Uniformly Decorated Molybdenum Carbide/Nitride Nanostructures on Biomass Templates for Hydrogen Evolution Reaction Applications.

Natural fibrils derived from biomass were used as a template to synthesize uniformly decorated nanoparticles (10-12 nm) of molybdenum carbide (Mo2C) and molybdenum nitride (Mo2N) supported on carbon. The nanoparticles have been synthesized through the carburization and nitridation of molybdenum on cotton fibrils, using a high-temperature solid-state reaction. The catalyst exhibits an onset potential of 110 mV and an overpotential of 167 mV to derive a cathodic current density of 10 mA cm-2. The electrocatalyst also demonstrates excellent long-term durability of more than 2500 cycles in acidic media with a Tafel slope value of 62 mV dec-1.

Introduction

The environmental concern regarding carbon-dioxide emission, global warming, and increasing trend of energy requirements has augmented the urge to invent and use greener and renewable energy sources.[1] In this context, hydrogen could be one of the most feasible and promising energy carriers for future energy needs.[2] High energy density and sustainability of hydrogen as compared to other energy sources might provide carbon-neutral and clean energy for future needs.[3] The hydrogen production through electrochemical water splitting could be one of the most effective, economically sustainable, and environment-friendly processes. In practice, Pt is still considered as one of the most effective, sustainable, and universally accepted electrocatalysts for hydrogen generation.[4] However, the expense and availability of Pt make it limited for extensive use. Non-noble electrocatalysts having similar or better catalytic activity with respect to Pt can be one of the best alternatives for electrochemical hydrogen generation.[5,6] At present, the primary emphasis is being given in developing stable, low-cost, and efficient electrocatalysts. Transition metals with well-known accessibility and remarkable tendency toward electrochemical water splitting make it suitable for hydrogen generation.[7,8] Metal carbides have been used as a supporting material to reduce the platinum loading and the overall cost of the hydrogen evolution reaction (HER) catalyst. However, Vrubel and Hu[9] have reported HER-active molybdenum carbide for the first time in 2012. During 2012–2014, a transition took place in HER research, when people have started working on non-noble metal-based molybdenum catalysts. Later on, many reports were published on the Mo-catalyst, showing remarkably good activity toward HER.[9−12] Recently, binary phases of molybdenum-based nanostructured materials have drawn remarkable attention because of high charge-/mass-transfer ability on the electrode–electrolyte interface and to minimize interface resistance and enhance the catalytic performance of the electrocatalyst.[13−15] A report by Chen et al.[16] shows that a composite of molybdenum carbide and molybdenum nitride provides excellent synergic effects and improves the catalytic activity. A supporting template like holey carbon provides easy pathways to the charge transfer in the heterojunction Mo2C/Mo2N/HGr nanostructure with outstanding catalytic activity and stability.[17] The d-band structure of the specific transition metals on heteroatom doping gets modified, which brings these composite into the category of near-noble metals by improving the charge-transfer process at the interface.[11,18−21] Besides this, a morphological variation using templates also plays a vital role in determining the catalytic activity.[22] Herein, we report an inexpensive synthesis of a nanostructured electrocatalyst supported on carbon derived from cotton fibrils as templates (Scheme ). The use of templates not only provides carbon support for the nanocatalyst formed in the solid-state reaction but also enhances the charge transfer through the catalyst interface during the reaction.[23,24] The reaction condition was improvised in such a way that both Mo2C and Mo2N nanoparticles were formed uniformly and simultaneously on the carbon support, which shows an overpotential of 167 mV in acidic media. This study will provide an exciting insight toward a biphasic molybdenum derivative catalyst grown on the natural biomass template and also introduce a new system for the fundamental study of nanohybrids with vast possibilities on the application on a realistic energy conversion system.

Scheme 1

Illustrating the Synthesis of the Carbon-Supported Mo2C/Mo2N Nanohybrid Electrocatalyst from the Biomass Template and Its Application in Hydrogen Evolution Reaction

