Rationally designed Water-Insertable Layered Oxides with Synergistic

Jun 20, 2019 - The oxidation states of the transition-metal elements at the catalyst ... and the corresponding specific surface areas (SBET) were obta...
1 downloads 0 Views 3MB Size
Download PDF
Research Article www.acsami.org

Cite This: ACS Appl. Mater. Interfaces 2019, 11, 25227−25235

Rationally designed Water-Insertable Layered Oxides with Synergistic Effect of Transition-Metal Elements for HighPerformance Oxygen Evolution Reaction Shiyong Chu,† Hainan Sun,† Gao Chen,† Yubo Chen,‡ Wei Zhou,† and Zongping Shao*,†,§

Downloaded via UNIV OF ROCHESTER on July 21, 2019 at 18:03:20 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

†

State Key Laboratory of Materials-Oriented Chemical Engineering, College of Chemical Engineering, Nanjing Tech University, Nanjing 210009, P.R. China ‡ School of Material Science and Engineering, Nanyang Technological University, Singapore 639798, Singapore § Department of Chemical Engineering, Curtin University, Perth, Western Australia 6845, Australia S Supporting Information *

ABSTRACT: Oxygen evolution reaction (OER) is a key step in many energy conversion and storage processes. Here, by rationally adding an appropriate amount of Mn into the lattice of a layered NaxCoO2 parent oxide, high solubility of iron into the NaxCoO2 oxide lattice is realized without the use of an extremely air-sensitive Na2O2 raw material, and the synergy created between the Co and Fe can boost the catalytic activity of the layered oxide for OER. Moreover, the water intercalation capability of the layered oxides can be utilized to make the oxide resemble mixed metal hydroxides, which will also bring a beneficial effect for OER. As a result, the as-developed Na0.67Mn0.5Co0.3Fe0.2O2 (CF-32) layered oxide with an optimal Co/Fe ratio and water intercalation shows high OER performance in alkaline media, overperforming the benchmark IrO2 catalyst. In 0.1 M KOH solution, the novel catalyst shows 0.39 V overpotential at 10 mA cm−2 and favorable stability. The excellent OER performance of CF-32 is due to the synergistic effect of transition-metal elements (Co and Fe) and water intercalation, leading to little charge transfer resistance, large amounts of exposed catalytic active sites, plenty of surface high oxidation state O22−/O− oxygen species, and hydroxide-rich surface. The facile synthesis and high OER performance of CF-32 enriches the non-noble metal family of OER catalysts and boosts the practical application of non-noble metal catalysts. KEYWORDS: oxygen evolution reaction, layered sodium transition-metal oxides, multi-metal synergy, water intercalation, water splitting

■

reduces their energy storage/conversion efficiency.8−10 Although noble metal oxides, such as IrO2 and RuO2, show favorable OER activity, their high price hinders further largescale development and application.11,12 Therefore, the development of cost-effective and efficient non-noble metal-based OER catalysts has become highly critical. Recently, many noble metal-free multi-metal oxides (spinel and perovskite crystallized oxide phases, and amorphous metal oxides) and some mixed transition-metal-based (oxy)hydroxides (such as layered hydroxides and perovskite hydroxides) were found to be highly attractive electrocatalysts for OER.13−15 A synergy could be present between the different metal elements in multi-metal oxides, which could effectively tune the electronic, defect and surface structure of the catalysts, and impact the adsorption energies of OH*, O*, OOH* intermediates of the existing active sites, thus greatly affecting the OER catalytic activity.16−18 Optimizing the atom

INTRODUCTION Renewable energies, such as solar and wind power, are becoming increasingly important in modern society because of concerns about the deterioration of the global environment from inefficient and excessive use of fossil fuels and the gradual depletion of fossil fuel resources.1,2 Because of the intermittent nature of such renewable energies, an energy storage/ conversion system is required for their utilization, and can be an electricity-to-electricity-type, such as rechargeable batteries, or an electricity-to-chemical-type, such as electrochemical water electrolyzers.3−5 For example, rechargeable metal−air batteries are an important type of high-energy-density electrochemical energy storage system that are believed to be highly promising for large-scale energy storage and as a power source for electric vehicles. On the other hand, hydrogen is an ideal clean fuel for the future, and electrochemical watersplitting is the most promising method for mass production of hydrogen by using renewable energy as an energy input.6,7 Both metal−air batteries and water-splitting involve the oxygen evolution reaction (OER), but the sluggish kinetics of conventional electrodes for OER in these systems significantly © 2019 American Chemical Society

