Rationally Dispersed Molybdenum Phosphide on Carbon Nanotubes

Research Article Cite This: ACS Sustainable Chem. Eng. XXXX, XXX, XXX−XXX

pubs.acs.org/journal/ascecg

Rationally Dispersed Molybdenum Phosphide on Carbon Nanotubes for the Hydrogen Evolution Reaction Alaaldin Adam,†,‡ Munzir H. Suliman,†,‡ Hatim Dafalla,§ Abdul R. Al-Arfaj,‡ Mohammad N. Siddiqui,‡ and Mohammad Qamar*,† †

Center of Excellence in Nanotechnology (CENT), ‡Department of Chemistry, and §Center for Engineering Research (CER), King Fahd University of Petroleum and Mineral, Academic Belt Road, Dhahran 31261, Saudi Arabia

ACS Sustainable Chem. Eng. Downloaded from pubs.acs.org by UNIV OF READING on 08/21/18. For personal use only.

S Supporting Information *

ABSTRACT: Molybdenum phosphide (MoP) is viewed as a potential electrocatalyst for the electrochemical hydrogen evolution reaction (HER). However, crystallization of MoP occurs at rather high temperature (>600 °C). At this temperature, coalescence and agglomeration, which affect the performance severely, become inevitable. Herein, an oxalate-guided nonhydrolytic method is demonstrated for the preparation of MoP with smaller particle size and better dispersion qualities onto the surface of carbon nanotubes (CNTs). Molybdenum is coordinated with the oxalate group using oxalic acid, which modifies the self-assembling of molybdenum at the molecular level and renders discrete nucleation and growth of MoP on CNTs. Phosphoric acid (crystalline) was used as a source of phosphorus. The method is simple with the potential to scale-up. A probable mechanism for the growth of MoP on CNTs is proposed. The as-derived MoP/CNT electrode exhibits excellent performance, outperforming most of the MoP-based electrocatalysts, for hydrogen evolution in both acidic and basic media. In addition, the electrode possesses excellent stability. The higher performance of the electrode is rationalized in terms of small particle size with uniform dispersion, high specific and electrochemically active surface area, electrical conductivity, interfacial charge transfer kinetics, and turnover frequency. Estimation of Tafel slope is consistent with electrochemical desorption of hydrogen gas following the Volmer−Heyrovsky mechanism as the rate-determining step. KEYWORDS: Nanomaterials, Electrocatalysts, Dispersion, Transition metal phosphide, Water electrolysis, Energy conversion

■

oxides,13−15 metal borides,16,17 metal phosphides,18−27 and so forth, have been explored aiming to reduce the overpotential of the hydrogen evolution reaction (HER).28−32 Although noticeable reduction in overpotential has been achieved, a tremendous opportunity exists to further push this threshold to a lower value (closer to Pt). Recently, transition metal phosphides are reported to show a high and stable HER with lower overpotential, and are endeavored as a potential alternative to noble-metal-based electrocatalysts.18,33,34 Fundamentally, the overpotential can be substantially lowered, on one hand, by choosing an appropriate composition of electrocatalysts. Because of a unique d-band electronic

INTRODUCTION Large-scale production of H2 via photoelectrochemical or electrochemical processes has great prospects due to its low energy consumption, high purity of H2, and environmentally benign features.1 Currently, platinum (Pt) is used as the electrocatalyst. Its high cost and instability in the electrolyzer units for hydrogen generation are still predominant challenges.2 Development of electrocatalysts alternative to Pt remains at the forefront of the technology development. Developed cathodes could be employed in the photoelectrochemical and/or electrochemical water electrolysis process. Consequently, a considerable amount of research is underway to develop cathodes consisting of transition-metalbased electrocatalysts for the hydrogen evolution reaction. A variety of materials consisting of nonprecious metals, such as metal carbides,3−5 metal sulfides,6−9 metal nitrides,10−12 metal © XXXX American Chemical Society

Received: March 26, 2018 Revised: July 19, 2018

A

DOI: 10.1021/acssuschemeng.8b01359 ACS Sustainable Chem. Eng. XXXX, XXX, XXX−XXX

Research Article

ACS Sustainable Chemistry & Engineering

oxalic acid and benchmark electrocatalyst Pt/C. The HER performance of the electrocatalysts was correlated to their high specific and electrochemically active surface area, low electrochemical impedance, and high charge transfer kinetics.

structure, molybdenum-based compositions remain at the forefront of low-cost electrocatalysts for the HER. Being one of the most active catalysts for hydrodesulfurization, recently molybdenum phosphide (MoP) has been studied for the HER. The first report demonstrating MoP as a potential electrocatalyst for the HER was documented by Wang and coworkers.35 It exhibits high performance even in its bulk form.35 Like molybdenum carbides, molybdenum phosphide exhibits a phase-dependent HER activity. For instance, Wang and coworkers prepared crystalline Mo3P and MoP, and reported that the metal-rich form (Mo3P) requires significantly higher overpotential (∼500 mV) as compared to stoichiometric MoP (125 mV) to drive 10 mA cm−2 current density.35 It was further revealed that phosphorization modifies the properties of the metal, and degree of phosphorization determines the activity and stability.35 Amorphous MoP has also been demonstrated to be highly active for the HER. Lewis and coworkers produced amorphous MoP with diameter of ∼3 nm, and showed that it required only 90 mV to deliver 10 mA cm−2 current density under acidic medium.36 MoP is found to be active under both acidic as well as basic media.37 For instance, Huang et al. demonstrated that MoP required 155 mV under acidic and 184 mV under basic media to generate 20 mA cm−2.37 Because of the superior performance and low cost, supported catalysts remain the favored model for further investigation. In the case of supported catalysts, coupling between active surfaces and support may induce unique metal−support interaction. This may have significant impact on critical electrode processes such as the interfacial equilibrium, adsorption and/or desorption of reactive species, and their interaction with electrocatalysts surface, electrical conductivity, and so forth. The synergistic chemistry between active sites and support could favorably be induced by achieving higher and uniform surface coverage of support by functional sites. However, uniform dispersion of MoP on support remains a challenging endeavor. In the absence of any driving force, self-assembling, agglomeration, nucleation, coalescence, and uncontrolled growth of functional sites are predominant, particularly at high synthesis temperature (>700 °C). This produces catalysts with low surface area and lessexposed active sites. Hence, developing a protocol which could warrant mitigation of agglomeration and coalescence, and render discrete growth of active sites on support, is crucial to obtain high-performance electrocatalysts. The aim of this study, therefore, is to develop a simple nonhydrolytic route to produce small and highly dispersed MoP onto the surface of carbon. This is achieved by controlling the self-assembly of molybdenum at the molecular levelmolybdenum is coordinated with the oxalate group, which seemed to (1) enhance the immobilization of Mo on the carbon nanotube (CNT) surface through interaction between carbon functionalities and oxygen, and (2) increase the intermolybdenum distance through steric hindrance. In addition, crystalline phosphoric acid is used as a nonhazardous phosphorus source (unlike the most popular trioctylphosphine, sodium hypophosphite). As the oxalic and phosphoric acids are cost-effective, the method is simple with the potential to scale-up. A plausible mechanism of molybdenum complexation with oxalate and concomitant growth of MoP is proposed. The electrocatalytic performance of oxalate-derived MoP/CNT was examined toward the HER in both acidic and basic media, and compared with that of MoP/CNT prepared in the absence of

