Porous Nickel–Iron Oxide as a Highly Efficient Electrocatalyst for Oxygen Evolution Reaction

Abstract

A porous Ni–Fe oxide with improved crystallinity has been prepared as a highly efficient electrocatalytic water oxidation catalyst. It has a small overpotential, a low Tafel slope, and an outstanding stability. The remarkably improved electrocatalytic performance is due to the porous structure, high extent homogeneous iron incorporation, ameliorative crystallinity, and the low mass transfer resistance. Water splitting for the generation of clean and sustainable energy resource represents one of the most promising processes toward environmental remediation.1 Oxygen evolution reaction (OER) is the bottleneck in water splitting as it is kinetically slow and requires high overpotential (η) to reach a substantial current density in water electrolysis.2 Broad attention has been therefore attracted to seek efficient OER electrocatalysts working at low overpotentials to reduce the energy input for water splitting.3 Regarding this, IrO2 and RuO2 were reported to have good electrocatalytic OER activity.[3] However, the use of scarce noble metals limited their widespread applications. Toward identifying cheap and efficient electrocatalysts for OER, transition metal oxides,4 hydroxides,5 and oxyhydroxides6 based on earth abundant elements have been extensively examined. Among them, mixed-metal materials containing iron group elements (Fe, Co, Ni) have been generally acknowledged as one of the most competent candidates for OER in aqueous media.[3],[4],[6],7 For the fabrication of iron group metal based materials as OER catalysts, one of the widely employed techniques is in situ electrodeposition method for the generation of amorphous films on the surface of electrodes. The structure and electrochemical activity of electrodeposited Ni, Co, Ni–Fe, Co–Fe, and Ni–Co films have been extensively studied.[4],6, 8 Our recent work reported the first Fe-based film prepared by a fast and simple cyclic voltammetry (CV) electrodeposition method as an OER electrocatalyst in neutral media with a high turnover frequency (TOF).[5] Besides the electrodeposition route, photochemical deposition method for a soft loading of mixed Ni–Fe–Co film on electrode with controlled metal stoichiometry has also been developed.[4] Alongside the deposition methods, mixed-metal materials containing iron group elements have also been synthesized via a variety of techniques such as spin-coating,9 pulsed-laser ablation,10 aerosol spray,11 template synthesis,[5],[7] solid-state reaction,12 as well as exfoliation of bulk materials13 toward improved electrochemical activity. Moreover, Ni–Fe layered double hydroxides (LDH) with the incorporation of carbon materials have been reported for significant enhancement of the OER activity.14 Typically, mixed Ni–Fe materials have shown excellent electrocatalytic activities in basic media.[6],[14] Ni is the active oxygen evolution center, and Fe incorporation will improve the conductivity of the material.[4],[6] It is universally acknowledged that high surface area will contribute to the activity improvement, whereas the poor crystallinity arisen from high surface area is adverse to the electrocatalytic stability. Thermal treatment of the materials for improved crystallinity can ameliorate their electrochemical stability.[7],12 However, thermal treatment has been rarely applied to Ni–Fe electrocatalysts as it can largely reduce the surface area and thus the activity of the materials.[4] Although considerable efforts have been devoted to fabricating electrode materials with optimized surface active sites via porous structure or 3D architecture,[5],7,[14],15 limited work has been documented on porous electrocatalysts with improved crystallinity for enhanced stability without the damage of their efficient OER performance. Thus, it is highly desirable to develop novel protocols for the synthesis of porous Ni–Fe materials with abundant surface active sites, and meanwhile, with improved crystallinity for ameliorative electrochemical stability. Herein, we report a facile synthesis of porous Ni–Fe mixed oxides in the presence of organic surfactant Tween 85. By using Tween, which has a perfect boiling point for the synthesis of the catalyst in this work, the porous structure can be created at low temperature from evaporation of the mixed surfactant. While in traditional method, those carbon-based surfactant templates are burnt out under high temperatures to get the porous structure. Heating treatment at elevated temperature not only requires large energy cost but also limits the application in thermal instable or unfavorable