Selecting Support Layer for Electrodeposited Efficient Cobalt Oxide

Here, we have grown cobalt oxide nanoflakes applying a simple one-step ... for fabrication of nanostructured carbonaceous paste were provided by Neutr...
0 downloads 0 Views 5MB Size
Research Article pubs.acs.org/journal/ascecg

Selecting Support Layer for Electrodeposited Efficient Cobalt Oxide/ Hydroxide Nanoflakes to Split Water Naimeh Naseri,*,† Ali Esfandiar,‡ Mohammad Qorbani,† and Alireza Z. Moshfegh†,§ †

Department of Physics, Sharif University of Technology, P.O. Box 11155-9161, Tehran, Iran Department of Physics, Kharazmi University, P.O. Box 37551-31979, Tehran, Iran § Institute for Nanoscience and Nanotechnology, Sharif University of Technology, P.O. Box 14588-89694, Tehran, Iran ‡

S Supporting Information *

ABSTRACT: Energy and environment crises motivated scientists to develop sustainable, renewable, and clean energy resources mainly based on solar hydrogen. For this purpose, one main challenge is the low cost flexible substrates for designing earth abundant efficient cocatalysts to reduce required water oxidation overpotential. Here, a systematic morphological and electrochemical study has been reported for cobalt oxide/hydroxide nanoflakes simply electrodeposited on four different commercially available substrates, titanium, copper sheet, steel mesh, and nickel foam. Remarkable dependence between the used substrate, morphology, and electrocatalytic properties of nanoflakes introduced flexible porous steel layer as the best substrate for samples with 499 mV overpotential, 5.3 Ω charge transfer resistance, and 0.03 S−1 turnover frequency. Besides, carbonaceous paste including carbon nanotubes and graphene sheets as the middle layer increased turnover frequency by 33%, effective surface interface nearly three times while it reduced 7.5% of resistance. Hence, optimizing the conductive nanostructured paste can lead to more efficient cobalt electrocatalysts exposing more active atomic surface sites. KEYWORDS: Co oxide/hydroxide nanoflakes, Oxygen evolution, Overpotential, Surface roughness, Electrocatalytic performance



oxygen evolution reactions8 according to the pH of the medium:

INTRODUCTION Growing concerns about limited fossil energy supplies and environmental aspects of greenhouse gases emission have attracted scientist’s attentions to finding renewable and carbon free energy resources. Hydrogen as an emission free energy carrier fulfills all desired criteria for future word and water is the most available and cheap H2 feedstock for in this scenario.1−3 At the same time, sun, the free charge source of life, is the ultimate solution for the energy crisis which provides 120 000 TW radiations per year exceeding the whole world energy demand.4 Considering solar hydrogen as the future energy carrier there are two main approaches in integrated solar water splitting devices: (i) capturing the light in proper semiconducting materials and conducting desired redox reaction (photoelectrochemical water splitting system)5,6 and (ii) coupling and efficient solar cell to provide electric power and split water in a proper electrolyzer.1,2,7 In both pathways, water is oxidized on the electrode surface involving transfer of four protons to the aqueous electrolyte and resulting in these two © XXXX American Chemical Society

2H 2O → 4H+ + 4e− + O2 (g) 4OH− → 2H 2O + 4e− + O2 (g)

(in acid) (in base)

One of the main challenges in the solar fuel generation is the lack of a suitable electrocatalyst for reducing high oxygen evolution reaction (OER) overpotential and resulting water splitting at the minimum thermodynamic required energy (1.23 V at pH = 0) with an acceptable stability under harsh oxidizing conditions.9,10 Naturally in the photosynthesis process of leaves, cubelike CaMn4O5 centers in photosystem II (PSII) act as active electrocatalysts to mediate water oxidation inspiring researchers Received: January 26, 2016 Revised: March 17, 2016

A

DOI: 10.1021/acssuschemeng.6b00178 ACS Sustainable Chem. Eng. XXXX, XXX, XXX−XXX

Research Article

ACS Sustainable Chemistry & Engineering to study similar biomimic complexes for OER.10,11 Besides, metals also catalyze OER and at the same time, they are changed to oxide phase due to high potential needed for water oxidation.12 There is a volcano relationship between required overpotentials as a function of binding energy between chemisorbed O and OH species on the variety of metal oxide surfaces.12 Although ruthenium and iridium are considered as the best OER catalysts in both acidic and basic conditions, they are one of the rarest elements on the earth which cannot be applied in large scale applications. Substantially, first row transition metal oxides show comparable catalytic efficiency and consequently, these low cost earth abundant materials have devoted considerable efforts to tune their structures and chemical compositions and optimize their electrocatalytic performance.10,12 Among them, cobalt is known as an active electrocatalyst for OER in different phases of cobalt hydroxide, (Co(OH)2), cobalt oxyhydroxide (CoOOH), and cobalt oxides (Co3O4 and CoO).13−16 Cobalt based catalyst can be synthesized using various methods including sol−gel process, solvothermal/hydrothermal, electrochemical routes, etc.15,16 Comparing these synthesis methods, electrochemical galvanostatic and/or potentiostatic approaches enable fabrication of uniform and designable structures with high purity due to absence of any undesired ions.17 Chou et al.15 have reported nearly similar electrocatalytic activity for nanoparticles with the same sizes but different oxide phases, CoO, Co3O4, under basic condition. Based on their report, these materials possess similar surface structures and exhibit analogous catalytic mechanisms leading to 473 and 497 mV potential measured at 10 mA cm−2 for ε-Co and the oxide phase, respectively. On the other hand, the shape and size of the Co3O4 nanoparticles shows strong influence on catalyst activity and the resulting overpotentials. It is found that smaller sizes pose higher effective surface interface and consequently better catalytic performance.18,19 Dong et al.18 have grown Co3O4 nanoparticles with variety of sizes between 3.5 to 70 nm simply by changing the ethanol to water ratio in the initial solution and found more efficient catalytic ozonation degradation of phenol for small sizes. Afterward, one other concern in electrocatalyst engineering is selecting a proper support layer or substrate to maximize catalyst/electrolyte interface and also facilitate charge transfer in OER. In this regard, it is found that high electronegative substrates such as gold for a thin layer of cobalt oxide film leads to formation of more active CoIV sites and, consequently, increases the turnover frequency as a criteria for catalyst activity.20 Additionally, immobilizing physically separated Co3O4 nanoclustures on mesoporous hosts such as Al2O3 is beneficial for the kinetics of water oxidation.21 In this regard, Yusuf et al. has grown Co3O4 nanoparticles inside SiO2- and γAl2O3 supports revealing both supported Co3O4 nanoparticles exhibited significant enhancement (50−80%) in oxygen evolution activity, compared with bare Co3O4 nanoparticles. They found that the major role of catalyst supports in Co3O4based water oxidation catalysts is to provide a medium to physically separate Co3O4 nanoclusters from aggregation, leading to superior photocatalytic activities. Meanwhile, cost, availability, and flexibility of substrates are also other determining factors in practical purposes for sustainable H2 production from water splitting reaction. Hence, low cost supports such as steel22 and carbon based materials13,23−27 are preferred for mass production of these electrocatalysts.

