursor is referred to as PtOi-CoAq The preparation of BCME was carried out in 20 mL glass vials. To prepare PtAq-CoAq, proper amounts of Pt and Co aqueous precursors were dissolved in Milli-Q water to give a final aqueous mixture containing 1 wt.% of Pt and 1 wt.% of Co. This was followed by the addition of Synperonic TM 91/5 and Isooctane to obtain a final composition of 20.64% aqueous phase, 27.36% surfactant, and 52% oil phase. The mixture was vortexed for few seconds, resulting in a transparent, fluid, optically isotropic liquid phase and kept at 27 • C. A 100 µL of NaBH 4 solution with concentration equivalent to twice the total sum of Co and Pt moles, was prepared and added to the BCME within 3 s, and vortexed during 5 s. Then the reaction mixture was left to stand unstirred at 27 • C during 48 h prior to recovery and isolation of the nanomaterials. Detailed descriptions of the synthesis method for the other Pt-Co 3 O 4 -HSs are found in the Supplementary document. In brief, PtOi-CoOi, PtAq-CoOi and PtOi-CoAq are prepared by dissolving each precursor in the appropriate solvent, with the final mixture containing 1 wt.% of Pt and 1 wt.% of Co. The pH of BCMEs containing Pt and Co precursors before the addition of the NaBH 4 and after the recovery of the solids were measured using a Thermo Scientific Orion Star A211 pH Benchtop Meter.

Recovery and Isolation of Pt-Co 3 O 4 -HSs and Pt-HS
Isopropanol was added to the BCME containing the Pt-Co 3 O 4 -HSs in order to break the microemulsion and separate the solid; this was followed by washings with isopropanol, Milli-Q water and finally Isopropanol. For Pt-HS, a solution containing HNO 3 : H 2 O at 50:50 wt. ratio was used as additional washing solvent. Milli-Q water was then used to wash Pt-HS, until a neutral pH was recorded. The Pt-Co 3 O 4 -HSs and Pt-HS were then dried at room temperature, recovered and used for further characterizations.

Characterization of Pt-Co 3 O 4 -HSs and Pt-HS
The structure and morphologies of the synthesized Pt-Co 3 O 4 -HSs and Pt-HS were analyzed by scanning electron microscopy (SEM) and Scanning transmission electron microscopy (STEM) with a Field Emission Model 200 Nova NanoSEM at working voltages of 15-18 kV and equipped with an energy dispersive X-ray detector. The crystalline structure was characterized using a X-ray diffractometer PANalytical Empyrean equipped with a CuK α cathode using a continuous scan mode from 10 to 80 • (2θ) with 0.001 step size. The surface analysis was carried out using X-ray photoelectron spectroscopy (XPS) with a Thermo Fisher Scientific Escalab 250 Xi with an Al-K α (1486.7 eV) X-ray source. Raman spectra were recorded with a Labram HR Evolution from Horiba, equipped with a 532 nm Nd: Yag laser (50 mW); the measurements were run with 10 s acquisition time, 50 accumulations. The metal content was measured using Inductively coupled plasma atomic emission spectroscopy (ICP-AES) with a Thermo Jarrell Ash iCAP 6000 equipment; appropriate sample treatment and dilution was carried out prior to ICP-AES measurement by open-beaker digestion with aqua regia.

Electrocatalytic Characteristics of Pt-Co 3 O 4 -HSs and Pt-HS
The electrocatalytic activity was evaluated using CH Instruments 440C Potentiostat/Galvanostat. The experiments were conducted at room temperature. A glassy carbon electrode (GCE) was used as the working electrode. It was polished using alumina powder suspensions of different particle sizes (1 µm, 0.3 µm and 0.1 µm). A three-electrode cell was assembled with a platinum mesh counter electrode and a Mercury/Mercurous oxide Reference Electrode (Hg|HgO, 20% KOH; 0.098 V vs. RHE). The oxygen evolution reaction was investigated by linear sweep voltammetry (LSV) in a range of 0 to 1.05 V vs. Hg|HgO in 0.1 M KOH at a scan rate of 1 mV/s. The Pt-Co 3 O 4 -HSs and Pt-HS were dispersed in a 2:1 (Isopropanol: water) mixture, followed by a vortex mixing after ultrasonication. The concentration of the Pt-Co 3 O 4 -HSs and Pt-HS suspension prepared were 10 mg mL −1 . 20 µL of the solution containing the HSs was drop-casted onto the GCE electrode with an exposed cross section diameter of 7.9 mm. The HSs were allowed to dry at room temperature, after which 5 µL of a 0.05 wt.% Nafion solution was added to fix the HSs to the surface of the GCE. The stability of the HSs was investigated by chronoamperometry using a potential step of 650 mV vs. Hg|HgO in 0.1 M KOH for 4000 s. IR correction (E corr = E RHE − iR u ) was done on all the potentials, where E corr is the iR u -compensated potential, E RHE is the potential obtained from experimentally measured potential and R u is the resistance of the electrolyte determined by the iR compensation function of the CH Instruments 440 C electrochemical workstation.

Conclusions
The synthesis of Pt-Co 3 O 4 -HSs and Pt-HS using bicontinuous microemulsion was successfully carried out by a facile, fast, reproducible soft chemical method under ambient temperature and without further treatments. The characteristics, properties and performance of the obtained materials were dependent on the type of Pt and Co precursor (hydrophilic or lipophilic) and their combination; the best material was the one synthesized using both Pt and Co hydrophilic precursors. Oxygen evolution reaction performances of Pt-Co 3 O 4 -HSs were shown to be significantly faster and higher than on Pt-HS. The overpotential required to achieve 10 mA cm −2 when PtAq-CoAq catalyst was used was 381 mV, compared to 627 mV obtained when Pt-HS was used as catalyst. The onset potential of oxygen evolution on Pt-HS at 1.56 V vs. RHE was much higher than all the Pt-Co 3 O 4 -HSs (E onset of PtAq-CoAq, PtAq-CoOi, PtOi-CoAq and PtOi-CoOi were 70, 60, 80, and 100 mV, respectively, more negative than that of Pt-HS). Additionally, the Tafel slopes of Pt-Co 3 O 4 -HSs are lower and the kinetics are faster than Pt-HS. The improvement in OER activity is due to the synergistic effect of electronegative Pt on Co 3 O 4 , which allows the formation of Co 4+ species. This current research provides an opportunity to explore bicontinuous microemulsion as a structure-directing template for metal-metal oxide catalyst that are efficient in water electrolysis.