■

EXPERIMENTAL SECTION

Synthesis of MoP/CNT Electrocatalyst. First, a homogeneous suspension consisting of carbon nanotubes (CNTs) and anhydrous ethanol was prepared in a closed vial. Then, a calculated amount of molybdenum chloride (MoCl5) was dissolved, and the solution was kept under stirring for 2 h. After complete dissolution of MoCl5, oxalic acid and crystalline phosphoric acid were added, and the solution was kept under stirring. The stirring was carried out at 70 °C for 12 h to achieve maximum complexation of molybdenum with the oxalate group (C2O42−). After reaction, the vial was opened, and ethanol was allowed to evaporate under a vigorous stirring, leaving behind a black powder. The resulting product was dried in an oven at 110 °C overnight, and finally carburized in a tubular furnace in the flow of H2/Ar mixture (10:90) at the desired temperature (with a heating rate of 5 °C min−1) for 2 h. The flow of mixture gas was maintained at 100 mL min−1 with a mass flow controller (Alicat) throughout the calcination process. After the reaction, the black product was collected and used as the electrocatalyst for the HER. For comparison, MoP/CNT was also synthesized without oxalic acid following identical conditions. Characterization. Morphological and detailed microstructural attributes of the materials were discerned with the aid of X-ray diffractometry (XRD, Rigaku MiniFlex), X-ray photoelectron spectroscopy (XPS, Thermo Scientific ESCALAB 250Xi), transmission and high-resolution transmission electron microscopy and selected area electron diffraction (TEM/HR-TEM, FEI Tecnai TF20) (SAED), and BET surface area analyzer (Micromeritics ChemiSorb 2750). Evaluation of Electrocatalytic Activity. A homogeneous ink solution was prepared by sonicating a suspension consisting of electrocatalyst (10 mg), water and isopropanol (30% v/v), and 37 μL of 1.66% w/w Nafion for approximately 30 min. A measured amount (16 μL) of ink was drop-cast on a precleaned glassy carbon (GC) disc electrode (5.0 mm diameter, 0.196 cm2, Pine Instruments), and the electrode was allowed to dry under air flow at ambient conditions. The deposition steps were repeated to achieve the desired catalyst loading on the GC electrode. The hydrogen evolution reaction (HER) was studied in a three-electrode cell assembly connected to a potentiostat (EG&G 273A). A saturated calomel electrode (mercury/ mercury chloride, SCE) and coiled platinum were used as the reference and counter electrode, respectively. A 0.5 M H2SO4 or 1.0 M KOH aqueous solution was used as working electrolyte. Linear sweep voltammetry was applied with a scan rate of 5 mV s−1. The SCE electrode was calibrated against a reversible hydrogen electrode (RHE), and its potential was converted into RHE potential. Unless mentioned otherwise, the current density was calculated against the geometric area of the glassy carbon electrode and was presented after iR correction. Before and during the cathodic measurement, highpurity H2 gas was used to remove the dissolved O2 from the solution. Electrochemical impedance spectroscopy (EIS) was performed in 0.5 M H2SO4 in the frequency range 105−0.01 Hz with ac amplitude of 10 mV. All the EIS data was normalized to the geometric surface area of the working electrode. Calculation of Turnover Frequency (TOF). For elucidation of the active sites, TOF was estimated using the following equation: TOF = JA /2FN

(1)

2

where J (A/cm ) is the geometric current density recorded during the linear sweep voltammetry (LSV) measurement in 0.5 M H2SO4, A is the geometric area of the glassy carbon electrode GC (0.196 cm2), 1/ 2 denotes that two electrons are required to form one hydrogen molecule from water, F is the Faradaic constant (96 485 C mol−1), and N is the number of active sites (mol). N was determined by carrying out CV measurements between −0.2 V and +0.6 VRHE in 1.0 B


Research Article


Figure 1. Schematic showing complexation/ion exchange, nucleation, and concomitant growth of crystalline MoP on the CNT surface.

Figure 2. XRD of MoP/CNT synthesized in the presence (a) and absence (b) of H2C2O4, and XPS signatures of C 1s, Mo 3d, and P 2p in oxalatederived MoP/CNT.