Moreover, because cobalt oxide/hydroxide activity is inherently limited by its poor electrical conductivity, conductive carbon based nanomaterials such as carbon nanotubes (CNTs)24−27 and graphene sheets23 with high conductivity, mechanical strength, flexibility, and also large surface area are promising candidates as support layers. In spite of the numerous reports in this field, no systematic comparison between various common substrates and the effect of the carbonaceous support layer on the electrocatalytic activity of cobalt oxide nanostructures was published. Here, we have grown cobalt oxide nanoflakes applying a simple one-step electrochemical method on different low cost and commercially available substrates: titanium and copper sheets, nickel foam, and steel mesh. A comprehensive study on structural and electrocatalytic properties of these samples has been reported for the first time to propose the best choice for future practical targets. Afterward, using a carbon paste including new nanostructured materials, graphene nanosheets, and multiwall CNTs (MWCNT), between the substrate and cobalt layer, the electrocatalytic performance of Co nanoflakes has been modified and investigated in detail.



EXPERIMENTAL DETAILS

Materials. All chemicals to grow cobalt oxide nanoflakes including cobalt chloride (CoCl2·6H2O, Sigma-Aldrich >98%), boric acid (H3BO3, Sigma, 99.5%), potassium hydroxide (KOH, Merck, > 99%), sodium nitrate (NaNO3, Merck, 99.9%), sulfuric acid (H2SO4, Merck, 98%), hydrogen peroxide (H2O2, Merck, 33%), and potassium permanganate (KMnO4, Merck, 99.9%) were purchased with high quality and purity. Water used in all aqueous solutions and washing procedures was deionized (DI, ∼18.2 MΩ cm−1). Commercially available MWCNTs utilized for fabrication of nanostructured carbonaceous paste were provided by Neutrino Co. with average diameters of 2−30 nm and lengths of 1−3 μm (functionalized with OH groups, purity >95%). Preparation of Working Electrodes. To fabricate cobalt based electrocatalysts, first a metallic cobalt layer was deposited electrochemically in galvanostat mode using a conventional three electrode system (SAMA, 500). For this purpose, Ag/AgCl, Pt, and four different substrates (Ti and Cu sheets, commercially available steel mesh (pore size of about 30 μm), and Ni foam (pore size of about 200 μm) with similar geometrical surface area) were applied as reference, counter, and working electrode, respectively. A mixture of aqueous solution of CoCl2 (0.5 M) and H3BO3 (0.5 M) was used as electrolyte. All depositions were performed at 5 mA/cm2 constant current density for 800 s resulting in an average thickness of about 4 μm. Schematic of electrodeposition setup has been presented in Figure 1a. Before characterizing structural and electrocatalytic properties of the samples, cobalt layers immersed in 1 M KOH solutions for about 2 h for activation by changing their surface from metallic to a mixture of oxide and hydroxide state. Afterward, cyclic voltammetry techniques in the same electrolyte was applied for five cyclic in the range of −0.2 to 0.6 V using the treated cobalt layers coated on various substrates as the working electrode to formation of oxide phase, mainly Co3O4 and Co(OH)2 on the surface. This method has been recently reported for cobalt based supercapacitors28 and also other reports on Co electrocatalyst.29 The final cobalt oxide/hydroxide electrocatalysts grown on different substrates have been named as Co/X where X refers to Cu, Ti, SM (steel mesh), and Ni (nickel foam). In another approach, to explore how a nanostructure carbonaceous support layer can effect on the sample’s electrocatalytic activities, a carbon based paste has been prepared by mixing graphene nanosheets provided by Hummers method (see the SI, Figure S130), commercially available MWCNTs, polyvinylidene fluoride (PDVF) as a bounder, and N-methyl-2-pyrrolidone (NMP) as solvent. More details on preparation the carbon paste is available in the SI. Then, the synthesized paste was drop casted on different substrates, dried, and, B

DOI: 10.1021/acssuschemeng.6b00178 ACS Sustainable Chem. Eng. XXXX, XXX, XXX−XXX

Research Article

ACS Sustainable Chemistry & Engineering

morphology as is clear in Figure 1. Figure 1b and c represent the surface structure of as-deposited cobalt layer on the copper substrate while Figure 1d and e show the same sample after activation process. According to these images, as-deposited cobalt layer consisted of packed grains with average size of 1 μm. Despite the wrinkles observed on them, each microparticles had a nearly smooth surface without any distinct flakes or branch formation. However, as the samples treated by potash, surface was roughened and the hemispherical micrograins replaced with groups of attached tablet like nanoflakes with average thickness and diameter of 100 and 400 nm, respectively. Based on Figure 1e, some of the formed flakes were totally exposed on the surface and as a result, their whole surface sites involved on the electrocatalytic reaction (solid circle in Figure 1e) while some flakes covered by others (dashed circle in Figure 1e) interfacing less active sites with the electrolyte. The formation of such flakelike structures was due to changing metallic to oxide/hydroxide phases as also observed elsewhere.31 To study the morphology dependence on the substrate, SEM pictures have been presented for activated sample grown on various substrates: Cu and Ti sheets, Ni foam, and steel mesh. Aside from the copper substrate which was discussed before, Figure 2a and b shows that formation of cobalt flakes on Ti

