Pt-Ni Aerogels as Unsupported Electrocatalysts for the Oxygen Reduction Reaction

The commercial feasibility of polymer electrolyte fuel cells (PEFCs) passes by the development of Pt-based, O2-reduction catalysts with greater activities and/or lower Pt-contents, as well as an improved stability. In an effort to tackle these requirements, unsupported bimetallic Pt-Ni nanoparticles (NPs) interconnected in the shape of nanochain networks (aerogels) were synthesized using a simple one-step reduction and gel formation process in aqueous solution. The products of this novel synthetic route were characterized by X-ray absorption spectroscopy to elucidate the materials’ structure. Using electrochemical experiments, we probed the surface composition of the as-synthesized aerogels and of equivalent materials exposed to acid, and concluded that a Ni-(hydr)oxide side phase is present in the aerogel with a larger Ni-concentration. Regardless of this initial surface composition, the Pt-Ni aerogels feature a ≈3-fold increase of surface-specific ORR activity when compared to a commercial platinum-on-carbon catalyst, reaching the mass-specific requirement for application in automotive PEFCs. © The Author(s) 2016. Published by ECS. This is an open access article distributed under the terms of the Creative Commons Attribution 4.0 License (CC BY, http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse of the work in any medium, provided the original work is properly cited. [DOI: 10.1149/2.0251609jes] All rights reserved.

State-of-the-art polymer electrolyte fuel cells (PEFCs) require large amounts of carbon-supported platinum nanoparticle (Pt/C) catalysts (≈0.4 mg Pt /cm 2 electrode ) 1 to account for the large overpotential of the oxygen reduction reaction (ORR). Thus, Pt contributes significantly to the fuel cell system cost, and progress to reduce its loading is crucial to meet the long-term PEFC cost target of 40 $/kW set by the U. S. Department of Energy (DOE). 2 One approach to reduce this excessive Pt-loading relies on increasing the catalysts' ORR activity, e.g. by alloying platinum with other metals like Ni, Cu and Co, to form materials which show up to one order of magnitude higher mass-specific activity than commercial Pt/C catalysts. 3 On the other hand, these carbon-supported materials suffer from significant carbonand Pt-corrosion during the standard operation of PEFCs, gradually compromising their efficiency and reliability. 4 To partially overcome stability issues, research focuses on unsupported materials (e.g. Ptcoated Ni, Co or Cu nanowires 5-7 ) besides extended metal surfaces (e.g. 3 M nanostructured thin film catalysts 8 ) or alternative supports (e.g. conductive metal oxides [9][10][11][12]. Naturally, those materials should be processable into catalytic layer architectures that provide reactant and product diffusion pathways similar to those in conventional Pt/C electrodes to guarantee high catalyst utilization and PEFC performance. 13 To meet the requirements mentioned above, un-supported bimetallic electrocatalysts with high surface area (up to ≈80 m 2 /g metal ) and nanochain network structure, referred to as aerogels, have been synthesized. [14][15][16] The synthetic routes to prepare such materials vary, but generally involve the use of stabilizing surfactants and/or organic solvents [17][18][19] that can poison the catalyst's surface and decrease its activity. In contrast to those approaches, our groups have developed a facile one-step synthesis for mono-and bimetallic aerogels based on the reduction of metal salts by NaBH 4 in aqueous solution without addition of stabilizing surfactants. 20 On the basis of previous work on Pt-Pd alloy aerogels prepared with the aforementioned method, 20 we have adapted our synthetic approach to the combination of a noble and a non-noble metal, targeting a reduced noble metal content while maintaining a high ORR activity. As we will demonstrate in this article, the combination of Pt and Ni yielded materials with a mass-normalized ORR activity ≈2-fold = These authors contributed equally to this work.
* Electrochemical Society Member. z E-mail: juan.herranz@psi.ch higher than that of a commercial Pt/C catalyst and that reaches the DOE target for automotive PEFC application. 21

Experimental
Pt-Ni hydrogels were prepared by a simple co-reduction route in aqueous solution under ambient conditions (room temperature, air). Briefly, Pt and Ni precursors were dissolved in water ( 4 , the color of the solution turned immediately from light yellow to dark brown. The solution was kept stirring for another 30 min. Afterwards, the reaction solution was divided and transferred to 100 ml vials. After about four days, black Pt x Ni hydrogel was formed at the bottom of the containers. The hydrogel parts obtained from the same synthesis were collected in a small vial and washed with water. For this, half of the supernatant was removed and replaced cautiously with fresh water. This step was repeated six times. Afterwards, the solvent was exchanged stepwise with acetone. Again, half of the supernatant was removed and replaced by acetone. This step was repeated 11 times. The resulting anhydrous gels were transferred to a critical point dryer (Critical Point Dryer 13200J-AB, SPI Supplies) operating with CO 2 .