■

M phosphate buffer solution with a scan rate of 20 mV s−1. While it is difficult to assign the observed peaks to a given redox couple, N should be proportional to the integrated charge over the whole potential range. Assuming a one electron redox reaction, the upper limit of active sites was calculated as given in the following equation:

RESULTS AND DISCUSSION

A plausible mechanism showing complexation and nucleation with concomitant growth of MoP/CNT is illustrated in Figure 1. A rapid reaction between molybdenum chloride and ethanol produces an intermediate product and HCl and/or C2H5Cl. Upon addition of oxalic acid, an ion-exchange reaction takes place between ethoxy and oxalate groups producing a

N = Q /2F where Q is voltammetric charge. C


Research Article


Figure 3. TEM of MoP/CNT synthesized in the absence (a) and presence (b−d) of H2C2O4. HR-TEM (e) and SAED (f) results of oxalatederived MoP/CNT. Also included is the elemental mapping of carbon, molybdenum, and phosphorus.

phosphides, though the amorphous phase as indicated by XRD patterns recorded a function of synthesis temperature (Figure S2). At higher temperature (≥650 °C), the amorphous phase crystallizes to produce crystalline MoP. All these processes occurred onto the surface of CNTs. Powder X-ray diffraction (XRD) results of MoP prepared with and without oxalic acid are compared in Figure 2. The diffraction peak at 26.1° corresponds to the 002 plane of CNTs, while diffractions centered at 27.9° (001), 31.2° (100), 43.1° (101), 57.5° (110), 65.0° (111), 67.0° (102), and 74.3° (201) are indexed to MoP with a hexagonal closed packed structure (JCPDS 24-0771)38 (P6m2) in which Mo is 6coordinated by P atoms. Compared to those of MoP/CNT prepared without oxalic acid, diffractions of oxalate-derived MoP/CNT are wide and less intense, suggesting restricted growth of nanoparticles. Details of chemical composition and oxidation states of electrocatalyst were obtained by X-ray photoelectron spectroscopy (XPS). Survey spectrum verified the presence of intended compositions (Figure S3). Signatures of C 1s, Mo 3d, and P 2p of as-synthesized MoP are included in Figure 2. In addition to the presence of graphitic carbon at 284.7 eV, the characteristic signal of C−O (at 285.7 eV) was also detected.39,40 The spectrum in the Mo 3d region suggests the presence of Mo in different oxidation states. Peaks centered at ∼235.6 and 233.4 eV account for Mo 3d3/2 and Mo 3d5/2 of Mo6+ (MoO3) spectral lines, which agrees well with those reported for MoP.35,41,42 Although located at a slightly higher value than that of metallic Mo, peaks at 228.4 and 231.6 eV account for

molybdenum−oxalate complex. Formation of molybdenum− oxalate complex was indicated by a change of solution color. It was confirmed by C13 and H1 nuclear magnetic resonance (NMR) spectroscopic techniques (Figure S1). The NMR spectrum of oxalic acid dissolved in D2O is shown in Figure S1a. The peak at 162.24 ppm is assigned to the resonance of the carbon of the oxalate group. For comparison, the NMR spectrum of the product obtained in the reaction between molybdenum chloride and oxalic acid in ethanol is presented in Figure S1b. A downfield shift in the resonance of carbon was noticed; the peak shifted to 166.77 ppm due to a likely interaction between oxalate (C2O42−) and the molybdenum (Mo) metal center, forming a molybdenum−oxalate complex. Moreover, only one peak for carbon was detected in both the cases. This suggested the presence of a similar coordination environment around all carbon coordinated to molybdenum, and a C2 symmetry center in the complex. The possibility of any proton coordinated to this molybdenum−oxalate complex was corroborated by H1 NMR. Only one resonance peak at 4.6 ppm corresponding to deuterated solvent was recorded, implying the absence of any proton in the molybdenum− oxalate complex. Because of the rather bulky nature of the oxalate group, a disseminated self-assembling of molybdenum with defined interatomic distance is likely. This discrete molecular assembly seems to mitigate the agglomeration of molybdenum and steer a controlled nucleation and growth. When heated, molybdenum−oxalate decomposes and reacts with PO43− forming molybdenum phosphate. In the presence of H2/Ar, phosphate is reduced and transformed into D


Research Article


Figure 4. Comparative polarization curves in acidic (a) and basic (b) media: MoP/CNT prepared in absence and presence of oxalic acid (A, B), respectively; Pt/C (C). Specific activity of electrodes in acidic (c) and basic (d) media. Current−time curve (e) and polarization curves (f) obtained before and after 1000 potentiodynamic sweeps.

MoP.35,41,42 Peaks at binding energies of 229.3 and 232.3 eV are attributed to oxides of molybdenum. The surface of MoP is susceptible to oxide (MoO2 and MoO3) formation when exposed to air atmosphere, as observed in previous studies.35,41,42 Deconvolution of the P 2p spectrum reveals two spectral signals at lower binding energy (peaks at 129.9 and 130.9 eV), corresponding to the phosphorus anion bonded to molybdenum in MoP, and at higher binding energy (peaks at 133.7 and 134.6 eV), accounting for phosphate species.35,41,42 The morphology and dispersion of MoP on CNTs are highlighted in Figure 3. The morphology of CNTs was retained at high processing temperature such as 700 °C. For comparison, morphological and dispersion qualities of MoP prepared in the absence and presence of oxalic acid were examined by TEM. The microscopy image shown in Figure 3a highlights the size, morphology, and homogeneity of MoP obtained without oxalic acid. Analysis reveals a wide particle size distribution, between 5 and 40 nm, of MoP on CNT. On the contrary, particles of oxalate-derived MoP were quasispherical and uniformly distributed with size in the range between 5 and 15 nm (Figure 3b,c). Field emission SEM (FESEM) and TEM images giving a broader view of the MoP dispersion are presented in Figures S4 and S5. Some of the MoP particles seemed to grow inside the carbon nanotubes (Figure 3d). Highly uniform and ultrafine size (∼2 nm) implies that MoP nucleation and growth occurred inside the tubular confinement. Microscopy observations reinforce our assumption that complexation of molybdenum with oxalate renders a discrete inter-molybdenum complex distance, and