Figure 1. (a) Schematic of the applied electrodeposition method. (b− e) SEM images of Co layer deposited on Cu substrate (b,c) before and (d,e) after activating with KOH. then, used as the working electrode for electrodepositing the Co layer. These samples are also labeled as Co/CP/X in which “CP” refers to carbonaceous paste. Characterization Techniques. Size and morphology of the prepared cobalt oxide/hydroxide layers were studied by scanning electron microscopy (SEM, TESCANVEGA3-SB) while surface chemical state of the samples was investigated by X-ray photoelectron spectroscopy (XPS). In the later technique, Al Kα was used as monochromatic radiation source (1486.6 eV). All binding energies were calibrated by fixing C(1s) peak at 285.0 eV with accuracy of ±0.1 eV. Chemical state analysis was performed by deconvoluting all peaks to different possible states using SDP software (version 4.1) with 80% Gaussian−20% Lorentzian peak fitting. X-ray diffraction patterns of a typical sample were obtained using a STOE X-ray diffractometer with Cu Kα radiation of 1.54 Å. Electrocatalytic performance of the samples was tested by linear sweep voltammetry (LSV) and cyclic voltammetry (CV) techniques all in 1 M KOH as an electrolyte using a three electrode system (Autolab PGSTAT302) with a scan rate of 1 mV/s. Reference and counterelectrodes were used similar to the electrodeposition step as described before. To study the charge transport mechanism in the fabricated electrocatalysts, electrochemical impedance spectroscopy (EIS) was applied at a bias of 0 V with 10 mV potential perturbation amplitude, and the best equivalent circuit was fitted to the obtained data using FRA software (version 4.9).

Figure 2. SEM images of Co nanoflakes deposited on (a,b) Ti sheet, (c) steel mesh, and (d) Ni foam (Co/X samples).

sheets led to thicker cylindrical plates with random orientation and about 500 nm in diameter. Between the plates were filled with thin tiny sequin like flakes with average thickness of about 10 nm. This structure has been formed uniformly as is clear in low magnification for Ti substrate as an instance. On the other hand, changing the flat substrate with porous ones resulted in different morphology. For the case of steel mesh (Figure 2c), grains with 300 nm average size were observed each consisted of sequin like nanoflakes with thickness and lateral size of 10 and 100 nm, respectively. These nanoflakes are aligned exposing high specific surface area



RESULTS AND DISCUSSIONS Morphology of the Samples. As mentioned before, all samples have been activated by immersing in KOH aqueous solution. SEM analysis has been used to explore how this pretreatment step influenced the sample’s structure and surface C

DOI: 10.1021/acssuschemeng.6b00178 ACS Sustainable Chem. Eng. XXXX, XXX, XXX−XXX

Research Article

ACS Sustainable Chemistry & Engineering

nanoplates with thickness of 150 nm and diameter of about 1 μm and also very thin nanoflakes distributed on the wrinkled surface between the plates. For the samples fabricated on Ti sheets covered by carbonaceous paste (Figure 3c), it seems that the detected cylindrical plates before (Figure 2b) were formed again but the surface covered with randomly formed nanoflakes. In contrast, applying carbon paste for Ni foam substrate has made no meaningful change on the Co flakes morphology as is clear from Figures 2d and 3d. Finally it can be concluded that using graphene/MWCT composite paste as the mediator layer reduced uniformity of cobalt nanoflakes, provided hierarchical topography, and, hence, more specific surface area with more active catalytic sites. Surface Chemical Composition. Considering similar electrodeposition procedure for all samples, surface chemical composition was mainly affected by activation step using strong basic solution. Because of strong correlation between catalytic performance as a surface interface phenomena and surface chemistry of layers, XPS analysis was used for electrodeposited cobalt layer before and after immersion in KOH To scan how this pretreatment step may influence on the surface chemical state of Co layers. XPS results in a large energy window (Figure S3) showed that just oxygen, cobalt, and a trace amount of carbon due to adsorbed pollutions and CO/CO2 existed on the surface of both samples confirming the purity of the grown electrocatalysts. More details about the surface chemical state were determined using high resolution XPS spectra in the core level O(1s) and Co(2P) energy ranges which are all presented in Figure 4. From Figure 4a and b, it can be seen that O(1s) deconvoluted to three different peaks at 530.0, 531.8, and 533.7 eV which were attributed to oxygen in the cobalt oxide lattice, −OH groups, and water or other oxygen containing adsorbed

to the electrolyte. For nickel foam support layer (Figure 2d), low density of plates with 50 nm in thickness and about 200 nm in diameter was distributed on a wrinkled layer. Hence, it is clear that different substrates caused formation of cobalt oxide/ hydroxide nanoflakes with different morphologies, orientations, and densities resulting in more or less active catalytic sites and various level of electrocatalytic activity for water oxidation. Meanwhile, cobalt nanoflakes have been also grown on carbonaceous support layer which was a paste including graphene nanosheets and multiwall carbon nanotubes. SEM pictures of steel mesh surface completely covered by final carbon paste (Figure S2) revealed that although the substrate had porosities in the microscale, utilized carbonaceous nanostructure paste provided more surface area and increase number of available surface sites for Co flakes formation in nanometric scales. Moreover, the paste was a uniform mixture of MWCNTs and also graphene nanosheets (opaque areas surrounded by dotted line in Figure S2). To investigate how using such a carbon based support layer can influence on the surface structure and morphology of the grown cobalt flakes, SEM analysis was applied for Co/carbon paste/substrate samples and the results have been shown in Figure 3. Comparing Figures 1d and 3a revealed that for copper

Figure 3. SEM images of Co nanoflakes deposited on (a) Cu, (b) steel mesh, (c) Ti, and (d) Ni foam in which all substrates are modified with carbon paste (Co/CP/X samples).

substrates, using the paste changed the morphology of samples significantly turning to a hierarchical structure including microparticles (∼1.5 μm) with tablet like flakes (average thickness of 200 nm and lateral size of 500 nm) and sequin like thin aligned flakes (average thickness of 50 nm and lateral size of 600 nm). These components have been specified by dotted, dashed, and solid circles in the Figure 3a, respectively. In the case of steel mesh substrate by comparing Figures 2c and 3b, the observed change in the deposited and activated Co layer was also remarkable. The grains with thin and tiny nanoflakes turned to a composite morphology including