systems. Ni–Fe OER catalyst is a thermal unfavorable system. The creation of porous structure under mild temperature is not only suitable for the formation of active phase, but also preferable for the reservation of high surface area. The sample after thermal treatment under 200 °C has a porous structure and improved crystallinity compared to traditional Ni–Fe hydroxides synthesized under room temperature. This porous Ni–Fe catalyst outperformed its bulk material counterpart (Tween free sample) for OER. In CV studies, it can highly efficiently catalyze water oxidation in a 0.1 m KOH aqueous solution. Small overpotentials of 328 and 420 mV are required to reach OER current densities of 10 and 50 mA cm−2, respectively. In controlled potential electrolysis (CPE) at η = 387 mV, an extremely stable current density of ≈10.2 mA cm−2 maintained, giving an extraordinary TOF value of 432 h−1. Besides the outstanding performance of the material, our synthesis method is suitable for large-scale production, and thus delivers a valuable contribution to electrocatalytic water splitting research. The Ni–Fe composites were obtained through coprecipitation strategy in the presence of Tween 85, a polysorbate surfactant with a boiling point above 100 °C. The solids were collected and subjected to mild heat treatment under elevated temperatures for the removal of Tween to construct pores in the material. X-ray diffraction (XRD) patterns for composites with different starting Ni–Fe ratios after calcination under 200 °C for 3 h (all the thermal treatment time is 3 h in this work), which is the thermal treatment condition toward the highest OER activity (see below), are shown in Figure 1A. Samples are denoted as Ni-X-Y, in which X stands for the starting Ni percentage and Y stands for the treatment temperature in degrees Celsius. The two samples with high nickel concentration tended to form β-Ni(OH)2 (JCPDS 742075) phase showing selected better crystallinity from (100) and (110) reflections at 2θ degree of 33° and 59°, respectively. Increased Fe incorporation in Ni-90-200 and Ni-85-200 samples resulted in a dominant NiO (JCPDS 780643) phase with mixed weak signals from NiFe2O4 (JCPDS 742081) observed in the latter sample. The peaks around 62° and 75° in these two samples were assigned to NiO, as the characteristic XRD peak of NiFe2O4 at 30° was not observed. The iron incorporation in the composites can promote the transfer of hydroxides to oxides under thermal treatment, which is consistent to the X-ray photoelectron spectroscopy (XPS) studies of Ni–Fe material by Smith et al.[4] Samples with further increased iron content existed as NiFe2O4 with mixed NiO, and it was a mixture of NiFe2O4 and γ-Fe2O3 (JCPDS 871166) for sample with even higher Fe content. For Ni-50-200 sample, which contains actually over 94% of iron as suggested by XPS, the crystallinity is quite poor with a dominant γ-Fe2O3 phase. With a starting Ni–Fe ratio of 85:15, which is the tested Ni–Fe ratio toward the highest OER activity (see below), the synergic effects of heat treatment on crystal structure can be referred from the XRD patterns shown in Figure 1B. The freshly synthesized composite showed a structure that can be indexed to α-Ni(OH)2 (JCPDS 380715). After heat treatment under 100 °C for 3 h, the α-Ni(OH)2 phase maintained with an emerging peak around 11° from the (003) reflection perpendicular to the α-Ni(OH)2 layers. This phenomenon resembles the structure conversion of electrodeposited Ni–Fe LDH with new XRD peak from (003) direction after aging in a hot solution for a period of time.[6] Due to the high content of Tween in our samples as determined by element analysis (≈40 wt%), the Ni–Fe layers in the freshly prepared samples are probably wrapped by Tween molecules leading to the formation of Ni–Fe hydroxide nanosheets with limited layers. Upon heat treatment, Tween molecules were released. As a consequence, the Ni–Fe layers assembled together causing the XRD peak at 11° due to the diffraction from Ni–Fe hydroxide interlayers. The XRD result of the sample Ni-85-200 without the addition of Tween shown in Figure S1 (Supporting Information) suggested a distinguishable crystalline phase of α-Ni(OH)2, which is consistent with the above discussions on the formation of assembled layers in Tween wrapped samples. With elevated temperature treatment, the hydroxides start to form a dominant NiO (JCPDS 780643) structure under 200 °C and an NiO/NiFe2O4 (JCPDS 742081) composites starting from 300 °C. As demonstrated in Figure 2A, the release of Tween upon heat treatment can be verified from infrared spectra of the samples studied in Figure 1B. The red line