Results and Discussion

The MoCot catalyst consisting of uniformly decorated molybdenum carbide/nitride nanostructures on the biomass template has shown good catalytic activity. Broadly, the reaction for the formation of Mo2C/Mo2N nanoparticles requires carbon, nitrogen, and a metal source at a moderately high temperature. It was observed that Mo2C was formed at a temperature as high as 750 °C, whereas Mo2N can be synthesized at a relatively low temperature of 650 °C. Continuing the reaction at 650 °C would lead to the formation of only nitride and the same reaction at 750 °C would lead to the formation of the only carbide. The stoichiometry of the initial precursor mixture is also very crucial as it determines the final composition of the catalyst. The reaction condition is maneuvered in such a way that both the nanoparticles can be formed at different steps of the same reaction. The critical parameters are temperature, hold-time, and gases, which are purged at an appropriate time. For example, ammonia was purged only during the formation temperature (650 °C) of Mo2N. The hexamethylenetetramine (HMT) acts both as a reducing agent and a carbon/nitrogen source during the annealing process.[25] Initially, Mo oxyanions decompose into MoO3, and then in the presence of the carbon/nitrogen source, it reacts and forms Mo2C and Mo2N nanoparticles onto the fibril substrate at high temperatures. One exciting aspect of this reaction is that the fibril structure of the natural cotton remains intact, and only the growth of the nanoparticles was observed to form onto the surface of the fiber. Detailed reaction conditions are mentioned in the Supporting Information S1 section. The objective toward the formation of both the nanomaterials on the same carbon support lies in the fact that it would result in a synergistic effect between the Mo2C and Mo2N nanoparticles and thereby increase the overall catalytic activity.[15] Series of reactions were performed to produce several compositions of Mo2C and Mo2N, out of which MoCot shows the best catalytic activity.

Structural information of the catalyst was obtained from powder X-ray diffraction (PXRD) measurements as provided in Figure a. The PXRD pattern of the catalyst after annealing exhibit a series of characteristic peaks located at approximately 34.3°, 37.9°, 39.4°, 52.1°, 61.5°, 69.5°, 72.4, 74.5°, and 75.4° which attributes to the diffractions of (100), (002), (101), (102), (110), (103), (200), (112), and (201) planes (PDF 35-0787) of Mo2C hexagonal system, respectively. The peaks at 37.3°, 43.4°, 63.1° and 75.7° are characteristics of γ-Mo2N (PDF 25-1366) and correspond to (111), (200), (220) and (311) planes, respectively. The XRD measurement confirmed that the Mo2C–Mo2N hybrid material (denoted as MoCot) was composed of hexagonal β-Mo2C and cubic γ-Mo2N. Broad hump in the PXRD pattern at around 25°–30° value can be related to the mixed amorphous carbon phase which appears only in a slow scan. No other impurities or characteristic peaks of oxides and so forth are observed. The XRD diffraction pattern indicates the co-existence of Mo2C and Mo2N in all the catalysts with variation in the composition of both. The catalyst was found to have 59.12% of Mo2C and 40.88% Mo2N nanoparticles, respectively, as calculated from Rietveld refinement (Figure b). Several other compositions were prepared, out of which two compositions (MoCot 1 & MoCot 2, Supporting Information S2, Figures S1 and S2) are mentioned in the Supporting Information, which has a closer overpotential with respect to MoCot.

Figure 1

Structural characterization (a) PXRD typically shows formation of both Mo2C–Mo2N phases and (b) compositional measurements: Rietveld analysis of the catalyst shows a composition 59.12 and 40.88% of Mo2C and Mo2N, respectively.

A close look toward scanning electron micrograph (Figure ) along with elemental mapping gives detailed morphological information about the nanostructures and confirms the existence of two different phases. The image clearly shows that the catalyst having well-decorated Mo2C and Mo2N nanoparticles developed on the fibril microstructure template, having different lengths. The carbon support inhibits the agglomeration of the metal carbide/nitride nanoparticles and thereby enhances the catalytic activity.

Figure 2

Morphology characterization: scanning electron microscopy (a) of the as-obtained MoCot catalyst and corresponding elemental mapping of the nanostructure showing the (b) carbon, (c) molybdenum, and (d) nitrogen elements that illustrate the uniform distribution of these on the fibril template. (e) SEM image shows the collective distribution of different elements (Mo, C, and N) on the fibril structure.

Transmission electron microscopy (TEM) images (Figure a–c) clearly show the nanoparticles of the order of 10–12 nm, at two different scales. The high-resolution (HR) TEM image (Figure c inset) shows the two adjoining phases of nanoparticles with an interface. The visible crystal fringe width of 0.26 nm corresponds to the (100) plane and 0.24 nm to (111) planes of Mo2C and Mo2N, respectively.[19,26]

Figure 3

Microscopic characterization of the MoCot catalyst: a transmission electron microscopic image of MoCot at (a) medium (b) low and inset-(c) high magnification: showing the fringe width 0.26 and 0.24 nm consistent with the (100) and (111) planes of Mo2C and Mo2N nanocrystals.