thus controlled nucleation and growth of MoP nanoparticles. This led to the formation of fine and uniformly dispersed MoP on CNT. The HR-TEM microscopy image and the electron diffraction (Figure 3e,f) show the polycrystalline nature of the sample with high degree of crystallinity. The interplanar distance was estimated to be 0.28 nm, which corresponds to the (100) plane. This is in good agreement with 0.28 nm, obtained from the XRD diffraction peak at 31.2°. Distribution of C, Mo, and O in oxalate-derived MoP/CNT was mapped, and results are included in Figure 3. Molybdenum seems to be evenly distributed throughout the surface. The performance of as-synthesized electrocatalysts was tested toward the hydrogen evolution reaction in acidic (0.5 M H2SO4), neutral (1.0 M phosphate buffer), and basic (1.0 M KOH) aqueous electrolytes. The current density was normalized to the geometric area of the glassy carbon electrode, unless otherwise stated. Before the linear sweep voltammograms were recorded, working electrodes were swept for 10 cycles between +0.1 and −0.3 VRHE to allow the interface to attain equilibrium, if any. The performance of oxalate-derived MoP/CNTs was compared with that of MoP/ CNT (obtained in the absence of oxalic acid) and commercial Pt/C. A comparative electrocatalytic performance was carried out under identical experimental conditions with the same mass of catalyst loaded on the GC electrode. Cathodic polarization (current density versus potential) profiles under acidic and basic media are shown in Figure 4a,b, respectively. Under acidic conditions, the overpotentials, recorded to produce a current density of 10 mA cm−2, for Pt/C, oxalatederived MoP/CNT, and MoP/CNT were ∼41, ∼114, and E


Research Article


Figure 5. Nyquist plots of MoP/CNT prepared in the absence (a) and presence (b) of oxalic acid. (c, d) Enlarged view of Nyquist plots showing electrochemical impedance and charge transfer behavior. (e) Bode plots. (f) Tafel slopes.

∼160 mV, respectively. The overpotential of oxalate-derived MoP/CNT outperforms most of the MoP-based electrocatalysts under acidic medium. Table S15 summarizes the electrocatalytic performance of MoP-based electrocatalysts for the HER. A similar activity trend, though requiring higher potential, was discerned under basic media. For the identical current density, respective potentials were calculated to be ∼65, ∼170, and 220 mV. As expected, commercial Pt/C exhibited the best HER activity. The oxalate-derived electrocatalyst required much lower potential as compared to MoP/ CNT prepared in the absence of oxalic acid. The performance of oxalate-derived MoP/CNT at pH = 7 is shown in Figure S6. As is evident, the oxalate-derived MoP/CNT exhibited poor performance in neutral medium. In all the cases, polarization curves show the presence of hysteresis before the onset potential. This hysteresis is highlighted by performing cyclic voltammetry, as shown in Figure S7. The reduction peaks centered at ∼−45 and ∼−70 mV in acidic and basic media, respectively, are indicative of a surface redox process. This could presumably be attributed to the reduction of surface PO43− species,43 which was confirmed by XPS. A similar hysteresis was also noticed in a previous study using MoP electrocatalyst.43 The performance of electrodes was also evaluated in terms of specific activity. Specific activity was estimated against BET surface areacurrent was normalized with respect to the specific surface area of the catalysts in lieu of the geometric area of the GC electrode. Specific surface area of MoP/CNT, prepared with and without oxalic acid, was measured to be 48.7 and 37.2 m2 g−1, respectively. Higher surface area could presumably be due to smaller particle size of MoP. Respective nitrogen sorption isotherms are shown in Figure S8. Potentiodynamic curves of specific activity (specific current density versus potential) under acidic and basic media are shown in Figure 4c,d. As could be seen, oxalate-derived MoP/ CNT exhibits better specific performance than MoP/CNT obtained without oxalic acid. The superior performance of oxalate-derived MoP/CNT could presumably be attributed to

smaller size and improved dispersion of active sites, as discerned through microscopy. In addition to specific surface area, the HER performance of the electrode was also correlated to electrochemical impedance, charge transfer kinetics, and electrochemically active surface area. Stability tests were performed under both acidic and alkaline media under identical experimental conditions: MoP/CNT with 50 wt % Mo, catalyst loading 1 mg cm−2, 0.5 M H2SO4 or 1.0 M KOH. Figure 4e,f demonstrates the durability of the oxalate-derived electrode under acidic conditions. Two assessment methods were pursued to evaluate the stability of the electrode: (1) chronoamperometry (between 0.1 and −0.285 VRHE with a scan rate of 20 mV s−1 for 1000 cycles) and (2) chronopotentiometry (at overpotential (η) of 141 mV for 24 h). The time-dependent profile of current density at constant applied potential of 141 mV is shown in Figure 4e. A steady current flow was recorded for 24 h without any noticeable drop in the current density (Figure 4e). Moreover, repolarization of the electrode after potentiostatic measurement did not indicate any noticeable modification in the requirement of η to produce equal magnitude of current density. On the other hand, in the chronoamperometry experiment, the used electrode was repolarized after 1000 scans, and the current density was quantified as a function of potential. The polarization curves (current density−potential curve) before and after continuous cycles are compared in Figure 4f. The required overpotential to produce similar current density remained almost intact after 1000 cycles, indicating excellent endurance of the electrode. Contrary to the excellent stability under acidic medium, the electrode was found to be unstable under alkaline. The chronoamperometry and chronopotentiometry results recorded in 1.0 M KOH solution are shown in Figure S9. For further insight about the electrodes’ structural and chemical stability in two electrolytes, the used electrocatalyst (collected after the durability test) was analyzed by XRD, XPS, and HRTEM, and the results are shown in Figures S10−S14. All the performed analyses confirm the structural and chemical stability of the electrode under acidic medium. On the F


Research Article


Figure 6. Cyclic voltammograms (a, b) and their corresponding plots (c, d) of the current density at 0.248 VRHE vs scan rate: MoP/CNT prepared in the absence of oxalic acid (a, c), and oxalate-derived MoP/CNT (b, d). Cyclic voltammograms (e) recorded between −0.2 and 0.6 VRHE and their corresponding TOF profiles vs overpotential (f).

