Electroanalytical Assessment of the Effect of Ni:Fe Stoichiometry and

Jun 19, 2017 - U.S. Naval Research Laboratory, Surface Chemistry Branch (Code ... with Ni:Fe stoichiometry as a function of surface area, crystallite ...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/Langmuir

Electroanalytical Assessment of the Effect of Ni:Fe Stoichiometry and Architectural Expression on the Bifunctional Activity of Nanoscale NiyFe1−yOx Jesse S. Ko,‡,# Christopher N. Chervin,‡ Mallory N. Vila,‡,# Paul A. DeSario,‡ Joseph F. Parker,‡ Jeffrey W. Long,‡ and Debra R. Rolison*,‡ ‡

U.S. Naval Research Laboratory, Surface Chemistry Branch (Code 6170), Washington, DC 20375, United States S Supporting Information *

ABSTRACT: Electrocatalysis of the oxygen evolution reaction (OER) and oxygen reduction reaction (ORR) was assessed for a series of Ni-substituted ferrites (NiyFe1−yOx, where y = 0.1 to 0.9) as expressed in porous, high-surface-area forms (ambigel and aerogel nanoarchitectures). We then correlate electrocatalytic activity with Ni:Fe stoichiometry as a function of surface area, crystallite size, and free volume. In order to ensure in-series comparisons, calcination at 350 °C/air was necessary to crystallize the respective NiyFe1−yOx nanoarchitectures, which index to the inverse spinel structure for Fe-rich materials (y ≤ 0.33), rock salt for the most Ni-rich material (y = 0.9), and biphasic for intermediate stoichiometry (0.5 ≤ y ≤ 0.67). In the intermediate Ni:Fe stoichiometric range (0.33 ≤ y ≤ 0.67), the OER current density at 390 mV increases monotonically with increasing Ni content and increasing surface area, but with different working curves for ambigels versus aerogels. At a common stoichiometry within this range, ambigels and aerogels yield comparable OER performance, but do so by expressing larger crystallite size (ambigel) versus higher surface area (aerogel). Effective OER activity can be achieved without requiring supercritical-fluid extraction as long as moderately high surface area, porous materials can be prepared. We find improved OER performance (η decreases from 390 to 373 mV) for Ni0.67Fe0.33Ox aerogel heat-treated at 300 °C/Ar, owing to an increase in crystallite size (2.7 to 4.1 nm). For the ORR, electrocatalytic activity favors Fe-rich NiyFe1−yOx materials; however, as the Nicontent increases beyond y = 0.5, a two-electron reduction pathway is still exhibited, demonstrating that bifunctional OER and ORR activity may be possible by choosing a nickel ferrite nanoarchitecture that provides high OER activity with sufficient ORR activity. Assessing the catalytic activity requires an appreciation of the multivariate interplay among Ni:Fe stoichiometry, surface area, crystallographic phase, and crystallite size. (OER).1,3 Significant progress has been made in exploring various metal oxides (e.g., NiOx and CoOx), mixed metal oxides (e.g., NiFeOx, CoFeOx, NiCoOx, and NiLaOx), and metal (oxy)hydroxides (e.g., NiOxHy, FeOxHy, and Ni0.71Fe0.29OxHy)4 that demonstrate OER activity and durability in alkaline electrolytes.5−8 The main goal with new OER electrocatalysts is to lower the overpotential required to realize device-relevant rates of OER activity in order to improve device efficiency and enhance long-term stability by avoiding high positive potentials that can corrode electrode components, such as the carbon that typically provides the electron-conductive and catalyst-dispersal functions in the electrode.9,10

1. INTRODUCTION Near the top of the (long) wish list for functional components that enable next-generation rechargeable metal−air batteries lies “bifunctional catalyst”. A simplifying desireone in which a single nanoscale catalytic material integrated within the electron, ion, and molecule transport−active air cathode minimizes the voltage penalty to reduce molecular oxygen during discharge and concomitantly minimizes the voltage penalty to evolve molecular oxygen during recharge. All while alternating each half reaction in the same air-breathing electrode structure at high activity and with operational durability. The efficient electrocatalysis of redox reactions involving molecular oxygen also determines the performance of many other electrochemical energy-storage and -conversion systems.1,2 While technologies such as hydrogen fuel cells are limited at low temperature by the recalcitrant two- and fourelectron oxygen-reduction reaction (ORR), electrolyzers require efficient turnover for the oxygen-evolution reaction © XXXX American Chemical Society

Special Issue: Fundamental Interfacial Science for Energy Applications Received: March 27, 2017 Revised: May 17, 2017

A

DOI: 10.1021/acs.langmuir.7b01046 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