The surface area, pore size and distribution, and porosity of the nanohybrid material were investigated by N2 adsorption/desorption measurements (Figure S3). The figure typically demonstrates the type-IV curve shape of the Brunauer–Emmett–Teller (BET) isotherm. As evident from the BET measurements, the nanohybrid material is having a surface area of 63 m2 g–1, in which mesopores are predominantly distributed over a range of 30 nm (Figure S3, inset). This distribution indicates the material to be mesoporous in nature, which favors mass transport and adsorption. Formation of uniform nanostructures is essentially acquired from the inherent morphology of the biomass templates. Hence, the template used in the reaction is of extreme importance.

X-ray photoelectron spectroscopy (XPS) further affirmed the chemical state of the catalyst. The wide range energy survey scan (Figure a) indicates the presence of C, Mo, and N elements in the hybrid structure. The survey XPS spectrum shows five characteristic signals which are located at 233.6, 285.8, 398.1, 416.5, and 532.2 eV, consistent with Mo 3d, C 1s, N 1s (Mo 3p3/2), Mo 3p1/2, and O 1s, respectively. The HR spectrum of Mo 3d (Figure b) shows splitting in the core level because of spin–orbit coupling into 3d5/2 and 3d3/2 peaks. These peaks with binding energy values at 228.4 and 231.5 eV are described as Mo 3d5/2 and Mo 3d3/2 of Mo2+ spectral lines, respectively, and they confirm the presence of Mo2C.[27,28] Other peaks at 235.1 and 235.9 eV are indicating the presence of Mo4O11 and MoO3[29−31] while 232.4 and 229.2 eV may be assigned to MoO2.[32] Surface oxidation of Mo2C (during the XPS measurements) is the reason for the indication of the presence of high concentration of molybdenum oxides. The Mo 3d5/2 peak of Mo2N appears at around 228.7 eV which is overshadowed because of the adjacent peak.[33] In C 1s, energy spectra (Figure c) were deconvoluted into two peaks centered at 283.8 and 284.7 eV, which shows the binding energy of Mo–C and C=C species, respectively.[27,34] Deconvolution of the N 1s and Mo 3p3/2 peaks depict two different Mo 3p3/2 peaks having binding energies 398.8 and 395.3 eV, respectively (Figure d). The binding energy of 395.3 eV may be attributed to Mo 3p3/2 of Mo2N[35] and the binding energy of 398.8 eV is due to Mo 3p3/2 of MoO3. The overlap of N 1s peaks of Mo2N and pyridinic-N with Mo 3p peaks might significantly enhance the electrochemical activation of graphitic carbon.[35] Because of the use of extrinsic nitrogen during the pyrolysis of biomass or performing the reaction under nitrogen-rich conditions, species may be generated with an abundance of pyridinic-N, pyrrolic-N, quaternary-N, and pyridone-N oxides.[36,37] Among the nitrogen species, the formation of pyridinic N dominates the formation of other species which is favorable to HER performance.

Figure 4

XPS of the MoCot catalyst: wide-scan survey spectra (a) and HR spectra of Mo 3d (b), C 1s (c), and N 1s (d) electron: experimental data (dotted curve) and fitting results (solid curve). The peaks are assigned by oxidation states of different elements with their corresponding binding energy.

The HER activity of all the catalysts was primarily measured by linear sweep voltammetry in 0.5 M H2SO4 saturated with argon at 25° centigrade using three-electrode cells. In order to perform a systematic study, Mo2C and Mo2N have also been synthesized with the same method as was used for MoCot, but varying the parameters and gases used in the reaction. The commercially available Pt/C (20 wt %) catalysts were also tested under similar conditions. As depicted in Figure , MoCot has an onset potential nearly 110 mV. The polarization curves at rate 5 mV/s were calculated considering the iR-compensation. From the polarization curve (Figure a), it can be concluded that MoCot is HER-active and has an overpotential of 167 mV to derive the current density of 10 mA/cm2. The comparison of the MoCot catalyst with that of Mo2C (218 mV) and Mo2N (465 mV) catalysts undoubtedly confirms that the catalytic activity MoCot is way too better than the other two. The Tafel slope value has a high significance not only in determining the possible reaction pathway but also in predicting the adsorption behavior of the catalyst for a quantitative determination of HER kinetics.[38] A description of the procedure for electrode fabrication is described in the Supporting Information S4.1. The LSV plot of the MoCot 1 & MoCot 2 composites is shown in Supporting Information S4.2 (Figure S5).