sample was synthesized from the same conditions without Tween. The freshly prepared sample with Tween before heat treatment displays distinct peaks from C–H stretching vibrations of methylene and methyl groups (2860, 2940 cm−1), CO/C–O stretching vibrations of ester/ether groups (1750, 1130 cm−1). Tween molecules were gradually removed from the sample after heat treatment, and only trace was left in the sample after 200 °C treatment. The O–H stretching vibration peaks (≈3480 cm−1) and scissoring vibration peaks (1700–1500 cm−1 from water, 1400–1300 cm−1 from structure hydroxyl groups) are obvious in all the samples, especially in samples under low-temperature treatment. Thermal gravimetric analysis (TGA) and derivative thermogravimetric (DTG) analysis studies of the above mentioned sample are shown in Figure 2B. Upon elevated heat treatment, the mixture starts to lose weight from a sequence of absorbed water, structure water, Tween, and eventually undergoes a phase transfer from hydroxide to oxide. Upon the removal of Tween in the samples, porous structure was obtained and demonstrated by transmission electron microscopy (TEM) images (Figure 3A–D), Brunauer–Emmett–Teller (BET) surface area, and pore size distribution measurements (Table S1, Supporting Information, and Figure 3E,F). After heat treatment, the bulk hydroxides (Figure 3A) turned into aggregates comprised of Ni–Fe domains (darker areas) and pores (lighter areas) as illustrated in Figure 3B,C. Higher temperature treatment converts the aggregates into more separated nanoparticles as shown in Figure 3D. Scanning electron microscope (SEM) and high-resolution TEM (HRTEM) images of the Ni-85-200 sample are provided in Figures S2 and S3 (Supporting Information). Well-dispersed electrocatalysts on indium tin oxide (ITO) electrode was observed from the SEM image. In HRTEM, the lattice can be fairly indexed to the dominant reflection peaks of NiO. Elemental mapping of the material was demonstrated via energy dispersive X-ray (EDX) spectroscopy, suggesting a homogeneous nickel/iron dispersion in the material as shown in Figures S4 and S5 (Supporting Information). BET surface areas and pore size distributions of the materials studied in TEM have been examined through N2 physical adsorption measurements. Figure 3E shows the pore size distribution of the aforementioned sample with a pore diameter around 3 nm. The pore size expanded with increasing temperature treatment. After calcination under 200 °C for 3 h, the pore volume of the sample with the addition of Tween during synthesis (red line, Figure 3E) is significantly higher than that from the sample without Tween (blue line, Figure 3E). The adsorption–desorption isotherm plot for the sample treated under 200 °C displays a type-IV isotherm with a closure at P/P0 around 0.4, suggesting the presence of small mesopores in the material. Integrally, we are capable to illustrate the evolution procedure of the Ni–Fe material under heat treatment in Figure 4. The freshly synthesized Tween wrapped Fe doped α-Ni(OH)2 sheets assembled into layered Fe doped α-Ni(OH)2 after heat treatment as the hydroxide sheets were less wrapped by Tween. Subsequently, it evolved into porous Fe-doped NiO with the complete removal of Tween under higher temperature. With further elevated temperature treatment, the pore expanded and separated nanoparticles were formed eventually. The actual Ni–Fe composition information of the samples determined by XPS analysis is summarized in Table S1 (Supporting Information) with the foregoing discussed physical properties. The Ni/Fe 2p spectra for the Ni-X-200 and Ni-85-Y samples are presented in Figure S6 (Supporting Information). The XPS results from the Ni-50-Y samples as displayed in gray line in Figure S6A,B (Supporting Information) indicate the very low nickel content, which is due to the much smaller solubility product from Fe(OH)3 (Ksp = 2.8 × 10−39) than Ni(OH)2 (Ksp = 5.5 × 10−16).16 FeIII will be precipitated from basic solutions much easier than NiII, leading to a low nickel content sample when base is not sufficient during synthesis. The Ni 2p and Fe 2p peaks for the Ni-85-200 sample can be fairly resolved to support the XRD results of the composite (Figure 5). A dominant Fe-incorporated NiO phase and trace amount of NiFe2O4 was suggested from the XRD results. The Ni 2p3/2 peak centered at 855.5 eV is higher than that from pure NiO around 854 eV. The high extent doping of FeIII with high electronegativity causes a higher valence state of NiII.