Figure 4. XPS window spectra of (a, b) O(1s) and (c, d) Co(2p) core level peak before (a, c) and after (b, d) activation with KOH. D

DOI: 10.1021/acssuschemeng.6b00178 ACS Sustainable Chem. Eng. XXXX, XXX, XXX−XXX

Research Article

ACS Sustainable Chemistry & Engineering compound, respectively.32,33 It can be confirm that the hydroxide phase which was the main portion of surface oxygen was related to Co(OH)2 species. Regarding Co(2p) XPS peaks in Figure 4c and d, four separate peaks have been fitted under the spectrum at 778.2, 780.0, 8781.9, and 785.5 eV which were attributed to cobalt in metallic zerovalent form, oxide phase, hydroxide phase, and cobalt satellite, respectively.32,33 According to the XPS spectra, immersion of Co layer in KOH solution caused two distinct effects on cobalt surface chemical state. First, for the fresh as deposited sample, 8.2% of surface Co was detected in metallic Co0 form which totally disappeared in the layers treated by KOH. Second, the observed satellite in Co(2p) spectra was attributed to existence of CoII and CoIII phases on the surface.28,34 By activating the Co layer in potash solution, the portion of satellite peaks under the Co(2p) peak increased from 25.1 to 38.7% approving this activation process was beneficial to complete oxide/hydroxide state formation. With this evidence, and also significant increase in oxide and hydroxide parts on O(1s) peaks from fresh to activated samples, it can be concluded that treating the surface by KOH resulted in transferring a metallic state to oxide/hydroxide which improves electrocatalytic activity of cobalt-based electrodes for water oxidation reaction.29 Moreover, for the as deposited samples, using the area surrounded under Co and O peak, surface concentrations were calculated 41.9 and 58.1%, respectively. After deconvolution of each peak, considering the percentage of hydroxide phase (85.1% under O(1s) and 54.8% for Co(2p)) and the stoichiometric form of Co(OH)2, the total portion of hydroxide form based on Co and O peaks was obtained 22.6 and 29.9%, respectively. These calculated values were nearly same confirming Co(OH)2 was mainly responsible for the observed surface hydroxide phase. Results similar to these have been also obtained for the samples after the activation step. Additionally, considering that the applied carbonaceous paste was deposited between substrate and Co layer, surface chemical composition of Co/X and Co/CP/X samples should be the same. The obtained XPS results from surface chemistry of fresh and KOHtreated samples have been summarized in Table 1.

octahedral sites. The hexagonal cell of brucite-like β-Co(OH)2 has been seen in the XRD pattern in Figure 5.37 It is obvious that applying carbon paste as a support layer made no remarkable change in the crystalline phase of Co electrocatalysts. Similar results have been observed for other substrates also. Electrochemical Detection of Surface. To elucidate electrochemical properties of the synthesized and activated cobalt nanostructured layer, a cyclic voltammetry technique was applied in −1.0 to +0.6 V window potential with different scan rates of 1, 50, and 200 mV/s in 1 M KOH electrolyte. The obtained results for the samples deposited steel mesh substrate modified with carbon paste presented in Figure 6 revealing

Table 1. Summarized Surface Chemical State Parameters for As-Deposited and Activated Co Layers

Figure 6. CV spectra of Co nanoflakes deposited on Cu substrate with different scan rates (1) 1, (2) 50, and (3) 200 mV s−1. (inset) Linear relation between the oxidation current of specified peak and scan rate.

sample

Oox (%)

OOH (%)

Co0 (%)

satellite (%)

as-deposited activated

7.1 15.8

65.9 85.1

8.2

25.1 38.7

Figure 5. XRD spectra of Co/CP/SM and Co/SM samples.

various oxidation peaks at −0.28, +0.16, and +0.45 V which were characteristics of cobalt species involved in the following surface reactions, respectively: Co(OH)2 + OH− → CoOOH + H 2O + e−

Crystalline Structure. In order to study the structure of prepared Co nanoflakes, XRD analysis was done which is typically shown for the samples deposited on steel mesh substrate in Figure 5. Subtracting the peaks related to the substrate and carbonaceous support layer, all observed peaks were attributed to cobalt in the oxide, hydroxide forms: Co3O4 in cubic form, CoOOH, and Co(OH)2.35−37 It is reported that cobalt hydroxides crystallize in two polymorphs, α and β. The α-cobalt hydroxides are isostructural with hydrotalcite-like compounds that consist of positively charged Co(OH)2−x layers and charge balancing anions (e.g., NO3−, CO32−, Cl−, etc.) in the interlayer gallery, while the β-form is a stoichiometric phase of the composition Co(OH)2 with brucite-like structure and consists of a hexagonal packing of hydroxyl ions with Co(II) occupying alternate rows of



Co3O4 + H 2O + OH → 3CoOOH + e −

CoOOH + OH → CoO2 + H 2O + e





(1) (2) (3)

From the inset plot, linearity behavior between current density of the peaks and the applied scan rates implies that the electron transfer process was surface controlled. 38 By comparing I−V characteristics of the bare substrate, substrate modified with just carbonaceous paste and also cobalt flakes in Figure S4, it was found that these all redox peaks disappeared for the samples without cobalt oxide/hydroxides compounds confirming all oxidation/reduction currents in a potential range of V ≤ +0.6 V were related to electrochemical activity of cobalt nanoflakes.28 In addition, difference between anodic and E

DOI: 10.1021/acssuschemeng.6b00178 ACS Sustainable Chem. Eng. XXXX, XXX, XXX−XXX

Research Article

ACS Sustainable Chemistry & Engineering Table 2. Calculated Electrocatalytic Parameters for Co Nanoflakes Grown on Different Support Layers electrocatalyst

CT (mF cm−2)

OP (mV)

CDL (mF cm−2)

RF

TOF (s−1)

RCT (Ω)

io (mA cm−2)