Received: April 15, 2019 Accepted: June 20, 2019 Published: June 20, 2019 25227

DOI: 10.1021/acsami.9b06560 ACS Appl. Mater. Interfaces 2019, 11, 25227−25235

Research Article

ACS Applied Materials & Interfaces

oxides are highly promising OER electrocatalysts that may have great potential for practical water-splitting systems and rechargeable metal−air batteries.

ratio of the various transition-metal elements to maximize the synergistic effect is key to obtaining the best OER activity from multi-metal oxides. Layered LiCoO2 is a commercial cathode material for lithium-ion batteries.19 Recently, it was also exploited as an OER electrocatalyst in alkaline media and showed excellent activity after properly tuning the oxidation state of cobalt based on electrochemical charge−discharge reactions.20 However, the electrochemical tuning process is too complicated and difficult for mass production; thus, many other strategies for adjusting the OER performance of layered LiCoO2 and similar materials have been reported.21−24 Interestingly, by doping with a certain amount of Fe into the lattice of LiCoO2, significantly improved OER activity was also observed, and the same phenomenon was also observed in other structural types of cobalt-based metal oxides, such as spinel-type ZnCo1.6Fe0.4O2 and perovskite-type LaCo0.9Fe0.1O3.25,26 This improvement was attributed to the synergy created between cobalt and iron in the oxide lattices. Considering the limited reserves and sharp increase in the price of lithium, Li-free layered oxides are more preferred as OER electrocatalysts. In contrast to lithium, sodium resources are abundant in the forms of NaCl and Na2CO3 and low in price.27−29 Layerstructured NaxCoO2 (0 < x ≤ 1) was also recently exploited as an OER catalyst, and the activity was improved through sodium extraction or Fe-doping.30,31 However, the activity was still not satisfactory. One main reason is the low solubility of iron in a NaxCoO2 oxide lattice as prepared by a conventional solid-state reaction or sol−gel method using nonoxidative and air-stable Na2CO3 or NaNO3 as the sodium source. Only 5% iron could be introduced into the layer-structured oxide, and an impurity phase of iron oxide started to appear when we substituted 10% of Co by iron in NaxCoO2.30 Although higher iron solubility in NaxCoO2 can be realized by using an oxidative and extremely air-sensitive Na2O2 sodium source, its strong water absorption behavior and easy reaction with CO2 in the surrounding atmosphere make the synthesis possible only in the glove box, which is problematic for mass production.32,33 Many-layered oxide cathodes for sodium ion batteries are air-sensitive, and they can react with H2O because of the large sodium layer spacing caused by the large sodium ion radius.34,35 This may also be a concern for application in OER. In this study, we report the realization of a high iron solubility in a NaxCoO2-layered oxide by applying nonoxidative and air-stable Na2CO3 or NaNO3 as sodium sources by co-doping with a proper amount of Mn; thus, the cobalt-toiron atom ratio can be well tailored to optimize the performance. In addition, the high sensitivity of layered oxides to water was turned into a beneficial property and was utilized for facile intercalation of water into the layers of the oxides, resulting in a further performance improvement. As a result, the as-developed layered oxide with the optimal composition of Na0.67Mn0.5Co0.3Fe0.2O2 (CF-32) showed a satisfactory OER activity. When tested as an OER catalyst in 0.1 M KOH solution, CF-32 showed a 0.39 V overpotential at 10 mA cm−2; its performance is substantially better than other layered materials of Na0.67MnO2, Na0.67CoO2, and Na0.67Mn0.5Co0.5O2, and even superior to that of the benchmark IrO2 noble metalbased catalyst. The synergistic effect between Co and Fe led to a low charge transfer resistance, a large amount of exposed active sites, and plenty of high oxidation state O22−/O− oxygen species, which are beneficial for promoting the OER. Layered