XAS spectra at the Pt L 3 and Ni K edges were recorded at the SuperXAS beamline of the Swiss Light Source (Paul Scherrer Institut, Villigen, Switzerland), whereby the monochromator energy was calibrated using Pt and Ni foils, respectively. Measurements were performed ex-situ in transmission mode on pouch bags made from conductive Kapton tape that were filled with aerogel powder. The quick-EXAFS (QEXAFS) method was used to increase time resolution; 22 100 and 600 quick-XAS spectra were recorded at the Pt L 3 and Ni K edges, respectively. These XAS data were first analyzed with the JAQ analyzer software which allows averaging of individual XAS spectra into a single spectrum. Subsequently, Athena of the Demeter software F999 package was used for normalization and background subtraction. 23 The fit of the EXAFS oscillations was performed with Artemis 23 using a face centered cubic Pt-Ni structure for the paths description. Amplitude reduction terms were calculated from EXAFS fits of Pt and Ni reference foils assuming a coordination number of 12 and amounted to S 0,Ni 2 = 0.97 and S 0,Pt 2 = 0.85, respectively. Electrolyte solutions were prepared from NaOH · H 2 O pellets (99.995% purity, TraceSELECT, Sigma Aldrich) or 60% HClO 4 (Kanto Chemical Co., Inc.) diluted in ultrapure water (18.2 M cm, Elga Purelab Ultra). High purity N 2 (5.0), O 2 (5.0) and CO (4.7) were purchased from Messer AG, Switzerland. The benchmark catalyst for this study is a Pt/Vulcan XC-72 electrocatalyst with a weight fraction of 30% purchased from E-TEK Inc. (Lot# C0160311). The electrochemical measurements in acid were performed in a house-made three electrode glass cell. A reversible hydrogen electrode was connected to the main cell compartment through a Luggin cappilary, and a gold mesh (99,99% metals basis, Alfa Aesar) held in a separate tube in contact with the main cell served as counter electrode (CE). The working electrode (WE) was a PTFE-made, interchangeable rotating ring-disc electrode (RRDE, Pine Research Instrumentation) equipped with a Ptring and a mirror-polished, glassy carbon disk insert (5 mm diameter, HTW Hochtemperatur-Werkstoffe GmbH). Before the measurements, the WE was mounted to a polyetheretherketone (PEEK) shaft that was attached to a modulated speed rotator (both Pine Research Instrumentation). A Biologic VSP-300 was used as potentiostat, in combination with the EC-Lab V10.44 software package. The ohmic drop was determined by electrochemical impedance spectroscopy, applying a 5 mV perturbation (100 kHz to 1 Hz) at 0.45 V RHE . The setup for measurements in alkaline electrolyte was described in detail in a previous work. 24 In brief, it consisted of a house-made polytetrafluoroethylene (PTFE) cell with a four-necked glass cover. A calomel reference electrode (ALS Co., Ltd.) was kept in a separated glass holder filled with electrolyte connected to the main compartment through a Luggin capillary consisting of a FEP tube closed with a 50 μm thick Nafion membrane (Ion Power, New Castle, DE). The setup was completed by a gas bubbler and a gold mesh counter electrode.
Thin-film electrodes were prepared by dispersing catalyst materials in mixtures of ultrapure water (18.2 M cm, Elga Purelab Ultra) and isopropanol (99.9%, Chromasolv Plus for HPLC, Sigma Aldrich), followed by 10 min sonication in an ultrasonic bath (USC100T, 45 kHz, VWR). Fractions of the inks were pipetted onto the glassy carbon inset of the RRDE to obtain catalyst loadings of 30 and 15 μg Pt /cm 2 electrode for Pt-Ni aerogels and Pt/C, respectively. Subsequently, coatings were dried under a gentle N 2 -flux.

Results and Discussion
Bimetallic aerogels with various Pt:Ni atomic ratios were synthesized by the simple co-reduction process of NiCl 2 · 6H 2 O and H 2 PtCl 6 with NaBH 4 in aqueous solution described in the Experimental section.