CNT possessed better electrical conductivity, translating into improved performance. In addition to smaller Rct, Nyquist plots indicated the presence of two semicircles or two time constants. This was corroborated by Bode plotscorresponding Bode plots obtained at η = 150 mV are shown in Figure 5e. Both Nyquist and Bode data constitute two semicircles, though it was much more discernible in oxalate-derived MoP/CNT. The semicircle recorded at higher frequency originates from surface porosity, while the near-semicircle appearing at lower frequency accounts for charge transfer involved in the hydrogen evolution reaction.45 For electrochemical hydrogen evolution reactions, EIS findings are predominantly explained by three types of electrical equivalent circuit models; one time constant model,46 two time constant parallel model,4,47 and two time constant serial models.48 On the basis of the electrochemical findings, a two time constant parallel model consisting of solution resistance (Rs) in series with two parallel constant phase element resistance was used to fit the experimental data (Figure S15). According to this equivalent circuit model, Rs is a collective resistance, which incorporates resistances from wiring (Rwiring), carbon support (Rcarbon), MoP (RMoP), and solution (Rsolution).45 Charge transfer resistance and capacitance are denoted by Rct and Cdl, respectively. Evaluation of Tafel slope could shed light on the operative mechanism in the HER. Under acidic conditions, the HER proceeds via two different pathways involving three probable reactions: (1) Volmer reaction (adsorption), (2) Tafel reaction (chemical desorption), and (3) Heyrovsky reaction (electrochemical desorption). The mechanism could be determined by estimating the Tafel value, which can be derived either from polarization curve or EIS measurements. Although the Tafel slope computed from the polarization curve is wellfounded,4,49,50 selection of an inappropriate region of the polarization curve could lead to ambiguous interpretation and consequently the HER mechanism. However, Tafel values

contrary, under alkaline conditions, the electrode’s structure and composition undergo a drastic change. The XRD diffractions of used MoP/CNT correspond to MoOPO4, and no phosphide phase was detected. This was corroborated by XPS. In the case of Mo 3d, peaks with binding energy of 233.24 and 236.79 eV, corresponding to Mo 3d5/2 and Mo 3d3/2, are attributed to Mo6+ of MoO3.44 For the P 2p spectrum, peaks at binding energy of 133.58 and 134.47 eV are assigned to phosphate species.44 The decrease in the electrode performance under alkaline conditions relates to the structural and chemical instability of the electrocatalyst. Bulk conductivity as well as surface-charge-resistance of MoP/CNT were probed by employing electrochemical impedance spectroscopy (EIS). Measurements were conducted in the frequency range 105−0.01 Hz with ac amplitude of 10 mV in 0.5 M H2SO4 aqueous electrolyte. Representative potential-dependent Nyquist plots (real versus imaginary impedance) are shown in Figure 5a,b, respectively, for MoP/ CNT prepared with and without oxalic acid. Plots were constructed as a function of overpotential (η = 0, 30, 50, 70, 90, 110, 130, and 150 mV). The semicircular dispersion of impedance is characteristic of an interfacial phenomenon occurring near the electrode surface. Shorter semicircular arcs imply faster electron transfer at the interface and vice versa. As is evident, surface-charge-resistance (Rct) was found to be a function of overpotential. Rct decreases with increasing overpotential, implying faster charge transfer kinetics at higher η. Furthermore, the Rct at all applied potentials (Table S14) of oxalate-derived MoP/CNT was substantially lower than that prepared in the absence of oxalic acid. For instance, Nyquist plots of both the electrocatalysts recorded at η = 0 and 150 mV are compared in Figure 5c,d. At 150 mV, Rct of oxalate-derived MoP/CNT was ∼9.9 Ω, whereas Rct of MoP/CNT obtained in the absence of oxalic acid was ∼19.5 Ω. This indicates that oxalate-derived MoP/ G


Research Article

ACS Sustainable Chemistry & Engineering obtained through EIS measurements are more persuasive.18,51 Tafel slopes derived from EIS results are shown in Figure 5f. The calculation of the semilogarithmic values of the inverse of Rct against η resulted in a linear relationship with a gradient, which is attributed to the Tafel slope of 51.6 and 54.5 mV dec−1 for MoP/CNT synthesized with and without oxalic acid, respectively. Such values imply that the Heyrovsky mechanism was operative, and the rate-determining step was more likely electrochemical desorption. The augmented performance of the oxalate-derived electrode was also correlated to enhanced electrochemically active surface area (ECSA). Since the electrochemical hydrogen evolution reaction is a predominantly surfacedictated process, usually high surface area (specific and ECSA) is desirable. The electrochemically active area was quantified through the measurement of double layer capacitance (Cdl) of cyclic voltammograms (CVs), which were recorded at different scan rates from 5 to 100 mV s−1. CVs together with corresponding plots for MoP/CNT prepared in the absence and presence of oxalic acid are shown in Figure 6. Cdl values were determined to be approximately 59 and 92 mF cm−2, respectively. Usually, higher Cdl is discerned as higher ECSA, inducing higher electrocatalytic performance. For further evaluation and comparison of the catalytic ability of as-prepared electrodes, turnover frequency (TOF) was calculated. TOF signifies the number of reactant molecules reacted per active site in unit time. Higher TOF values denote higher catalytic ability of the electrode. Since electrocatalytic performance depends on the catalyst mass deposited on the working electrode (catalyst loading) and is usually investigated with different mass loadings, estimation of TOF would allow the comparison of catalytic performance of as-prepared electrodes with those reported in the literature. First, the number of active sites was quantified by electrochemical method.52 Figure 6e compares the cyclic voltammograms of MoP/ CNT electrodes recorded between −0.2 V and +0.6 VRHE. Experiments were carried out in neutral phosphate buffer solution (pH = 7) with a scan rate of 20 mV s−1. The number of active sites for oxalate-derived MoP/CNT and MoP/CNT was estimated to be 1.1 × 10−8 and 4.7 × 10−9 mol, respectively. The number of active sites was used to determine TOF using eq 1, and the resulting TOF values were plotted against potential (Figure 6f). At η = 114 mV (the potential to drive a geometric current density of 10 mA), the electrode comprising oxalate-derived MoP/CNT exhibits significantly higher TOF (0.93 s−1) as compared to MoP/CNT (0.6 s−1). This could presumably be ascribed to uniform dispersion of small MoP on carbon nanotubes, rendering more functional sites accessible for H+.

and electrochemically active surface area, small surface-chargeresistance, high electron transfer kinetics, and high turnover frequency. As a result, the electrode outperformed most MoPbased electrocatalysts for the HER, requiring only 114 mV to produce a current density of 10 mA cm−2 in acidic electrolyte. In addition, the electrode showed excellent stability. On the contrary, the electrode was found to be unstable in basic medium. Elucidation of the rate-determining step indicated that the electronic structure of MoP/CNT could be modulated to facilitate the electrochemical desorption of H2 gas and achieve even higher efficiency. As the disclosed method is simple yet effective with potential to scale-up, it is likely to open new avenues for discrete growth and controlled dispersion of active sites on the support for the development of highly active non-noble-metal-based catalysts for a variety of applications.