■

EXPERIMENTAL SECTION

Catalyst Preparation. In this study, the layered oxides with the nominal composition of NaxMnyCo1−y−zFezO2 were prepared using the two classical preparation methods of solid-state reaction and sol− gel synthesis. For the solid-state reaction, the preparation of the CF32 catalyst is described as an example. For a detailed preparation process, stoichiometric raw materials of Na2CO3 (5% sodium excess), Mn2O3, Co2O3, and Fe2O3 were thoroughly mixed in liquid medium acetone under high-energy ball milling (Fritsch Pulverisette 6) at 400 rpm for 2 h. The ball-milled suspension was dried under a hightemperature sodium lamp and then pressed into a disk and sintered at 900 °C for 12 h. The sintered disk was ground into powder, washed with deionized water for 5 min, and dried for later use. For the sol− gel method preparation process, the preparation process of the CF-32 catalyst is also described as an example. Stoichiometric NaNO3 (5% sodium excess), Mn(CH3COO)2, Co(NO3)2−6H2O, and Fe(NO3)2−6H2O were dissolved in deionized water, and then the complexing agents of ethylene diamine tetra acetic acid (EDTA) and citric acid (CA) were added to the solution. The molar ratio of all metal ions in the solution, EDTA and CA was 1:1:2. A certain amount of aqueous ammonia was also added to the solution until the pH of the solution was approximately 6. After stirring at 90 °C for several hours, the above transparent solution became a clear gel. The gel was then placed in an electric oven and heated at 250 °C for 5 h to form a fluffy black precursor, and the precursor was then calcined at 900 °C for 12 h to form the final product. The powder was also washed in deionized water for 5 min and then dried for later use. The Na0.67Mn0.5Co0.5−xFexO2 with different Co/Fe ratios (x = 0, 0.1, 0.2, 0.3, 0.4, and 0.5) were denoted as CF-50, CF-41, CF-32, CF-23, CF14, and CF-05, respectively. Basic Characterizations. The room-temperature powder X-ray diffraction (XRD) of the synthesized powders was conducted on a diffractometer (Rigaku Smartlab) with Cu Kα radiation at the voltage and current of 40 kV and 40 mA, respectively, and the diffraction data were collected for the 2θ range of 10°−70° with a scanning step size of 0.02°. The Rietveld refinement method based on the EXPGUI interface and GSAS program was used to refine the XRD pattern of CF-32 to achieve the crystal structure. The oxidation states of the transition-metal elements at the catalyst surface and the surface oxygen species were characterized by X-ray photoelectron spectroscopy (XPS) using a Thermo ESCALAB 250 spectrometer equipped with Al Kα radiation. Fits of the XPS spectra were performed with the public software package XPSPEAK to estimate the chemical species. All species were fitted using a Shirley background. All spectra were shifted relative to the binding energy of the carbon 1s sp3 (assigned to 284.8 eV) to compensate for any offset during the measurement. The relative content (molar fraction) of the different kinds of oxygen species over total surface oxygen amount can be estimated from the relative area of these subpeaks. A Hitachi S-4800 field emission scanning electron microscope was used to obtain the microscopic morphology of the catalysts. Inductively coupled plasma-atomic emission spectroscopy (Optima 7000 DV) was used to confirm the metal elements’ composition of the CF-32 and CF-50 samples. The lattice distance of the CF-32 catalyst was further analyzed by an FEI Tecnai G2T20 transmission electron microscope at a 200 kV electron beam acceleration voltage. The nitrogen isothermal adsorption− desorption curves of the catalysts were obtained using a Quantachrome automatic nitrogen adsorption−desorption instrument (AutoSorb-iQ3), and the corresponding specific surface areas (SBET) were obtained by analyzing the curves based on the Brunauer− Emmett−Teller method. The electrical conductivity of CF-50 and CF-32 samples in the temperature range of 25−500 °C was measured by a four-probe DC method using a Keithley 2420 digital source meter. 25228

DOI: 10.1021/acsami.9b06560 ACS Appl. Mater. Interfaces 2019, 11, 25227−25235

Research Article

ACS Applied Materials & Interfaces Electrochemical Measurements. The electrochemical performances of the catalysts for OER at room temperature were measured using a rotating disk electrode (RDE), where an electrochemical workstation (CHI 760e) was integrated for RDE control and data acquisition. The test was performed in an oxygen-saturated 0.1 M KOH solution and the RDE was maintained at a rotation speed of 1600 rpm. All the potentials in this study were corrected to the reversible hydrogen electrode (RHE). For the linear sweep voltammetry (LSV) and CV tests, a three-electrode cell configuration was applied with a carbon rod as the counter electrode, Ag/AgCl (3.5 M KCl) as the reference electrode, and catalyst-covered glassy carbon (GC) as the working electrode. To prepare the working electrodes, 5 μL of catalyst ink was drop-coated onto the GC substrate (0.196 cm2) and then dried at room temperature for 1 h. To obtain a homogeneous ink, 10 mg of catalyst, 10 mg of Super P carbon, 1 mL of ethanol, and 0.1 mL 5 wt % Nafion solution were well mixed under ultrasonic dispersion for 1 h. The LSV mode was used to record the OER polarization curves with a scan range of 0.2−1 V versus Ag/ AgCl and a sweep rate of 5 mV s−1. The stability of the catalysts was evaluated by chronopotentiometry (CP) mode with a current density of 5 mA cm−2. The electrochemical impedance spectroscopy (EIS) test was conducted in the frequency range of 100 000−0.1 Hz at 0.7 V versus Ag/AgCl. Cyclic voltammogram (CV) mode was conducted at different sweeping rates (20, 40, 60, 80, 100, 120, 140, and 160 mV s−1) in the voltage range of 0.2−0.3 V versus Ag/AgCl to obtain the electrochemical double-layer capacitance (Cdl).