[4],[6] The small orange line peak with binding energy of 857.0 eV is assigned to the trace amount of isolated NiFe2O4, whose weak signals in XRD analysis can also be observed.17 The NiFe2O4 content is around 3.1 mol%. The Fe 2p3/2 peak is well fitted into the FeIII signals at 710.2 and 715.6 eV based on the method developed for the spectrum of ferric with a nickel environment.18 The electrocatalytic performance of the porous Ni–Fe materials for water oxidation was tested. CVs of the Ni-X-200 samples are shown in Figure 6A. Pure nickel material has poor OER activity and displays distinct self-redox peaks due to the oxidation of NiII to NiIII at 1.45 V versus reversible hydrogen electrode (RHE, all potentials reported in this work are vs RHE).[5],[6] The addition of two percent of Fe during synthesis will significantly suppress the oxidation of is NiII self-redox peaks from samples with higher percentage of Fe It is that FeIII with high electronegativity can the valence state of nickel which to be further With the on iron incorporation the transfer in the material for higher OER The OER activity is remarkably improved as the starting Fe content increased to of the Fe content to the OER and is water oxidation peak from the sample with or higher starting Fe content. Thermal treatment effects on the electrocatalytic activity have also been studied for the Ni-85-Y samples (Figure Thermal treatment of the freshly prepared sample at 100 °C for 3 h can improve its OER performance For the Ni-85-200 sample, the current can reach as high as mA cm−2 at V = Thermal treatment under further increased temperature will remarkably the OER probably due to the surface area as well as structure of the material. As by XRD the mixture was in the form of Fe-doped NiO after 200 °C After higher temperature treatment, it turned into a mixture of NiO and NiFe2O4, which is reported to have excellent OER As for Ni-85-200 sample, the trace amount of NiFe2O4 in the sample is not to the OER performance the Fe content is a significant for the activity of Ni–Fe Thus, the higher activity from the Ni-85-200 is due to the and doped Fe content in the NiO phase. For Figure shows the CV of the porous Ni-85-200 sample, the CV from the bulk sample synthesized without Tween, and the CV from carbon The porous sample outperformed the bulk Ni–Fe hydroxide that is reported extensively as an efficient OER In the of electrocatalyst was the electrode for CV from the same condition showed a less active performance compared to the porous Ni–Fe an potential for OER was observed. The obtained porous Ni–Fe oxides can efficiently catalyze water oxidation with small overpotentials of 328 and 420 mV to reach current densities of 10 and 50 mA cm−2, respectively. better our a of the current densities from the samples the addition of Tween during synthesis is shown in Figure For the freshly prepared samples, the addition of Tween can suppress the OER which is due to the of surface active sites by large Tween With elevated heat treatment, the Tween molecules were removed a porous structure. As a consequence, it is to that the samples with Tween during synthesis the bulk material samples after heat treatment. It is also that calcination to °C the structure less significant and Tween samples, and thus resulted in OER The sample with Tween has the same α-Ni(OH)2 structure to the Ni-85-200 sample without Tween. The higher performance from the is probably due to the less assembled α-Ni(OH)2 nanosheets as demonstrated from the XRD reflection from (003) around 11° as shown in Figure and Figure S1 (Supporting Information). Tween wrapped sample can be from Ni(OH)2 to NiO under a probably due to its This to the high surface area of the active Ni-85-200 sample. The of the CV studies for samples with or without Tween are shown in Figure (Supporting Information). The self-redox peaks due to the oxidation of NiII to NiIII were with the of thermal treatment temperature. This is consistent with the increased valence state of Ni in the nickel after under high temperature as studied by et this thermal treatment causes a Ni–Fe in the samples after which to the of electrocatalytic activity. the surface area after it is to the highest activity from the sample after thermal treatment under 200 °C. As shown in Figure at V was in 0.1 m KOH aqueous solution using with the porous Ni-85-200 catalyst (red or without thermal treatment and the catalyst without Tween (blue In a stable current density at ≈10.2 mA cm−2 maintained from the porous sample, its outstanding electrocatalytic Moreover, which is widely reported for Ni–Fe based materials in water is required to reach the high current density for this porous catalyst. stability are to the improved crystallinity of our sample as a of thermal The results from the same sample without thermal