0.66, and 0.9 were synthesized by stirring for 15 min the respective stoichiometric amounts of NiCl2·6H2O and FeCl3·6H2O in 53 mL of ethanol (Warner-Graham Company; anhydrous, 200 proof) and subsequently adding 22 mL of propylene oxide (C3H6O, SigmaAldrich, ReagentPlus ≥99%). Warning: propylene oxide is volatile and a probable human carcinogen (IARC Group 2B carcinogen) and should only be handled in a f ume hood with proper protective equipment. The mixture was then stirred for 20 min and sat unstirred overnight to complete gelation. In order to prepare ambigels (wet gels dried at ambient pressure), the NiyFe1−yOx gels were thoroughly washed with hexanes for 3 d before drying under a flowing N2 atmosphere at 50 °C. For processing aerogels, the NiyFe1−yOx gels were washed with copious amounts of acetone for 6 d before placing in an autoclave, rinsing with liquid CO2 until no acetone odor was perceived, and increasing temperature to take CO2 supercritical.17 The ambigels and aerogels were heated for 2 h at 350 °C under static air, with a ramp rate of 2 °C min−1. Note that during the stirring, aging, and washing steps, the container was covered with Parafilm to prevent evaporation of the solvent and the epoxide. 2.2. Materials Characterization of Nickel Ferrite Aerogels and Ambigels. The physicochemical properties of the NiyFe1−yOx ambigel and aerogel series were determined using scanning electron microscopy, energy-dispersive spectroscopy, X-ray diffraction, and N2 physisorption. The respective samples were imaged using scanning electron microscopy (SEM; Carl Zeiss Leo Supra MM microscope operating at 15 keV) by affixing the respective powders onto aluminum stubs with double-sided carbon tape. The chemical stoichiometry of the NiyFe1−yOx samples was verified to be within 10% of the targeted value using energy-dispersive spectroscopy (EDS; JSM-7600F JEOL microscope operating at 15 keV; 2000× magnification). Powder X-ray diffraction was performed using a Rigaku SmartLab X-ray diffractometer with a Cu Kα radiation source (λ = 1.5406 Å). Scans were recorded from 20−80°2θ with a 0.02° step size and an integration time of 1 s per step. The surface area and pore size distributions were measured by N2-porosimetry using a Micromeritics ASAP 2020 porosimeter. All samples were degassed with N2 under vacuum for 12 h at 80 °C before obtaining an isotherm. The Brunauer−Emmet−Teller (BET) model was used to determine the surface area of all samples and the pore size distribution was calculated according to the Barrett−Joyner−Halenda (BJH) analysis using a cylindrical geometry. 2.3. Electroanalysis of OER and ORR at Nickel Ferrite Aerogels and Ambigels. All electrochemical characterizations were performed with a Gamry Reference 600 potentiostat using a three-electrode configuration with 1 M KOH electrolyte continuously purged with O2, a gold-foil counter electrode, and all potentials referenced to a frit-isolated Hg/HgO electrode (138 mV vs NHE). Glassy-carbon (GC) rotating disk electrodes (5 mm diameter, nominal area = 0.196 cm−2, Pine Instruments) were coated with a thin, uniform composite film (mass loading of ∼0.04 mg cm−2). To prepare the working electrode, the GC electrodes were first polished with a 1.0 μm and then a 0.05 μm alumina slurry. In order to prepare the electrocatalyst film, equal amounts of electrocatalyst powder and Vulcan-XC72 carbon (VC; Cabot) were ground together in a mortar and pestle for several minutes, and then 5 mg of this mixture were suspended in a multicomponent solution (3 mL) containing Nafion perfluorinated resin (5 wt % solution of Nafion in aliphatic alcohols and water; Aldrich), isopropanol, and water (18 MΩ cm). The v:v ratio of this solution was adjusted for architectural type, namely, ambigel: 0.4 vol % Nafion, 20 vol % isopropanol, and 79.6 vol % water (0.4/20/79.6); aerogel: 0.8 vol % Nafion, 20 vol % isopropanol, and 79.2 vol % water (0.8/20/79.2). The suspension was sonicated for 1 h, stirred for 1 h, and then sonicated for an additional hour to produce a homogeneous ink. A 10 μL aliquot of this suspension was deposited onto an inverted GC electrode, which was then rotated at 600 rpm (Pine Instruments MSRX Speed Control Rotator) for 30 min to dry the suspension uniformly in ambient air. Electrodes were prepared in triplicate for each NiyFe1−yOx electrocatalyst material to gauge catalytic reproducibility.

The Ni-substituted inverse spinel ferrites (NiyFe1−yOx, where the Ni:Fe stoichiometry is ≥1) exhibit promising OER activity in aqueous alkaline electrolytes, operating at 10 mA cm−2 with less than 400 mV overpotential, while also being based on relatively low-cost metals.3,5,7 The OER activity of Ni-rich ferrites is attributed to in situ formation of oxyhydroxide species at the Ni.11,12 The addition of Fe into NiOx has been shown to drastically lower the overpotential of OER, and even biphasic NiFe2O4/NiO materials can show comparable activity.3−5,13,14 Bifunctional ORR+OER performance may even be feasible with nickel ferrites,8 as recently reported for the catalyst synthesized either in a three-dimensionally ordered mesoporous form15 or cross-linked with multiwall carbon nanotubes.16 We recently showed that when one designs NiFe2Ox as an aerogel with 11 nm crystallites, an Fe-rich stoichiometry (Ni:Fe = 0.5) can catalyze OER in alkaline electrolyte at rates and overpotentials comparable to Ni-rich, nonporous ferrite nanoparticles.17 The activity for OER correlates with the physisorption-accessible surface area of the crystallized nanomaterial (increasing from dense nanoparticulate to xerogel to aerogel) rather than the hydroxyl content of the temperature/ atmosphere-processed nanoscale catalyst. When molecularly accessible surface area matters, the bicontinuity of aerogel pore and solid networks ensures that the high surface area of the covalently bonded nanoparticulate network is accessible to reactants via the three-dimensionally open, high mesoporous volume (>0.5 cm3 g−1).18 While aerogels are typically fabricated by removing the pore fluid in the wet gel using supercritical fluid (SCF) extraction,19 gels filled with a nonpolar, low surface−tension pore fluid can be dried under ambient-pressure conditions to produce a dry solid that we designated in 2000 as “ambigel”.20 This benchtop protocol allows laboratories without SCF equipment to prepare aerogel-like nanoarchitectures that minimize pore collapse and densification relative to xerogels, which are prepared by ambient drying from high surface−tension pore fluids such as acetone, water, and alcohols.21 In this report, we use the synthetic flexibility afforded by epoxide-based sol−gel chemistry and ambient pressure vs supercritical drying to tune both the Ni:Fe stoichiometry (y = 0.1 to 0.9 in NiyFe1−yOx) as well as the free volume and specific surface area to create ten stoichiometric/nanoarchitectural variants. We then examine the OER activity in alkaline electrolyte for the resulting series of catalysts. Activity is assessed for catalyst+carbon composite inks deposited on glassy-carbon rotating disk electrodes using the overpotential necessary to reach a current density of 10 mA cm−2, a widely used performance metric for OER activity.5−7 We find that ambigel formulations of ferrites where Ni:Fe is ≥0.5 provide effective OER activity approaching that obtained by aerogel formulation. We also explore the ORR activity of our series of nanoarchitectured NiyFe1−yOx catalysts, identifying particular stoichiometries and nanoarchitectures that maximize bifunctional activity in this series.