The composition of the Pt-Ni aerogels produced in this manner was determined by inductively coupled plasma-optical emission spectrometry (ICP-OES) and compared to the values expected on the basis of the initial precursor ratios. As Figure SI-1 shows, a decrease in the extent of reduction of the Ni-precursor is observed upon preparation of Pt-Ni aerogels with x(Ni) ICP-OES -values above 25 at. %. Motivated by this result, two samples were selected for further characterization and analysis: one with a final Ni content ≤25 at. % (Pt 3 Ni, x(Ni) expected ≈ x(Ni) ICP-OES ≈ 25 at. %) and one with >25 at. % Ni (Pt 1.5 Ni, x(Ni) expected ≈ 50 at. % vs. x(Ni) ICP-OES ≈ 40 at. %). Fig. 1 shows representative transmission electron microscopy (TEM) images of the Pt 3 Ni and Pt 1.5 Ni aerogels that look similar to those obtained for aerogels with other Pt-Ni compositions (not shown here). In agreement with previous work, 20 the materials consist of a 3D structure of nanochains that are composed of interconnected/fused nanoparticles. The quality of the nanochain network is supported by the fact that no isolated nanoparticles can be observed. Additionally, nanochain diameters from TEM images based on measurement of the chain's width at different spots are on average ≈5 nm for both samples (cf. Table I).
To highlight the reproducibility of our synthesis, the good agreement between nanochain diameter distributions of different batches is illustrated in Fig. SI-2. Moreover, the surface areas of the aerogels were analyzed by N 2 -physisorption measurements (cf. Figure SI-3). Both materials exhibited a type II isotherm and a slight hysteresis indicating the presence of macropores and some mesopores. 25 Surface areas were estimated from Brunauer-Emmett-Teller (BET) plots to 55 m 2 /g Pt and 58 m 2 /g Pt for Pt 3 Ni and Pt 1.5 Ni, respectively (cf. Table I).
The samples were further characterized by X-ray diffraction (XRD) (see  Table I coincide well with the average nanochain diameters derived from TEM images. The crystallite sizes and nanochain diameters from XRD and TEM analyses were used to calculate surface areas assuming cylindrical nanowires and compared to the respective values derived from N 2physisorption and electrochemical measurements. The latter electrochemical surface areas (ECSAs) were determined by integration of the Surface area values based on the assumption of nanowires with a cylindrical geometry of diameter equal to the one averaged from XRD or TEM measurements, along with alloy density values of 19.0 (Pt 3 Ni) and 17.3 g/cm 3 (Pt 1.5 Ni). 2 Extracted from the integration of H upd desorption peaks (cf. Fig. SI-5) using the double layer current as an extended baseline and averaged on the basis of 3 independent measurements. 3 Based on the integration of the charges for the CO oxidation peak using the subsequent CV as a baseline (cf.  Table I, the ECSAs of both aerogels are similar (≈30 m 2 /g Pt ) and ≈20-40% lower than the surface areas obtained from BET plots or estimated from the XRD-or TEM-derived diameters on the basis of a cylindrical geometry approximation. This difference most likely results from the interconnectivity between the nanowires that form the tridimensional network and from the lack of wettability of micropores in the nanostructures, which are accessible to the N 2 -adsorbate used in the physisorption measurements.