treatment showed a much smaller current density and a of activity after 10 of and after CVs were and provided in Figure (Supporting Information) for the samples with or without thermal treatment. The stable performance of the Ni-85-200 is further from the SEM and images of the sample after electrolysis (Figure Supporting which suggested and phase. to the sample under room the porous sample showed much better stability due to its improved crystallinity after thermal treatment. With the Tween the mild heat treatment the high surface area porous structure for efficient and on the the crystallinity of the material for stability. For the sample without Tween during synthesis showed a much smaller current density of mA cm−2 with a gradually increased current after the h A of the evolution of oxygen during electrolysis under V is provided S1, Supporting Information). The amount of evolved oxygen was using a h of with of which a of A high TOF value of 432 based on the nickel amount was which is those highest reported for It be that the OER performance of metal can be improved in or from Thus, of the OER performance be based on the same Figure shows the Tafel plot of the porous sample and the bulk sample. A small Tafel of mV for the porous catalyst is a higher value of mV is obtained for the bulk catalyst without which has the same NiO phase as Ni-85-200 with Tween. The electrochemical spectroscopy was referred for more information of the electrocatalysts as shown in the in Figure As from Figure for the of the porous sample under different the first in the represents the transfer from the bulk to the catalyst which is of the potential The in the represents the from the mass transfer during electrocatalytic which is the of the reaction The porous electrocatalyst has a much mass transfer in OER as shown in Figure which is to the porous structure. The of the Ni-85-Y samples are provided in Figure (Supporting Information) for The mass transfer is from the Ni-85-200 sample, which is the most active catalyst studied in this In a porous Ni–Fe oxide has been prepared with the of Tween Upon mild thermal treatment, the polysorbate was removed gradually from the composite the porous structure material with large surface area. the crystallinity was improved and a homogeneous high extent Fe doping in NiO was by heat treatment. The material is highly efficient for electrocatalytic water the porous sample, small overpotentials of 328 and 420 mV are required for OER to reach current densities of 10 and 50 mA cm−2, respectively. The current density of ≈10.2 mA cm−2 under V The remarkably improved electrocatalytic performance compared to the bulk material is due to the porous structure, homogeneous Fe incorporation, ameliorative crystallinity, and the low mass transfer resistance. of The materials were obtained through coprecipitation strategy with the presence of Tween 85, a polysorbate surfactant with a boiling point above 100 °C. The polysorbate of 10 was first in 50 of m solution at °C under and a mixed aqueous solution containing and at different ratios metal concentration is volume is was into the hot basic solution The was under the temperature for The solids were collected and by in a sequence of water and water The obtained solids were in an at °C and were subjected to further thermal treatment under different temperatures as The bulk samples for were synthesized from the same procedure the addition of Tween. electrochemical were out using a at °C. CVs were obtained in of 0.1 m KOH aqueous solution using a with a electrode as the working as the and as the The working electrode was prepared through a Typically, of sample and of solution were in of solution at volume ratio of by for h to form a homogeneous of the mixture was the carbon for was for all potentials were reported versus the based on the = × for Tafel were by controlled potential electrolysis in of 0.1 m KOH solution at a variety of applied potentials in The stable were at applied potentials from to V in mV for with the solution was at V in a with working electrode × with the catalyst of the The was on an electrode over a frequency from 0.1 to at the of the of mV under different potentials from to The TOF value was by that nickel was in the TOF = × is the the are required for the generation of one oxygen is and is the of the nickel on the electrode based on the amount of catalyst. of in was by using a The are for support from the for the the of the of and the of under As a to our and this information by the materials are and be for but are not or support from information than be to the The is not for the content or of information by the than be to the for the