2. MATERIALS AND METHODS 2.1. Preparation of Nickel Ferrite Aerogels and Ambigels. NiyFe1−yOx ambigels and aerogels were prepared using an epoxideinitiated sol−gel process, similar to previously published protocols.17,22−24 Briefly, nickel(II) chloride hexahydrate (NiCl2·6H2O, Sigma-Aldrich, 99.9%) and iron(III) chloride hexahydrate (FeCl3· 6H2O, Sigma-Aldrich, 97%) were used as the sources of Ni and Fe for the synthesis. Amorphous gels of NiyFe1−yOx where y = 0.1, 0.33, 0.5, B

DOI: 10.1021/acs.langmuir.7b01046 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

(pO2, e.g., heating under flowing Ar) crystallizes NiFe2Ox into the inverse spinel structure with high OER activity. However, heating at 400 °C in low pO2 extrudes Ni metal out of the Ferich 1Ni:2Fe nanoscale oxide and detrimentally affects its OER activity.17 Heating the Ni-rich series at 300 °C/low pO2 does extrude Ni metal (Figure S1), so for this study, we chose to heat the entire NiyFe1−yOx series for both architectural types at 350 °C under static air. As verified by XRD (Figure 1), calcining at 350 °C/air

Current−potential measurements for electrocatalytic O2 evolution (OER) and reduction (ORR) at the NiyFe1−yOx series were obtained using linear-sweep voltammetry (LSV). All measured and applied potentials were converted to overpotentials (η) relative to the thermodynamic OER potential using the equation

E = [1.23 − 0.138 − (0.059 × pH)] V

(1)

where 138 mV vs NHE is the potential of the Hg/HgO reference electrode, 1.23 V vs NHE is the potential for O2 evolution at pH 7, and the pH of the 1 M KOH electrolyte was measured prior to each experiment (Fisher Scientific AB15 pH Meter). The LSV measurements were carried out at 10 mV s−1, which is a rate sufficiently slow to mirror steady-state conditions from measurements obtained using chronopotentiometry or chronoamperometry.25 All electrochemical OER and ORR potentials were iR compensated for uncompensated resistance as determined by measuring the impedance at the opencircuit potential. The Koutecky−Levich theory was also applied to determine the number of electrons (ne) passed during ORR, eq 226 1/J = 1/JL + 1/JK = 1/Bω1/2 + 1/JK

(2) −2

where J is the experimental current density (mA cm ), JK is the kinetics-limited current density (JK = neFkCO), JL is the diffusionlimited current density, and B is defined as

B = 0.62neFCO(DO)2/3 υ−1/6

(3) −1

where F is the Faraday constant (96 485 C mol ), CO is the bulk concentration of O2 (7.8 × 10−7 mol cm−3), DO is the diffusion coefficient of O2 (1.8 × 10−5 cm2 s−1), and υ is the kinematic viscosity of water (0.01 cm2 s−1).27 In aqueous alkaline solutions, the ORR mechanism is mainly dictated by two pathways: (i) the four-electron reduction pathway (E° = 0.401 V vs SHE)

O2 + 2H 2O + 4e− ↔ 4OH−

Figure 1. X-ray diffraction patterns of the NiyFe1−yOx series for (a) ambigels and (b) aerogels with corresponding references: NiFe2O4 (ICDD #00-074-2081); γ-Fe2O3 (ICDD #00-039-1346); NiO (ICDD #00-078-0643).

(4)

or (ii) the two-electron reduction pathway (E° = −0.076 V vs SHE)

O2 + 2H 2O + 2e− ↔ HO2− + OH−

rather than at either 300 °C/low pO2 or 300 °C/air suffices to crystallize all of the amorphous, as-prepared NiyFe1−yOx nanoarchitectures throughout the stoichiometric range of interest without giving rise to extruded Ni metal (Figure S2). Although the catalytic activity for the oxide in a given stoichiometry and architecture may not be optimized by heating in air to 350 °C, we deemed it a necessary protocol in order to ensure in-series comparisons were made with crystalline materials. Unambiguous phase assignment for these materials is complicated by their nanoscale nature (which induces peak broadening) and by the fact that the potential oxide crystal habits in this Ni−Fe−O composition space have phasecommensurate reflections. The positions of the reflections from 30° to 65° 2θ for γ-Fe2O3 (ICDD #00-039-1346), NiFe2O4 (ICDD #00-074-2081), and NiO (ICDD #00-0780643) are difficult to distinguish for bulk polycrystalline forms and difficult-to-impossible to distinguish in the presence of peak broadening (Figure 1). One key difference that can be observed between the phase-commensurate reflections occurs between 70° and 80° 2θ where NiO has higher-order X-ray reflections (75° and 79° 2θ) with modest rather than trace intensity. The primary structural distinction for this series of NiyFe1−yOx nanoarchitectures is whether the material crystallizes as either inverse spinel (γ-Fe2O3 or NiFe2O4) or rock salt NiO (Schematic 1). The Fe-rich nanoarchitectures (1Ni:9Fe and 1Ni:2Fe) exhibit the inverse spinel structure with a comparable crystallite size for ambigel and aerogel (14.4 and 14.1 nm, respectively,

(5)

for the formation of hydroperoxyl species; this two-electron pathway can be followed by a further two-electron reduction of the hydroperoxyl species (E° = 0.878 V vs SHE)

HO2− + H 2O + 2e− ↔ 3OH−

(6)

All polarization curves were recorded at varying rotation speeds from 800 to 2400 rpm (ambigels, Figure S10; aerogels, Figure S11). The ne was determined by measuring the slope (1/B) from a plot of the inverse current vs the inverse square of the rotational speed, ω−1/2 (Figures S12a,b). As a control experiment, Koutecky−Levich analysis was also applied to the data obtained at an unmodified glassy carbon substrate (Figure S13).