layered NaxMnyCo1−y−z Fez O 2 pure phase with a high concentration Fe could be prepared without the use of extremely air-sensitive Na2O2 as the sodium source. Thus, tailoring the cobalt-to-iron atom ratio to achieve the maximum OER performance of the layered oxide could be realized. To support the above assumption, three different materials with nominal compositions of Na0.67Mn0.33Co0.33Fe0.33O2 (Mn0.33), Na0.67Mn0.5Co0.25Fe0.25O2 (Mn0.5), and Na0.67Mn0.67Co0.17Fe0.17O2 (Mn0.67) were prepared by the solid-state reaction method. As shown in Figure 1a, when the

■

RESULTS AND DISCUSSION As mentioned, the synergy created between cobalt and iron in composite oxides could lead to improved electrocatalytic activity. For the case of layered NaxCoO2, although improved OER activity was indeed observed by introducing iron into the cobalt sites, the solubility of iron was very limited by using Na2CO3 or NaNO3 as the sodium source. As shown in Figure S1, single-phase layer-structured Na0.67CoO2 was successfully prepared by the solid-state reaction; however, the impurity phase of NaFeO2 (PDF no. 01-076-2299) appeared in the 10% Fe-substituted Na0.67Co0.9Fe0.1O2 (Figure S2), in good agreement with our previous report.30 For the sample with the nominal composition of Na0.67Co0.5Fe0.5O2, the XRD pattern in Figure S3 demonstrated that, in addition to the main phase of Na0.67CoO2 (PDF no. 01-087-0274), a high content of rhombohedral NaFeO2 (PDF no. 01-076-2299) and orthorhombic NaFeO2 (PDF no. 01-074-1351) phases also appeared. The sample prepared by the sol−gel method by using NaNO3 as the sodium source gave the same results (Figure S4). For the sodium stoichiometric layered oxide with O3 type, the synthesis of plenty of Fe-substituted NaCoO2 by the sol−gel method failed, and the iron was prone to form layered oxide NaFeO2 instead of forming a solid solution of cobalt and iron at the cobalt sites (Figure S5). These results indicate that the typical solid-state and sol−gel preparation techniques are not able to introduce a desired amount of Fe into the Na0.67CoO2 oxide lattice by using nonoxidative and air-stable sodium sources (e.g., Na2CO3 and NaNO3). This may be because the high-spin Fe3+ (0.645 Å) in the layered sodium transition-metal oxide structure shows an obviously larger ionic radius than Co3+ (0.545 Å) and Co4+ (0.53 Å).36,37 It is well known that manganese usually forms trivalent and tetravalent mixed valence layered oxide and has an ionic radius ranging from 0.53 to 0.645 Å, which covers the whole range of ionic radius sizes from Co4+ (0.53 Å) to high-spin Fe3+ (0.645 Å).38,39 Interestingly, Na0.67MnO2 (Figure S6) shows an identical layered lattice structure as Na0.67CoO2. This suggests that by selecting manganese as a co-dopant with iron, a Co/Fe

Figure 1. XRD patterns of (a) Na0.67MnxCo(1−x)/2Fe(1−x)/2O2 (x = 0.33, 0.5, 0.67) and (b) Na0.67Mn0.5Co0.5−xFexO2 (x = 0, 0.1, 0.2, 0.3, 0.4, and 0.5) synthesized by the solid-state method.

Mn content was 0.33, that is, for the sample with the nominal composition of Na0.67Mn0.33Co0.33Fe0.33O2, a small amount of impurity phase appeared in addition to the main layered oxide phase. However, phase-pure layered oxide phases were obtained for both Na 0 . 6 7 Mn 0 . 5 Co 0 . 2 5 Fe 0 . 2 5 O 2 and Na0.67Mn0.67Co0.17Fe0.17O2. This suggests that the strategy to increase the solubility of iron in layered oxide NaxCoO2 by introducing Mn as a co-dopant is highly effective. To further confirm the effectiveness of introducing Mn as a co-dopant to increase the solubility of iron in the layered oxide lattice, Na0.67Mn0.5Co0.5−xFexO2 with different Co/Fe ratios (x = 0, 0.1, 0.2, 0.3, 0.4, and 0.5) were synthesized by the solid-state reaction method with Na2CO3 as the sodium source and their structures were characterized by XRD. As shown in Figure 1b, all the Na0.67Mn0.5Co0.5−xFexO2 samples with different ratios of cobalt to iron exhibited the same pure layered crystal structure with no. 00-027-0751, especially without the impurity phase of sodium iron oxide. The corresponding (002) peak of ironintroduced Na0.67Mn0.5Co0.5−xFexO2 shifted to lower angles compared to that of the iron-free Na0.67Mn0.5Co0.5O2 (Figure 1b), implying a lattice expansion because of the larger size of the Fe3+ (0.645 Å) than that of Co3+ (0.61 Å). Similarly, Na0.67Mn0.5Co0.5−xFexO2 oxides prepared by the sol−gel method with NaNO3 as the sodium raw material demonstrated the same structural feature (Figure S7). 25229