Co/Cu Co/Ti Co/SM Co/Ni Co/CP/Cu Co/CP/Ti Co/CP/SM Co/CP/Ni

721.3 420.0 392.5 640.4 745.6 732.2 590.8 1330.1

600 615 499 539 576 601 500 537

81.5 14.2 81.5 127.5 131.4 63.4 247.4 357.3

745 127 828 1138 1173 566 2209 3351

0.06 0.02 0.03 0.17 0.03 0.03 0.04 0.04

37.2 112.3 5.3 11.2 17.2 65.2 4.7 9.8

0.3 0.1 1.6 1.5 0.8 0.2 10.8 4.3

cathodic recorded branches of the CV spectra is proportional to the equivalent capacitance of the layer in junction with electrolyte.28,38 Hence, it is clear that applying CNT−graphene carbon paste increased the capacitance of the layer as a result of increasing its specific surface area. This enhancement was more significant in the presence of Co flakes due to simultaneous oxidation/reduction reactions. Similar results have been also observed for the samples grown on other substrates. To estimate electrochemical capacitance of the nanostructured electrode, one can use the following relation: V

CT =

∫V f J(V )dV i

ϑ(Vf − Vi )

(4)

In which CT, Vi, Vf, J(V), and ν are total capacitance of the layer (mF cm−2), staring and final potentials in the test window (V), current density response of the sample at different applied bias (mA cm−2), and the scan rate of voltage sweeping (V s−1), respectively.28,38 According to this equation, total capacitance of the cobalt nanoflakes deposited on different support layers have been calculated and listed in Table 2. These values reflected formation of electrical double layer on the surface as well as redox reactions occurred on the layer/electrolyte interface which will be discussed later more. Although these values were not directly related to electrocatalytic activity of the grown Co nanoflakes, they obtained high total capacitance, illustrating that these samples can be also a promising candidate to use as supercapacitors with some surface modification.28 In this regard, it is obvious that utilizing nanocomposite carbonbased paste also enhance the capacitance of the layers significantly, especially when high surface area substrates are applied such as Ni foam. Electrocatalytic Performance of the Samples. One activity metric to evaluate an electro-catalyst performance which is usually reported in the literatures is overpotential (OP). This quantity refers to required applied potential to obtain specific current density that is contracted 10 mA cm−2 to unify all reported values.9 The overpotentials for water oxidation on the Co nanoflakes surface has been extracted from I−V curves in which the applied voltage swept with 1 mV s−1 scan rate in the range of 0.0 to 1.2 V (vs Ag/AgCl reference electrode). According to Table 2 and Figures 7a and b, the lowest OP was measured for catalysts deposited on steel mesh substrate ∼500 mV while the highest value was recorded for the samples synthesized on Ti sheet (∼615 mV). On the other hand, applying carbonaceous paste reduces these required potentials in the case of Cu and Ti substrates which can be attributed to facilitated charge transport processes in the presence of CNT and graphene nanostructures. Turnover Frequency. Another parameter to determine how the electrocatalyst is efficient is turnover frequency (TOF)

Figure 7. I−V characteristics of Co nanoflakes deposited on different substrates (a) without and (b) with carbon support layer.

which refers to the rate of electron delivery per surface atom per second.9,13,15 Measuring this factor is more beneficial when attempting compare intrinsic activity of the electrocatalysts with different specific surface area. TOF (s−1) can be calculated according to following formula:

TOF =

J(A)NA nFRFN

(5)

Here, J(A), NA, F, RF, and N are current density obtained from unite area of the electrode (A m−2), Avogadro number, Faraday constant, roughness factor, and the total number of cobalt atoms exposed to the surface in the smooth Co layer (1.25 × 1015),15 respectively. Furthermore, n is the number of electron transferred for evolution of one molecule of O2 which equals 4. To estimate roughness factor which is the ratio of threedimensional real area to two-dimensional projected area (due to some technical problems in using BET and AFM analysis for these samples), one can use the method proposed by Chou et al.15 which is based on measuring double layer capacitance of cobalt-based electrodes: RF =

C DL CSL

(6)

CDL and CSL are double layer capacitance of Co nanoflake electrodes and double layer capacitance of a smooth Co layer both in mF cm−2. CSL has been reported as 112 μF cm−2 in the literature, while CDL has been determined based on CV curves F

DOI: 10.1021/acssuschemeng.6b00178 ACS Sustainable Chem. Eng. XXXX, XXX, XXX−XXX

Research Article

ACS Sustainable Chemistry & Engineering in a proper window potential without any redox peaks and using eq 4. Calculated RF values have been listed in Table 2 for all Co/X and Co/CP/X samples. It can be seen that Co nanostructured catalysts grown on a Cu sheet represented high roughness factor which was in coincidence with thin plate form structures observed with high density in the SEM image (Figure 1d). Even higher RF was determined for the sample deposited on steel mesh due to more vertically oriented and cross-linked flakes in the unit area with lower thickness and higher aspect ratio. However, roughness factor was remarkably reduced in the Co/Ti layer because the plates became thicker and completely packed exposing less specific surface interface. According to the SEM picture of the Co/Ni sample at large scale (Figure S5) and the reported electrochemical activity of Ni species in KOH electrolyte,39,40 the highest RF was recorded for this system, but this result was attributed to the high surface area introduced by Ni foam with big holes which was not completely and efficiently covered by the Co layer. As expected from morphology analysis, introducing nanostructured carbon paste provided Co nanoflakes with hierarchical surface topography and led to significant increase in RF values for all samples as high as 57, 345, 167, and 194% for Cu, Ti, steel, and nickel substrates, respectively. To understand whether these high obtained roughness factors were originated from nanoflakes or reflected the substrate morphology (bare substrate and/or substrates modified with carbonaceous interlayer), CDL and RF values have been also measured for X and CP/X electrodes which have been reported in Table 1 of the Supporting Information. It is obvious that, for all systems, these values were dramatically lower in comparison with the same system with cobalt nanoflakes on the surface. Finally, turnover frequencies have been determined based on these obtained roughness factors which are listed in the table. First, it should be noted that the high TOF value obtained for systems on Ni foam was not reliable due to uncovered Ni sites participating in electrocatalytic response. Afterward, Co nanoflakes grown on Cu substrates with 0.06 s−1 turnover frequency revealed more active catalytic sites as compared with one on SM and Ti substrates. However, when Co nanoflakes were fabricated on CP/Cu support layer, TOF value reduced sharply to 0.03 S−1 which implied that in spite of the dramatic increase in surface interface, most of new atomic surface sites were catalytically inactive. As a result, the best intrinsic activity of Co nanostructures was detected for Co/CP/SM system with high surface interface and also efficient atomic surface sites representing 0.04 S−1 turnover frequency which was higher than similar Co based electrocatalysts reported elsewhere.15 Interface Electron Transfer. In another approach, electrocatalyst performance can be investigated by studying charge transfer phenomena in an electrode/electrolyte interface which is possible using two different methods. First, by using electrochemical impedance spectroscopy and obtained Nyquist plots in Figure 8a and b, fitting the best equivalent circuit41 to data points (inset scheme in Figure 8a), charge transfer resistance (RCT) has been measured which represented how charge exchange is difficult for water oxidation on each sample’s surface. These values have been listed in Table 2. It is found that for both Co/X and Co/CP/X systems, the highest and lowest resistances were observed for Ti and SM substrates, respectively. In addition, for all substrates, charge transfer processes facilitated in the presence of carbon paste middle layer possibly due to high electron mobility in CNTs and graphene nanosheets.