3. RESULTS AND DISCUSSION 3.1. Physicochemical Properties. The series of NiyFe1−yOx nanoarchitectures prepared using the epoxideinitiated sol−gel method with y set to 0.1, 0.33, 0.5, 0.67, and 0.9 was thermally treated and analyzed by X-ray diffraction and N2 porosimetry to determine crystallinity, pore volume, and surface area as a function of Ni content. The thermal treatment conditions were chosen to induce sufficient crystallinity to ensure high catalytic activity,17 while avoiding calcination temperatures that create particle over-ripening or Ni metal phase segregation. As established in our previous study of NiFe2Ox aerogel,17 calcining at a temperature of 300 °C in air yields a poorly crystalline material with high specific surface area, but poor OER activity, whereas thermally processing at the same temperature under low partial pressure of oxygen C

DOI: 10.1021/acs.langmuir.7b01046 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

throughout the entire Ni:Fe stoichiometric range. For the aerogel series, the pore volume is at least 2× larger, but the lowest pore volumes (1.09 and 0.64 cm3 g−1) occur for the extremes (1Ni:9Fe and 9Ni:1Fe, respectively). The pore volume in the intermediate range is comparable for 1Ni:1Fe and 2Ni:1Fe aerogels at ∼1.35 cm3 g−1, while NiFe2O4 exhibits the highest pore volume (1.88 cm3 g−1) in the aerogel series. Field-emission scanning electron microscopy also supports the differences between the architectural expressions with respect to free volume and specific surface area. The highly textured, fluffier, pore-solid architecture of the aerogel (Figure 2a) vs the

Scheme 1. Crystallographic Representation of an (a) Inverse Spinel and (b) Rock Salt Structure

Table 1). The most Ni-rich nanoarchitecture (9Ni:1Fe) indexes to the rock salt NiO phase with equivalent crystallite sizes for ambigel and aerogel (3.6 and 3.7 nm, respectively) but with fourfold smaller crystallites compared to Fe-rich nanoarchitectures. For intermediate Ni:Fe stoichiometry, i.e., 1Ni:1Fe and 2Ni:1Fe, the nanoarchitecture-invariant comparability in crystallite size is lost, with the crystallite size consistently smaller when the final pore−solid architecture is aerogel (Table 1). Evidence of coexistence of inverse spinel and rock salt phases can be seen in the diffraction patterns of the ambigels with intermediate Ni:Fe stoichiometry. The 2Ni:1Fe ambigel combines features of the sharp reflections seen for the ∼14 nm crystallites in the inverse spinel-dominant Fe-rich ambigels combined with the broad reflections characteristic of the small crystallites in the NiO-dominant 9Ni:1Fe ambigel. Using the broad reflections characteristic of the small NiO crystallites as a marker, one can discern evidence of two-phase coexistence for even the 1Ni:1Fe ambigel. The broad reflections/small crystallite size innate to the aerogels does not allow an assignment of either single-phase inverse spinel or even twophase inverse spinel/rock salt coexistence for these nanoarchitectures. Ambigels and aerogels are differentiated by their respective porosity and specific surface area, with aerogel expressions typically conferring higher free volume, higher specific surface area, and a broader pore size distribution (Figure S3). The specific surface area of the NiyFe1−yOx aerogels in the intermediate stoichiometric range is indeed higher (1.5− 2.7×) relative to the respective ambigel stoichiometric counterpart (Table 1). Within this range (1Ni:2Fe, 1Ni:1Fe, and 2Ni:1Fe), the specific surface area for both types of nanoarchitectures monotonically increases with increasing Ni content (Figure S4). The porosity of the NiyFe1−yOx nanoarchitectures also fundamentally differs (Table 1) with the ambigel series exhibiting a comparable pore volume (∼0.29 cm3 g−1)

Figure 2. Scanning electron micrographs of Ni2FeOx (2Ni:1Fe) (a) aerogel and (b) ambigel.

more compact configuration of the ambigel is micrographically discernible (Figure 2b). Lower magnification micrographs verify the uniformity of the two respective morphologies across square micrometers (Figure S5). 3.2. Electroanalytical Assessment of OER Activity. The OER activity of the NiyFe1−yOx series was determined at a rotating disk working electrode using linear-sweep voltammetry (LSV) at 10 mV s−1 (Figures 3a,b). As verified by chronoamperometry, LSV at this scan rate provides an accurate representation of steady-state OER conditions.17,25 As established in previous reports for thin-film electrocatalysts5−7 and carbon+catalyst inks,17 the overpotential (η) measured at a current density of 10 mA cm−2 defines an electroanalytical figure-of-merit (FOM) for OER activity that facilitates comparison of different catalyst types. This geometric areanormalized performance metric, however, fails to capture how efficiently the catalyst is utilized,28 so we also report corresponding voltammograms normalized to the mass of the electrocatalyst (Figure S6). The mass-normalized data follow the same trends seen in Figure 3, indicating that catalyst particles throughout the carbon+catalyst composite coating are fully accessible.