Figure 5

Electrochemical measurements of specific electrocatalysts for hydrogen evolution in 0.5 M H2SO4 acidic medium. (a) Polarization curves (iR-corrected) of MoCot compared with the other electrode (b) the corresponding Tafel plots derived from the curve (c) EIS Nyquist plot (with corresponding equivalent circuit) of the electrode@50 mV, @150 mV, and @250 mV (inset shows zoomed Nyquist plot of the electrode @50 mV) (d) figure shows the LSV stability curve at 1st, 2500th, and 3000th cycle (inset shows the potential testing at constant current density 10, 20, and 30 mA cm–2).

Theoretically, for the Tafel slope, the linear portions must fit with equation η = b log(j) + a, where a is the Tafel constant, b is the Tafel slope, j is the current density, and η is the overpotential.[39] Here, we have calculated Tafel plots for MoCot and of pure Mo2C, Mo2N, and PtC (20%) as shown in Figure b. The Tafel slope values of 62, 85, 196 mV/dec were measured for MoCot, Mo2C, and Mo2N, respectively. The same for commercially available Pt/C catalyst was found to be ∼35 mV/dec. However, with respect to Mo2C and Mo2N, MoCot have a lower value suggesting that the reaction proceeds through a mixed Volmer–Heyrovsky process and thus the recombination step is the rate-determining step.

Transport properties and interfacial interactions of the concerned catalyst were investigated by electrochemical impedance spectroscopy (EIS) with an overpotential of 50, 150, and 250 mV in 0.5 M H2SO4 aqueous solution (Figure c). The Nyquist plot reveals the effective resistance shown by the electrocatalyst resulting from the combined effects of Ohmic resistance and reactance. Simulating and simultaneous fitting of the experimental curve with the equivalent circuit provides information regarding the contribution of each component. Here, the fitting of the plot with the equivalent circuit shows two time-constant components Rct-CPEdL and Rp-CPEL along with an uncompensated solution resistance Rs (in series) describing the response of HER on the catalyst-modified electrode. CPE is the constant phase angle element, which represents the double-layer capacitance of the solid electrode. Rct-CPEdL is related to the charge-transfer process, and Rp-CPEL can be correlated with surface porosity.[26,40] The charge transfer resistance varies inversely with respect to the potential applied. A charge transfer resistance (Rct) of 20 Ω is obtained in the low-frequency region when a potential of 50 mV is applied. The smaller value of Rct implies a larger number of active sites present in the catalyst, and hence, the charge transfer between the electrode and electrolyte would be high, which favors the HER process. The durability measurement (Figure d) indicates a high stability performance of the catalyst and shows no significant change in the overpotential of the catalyst even after 2500 LSV cycles. The stability trend (inset Figure d) measured at higher current densities, that is, 20 and 30 mA/cm2, are similar to that of 10 mA/cm2, suggesting that the electrocatalytic activity profile does not change at high current density. The stability of the MoCot composite was further verified by X-ray analysis of the catalyst after 2500 cycles (Supporting Information S5, Figure S8). The composition was found to be consistent with the one recorded before the reaction. Thus, the catalyst demonstrates good stability toward the HER process.

Conclusions

In this work, nanoparticles of molybdenum carbide (Mo2C) and molybdenum nitride (Mo2N) with size 10–12 nm were grown uniformly on biomass (used as a template obtained from cotton) by manipulating the reaction conditions. Formation of uniform nanostructures was acquired from the inherent structure of the biomass templates. Hence, the nature of the template used in the reaction is of tremendous importance. The catalytic activity increased abruptly when both the nanoparticles were grown on the same carbon platform through the carburization and nitridation of the molybdenum precursor by annealing at an appropriate temperature. The activation caused possibly due to the synergistic effect of C–N present in the metal carbide/nitride nanoparticles through the carbon layers along the surface. The catalyst demonstrates long-term stability and high conductivity in acidic media and exhibits an onset potential of 110 mV and an overpotential of 167 mV to derive a cathodic current density 10 mA cm–2. The electrocatalyst also exhibits high stability of withstanding more than 2500 cycles in acidic media with a Tafel slope value of 62 mV dec–1. This study will open up an area of exploration for developing nanostructured binary composites on biomass as templates for the HER application. The process is inexpensive and produces highly stable electrocatalysts. The methodology can be applied to numerous kinds of biomass having different morphologies as the later plays a crucial role in catalytic activity.