Figure 8. Nyquist plots of Co nanoflakes deposited on different substrates (a) without and (b) with carbon paste.

The second way to probe electron transfer is based on Tafel plots extracted from I−V curves presented in Figure 7. According to the well-known Tafel equation,38,42 the intercept in a log(i)−V plot, called exchange current density (io), can be measured which also reflects surface electron transfer rate. Hence, the higher the io value, the easier the electron transfer on the surface. It can be seen that Co nanoflakes deposited on steel mesh showed more tendency to transfer the charge on the surface with an exchange current density of 1.5 mA cm−2 which increased to 10.8 mA cm−2 for ones grown on CP middle layer. The improved electron transfer recorded in Tafel plots by incorporating carbon paste was in agreement with reduced RCT in Nyquist plots which was discussed before. Consequently, not only SM substrate hosted Co nanoflakes with more efficient surface charge transfer but also introducing CNT−graphene mediator facilitated the process well. Effect of Growth Parameters. Generally, there are two different determinant factors in electrocatalyst performance: TOF which is related to intrinsic activity of each surface atom and overpotentials which is attributed to I−V response of an electrode depending on both catalytic characters of surface sites as well as number of active sites (surface interface). In this systematic comparison between various substrates with and without modification of the nanocomposite carbon-based paste, it was revealed that, although TOF values were nearly same (considering TOF unreliable for Ni substrate due to incomplete coverage), the highest one was obtained for Co/CP/SM electrocatalyst representing that atomic sites on the grown thin nanoflakes were efficiently involved in water oxidation reaction. On the other hand, by simple electrochemical techniques and measuring double layer capacitance of the electrodes, surface roughness of the samples was calculated which reflected how large the catalyst/electrolyte interface is. In this regard, the Rf value for the flakes grown on steel mesh substrate was interestingly high for both carbon supported and unsupported flakes representing a high effective surface interface between G

DOI: 10.1021/acssuschemeng.6b00178 ACS Sustainable Chem. Eng. XXXX, XXX, XXX−XXX

Research Article

ACS Sustainable Chemistry & Engineering flakes and water for O2 evolution. These observed enhancements in TOF and surface roughness for Co nanoflakes fabricated on steel mesh were mainly responsible for lowest obtained overpotential and charge transfer resistance. Hence, the charge transfer phenomenon facilitates growth of tiny flakes on SM and CP/SM leading to low required overpotentials for splitting water into O2 and H2. Moreover, CNT−graphene pastes with high conductivity exposed more surface area for flake formations and also facilitated electron transports from Co species to conductive substrates. Although applying this carbon paste made the flakes disorder, it lead to a hierarchy structure with greater roughness factor. Correlations between flake aspect ratio, size, density, and roughness factor and the electrocatalytic performance of Co can be illustrated more obviously by extending the electrodeposition time to 2000 s leading to larger flakes with a greater length to thickness ratio. Figure 9a and b shows SEM images of

based on the obtained XPS and XRD results. Electrochemical studies illustrated steel mesh as the substrate leads to lowest overpotential of 498 mV for 10 mAcm−2 current density and also least value for charge transfer resistance, 5.3 Ω at the electrode and electrolyte interface. Meanwhile, measuring double electrical layer capacitance based on CV analysis in a proper window potential revealed high roughness factor for this sample as compared with others confirming significant effective catalyst/water interface. Consequently, low cost and flexible steel mesh was the most suitable substrate for fabrication of efficient Co nanoflakes at a commercial scale. Moreover, carbonaceous paste including MWCNT and graphene as a middle layer between the flakes and substrates, facilitated the electron transport phenomenon, increased the surface interface significantly, and decreased RCT and OP values. In the most prominent case, for the SM substrate, applying carbonaceous support layer led to a 200% increase in double layer capacitance which reflects effective 3D surface interface and well as 12% reduction in charge transfer resistance. Hence, steel mesh modified with carbon paste can be the best support layer for electrocatalyst Co nanostructures. To optimize such a system and shed light on the exact improvement mechanism, more detailed studies on surface structure are required which could clarify the relation between growth parameters such as deposition time, deposition current density, CNT−graphene ratio, and electrocatalyst performance.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acssuschemeng.6b00178. Experimental details about the carbon-based paste, Figures S1−S5, and Table 1 (PDF)

Figure 9. SEM images of Co nanoflakes deposited on Cu substrate with a 2000 s electrodeposition procedure.

these flakes deposited on a Cu substrate after activated in KOH solution. Circular flakes with ∼100 nm thickness and ∼1 μm diameter were grown on the surface with high surface density representing TOF of 0.23 s−1 at 480 mV overpotential with an order of magnitude RF higher than that for Co flakes deposited for 800 s. This brief result suggested that by optimizing growth parameters, such as deposition time and current density, and the carbonaceous paste composition (CNT to graphene weight ratio), the structure and morphology of nanoflakes can be tuned to improve electrocatalytic activity more. Finally, the last question is how these synthesized electrodes are stable by passing the time? To answer, in one experiment, overpotentials, TOF values, and charge transfer resistances for all reported samples have been measured again using LSV and EIS methods, after keeping them for three mounts in standard conditions in the laboratory. Less than 10% change in the recorded parameters has been observed for the samples. In the second experiment, each working electrode run for 2 h continuously applying constant bias (∼ its overpotential) and the obtained current density was recorded by chronoamperometric technique. The observed decay in oxidation currents was not more than 5% which confirmed reasonable stability of the fabricated Co nanoflakes as water oxidation electrocatalyst.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel.: +98-21-6616-4565. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors would like to thank Research and Technology Council of Sharif University of Technology and Iran National Science Foundation (INSF) for financial support. Useful assistance of Mr. S Rafiee for XPS measurements and Mr. Dehnavi for SEM analysis is also greatly acknowledged.