Table 1. Summary of Physicochemical Properties of NiyFe1−yOx Ambigels and Aerogels Ambigel

Fe-Rich ↑ Ni-Rich ↓

Aerogel

material stoichiometry

Ni/Fe ratioa

average crystallite sizeb (nm)

BET surface area (m2 g−1)

BJH pore volume (cm3 g−1)

Ni/Fe ratioa

average crystallite sizeb (nm)

BET surface area (m2 g−1)

BJH pore volume (cm3 g−1)

1Ni:9Fe 1Ni:2Fe 1Ni:1Fe 2Ni:1Fe 9Ni:1Fe

0.12 0.48 0.91 1.7 8.1

6.1 13.1 14.7 9.2 3.6

73 70 121 153 166

0.24 0.28 0.31 0.27 0.33

0.12 0.5 1.1 2 10

14.1 4.9 2.2 2.7 3.7

107 187 223 231 204

1.09 1.88 1.4 1.3 0.64

a Ni:Fe stoichiometry was verified using energy-dispersive spectroscopy (EDS). bCrystallite sizes were calculated from the full-width at halfmaximum of the (440) reflection of inverse spinel NiFe2O4 and/or (220) reflection of rock salt NiO (see Figure 1) using the Scherrer equation.

D

DOI: 10.1021/acs.langmuir.7b01046 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Figure 3. Oxygen-evolution activity assessed at glassy-carbon rotating disk electrodes for the 350 °C/air-treated NiyFe1−yOx series prepared as coatings deposited from carbon + catalyst inks. (a,b) Linear-sweep voltammetry at 10 mV s−1 in O2-purged 1 M KOH with the electrode rotating at 1600 rpm for (a) ambigels and (b) aerogels; insets expand the measured current density over the overpotential range of 320−420 mV. (c) Specific surface area versus current density at an overpotential of 390 mV for the NiyFe1−yOx ambigel and aerogel series.

Of note is the fact that for each Ni:Fe stoichiometry in the intermediate range, the lower-surface-area ambigel achieves roughly the same current density for OER at 390 mV as that obtained at the higher-surface-area aerogel. The competitive performance of an ambigel compared to an aerogel highlights the advantage of using a simple processing approach (bypassing the need for SCF extraction) to obtain a porous electrocatalyst with sufficient surface area to provide reasonably high OER activity. One possible explanation for the high performance seen at 1Ni:1Fe and 2Ni:1Fe NiyFe1−yOx ambigels may arise from the increased crystallite size of these materials relative to their aerogel cousins. To reach comparable performance, the aerogels appear to compensate for their >2.5× smaller crystallite size with >50% increases in specific surface area. This interplay of crystallite size and surface area made possible by expressing the catalyst as a mesoporous nanoarchitecture offers additional design, synthetic, and processing flexibility in the effort to achieve high OER activity. After benchmarking the OER activity for 350 °C/air-treated NiyFe1−yOx nanoarchitectures, the as-prepared 2Ni:1Fe ambigels and aerogels were down-selected for heat treatment under argon to impose more crystalline order, in keeping with our previous study that found high OER activity for 300 °C/Artreated NiFe2O4 aerogels with ∼11 nm inverse spinel crystallites.17 The contrast in OER activity between 350 °C/ air and 300 °C/Ar treatments (Figure 4) finds little improvement for the ambigel (η improves only slightly from 389 to 383 mV, respectively), but significant improvement for the aerogel (η decreases from 390 to 373 mV, respectively).

The Fe-rich ambigels (1Ni:9Fe and 1Ni:2Fe) cannot reach the FOM at η < 500 mV, whereas the Ni-rich ambigels (9Ni:1Fe and 2Ni:1Fe) can reach 10 mA cm−2 at ≤400 mV. Activity improves (η decreases) for the NiyFe1−yOx ambigels at y > 0.5 with 2Ni:1Fe achieving the lowest η for the ambigel series at 389 mV. The trend is similar for the aerogel series with 2Ni:1Fe also exhibiting the lowest η at a comparable 392 mV. Unlike the NiFe2Ox ambigels, FOM can be assessed for the aerogel expression, albeit at a high η (467 mV), which is consistent with our previous finding that expressing NiyFe1−yOx in an aerogel architecture markedly improves OER activity over less porous expressions.17 A recent study of nickel ferrites with varying Ni composition, but not expressed in high surface area forms, found a different stoichiometric performance trend, with consistently low OER activity for both the most Ni-rich and Ferich materials (1Ni:9Fe and 9Ni:1Fe, respectively).29 An OER activity−surface area trend was found in our previous study of 300 °C/Ar heat-treated, nanocrystalline NiFe2Ox, where the aerogel yielded the highest current density at 360 mV compared to xerogel and dense nanoparticulate counterparts.17 For the 350 °C/air-crystallized intermediate stoichiometric nanoarchitectures (1Ni:2Fe, 1Ni:1Fe, and 2Ni:1Fe), a strong linear correlation is again observed between current density and specific surface area (Figure 3c), but with one working curve expressed for the ambigels and a different one for the aerogels. The BET surface area-normalized OER activity (Figure S7) follows the activity trends seen for both geometric area-normalized and mass-normalized activity (summarized in Table S2), indicating consistent access to the catalytic centers. E