■

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acssuschemeng.8b01359. 13 C NMR, temperature-dependent evolution in XRD patterns, XPS survey spectra, FESEM and TEM images, HER performance, CV, nitrogen adsorption−desorption isotherms, stability test, HR-TEM images, electrical equivalent circuit model exercised to fit the electrochemical impedance (EIS) results, EIS fitting values (Rs, Rp, and Rct), and comparison of electrocatalysts (MoP) for the HER (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Mohammad Qamar: 0000-0002-5351-9872 Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors acknowledge the support provided by the Center of Excellence in Nanotechnology through NT-2018-MQ at King Fahd University of Petroleum and Minerals (KFUPM). The authors acknowledge the help of Dr. Adam Seliman for NMR analysis and discussion.

■

■

CONCLUSIONS In summary, an oxalate-guided nonhydrolytic method is reported to obtain small and highly dispersed nanoparticles of MoP on a CNT surface. The coordination between Mo and oxalate group seemed to (1) enhance the immobilization of Mo on the CNT surface through interaction between carbon functionalities and oxygen of oxalate, and (2) increase the inter-molybdenum distance through steric hindrance. As a result, formation of small and well-dispersed MoP on CNTs was discerned. Smaller particle size and improved dispersion of MoP on CNTs imparts benign attributes such as high specific

REFERENCES

(1) Turner, J. A. Sustainable Hydrogen Production. Science 2004, 305, 972−974. (2) Harnisch, F.; Sievers, G.; Schröder, U. Tungsten Carbide as Electrocatalyst for the Hydrogen Evolution Reaction in pH Neutral Electrolyte Solutions. Appl. Catal., B 2009, 89, 455−458. (3) Michalsky, R.; Zhang, Y. J.; Peterson, A. A. Trends in the Hydrogen Evolution Activity of Metal Carbide Catalysts. ACS Catal. 2014, 4, 1274−1278. (4) Chen, W. F.; Wang, C. H.; Sasaki, K.; Marinkovic, N.; Xu, W.; Muckerman, J. T.; Zhu, Y.; Adzic, R. R. Highly Active and Durable

H


Research Article


Hydrogen-Evolving Cathode over the Wide Range of pH 0−14. J. Am. Chem. Soc. 2014, 136, 7587−7590. (23) Liu, T.; Xie, L.; Yang, J.; Kong, R.; Du, G.; Asiri, A. M.; Sun, X.; Chen, L. Self-Standing CoP Nanosheets Array: A Three-Dimensional Bifunctional Catalyst Electrode for Overall Water Splitting in both Neutral and Alkaline Media. ChemElectroChem 2017, 4, 1840−1845. (24) Pan, Y.; Sun, K.; Liu, S.; Cao, X.; Wu, K.; Cheong, W. C.; Chen, Z.; Wang, Y.; Li, Y.; Liu, Y.; Wang, D. Core-Shell ZIF-8@ ZIF67 Derived CoP Nanoparticles-Embedded N-doped Carbon Nanotube Hollow Polyhedron for Efficient Over-all Water Splitting. J. Am. Chem. Soc. 2018, 140, 2610−2618. (25) Pan, Y.; Hu, W.; Liu, D.; Liu, Y.; Liu, C. Carbon Nanotubes Decorated with Nickel Phosphide Nanoparticles as Efficient Nanohybrid Electrocatalysts for the Hydrogen Evolution Reaction. J. Mater. Chem. A 2015, 3, 13087−13094. (26) Pan, Y.; Liu, Y.; Lin, Y.; Liu, C. Metal Doping Effect of the M− Co2P/Nitrogen-Doped Carbon Nanotubes (M= Fe, Ni, Cu) Hydrogen Evolution Hybrid Catalysts. ACS Appl. Mater. Interfaces 2016, 8, 13890−13901. (27) Pan, Y.; Lin, Y.; Chen, Y.; Liu, Y.; Liu, C. Cobalt PhosphideBased Electrocatalysts: Synthesis and Phase Catalytic Activity Comparison for Hydrogen Evolution. J. Mater. Chem. A 2016, 4, 4745−4754. (28) Ji, X.; Liu, B.; Ren, X.; Shi, X.; Asiri, A. M.; Sun, X. P-Doped Ag Nanoparticles Embedded in N-Doped Carbon Nanoflake: An Efficient Electrocatalyst for the Hydrogen Evolution Reaction. ACS Sustainable Chem. Eng. 2018, 6, 4499−4503. (29) Anjum, M. A. R.; Jeong, H. Y.; Lee, M. H.; Shin, H. S.; Lee, J. S. Efficient Hydrogen Evolution Reaction Catalysis in Alkaline Media by All-in-One MoS2 with Multifunctional Active Sites. Adv. Mater. 2018, 30, 1707105. (30) Anjum, M. A. R.; Lee, M. H.; Lee, J. S. BCN NetworkEncapsulated Multiple Phases of Molybdenum Carbide for Efficient Hydrogen Evolution Reactions in Acidic and Alkaline Media. J. Mater. Chem. A 2017, 5, 13122−13129. (31) Raza, M. A.; Ahmed, R.; Saleem, A.; Din, R. U. Fabrication of Carbon−Polymer Composite Bipolar Plates for Polymer Electrolyte Membrane Fuel Cells by Compression Moulding. Nucleus 2009, 46, 351−356. (32) Ham, D. J.; Lee, J. S. Transition Metal Carbides and Nitrides as Electrode Materials for Low Temperature Fuel Cells. Energies 2009, 2, 873−899. (33) Callejas, J. F.; Read, C. G.; Roske, C. W.; Lewis, N. S.; Schaak, R. E. Synthesis, Characterization, and Properties of Metal Phosphide Catalysts for the Hydrogen-Evolution Reaction. Chem. Mater. 2016, 28, 6017−6044. (34) Xiao, P.; Chen, W.; Wang, X. A Review of Phosphide-Based Materials for Electrocatalytic Hydrogen Evolution. Adv. Energy Mater. 2015, 5, 1−13. (35) Xiao, P.; Sk, M. A.; Thia, L.; Ge, X.; Lim, R. J.; Wang, J. Y.; Lim, K. H.; Wang, X. Molybdenum Phosphide as an Efficient Electrocatalyst for the Hydrogen Evolution Reaction. Energy Environ. Sci. 2014, 7, 2624−2629. (36) McEnaney, J. M.; Crompton, J. C.; Callejas, J. F.; Popczun, E. J.; Biacchi, A. J.; Lewis, N. S.; Schaak, R. E. Amorphous Molybdenum Phosphide Nanoparticles for Electrocatalytic Hydrogen Evolution. Chem. Mater. 2014, 26, 4826−4831. (37) Chen, Z.; Lv, C.; Chen, Z.; Jin, L.; Wang, J.; Huang, Z. Molybdenum Phosphide Flakes Catalyze Hydrogen Generation in Acidic and Basic Solutions. Am. J. Anal. Chem. 2014, 5, 1200−1213. (38) Deng, C.; Ding, F.; Li, X.; Guo, Y.; Ni, W.; Yan, H.; Sun, K.; Yan, Y. M. Templated-Preparation of a Three-Dimensional Molybdenum Phosphide Sponge as a High Performance Electrode for Hydrogen Evolution. J. Mater. Chem. A 2016, 4, 59−66. (39) Wang, S.; Wang, J.; Zhu, M.; Bao, X.; Xiao, B.; Su, D.; Li, H.; Wang, Y. Molybdenum-Carbide-Modified Nitrogen-Doped Carbon Vesicle Encapsulating Nickel Nanoparticles: A Highly Efficient, LowCost Catalyst for Hydrogen Evolution Reaction. J. Am. Chem. Soc. 2015, 137, 15753−15759.