REFERENCES

(1) www.energy.gov (accessed April 10, 2015). (2) Chen, X.; Li, C.; Gratzel, M.; Kostecki, R.; Mao, S. S. Nanomaterials for Renewable Energy Production and Storage. Chem. Soc. Rev. 2012, 41, 7909−7937. (3) Kao, L. C.; Liou, S. Y. H.; Dong, C. L.; Yeh, P. H.; Chen, C. L. Tandem Structure of QD Cosensitized TiO2 Nanorod Arrays for Solar Light Driven Hydrogen Generation. ACS Sustainable Chem. Eng. 2016, 4, 210−218. (4) Yan, K.; Wu, G. Titanium Dioxide Microsphere-Derived Materials for Solar Fuel Hydrogen Generation. ACS Sustainable Chem. Eng. 2015, 3, 779−791. (5) Naseri, N.; Yousefi, M.; Moshfegh, A. Z. A Comparative Study on Photoelectrochemical Activity of ZnO/TiO2 and TiO2/ZnO Nano-



CONCLUSIONS In summary, using a low cost, scalable, and simple electrodeposition method, Co nanoflakes have been grown with a variety of morphological properties on different substrates. Formation of cobalt oxide/hydroxide phases has been approved H

DOI: 10.1021/acssuschemeng.6b00178 ACS Sustainable Chem. Eng. XXXX, XXX, XXX−XXX

Research Article

ACS Sustainable Chemistry & Engineering layer Systems under Visible Irradiation. Sol. Energy 2011, 85, 1972− 1978. (6) Dong, Y.; Wu, R.; Jiang, P.; Wang, G.; Chen, Y.; Wu, X.; Zhang, C. Efficient Photoelectrochemical Hydrogen Generation from Water Using a Robust Photocathode Formed by CdTe QDs and Nickel Ion. ACS Sustainable Chem. Eng. 2015, 3, 2429−2434. (7) Reece, S. Y.; Hamel, J. A.; Sung, K.; Jarvi, T. D.; Esswein, A. J.; Pijpers, J. H.; Nocera, D. G. Wireless Solar Water Splitting Using Silicon-based Semiconductors and Earth-abundant Catalysts. Science 2011, 334, 645−648. (8) Friebel, D.; Bajdich, M.; Yeo, B. S.; Louie, M. W.; Miller, D. J.; Casalongue, H. S.; Mbuga, F.; Weng, T. C.; Nordlund, D.; Sokaras, D.; Alonso-Mori, R.; Bell, A.; Nilsson, A. On the Chemical State of Co Oxide Electrocatalysts During Alkaline Water Splitting. Phys. Chem. Chem. Phys. 2013, 15, 17460−17467. (9) 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. (10) Singh, A.; Spiccia, L. Water Oxidation Catalysts Based on Abundant 1st Row Transition Metals. Coord. Chem. Rev. 2013, 257, 2607−2622. (11) Zhang, H.; Lin, C.; Du, F.; Zhao, Y.; Gao, P.; Chen, H.; Jiao, Z.; Li, X.; Zhao, T.; Sun, Y. M. Enhanced Interactions between Gold and MnO2 Nanowires for Water Oxidation: A Comparison of Different Chemical and Physical Preparation Methods. ACS Sustainable Chem. Eng. 2015, 3, 2049−2057. (12) Li, J.; Wu, N. Semiconductor-based Photocatalysts and Photoelectrochemical Cells for Solar Fuel Generation: a Review. Catal. Sci. Technol. 2015, 5, 1360−1384. (13) Han, A.; Du, P. Facile Deposition of Cobalt Oxide Based Electrocatalyst on Low-cost and Tin-free Electrode for Water Splitting. J. Energy Chem. 2014, 23, 179−184. (14) Bloomfield, A. J.; Sheehan, S. W.; Collom, S. L.; Anastas, P. T. Performance Enhancement for Electrolytic Systems through the Application of a Cobalt-based Heterogeneous Water Oxidation Catalyst. ACS Sustainable Chem. Eng. 2015, 3, 1234−1240. (15) Chou, N. H.; Ross, P. N.; Bell, A. T.; Tilley, T. D. Comparison of Cobalt-based Nanoparticles as Electrocatalysts for Water Oxidation. ChemSusChem 2011, 4, 1566−1569. (16) Deng, X.; Tüysüz, H. Cobalt-Oxide-Based Materials as Water Oxidation Catalyst: Recent Progress and Challenges. ACS Catal. 2014, 4, 3701−3714. (17) Zhang, N.; Fan, Y.; Fan, H.; Shao, H.; Wang, J.; Zhang, J.; Cao, C. Cross-Linked Co3O4 Nanowalls Synthesized by Electrochemical Oxidation of Metallic Cobalt Layer for Oxygen Evolution. ECS Electrochem. Lett. 2012, 1, H8−H10. (18) Dong, Y.; He, K.; Yin, L.; Zhang, A. A Facile Route to Controlled Synthesis of Co3O4 Nanoparticles and Their Environmental Catalytic Properties. Nanotechnology 2007, 18, 435602−607. (19) Esswein, A. J.; McMurdo, M. J.; Ross, P. N.; Bell, A. T.; Tilley, T. D. Size-dependent Activity of Co3O4 Nanoparticle Anodes for Alkaline Water Electrolysis. J. Phys. Chem. C 2009, 113, 15068−15072. (20) Yeo, B. S.; Bell, A. T. Enhanced Activity of Gold-Supported Cobalt Oxide for the Electrochemical Evolution of Oxygen. J. Am. Chem. Soc. 2011, 133, 5587−5593. (21) Yusuf, S.; Jiao, F. Effect of the Support on the Photocatalytic Water Oxidation Activity of Cobalt Oxide Nanoclusters. ACS Catal. 2012, 2, 2753−2760. (22) Koza, J. A.; He, Z.; Miller, A. S.; Switzer, J. A. Electrodeposition of Crystalline Co3O4 - a Catalyst for the Oxygen Evolution Reaction. Chem. Mater. 2012, 24, 3567−3573. (23) Liang, Y.; Li, Y.; Wang, H.; Zhou, J.; Wang, J.; Regier, T.; Dai, H. Co3O4 Nanocrystals on Graphene as a Synergistic Catalyst for Oxygen Reduction Reaction. Nat. Mater. 2011, 10, 780−786. (24) Wu, J.; Xue, Y.; Yan, X.; Yan, W.; Cheng, Q.; Xie, Y. Co3O4 Nanocrystals on Single-walled Carbon Nanotubes as a Highly Efficient Oxygen-evolving Catalyst. Nano Res. 2012, 5, 521−530. (25) Lu, X.; Zhao, C. Highly Efficient and Robust Oxygen Evolution Catalysts Achieved by Anchoring Nanocrystalline Cobalt Oxides onto