DOI: 10.1021/acs.langmuir.7b01046 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

with increasing Ni content, whereas for the aerogels, η increases with increasing Ni content. Rather than evaluate ORR performance only by the measured η, we analyzed the data using Koutecky−Levich theory to determine the number of electrons passed (ne).26,27 In general, the Fe-rich ambigels and aerogels (1Ni:9Fe and 1Ni:2Fe) yield ne > 2 (∼2.6), indicating that reduction beyond the two-electron formation of hydroperoxyl species occurs, indicating a pathway exists to electrogenerate OH− from the peroxide. We posit that the higher activity of the Fe-rich samples arises from a heterogeneous Fenton reaction involving Fe2+ and H2O2 species, such as reported for Fe2+ or Co2+substituted magnetite in which the presence of Ni2+ inhibits H2O2 reactions.31 As the Ni-content increases beyond y = 0.5, ORR activity slightly decreases for both types of nanoarchitectures, with ne ∼ 1.9, indicating a two-electron reduction pathway dominates. The most Ni-rich ambigel and aerogel also follow the two-electron reduction pathway (Table 2). Bifunctional ORR+OER activity may be possible by choosing a nickel ferrite nanoarchitecture that provides high OER activity with sufficient ORR activity. A bifunctional performance metric is assessed by taking the difference in FOM-specified overpotential for OER and ORR (ΔEOER−ORR). Some of the most noted metal oxide bifunctional electrocatalysts exhibit ΔEOER−ORR on the order of 0.98 V,8,9 with one of the bestperforming bifunctional catalysts demonstrating 0.84 V (NiCo2O4).8,9,32 For the FOM determined in this study (with ηORR taken at −1 mA cm−2 and ηOER taken at 10 mA cm−2), the 1Ni:1Fe aerogel has the lowest ΔEOER−ORR at 895 mV, followed by 898 mV for the 2Ni:1Fe ambigel. The ΔEOER−ORR for all other NiyFe1−yOx nanoarchitectures remains ≤0.98 V (Table 2). Note that ηORR is typically determined at −3 mA cm−2,9 so a −1 mA cm−2 threshold for the NiyFe1−yOx series will poise ηORR at a lower value, increasing ΔEEOER−ORR. However, the mass-normalized activity we obtain for ORR is competitive with literature reports for ferrites.16

Figure 4. Oxygen-evolution activity assessed for the 2Ni:1Fe series heated under argon at 350 °C prepared as carbon + catalyst inks coating glassy-carbon rotating disk electrodes. (a,b) Linear-sweep voltammetry at 10 mV s−1 in O2-purged 1 M KOH with the electrode rotating at 1600 rpm for (a) ambigels and (b) aerogels; insets expand the measured current density over the overpotential range of 320−420 mV.

The XRD analysis of the 300 °C/Ar-treated 2Ni:1Fe nanoarchitectures may explain the marked improvement in OER activity of aerogel over ambigel. The crystallite size increases from 2.7 to 4.1 nm for the aerogel, although the peaks are still too broad to assign to inverse spinel or rock salt, but Ni metal reflections are not present (Figure S8). Our prior work with ferrite aerogels found that argon atmosphere and moderate temperature readily crystallize the amorphous particles in the covalently bonded network,17,23,24,30 in keeping with the increase in crystallite size upon a 300 °C/Ar treatment. However, a 300 °C/Ar treatment of 2Ni:1Fe ambigel decreases crystallite size from 9.2 to 2.9 nm and now Ni metal reflections are present in addition to inverse spinel and rock salt reflections (Figure S8). The decreased crystallite size coupled with extrusion of Ni metal indicates that the ferrite particles in the covalently bonded network are unable to sustain biphasic crystallization, further degrading OER activity. 3.3. Electroanalytical Assessment of ORR Activity. In order to assess the bifunctionality of the NiyFe1−yOx nanoarchitectures, each material’s respective η for ORR was measured at −1 mA cm−2 (Figure 5; Table 2); corresponding

4. LESSONS LEARNED Within this study focused on Ni:Fe stoichiometry and pore− solid architectural expression, crystallite size matters and if one cannot get large enough crystalliteswhich for OER in the nickel ferrite system seems to be on the order of 10 nmthen a way needs to be found to express high surface area, preferably 100 m2 g−1 or higher. Getting both is even better. As we reported earlier, the high surface area innate to aerogels does allow the designer to step away from a Ni-rich stoichiometry (e.g., 2Ni:1Fe), where NiFe2Ox aerogel with 11 nm inverse spinel crystallites performed competitively to literature reports for Ni2FeOx.17 This hand-off between crystallite size and surface area is exemplified by the matched ambigel-vs-aerogel OER activity at a given Ni:Fe stoichiometry across the entire Ni:Fe stoichiometric range of this study. In the realm of bifunctionality, competitive ΔEOER−ORR with 1Ni:1Fe aerogels is feasible, but as surface area decreases, such as occurs upon moving to the ambigel expression, Ni-rich stoichiometry can be used to bring back OER activity. Note that even with a two-phase coexistence of inverse spinel and rock salt crystal habits in the Ni-rich nanoarchitectures, effective bifunctionality is attainedas long as treatment conditions do not induce extrusion of Ni metal. The sol−gel route to porous nanoarchitectures affords the catalyst designer an extra dimension of control when tailoring electrocatalysts to the demands inherent to reducing and

Figure 5. Oxygen-reduction activity assessed at glassy-carbon rotating disk electrodes for the 350 °C/air treated NiyFe1−yOx series prepared as coatings deposited from carbon + catalyst inks. Linear-sweep voltammetry at 10 mV s−1 in O2-purged 1 M KOH with the electrode rotating at 1600 rpm for (a) ambigels and (b) aerogels.

voltammograms normalized to the mass of the electrocatalyst are shown in Figure S9. For the two Ni-rich compositions (2Ni:1Fe and 9Ni:1Fe), both ambigels and aerogels exhibit a relatively constant η of 509 mV. The remaining, more Fe-rich ambigels and aerogels (1Ni:9Fe, 1Ni:2Fe, and 1Ni:1Fe) show opposing trends in overpotential: for the ambigels, η decreases F

DOI: 10.1021/acs.langmuir.7b01046 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir Table 2. Oxygen-Evolution and -Reduction Activity of NiyFe1−yOx Ambigels and Aerogels Ambigel material stoichiometry Fe-rich ↑ Ni-rich ↓ a