To elucidate the aerogels' electronic and local structure, X-ray absorption spectroscopy (XAS) at the Ni K edge (8333 eV) was used to characterize the Pt 3 Ni and Pt 1.5 Ni aerogels, as well as a Ni(OH) 2 (β-phase) reference sample. 30 The X-ray absorption near edge structure (XANES) region within ±30 eV of the absorption edge reveals significant differences among the Pt-Ni aerogels (see Fig. 2A). The position of the absorption edge of the Pt 1.5 Ni aerogel is shifted to higher energy and the white line intensity is increased compared to that of the Pt 3 Ni aerogel, suggesting a partial oxidation of Ni in the Pt 1.5 Ni aerogel. 31 Considering the resembling nanochain diameters of both materials and the 2-fold larger Ni-concentration of the Pt 1.5 Ni sample, this difference in the Ni-oxidation states cannot be exclusively assigned to the Ni on the aerogels' surfaces (which would get oxidized upon contact with air). Alternatively, this discrepancy could be additionally related to the formation of Ni-(hydr)oxide sidephase(s) in the Pt 1.5 Ni aerogel, implying that only a fraction of the Ni is involved in the alloying with Pt. This hypothesis is further supported by the good agreement between the XANES recorded for the Pt 1.5 Ni aerogel and the corresponding fit as a linear combination of the Pt 3 Ni aerogel and Ni(OH) 2 spectra, displayed in Figure 2B. The best fit was obtained for relative weights of 75% and 25% for Pt 3 Ni and Ni(OH) 2 respectively, confirming that a significant amount of the Ni in the Pt 1.5 Ni aerogel is present in the form of a Ni-(hydr)oxide phase. Surface-sensitive X-ray photoelectron spectroscopy (XPS) in the Ni 2p 3/2 region (cf. Fig. SI-6) also corroborates this observation. The sharp peak at ≈852.6 eV in the spectrum of Pt 3 Ni is associated with Ni 0 , 32-34 whereas the Pt 1.5 Ni spectrum only contains a minor shoulder at the same energy, pointing at a larger content of Ni 0 in Pt 3 Ni when compared to Pt 1.5 Ni. Moreover, this finding is in qualitative agreement with previous work by Park et al., 26 who investigated carbon-supported Pt 3 Ni and PtNi nanoparticles of 3-4 nm diameter prepared by an aqueous-based approach similar to the one used here.
Next, the local structure of both Pt-Ni aerogels was studied by additionally analyzing the extended X-ray absorption fine structure (EXAFS) spectra collected at both the Pt L 3 (11564 eV) and the Ni K (8333 eV) edge. The outcome of the simultaneous first shell fit of the Pt L 3 and Ni K EXAFS spectra of Pt 3 Ni is summarized in Table II, and the Fourier transformed (FT) EXAFS spectra and corresponding fits are shown in Figure 3. The close agreement between the experimental data and the corresponding fits at the Pt L 3 and Ni K edge is discernable from Fig. 3 and confirmed by the low R-factor value. 35 Additionally, the properties of our Pt 3 Ni aerogel and those reported whereby CN Pt-Ni and CN Pt-Pt refer to the Pt-to-Ni and Pt-to-Pt coordination numbers, respectively, and x Ni stands for the molar concentration of Ni in the alloy (i.e. 0.25). This calculation yields a value of 0.0 ± 0.3 that is indicative of a homogeneously random alloy. 37 Our attempt at fitting the first shell EXAFS of Pt 1.5 Ni considering Pt-Pt, Pt-Ni, Ni-Pt and Ni-Ni scattering contributions was unsuccessful (not shown here), possibly due to the sample inhomogeneity caused  by the presence of a Ni-(hydr)oxide sidephase inferred above. Again, this finding is endorsed upon comparison of the Ni K edge Fourier transformed EXAFS for both aerogels (see Fig. 4), in which Pt 1.5 Ni displays a first coordination shell between 1.5-1.8 Å indicative of Ni-O scattering 31 that is only minor for the Pt 3 Ni aerogel. In a final effort to verify the hypothesized presence of two separate phases in the Pt 1.5 Ni aerogel, the surface compositions of the as prepared Pt 3 Ni and Pt 1.5 Ni aerogels were investigated electrochemically by cyclic voltammetry (CV). Since Ni is thermodynamically prone to dissolution in acid electrolyte, 39 the measurements were initially carried out in alkaline medium (0.1 M NaOH) to prevent its leaching from the alloy surface. 27 The presence of Pt and Ni on the catalyst surface is indicated by element specific features in the CVs, whereby the peaks between ≈0.05-0.4 V vs. reversible hydrogen electrode (V RHE ) are associated to H upd on polycrystalline Pt. 40 For Ni, the oxidation of Ni(OH) 2 to NiOOH at ≈1.4 V RHE and its quasi-reversible, subsequent reduction in the negative going scan provides such a characteristic feature. 41 According to the solid curves in Fig. 5 (CVs in alkaline electrolyte between 0.05 and 1.6 V RHE ), both aerogels exhibit Pt-related H upd features, but only Pt 1.5 Ni displays pronounced Ni(OH) 2 oxidation/NiOOH reduction peaks as well. Consequently, the Pt 3 Ni surface contains but traces of Ni, as one would expect from the nominal composition (Pt:Ni = 3:1) and alloy homogeneity derived from the XAS-analysis, whereas Pt 1.5 Ni contains a significant amount of surface Ni possibly in the form of a (hydr)oxide phase, further supporting the hypothesis that not all Ni is present as a Pt-Ni alloy phase.