Mildly Oxidized Multiwalled Carbon Nanotubes. J. Mater. Chem. A 2013, 1, 12053−12059. (26) Menezes, P. W.; Indra, A.; González-Flores, D.; Sahraie, N. R.; Zaharieva, I.; Schwarze, M.; Strasser, P.; Dau, H.; Driess, M. HighPerformance Oxygen Redox Catalysis with Multifunctional Cobalt Oxide Nanochains: Morphology-Dependent Activity. ACS Catal. 2015, 5, 2017−2027. (27) Liang, Y.; Wang, H.; Diao, P.; Chang, W.; Hong, G.; Li, Y.; Gong, M.; Xie, L.; Zhou, J.; Wang, J.; Regier, T. Z.; Wei, F.; Dai, H. Oxygen Reduction Electrocatalyst Based on Strongly Coupled Cobalt Oxide Nanocrystals and Carbon Nanotubes. J. Am. Chem. Soc. 2012, 134, 15849−15857. (28) Qorbani, M.; Naseri, N.; Moshfegh, A. Z. Hierarchical Co3O4/ Co(OH)2 Nanoflakes as a Supercapacitor Electrode: Experimental and Semi-Empirical Model. ACS Appl. Mater. Interfaces 2015, 7, 11172− 11179. (29) Liu, Y. C.; Koza, J. A.; Switzer, J. A. Conversion of Electrodeposited Co(OH)2 to CoOOH and Co3O4, and Comparison of Their Catalytic Activity for the Oxygen Evolution Reaction. Electrochim. Acta 2014, 140, 359−365. (30) Esfandiar, A.; Akhavan, O.; Irajizad, A. Melatonin as a Powerful Bio-antioxidant for Reduction of Graphene Oxide. J. Mater. Chem. 2011, 21, 10907−914. (31) Zhao, J.; Zou, Y.; Zou, X.; Bai, T.; Liu, Y.; Gao, R.; Wang, D.; Li, G. D. Self-template Construction of Hollow Co3O4 Microspheres from Porous Ultrathin Nanosheets and Efficient Noble Metal-free Water Oxidation Catalysts. Nanoscale 2014, 6, 7255−7262. (32) McIntyre, N. S.; Cook, M. G. X-Ray Photoelectron Studies on Some Oxides and Hydroxides of Cobalt, Nickel, and Copper. Anal. Chem. 1975, 47, 2208−2213. (33) Yao, K. X.; Zeng, H. C. Architectural Processes and Physicochemical Properties of CoO/ZnO and Zn1‑xCoxO/Co1‑yZnyO Nanocomposites. J. Phys. Chem. C 2009, 113, 1373−1385. (34) Petitto, S. C.; Marsh, E. M.; Carson, G. A.; Langell, M. A. Cobalt Oxide Surface Chemistry: the Interaction of CoO(100), Co3O4(110) and Co3O4(111) with Oxygen and Water. J. Mol. Catal. A: Chem. 2008, 281, 49−58. (35) Yang, J.; Liu, H.; Martens, W. N.; Frost, R. L. Synthesis and Characterization of Cobalt Hydroxide, Cobalt Oxyhydroxide, and Cobalt Oxide Nanodiscs. J. Phys. Chem. C 2010, 114, 111−119. (36) Chou, S. L.; Wang, J. Z.; Liu, H. K.; Dou, S. X. Electrochemical Deposition of Porous Co(OH)2 Nanoflake Films on Stainless Steel Mesh for Flexible Supercapacitors. J. Electrochem. Soc. 2008, 155, A926−A929. (37) Leng, L.; Zeng, X.; Song, H.; Shu, T.; Wang, H.; Liao, S. Pd nanoparticles decorating flower-like Co3O4 nanowire clusters to form an efficient, carbon/binder-free cathode for Li−O2 batteries. J. Mater. Chem. A 2015, 3, 15626−15632. (38) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications; Wiley: New York, 2000; Chapter 3, pp 87−136. (39) Lu, Q.; Lattanzi, M. W.; Chen, Y.; Kou, X.; Li, W.; Fan, X.; Unruh, K. M.; Chen, J. G.; Xiao, J. Q. Supercapacitor Electrodes with High-Energy and Power Densities Prepared from Monolithic NiO/Ni Nanocomposites. Angew. Chem., Int. Ed. 2011, 50, 6847−6850. (40) Li, X.; Walsh, F. C.; Pletcher, D. Nickel Based Electrocatalysts for Oxygen Evolution in High Current Density, Alkaline Water Electrolysers. Phys. Chem. Chem. Phys. 2011, 13, 1162−1167. (41) Naseri, N.; Kim, H.; Choi, W.; Moshfegh, A. Z. First Implementation of Ag Nanoparticle Incorporated WO3 Thin Film Photoanode for Hydrogen Production. Int. J. Hydrogen Energy 2013, 38, 2117−2125. (42) Naseri, N.; Yousefzadeh, S.; Daryaei, E.; Moshfegh, A. Z. Photoresponse and H2 Production of Topographically Controlled PEG Assisted Sol-gel WO3Nanocrystalline Thin Films. Int. J. Hydrogen Energy 2011, 36, 13461−13472.

I

DOI: 10.1021/acssuschemeng.6b00178 ACS Sustainable Chem. Eng. XXXX, XXX, XXX−XXX