Nanostructured Molybdenum Carbide Electrocatalysts for Hydrogen Production. Energy Environ. Sci. 2013, 6, 943−951. (5) Zoltowski, P. Hydrogen Evolution Reaction on Smooth Tungsten Carbide Electrodes. Electrochim. Acta 1980, 25, 1547−1554. (6) Yan, Y.; Xia, B.; Xu, Z.; Wang, X. Recent Development of Molybdenum Sulfides as Advanced Electrocatalysts for Hydrogen Evolution Reaction. ACS Catal. 2014, 4, 1693−1705. (7) Li, D. J.; Maiti, U. N.; Lim, J.; Choi, D. S.; Lee, W. J.; Oh, Y.; Lee, G. Y.; Kim, S. O. Molybdenum Sulfide/N-Doped CNT Forest Hybrid Catalysts for High-Performance Hydrogen Evolution Reaction. Nano Lett. 2014, 14, 1228−1233. (8) Voiry, D.; Yamaguchi, H.; Li, J.; Silva, R.; Alves, D. C.; Fujita, T.; Chen, M.; Asefa, T.; Shenoy, V. B.; Eda, G.; Chhowalla, M. Enhanced Catalytic Activity in Strained Chemically Exfoliated WS2 Nanosheets for Hydrogen Evolution. Nat. Mater. 2013, 12, 850−855. (9) Pan, Y.; Lin, Y.; Liu, Y.; Liu, C. A Novel CoP/MoS2-CNTs Hybrid Catalyst with Pt-Like Activity for Hydrogen Evolution. Catal. Sci. Technol. 2016, 6, 1611−1615. (10) Ojha, K.; Saha, S.; Kumar, B.; Hazra, K. S.; Ganguli, A. K. Controlling the Morphology and Efficiency of Nanostructured Molybdenum Nitride Electrocatalysts for the Hydrogen Evolution Reaction. ChemCatChem 2016, 8, 1218−1225. (11) Cao, B.; Veith, G. M.; Neuefeind, J. C.; Adzic, R. R.; Khalifah, P. G. Mixed Close-Packed Cobalt Molybdenum Nitrides as NonNoble Metal Electrocatalysts for the Hydrogen Evolution Reaction. J. Am. Chem. Soc. 2013, 135, 19186−19192. (12) Liu, Q.; Xie, L.; Qu, F.; Liu, Z.; Du, G.; Asiri, A. M.; Sun, X. A Porous Ni3N Nanosheet Array as a High-Performance Non-NobleMetal Catalyst for Urea-Assisted Electrochemical Hydrogen Production. Inorg. Chem. Front. 2017, 4, 1120−1124. (13) Wu, R.; Zhang, J.; Shi, Y.; Liu, D.; Zhang, B. Metallic WO2− Carbon Mesoporous Nanowires as Highly Efficient Electrocatalysts for Hydrogen Evolution Reaction. J. Am. Chem. Soc. 2015, 137, 6983−6986. (14) Wang, H.; Lee, H. W.; Deng, Y.; Lu, Z.; Hsu, P. C.; Liu, Y.; Lin, D.; Cui, Y. Bifunctional non-noble metal oxide nanoparticle electrocatalysts through lithium-induced conversion for overall water splitting. Nat. Commun. 2015, 6, 7261. (15) Xie, L.; Ren, X.; Liu, Q.; Cui, G.; Ge, R.; Asiri, A. M.; Sun, X.; Zhang, Q.; Chen, L. Ni (OH)2-PtO2 Hybrid Nanosheet Array with Ultralow Pt Loading Toward Efficient and Durable Alkaline Hydrogen Evolution. J. Mater. Chem. A 2018, 6, 1967−1970. (16) Chen, Y.; Yu, G.; Chen, W.; Liu, Y.; Li, G. D.; Zhu, P.; Tao, Q.; Li, Q.; Liu, J.; Shen, X.; Li, H. Highly Active, Nonprecious Electrocatalyst Comprising Borophene Subunits for the Hydrogen Evolution Reaction. J. Am. Chem. Soc. 2017, 139, 12370−12373. (17) Jothi, P. R.; Zhang, Y.; Scheifers, J. P.; Park, H.; Fokwa, B. P. Molybdenum Diboride Nanoparticles as a Highly Efficient Electrocatalyst for the Hydrogen Evolution Reaction. Sustain. Energy Fuels 2017, 1, 1928−1934. (18) Shi, Y.; Zhang, B. Recent Advances in Transition Metal Phosphide Nanomaterials: Synthesis and Applications in Hydrogen Evolution Reaction. Chem. Soc. Rev. 2016, 45, 1529−1541. (19) Popczun, E. J.; McKone, J. R.; Read, C. G.; Biacchi, A. J.; Wiltrout, A. M.; Lewis, N. S.; Schaak, R. E. Nanostructured Nickel Phosphide as an Electrocatalyst for the Hydrogen Evolution Reaction. J. Am. Chem. Soc. 2013, 135, 9267−9270. (20) Wang, X.; Kolen’ko, Y. V.; Bao, X. Q.; Kovnir, K.; Liu, L. OneStep Synthesis of Self-Supported Nickel Phosphide Nanosheet Array Cathodes for Efficient Electrocatalytic Hydrogen Generation. Angew. Chem., Int. Ed. 2015, 54, 8188−8192. (21) Laursen, A. B.; Patraju, K. R.; Whitaker, M. J.; Retuerto, M.; Sarkar, T.; Yao, N.; Ramanujachary, K. V.; Greenblatt, M.; Dismukes, G. C. Nanocrystalline Ni5P4: A Hydrogen Evolution Electrocatalyst of Exceptional Efficiency in Both Alkaline and Acidic Media. Energy Environ. Sci. 2015, 8, 1027−1034. (22) Tian, J.; Liu, Q.; Asiri, A. M.; Sun, X. Self-Supported Nanoporous Cobalt Phosphide Nanowire Arrays: An Efficient 3D I