Experimental Section

Synthesis of the Catalyst

Chemicals used in the following reactions were purchased and used without further purification. Ammonium molybdate tetrahydrate ((NH4)6·Mo7O24·4H2O), Sigma-Aldrich 99.0%; N,N-dimethylformamide anhydrous, 99.8%, Sigma-Aldrich; Nafion perfluorinated resin solution, 5 wt % in lower aliphatic alcohols; HMT (C6H12N4) 95%, TCI; ammonium hydroxide NH3·H2O (25%); sterilized cotton; deionized water was used in all preparations.

The synthesis of the MoCot catalyst was carried out in two steps. First, the preparation of the precursor and second, the calcination of the cotton-soaked metal precursor by temperatures controlled annealing in the ammonia and argon atmosphere at different time periods.

Synthesis of the Catalyst Precursor

For the precursor, 0.01 mol of HMT and 0.001 mol of ammonium molybdate tetrahydrate [(NH4)6Mo7O24·4H2O] were mixed into a solution of 2.5 mL of ammonia water (25%). The mixture was stirred until it became transparent. Then, 1 g of finely chopped cotton was dipped and soaked into the solution mentioned above. The soaked cotton, along with the remaining solution, was transferred into a hydrothermal bomb and heated to 100 °C into an oven for 12 h. The product obtained was then dried in a vacuum oven at 100 °C for 6 h to yield a greenish-white mass.

Temperature-Programmed Calcination of the Precursor (Second Step)

This process is specifically designed for the synthesis of MoCot. The furnace was purged with argon initially for 20 min to make an inert environment for the sample before the reaction. The cotton-soaked catalyst precursor (greenish-white mass) was heated to 650 °C with a ramp rate of 7°/min in the tube furnace under argon with a flow of 30 mL/min. The reaction was held for 1 h under ammonia at a flow rate of 20 mL/min, and after that, the gas was switched to argon and the temperature increased to 750 °C and held for another 2.5 h. The MoCot catalyst was obtained by gradually cooling the product under argon. The simultaneous formation of the Mo2C and Mo2N nanoparticles is highly sensitive not only toward the initial stoichiometry of the catalyst precursor but also toward the temperature manipulation with respect to the time and nature of the gases used.

All these parameters were judiciously optimized to obtain various compositions of Mo2C and Mo2N on the carbon template (Mo2C, Mo2N, MoCot 1, and MoCot 2, discussed in Supporting Information S1).

Physical Characterization

XRD of the solid was done using a Bruker Eco D8 ADVANCE X Powder X-ray diffractometer using the Ni filter employing Cu Kα radiation (λ = 1.54056 Å, 40 kV and 25 mA) in the 2θ range of 5°–80° with an increment of 0.00190/step. The surface area and pore size analysis of the samples were examined by nitrogen physisorption at 77 K in Autosorb IQ Quantachrome instrument using the BET and the Barrett–Joyner–Halenda equation. Surface morphology was studied using a scanning electron microscope from JEOL (JSM IT-300) equipped with energy-dispersive X-ray diffractometer (Bruker), and TEM images were acquired using a JEOL-2100 operated at 200 kV. XPS was performed on ESCALab: 220-IXL with Mg Kα nonmonochromated X-ray beam having photon energy 1253.6 eV.

Electrochemical Measurements

CHI 760E electrochemical workstation setup was used at room temperature for all the electrochemical measurements. Herein, we have used a conventional three-electrode cell system containing a carbon-based electrode, Ag/AgCl electrode, and glassy carbon electrode as a counter, reference, and working electrode, respectively. The catalyst ink was prepared by dispersing 5 mg of electrocatalyst powder into 500 μL of dimethylformamide containing 5 μL of 5 wt % Nafion and keeping it for 120 min under ultrasonication to form a homogeneous suspension. A dispersed ink of 5 μL was coated on to the surface of the working electrode employing drop-casting and dried under vacuum. The overpotential and other electrochemical measurements were calculated with respect to the Ag/AgCl electrode and further referenced with the reversible hydrogen electrode (RHE), according to the equation E (RHE) = E (Ag/AgCl) + 0.059 pH + 0.197 V. For all the measurements, polarization curves were collected under an argon atmosphere at a scan rate of 5 mV s–1 containing 0.5 M H2SO4 solution unless otherwise mentioned. The overpotential was further corrected to eliminate the Ohmic drop according to the equation η corrected = η – iRs whereas Rs is solution resistance. EIS measurements were performed at particular values of the overpotential, that is, η = 50, 150, and 250 mV under a frequency range of 100 000–1 Hz.