To probe the stability of this surface Ni species in acid, the experiment was continued by immersing each electrode in a separated electrochemical cell filled with 0.1 M HClO 4 electrolyte, followed by the recording of 25 potential cycles between 0.05 and 1.0 V RHE . Subsequently, CVs were again recorded in 0.1 M NaOH to identify possible changes in the surface. Interestingly, at this point both materials display very similar voltammetric profiles, with a complete absence of redox transitions at ≈1.4 V RHE that indicates the complete dissolution of the Ni-(hydr)oxide sidephase upon conditioning in acid medium.
After this structural characterization, the ORR activity of both Pt-Ni aerogels in acid medium (0.1 M HClO 4 ) was determined using rotating disk electrode (RDE) voltammetry, whereby a commercial Pt/C catalyst (30% Pt/Vulcan XC-72) served as benchmark. The ECSA-and mass-specific ORR activities at 0.95 V RHE are shown in Fig. 6. The average ECSA values from H upd and CO stripping analysis summarized in Table I  of 0.95 V RHE was chosen instead of 0.9 V RHE because the latter was too close to the diffusion-limited regime to allow for accurate correction of mass transport losses (cf. polarization curves in Figure  SI-7). 42 Moreover, mitigating such high currents by reducing the catalyst loading on the electrode was not possible, since minimal loadings of 30 μg Pt /cm 2 electrode were required for both aerogels to fully cover the glassy carbon disk.
As displayed in Fig. 6A, both aerogels show a ≈3-fold increase of the ECSA-specific activity vs. Pt/C that can be partially assigned to their larger particle size (cf. average aerogel diameters of 5.5 nm listed in Table I, vs. the 3.4 nm of the 30% Pt/C benchmark estimated in Ref. 43). However, the limited sensitivity of the ORR-kinetics to the Pt particle size within the corresponding range of ECSA-values (≈50 m 2 /g Pt for Pt/C vs. ≈30 m 2 /g Pt for the PtNi-aerogels, see Fig. 6A) implies that the aerogels' larger size can only account for an activity increase of ≈20%. 4,27 Thus, the great majority of the observed ≈3fold activity enhancement can be explained by a down shift of the d-band center through alloying with Ni. 44 According to literature, this shift decreases the fractional coverage of spectator species at a given potential and increases the number of free sites for O 2 adsorption and subsequent ORR activity. 45 Upon conversion of these figures into mass-specific ORR activities (Fig. 6B), Pt 3 Ni and Pt 1.5 Ni reach the DOE ORR activity target at 0.95 V RHE that was extrapolated from the reported value 21 of 440 A/g Pt at 0.9 V RHE assuming a Tafel slope of 60 mV dec −1 consistent with our measurements (cf. Fig. SI-7). 46 Following the same approach, the ORR activity values at 0.95 V RHE for both Pt-Ni aerogels (≈62 ± 5A/g Pt ) can be extrapolated to ≈422 ± 34 A/g Pt at 0.9 V RHE , in good agreement with recent reports on de-alloyed Pt-Ni/C catalysts based on RDE 47 and fuel cell 48 tests. The almost identical ECSA and ORR activity values for the two aerogels agree with the above conclusion that their surface compositions become much alike after a few conditioning cycles in 0.1 M HClO 4 (see Fig. 5), in which surface-Ni and Ni(OH) 2 sidephases dissolve into the acid electrolyte.  Reported ECSA values are averaged from results for H upd and CO stripping analysis (cf. Table I).

Conclusions
In summary, we have presented a new facile synthesis for the preparation of Pt-Ni bimetallic aerogels. Extensive sample characterization suggested complete alloy formation in the case of Pt 3 Ni, along with limited alloying upon targeting larger Ni-contents using this synthetic approach. Electrochemical measurements of the aerogels' catalytic activity toward oxygen reduction revealed a 3-fold ECSA-specific activity increase when compared to commercial Pt/C catalysts that is in agreement with reports in the literature. Additionally, the massspecific ORR activity of the Pt-Ni aerogels is comparable to that of state-of-the-art, de-alloyed Pt-Ni/C materials, and meets the DOEtarget value for automotive FC application. Future work will therefore focus on investigating the stability of these Pt-Ni aerogels under PEFC-relevant working conditions and comparing their behavior to that of commercial Pt/C catalysts, as to verify the stability enhancement expected from the absence of a carbon support in these novel materials. 49