Research Article

ACS Sustainable Chemistry & Engineering (40) Cui, W.; Cheng, N.; Liu, Q.; Ge, C.; Asiri, A. M.; Sun, X. Mo2C Nanoparticles Decorated Graphitic Carbon Sheets: BiopolymerDerived Solid-State Synthesis and Application as an Efficient Electrocatalyst for Hydrogen Generation. ACS Catal. 2014, 4, 2658−2661. (41) Cui, W.; Liu, Q.; Xing, Z.; Asiri, A. M.; Alamry, K. A.; Sun, X. MoP Nanosheets Supported on Biomass-Derived Carbon Flake: OneStep Facile Preparation and Application as a Novel High-Active Electrocatalyst Toward Hydrogen Evolution Reaction. Appl. Catal., B 2015, 164, 144−150. (42) Anjum, M. A. R.; Lee, J. S. Sulfur and Nitrogen Dual-Doped Molybdenum Phosphide Nanocrystallites as an Active and Stable Hydrogen Evolution Reaction Electrocatalyst in Acidic and Alkaline Media. ACS Catal. 2017, 7, 3030−3038. (43) Kibsgaard, J.; Jaramillo, T. F. Molybdenum Phosphosulfide: An Active, Acid-Stable, Earth-Abundant Catalyst for the Hydrogen Evolution Reaction. Angew. Chem., Int. Ed. 2014, 53, 14433−14437. (44) Wang, T.; Du, K.; Liu, W.; Zhu, Z.; Shao, Y.; Li, M. Enhanced Electrocatalytic Activity of MoP Microparticles for Hydrogen Evolution by Grinding and Electrochemical Activation. J. Mater. Chem. A 2015, 3, 4368−4373. (45) Cui, W.; Ge, C.; Xing, Z.; Asiri, A. M.; Sun, X. NixSy-MoS2 Hybrid Microspheres: One-Pot Hydrothermal Synthesis and their Application as a Novel Hydrogen Evolution Reaction Electrocatalyst with Enhanced Activity. Electrochim. Acta 2014, 137, 504−510. (46) Yan, X.; Tian, L.; He, M.; Chen, X. Three-Dimensional Crystalline/Amorphous Co/Co3O4 Core/Shell Nanosheets as Efficient Electrocatalysts for the Hydrogen Evolution Reaction. Nano Lett. 2015, 15, 6015−6021. (47) Liao, L.; Wang, S.; Xiao, J.; Bian, X.; Zhang, Y.; Scanlon, M. D.; Hu, X.; Tang, Y.; Liu, B.; Girault, H. H. A Nanoporous Molybdenum Carbide Nanowire as an Electrocatalyst for Hydrogen Evolution Reaction. Energy Environ. Sci. 2014, 7, 387−392. (48) Kucernak, A. R.; Sundaram, V. N. N. Nickel Phosphide: The Effect of Phosphorus Content on Hydrogen Evolution Activity and Corrosion Resistance in Acidic Medium. J. Mater. Chem. A 2014, 2, 17435−17445. (49) Chen, W. F.; Muckerman, J. T.; Fujita, E. Recent Developments in Transition Metal Carbides and Nitrides as Hydrogen Evolution Electrocatalysts. Chem. Commun. 2013, 49, 8896−8909. (50) Morales-Guio, C. G.; Stern, L. A.; Hu, X. Nanostructured Hydrotreating Catalysts for Electrochemical Hydrogen Evolution. Chem. Soc. Rev. 2014, 43, 6555−6569. (51) Vrubel, H.; Moehl, T.; Grätzel, M.; Hu, X. Revealing and Accelerating Slow Electron Transport in Amorphous Molybdenum Sulphide Particles for Hydrogen Evolution Reaction. Chem. Commun. 2013, 49, 8985−8987. (52) Wang, J.; Xu, F.; Jin, H.; Chen, Y.; Wang, Y. Non-Noble MetalBased Carbon Composites in Hydrogen Evolution Reaction: Fundamentals to Applications. Adv. Mater. 2017, 29, 1605838− 1605875.

J


Rationally Dispersed Molybdenum Phosphide on Carbon Nanotubes

Recommend Documents