OER η at 10 mA cm−2 (mV)

1Ni:9Fe 1Ni:2Fe 1Ni:1Fe 2Ni:1Fe 9Ni:1Fe

b b

397 389 401

Aerogel

ORR η at −1 mA cm−2 (mV)

ORR ne

482 497 523 509 506

2.9 2.4 1.6 1.9 2.0

ΔEOER−ORRa (mV)

OER η at 10 mA cm−2 (mV)

b

b

b

467 404 392 403

920 898 907

ORR η at −1 mAcm−2 (mV)

ORR ne

529 512 491 520 501

2.5 2.4 2.2 1.8 2.2

ΔEOER−ORRa (mV) b

979 895 912 904

OER potentials were measured at 10 mA cm−2; ORR potentials were measured at −1 mA cm−2. bElectrocatalyst could not reach 10 mA cm−2. Oxides and (Oxy)hydroxides: Activity Trends and Design Principles. Chem. Mater. 2015, 27, 7549−7558. (5) McCrory, C. C. L.; Jung, S.; Peters, J. C.; Jaramillo, T. F. Benchmarking Heterogeneous Electrocatalysts for the Oxygen Evolution Reaction. J. Am. Chem. Soc. 2013, 135, 16977−16987. (6) McCrory, C. C. L.; Jung, S.; Ferrer, I. M.; Chatman, S. M.; Peters, J. C.; Jaramillo, T. F. Benchmarking Hydrogen Evolving Reaction and Oxygen Evolving Reaction Electrocatalysts for Solar Water Splitting Devices. J. Am. Chem. Soc. 2015, 137, 4347−4357. (7) Jung, S.; McCrory, C. C. L.; Ferrer, I. M.; Peters, J. C.; Jaramillo, T. F. Benchmarking Nanoparticulate Metal Oxide Electrocatalysts for the Alkaline Water Oxidation Reaction. J. Mater. Chem. A 2016, 4, 3068. (8) Osgood, H.; Devaguptapu, S. V.; Xu, H.; Cho, J.; Wu, G. Transition Metal (Fe, Co, Ni, and Mn) Oxides for Oxygen Reduction and Evolution Bifunctional Catalysts in Alkaline Media. Nano Today 2016, 11, 601−625. (9) Lee, D. U.; Xu, P.; Cano, Z. P.; Kashkooli, A. G.; Park, M. G.; Chen, Z. Recent Progress and Perspectives on Bi-Functional Oxygen Electrocatalysts for Advanced Rechargeable Metal−Air Batteries. J. Mater. Chem. A 2016, 4, 7107. (10) Wang, Z.-L.; Xu, D.; Xu, J.-J.; Zhang, X.-B. Oxygen Electrocatalysts in Metal-Air Batteries: From Aqueous to Nonaqueous Electrolytes. Chem. Soc. Rev. 2014, 43, 7746−7786. (11) Zhang, X.; Xu, H.; Li, X.; Li, Y.; Yang, T.; Liang, Y. Facile Synthesis of Nickel−Iron/Nanocarbon Hybrids as Advanced Electrocatalysts for Efficient Water Splitting. ACS Catal. 2016, 6, 580−588. (12) Trotochaud, L.; Ranney, J. K.; Williams, K. N.; Boettcher, S. W. Solution-Cast Metal Oxide Thin Film Electrocatalysts for Oxygen Evolution. J. Am. Chem. Soc. 2012, 134, 17253−17261. (13) Liu, G.; Gao, X.; Wang, K.; He, D.; Li, J. Uniformly Mesoporous NiO/NiFe2O4 Biphasic Nanorods as Efficient Oxygen Evolving Catalyst for Water Splitting. Int. J. Hydrogen Energy 2016, 41, 17976−17986. (14) Smith, A. M.; Trotochaud, L.; Burke, M. S.; Boettcher, S. W. Contributions to Activity Enhancement via Fe Incorporation in Ni(oxy)hydroxide/borate Catalysts for Near-neutral pH Oxygen Evolution. Chem. Commun. 2015, 51, 5261−5263. (15) Li, Y.; Guo, K.; Li, J.; Dong, X.; Yuan, T.; Li, X.; Yang, H. Controllable Synthesis of Ordered Mesoporous NiFe2O4 with Tunable Pore Structure as a Bifunctional Catalyst for Li−O2 Batteries. ACS Appl. Mater. Interfaces 2014, 6, 20949−20957. (16) Li, P.; Ma, R.; Zhou, Y.; Chen, Y.; Liu, Q.; Peng, G.; Liang, Z.; Wang, J. Spinel Nickel Ferrite Nanoparticles Strongly Cross-Linked with Multiwalled Carbon Nanotubes as a Biefficient Electrocatalyst for Oxygen Reduction and Oxygen Evolution. RSC Adv. 2015, 5, 73834. (17) Chervin, C. N.; DeSario, P. A.; Parker, J. F.; Nelson, E. S.; Miller, B. W.; Rolison, D. R.; Long, J. W. Aerogel Architectures Boost Oxygen-Evolution Performance of NiFe2Ox Spinels to Activities Commensurate with Nickel-Rich Oxides. ChemElectroChem 2016, 3, 1369−1375. (18) Rolison, D. R. Catalytic NanoarchitecturesThe Importance of Nothing and the Unimportance of Periodicity. Science 2003, 299, 1698−1701. (19) Hüsing, N.; Schubert, U. AerogelsAiry Materials: Chemistry, Structure, and Properties. Angew. Chem., Int. Ed. 1998, 37, 22−45.

evolving oxygen. In the NiyFe1−yOx ambigel-vs-aerogel series, one variable does not provide a definitive predictor of catalytic performance, whether that variable is surface area, crystallite size, Ni:Fe stoichiometry, or single-phase purity. Significant interplay is at work between all fournot surprising in light of the mechanistic complexity as molecular oxygen is cycled electrochemically, but a hint that much remains to be done to design the optimal catalytic nanoscale interphase.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.langmuir.7b01046. Summary of crystal structure properties; XRD patterns; pore size distribution plots; trend plots; SEM; linearsweep voltammograms (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Jesse S. Ko: 0000-0001-6965-4174 Paul A. DeSario: 0000-0003-2964-4849 Jeffrey W. Long: 0000-0002-5184-5260 Debra R. Rolison: 0000-0003-0493-9931 Author Contributions #

J.S.K. is an NRL-NRC Postdoctoral Associate (2016−2018). M.N.V. is a Naval Research Enterprise Internship Program (NREIP) undergraduate research student (2016−2017).

Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS This work was supported by the U.S. Office of Naval Research. REFERENCES

(1) Hong, W. T.; Risch, M.; Stoerzinger, K. A.; Grimaud, A.; Suntivich, J.; Shao-Horn, Y. Toward the Rational Design of NonPrecious Transition Metal Oxides for Oxygen Electrocatalysis. Energy Environ. Sci. 2015, 8, 1404. (2) Lee, J.; Jeong, B.; Ocon, J. D. Oxygen Electrocatalysts in Chemical Energy Conversion and Storage Technologies. Curr. Appl. Phys. 2013, 13, 309−321. (3) Suen, N.-T.; Hung, S.-F.; Quan, Q.; Zhang, N.; Xu, Y.-J.; Chen, H. M. Electrocatalysis for the Oxygen Evolution Reaction: Recent Development and Future Perspectives. Chem. Soc. Rev. 2017, 46, 337− 365. (4) Burke, M. S.; Enman, L. J.; Batchellor, A. S.; Zou, S.; Boettcher, S. W. Oxygen Evolution Reaction Electrocatalysts on Transition Metal G

DOI: 10.1021/acs.langmuir.7b01046 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir (20) Long, J. W.; Swider-Lyons, K. E.; Stroud, R. M.; Rolison, D. R. Design of Pore and Matter Architectures in Manganese Oxide ChargeStorage Materials. Electrochem. Solid-State Lett. 2000, 3, 453−456. (21) Rolison, D. R.; Dunn, B. Electrically Conductive Oxide Aerogels: New Materials in Electrochemistry. J. Mater. Chem. 2001, 11, 963−980. (22) Gash, A. E.; Tillotson, T. M.; Satcher, J. H., Jr.; Poco, J. F.; Hrubesh, L. W.; Simpson, R. L. Chem. Mater. 2001, 13, 999−1007. (23) Long, J. W.; Logan, M. S.; Rhodes, C. P.; Carpenter, E. E.; Stroud, R. M.; Rolison, D. R. Nanocrystalline Iron Oxide Aerogels as Mesoporous Magnetic Architectures. J. Am. Chem. Soc. 2004, 126, 16879−16889. (24) Pettigrew, K. A.; Long, J. W.; Carpenter, E. E.; Baker, C. C.; Lytle, J. C.; Chervin, C. N.; Logan, M. S.; Stroud, R. M.; Rolison, D. R. Nickel Ferrite Aerogels with Monodisperse Nanoscale Building BlocksThe Importance of Processing Temperature and Atmosphere. ACS Nano 2008, 2, 784−790. (25) Louie, M. W.; Bell, A. T. An Investigation of Thin-Film Ni-Fe Oxide Catalysts for the Electrochemical Evolution of Oxygen. J. Am. Chem. Soc. 2013, 135, 12329−12337. (26) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications, 2nd ed.; John Wiley & Sons: New York, NY, 2001; Chapter 12. (27) Qiao, J.; Xu, L.; Ding, L.; Shi, P.; Zhang, L.; Baker, R.; Zhang, J. Effect of KOH Concentration on the Oxygen Reduction Kinetics Catalyzed by Heat-Treated Co-Pyridine/C Electrocatalysts. Int. J. Electrochem. Sci. 2013, 8, 1189−1208. (28) DeSario, P. A.; Chervin, C. N.; Nelson, E. S.; Sassin, M. B.; Rolison, D. R. Competitive Oxygen Evolution in Acid Electrolyte Catalyzed at Technologically Relevant Electrodes Painted with Nanoscale RuO2. ACS Appl. Mater. Interfaces 2017, 9, 2387−2395. (29) Jiang, J.; Zhang, C.; Ai, L. Hierarchical Iron Nickel Oxide Architectures Derived from Metal-Organic Frameworks as Efficient Electrocatalysts for Oxygen Evolution Reaction. Electrochim. Acta 2016, 208, 17−24. (30) Long, J. W.; Logan, M. S.; Carpenter, E. E.; Rolison, D. R. Synthesis and Characterization of Mn−FeOx Aerogels with Magnetic Properties. J. Non-Cryst. Solids 2004, 350, 182−188. (31) Costa, R. C. C.; Lelis, M. F. F.; Oliveira, L. C. A.; Fabris, J. D.; Ardisson, J. D.; Rios, R. R. V. A.; Silva, C. N.; Lago, R. M. Novel Active Heterogeneous Fenton System Based on Fe3−xMxO4 (Fe, Co, Mn, Ni): The Role of M2+ Species on the Reactivity towards H2O2 Reactions. J. Hazard. Mater. 2006, B129, 171−178. (32) Prabu, M.; Ketpang, K.; Shanmugam, S. Hierarchical Nanostructured NiCo2O4 as an Efficient Bifunctional Non-Precious Metal Catalyst for Rechargeable Zinc-Air Batteries. Nanoscale 2014, 6, 3173− 3183.

H

DOI: 10.1021/acs.langmuir.7b01046 Langmuir XXXX, XXX, XXX−XXX