DOI: 10.1021/acsami.9b06560 ACS Appl. Mater. Interfaces 2019, 11, 25227−25235

Research Article

ACS Applied Materials & Interfaces

Figure 2. (a) Typical SEM image of CF-32 particles. A layered structure can be easily seen in the SEM image. (b) Refined XRD patterns of CF-32 oxide. (c,d) HRTEM images of CF-32, and the lattice spacing of 0.562 and 0.248 nm corresponds to the (002) and (100) crystal planes, respectively. (e) Schematic diagram of the CF-32 crystal structure. The red, yellow, and blue spheres represent oxygen, sodium, and transition metals (Mn, Co, and Fe), respectively.

Figure 3. (a) LSV curves of the CF-50 and CF-32 catalysts on an RDE in oxygen-saturated 0.1 M KOH solution. The rotational speed of the RDE was 1600 rpm, and the sweep speed of the LSV was 5 mV s−1. The LSV plots of the blank electrode with only Super P conductive carbon and binder as a reference is also shown in the figure. (b) Tafel slope plots of CF-50 and CF-32 catalysts based on LSV curves. (c) Mass activity and specific activity histogram of CF-50 and CF-32 catalysts at the overpotential of 0.44 V (1.67 V vs RHE). Error bars are the standard deviation of three measurements. (d) Chronopotentiometric test (CP) curves of CF-50 and CF-32 catalysts at a current density of the 5 mA cm−2 disk with the same test environment as LSV (oxygen-saturated 0.1 M KOH and 1600 rpm).

Figure 2a shows a typical SEM image of the synthesized CF32 sample comprising micron-sized particles that were stacks of tens of nanometer-scale sheets. This suggests the layered structure nature of the oxide. The crystal structure of the assynthesized CF-32 oxide was further analyzed by Rietveld refinement of its XRD pattern, which was refined based on a layer structure with a P63/mmc space group (Figure 2b). The good fitting suggests a high purity and the sharp diffraction peaks suggest good crystallinity of the synthesized oxide. The ICP characterization of CF-50 and CF-32 confirmed that the atomic ratios of the as-prepared samples were consistent with

the target product (Table S1). Figure 2c,d shows the HRTEM images of the CF-32 particle, where clear diffraction planes with interplanar spacing of 0.562 and 0.248 nm that correspond to the (002) and (100) crystal planes, respectively, of the layer-structured CF-32 oxide can be observed. Figure 2e then gives the atom arrangement of the layer-structured CF-32 oxide, which consists of transition-metal layers based on the [TMO6] (TM = Mn, Co, and Fe) octahedron unit and sodium layers based on the [NaO6] prism unit. The effects of iron doping content tailoring were then investigated by electrochemical methods based on an RDE. 25230

DOI: 10.1021/acsami.9b06560 ACS Appl. Mater. Interfaces 2019, 11, 25227−25235

Research Article

ACS Applied Materials & Interfaces

Figure 4. (a) EIS curves of CF-50 and CF-32 electrodes at a potential of 0.7 V (vs Ag/AgCl). (b) Linear fit curves of capacitive currents vs different CV sweeping speeds (20, 40, 60, 80, 100, 120, 140, and 160 mV s−1) of CF-50 and CF-32. (c) XPS high-resolution spectra of O 1s. (d) Surface oxygen species content of the CF-32 and CF-50 catalysts based on XPS peak fitting.

The OER catalytic performance was evaluated by linear sweep voltammetric (LSV) polarization curves with a sweep speed of 5 mV s−1 (Figure 3a). The tests were performed in an oxygensaturated 0.1 M KOH electrolyte with an RDE rotation speed of 1600 rpm at room temperature. Figure 3a shows the LSV curves of the CF-50, CF-32 electrodes and the blank electrode with only Super P conductive carbon and Nafion binder. Considering the negligible catalytic activity at a potential of