Three-dimensionally ordered macro-/mesoporous Ni as a highly efficient electrocatalyst for the hydrogen evolution reaction

Three-dimensionally (3D) ordered macro-/mesoporous (3DOM/m) Ni is fabricated by the chemical reduction deposition method using lyotropic liquid crystals (LLC) to template the mesostructure within the regular voids of a colloidal crystal (opal). The thereby achieved structural advantages of combining well-ordered bicontinuous mesopores with 3D interconnected periodic macropores, such as abundant exposed catalytically active sites, efficient mass transport, and high electrical conductivity, make this non-noble metal structure an excellent hydrogen evolution reaction (HER) electrocatalyst. The 3DOM/m Ni exhibits a low onset overpotential of 63 mV (vs. RHE) and a small Tafel slope of 52 mV per decade, as well as a long-term durability in alkaline medium. These distinct features of the 3DOM/m Ni render it a promising alternative to Pt-based HER electrocatalysts.


Introduction
Hydrogen, as a clean and appealing energy carrier, is advocated to be an ideal candidate to replace carbon-based fuels. 1 Sustainable hydrogen production by electrocatalytic water splitting is one of the important clean-energy technologies having great potential for using renewable energy sources. 2,3 The efficiency of the hydrogen evolution reaction (HER) in water electrolysis strongly depends on the electrocatalyst. To date, Pt and other precious metals are still the most effective HER electrocatalysts because they require low overpotentials and have fast kinetics for driving the HER. 4,5 However, the high cost and limited supply of noble metals make them less attractive for commercial applications. [6][7][8] Therefore, the substitution of rare and expensive noble metal electrocatalysts by earth-abundant materials would represent a signicant step for the hydrogenbased energy industry. 9,10 Among non-precious metal electrocatalysts, Ni and Ni-based metals have been shown to be promising alternatives to Ptbased catalysts for the HER in alkaline electrolytes because of their relatively high activity and stability. 11 To achieve a more active catalyst, one efficient way is to nanostructure the catalyst to signicantly increase the number of catalytically active sites. A variety of nano or nanoporous structured Ni-based catalysts have been developed to catalyze the HER, such as threedimensionally (3D) nanoporous Ni-Mo, 12 IrNiN nanoparticles, 13 nanoporous Ni lms, 14 NiSe nanobers, 15 3D CoNiCu nanonetworks, 16 Ni-Mo nanopowders, 17 and Ni 2 P nanoparticles. 18 Owing to their periodic pore structure, uniform pore size in the mesoscale regime, high surface area and high electrical conductivity, 19 ordered mesoporous metals have shown superior performances for applications in diverse elds, including catalysis, magnetics, adsorption, and energy conversion and storage. [20][21][22][23][24] Several ordered mesoporous metals such as Pt, 25,26 Au, 27 Ru, 28 Ni 29 and their alloys with other metals 30,31 have been synthesized using different approaches based on hard templates (e.g. mesoporous silica) or so templates (e.g. lyotropic liquid crystals (LLC) made of nonionic surfactants). However, the small windows of mesopores, which make it difficult for the pores to effectively transport guest molecules to the active sites located in internal mesopores, signicantly limit applications of mesoporous metals. To overcome this problem, a promising strategy is to build hierarchical porous metallic nanostructures with interconnected macropores and mesopores, in which the macropores can facilitate mass transport and offset the diffusion limitations present in purely mesoporous metals while the mesopores enhance the surface area of the metals. [32][33][34] Dual-template synthesis methods, which employ a combination of both colloidal crystal (opal) templating (hard-templating) and surfactant templating (so-templating) techniques, are commonly used to fabricate hierarchical porous materials having 3D ordered macro-and mesoporous structures. 35,36 Due to their ordered and tunable pore structures on both macroscale and mesoscale, 3D ordered macro-/mesoporous (3DOM/m) materials are of great interest for applications in catalysis, sensing, adsorption, and energy devices. [37][38][39][40] Reports have so far mainly focused on the fabrication of 3DOM/ m carbon, silica and metal oxides, 41,42 but very few involved the production of 3DOM/m metals, even though nanometallic materials are among the most promising photoelectric and catalytic materials. 43 Recently, 3DOM/m Pt materials were prepared by electrochemical or chemical reduction of Pt ions in the LLC lled inside the opals. 44,45 The LLC was formed by inltrating the ethanol/water solution of a nonionic surfactant and H 2 PtCl 6 into the void spaces of the opal, followed by preferential evaporation of ethanol to allow self-assembly of the surfactant molecules inside the opal. The LLC-lled opal was then used as a dual-template for the fabrication of 3DOM/m Pt with the LLC and opal as a so template for mesopores and a hard template for macropores, respectively.
To our knowledge, no metal with the 3DOM/m structure, except Pt, has been previously prepared and studied. Herein, we extend the dual-templating technique to synthesize 3DOM/m Ni monoliths. To accomplish this we used the poly(methyl methacrylate) (PMMA) opal as the macroporous mold and the LLC of the nonionic surfactant Brij 58 (C 16 H 33 (OCH 2 CH 2 ) 20 OH) as the meso-structural template. A high concentration aqueous precursor solution containing Brij 58 and the nickel salt (NiCl 2 $6H 2 O) used as the Ni source was rst inltrated into the pores of the PMMA opal monolith and then self-assembled to form the LLC. Subsequent chemical reduction deposition of Ni and removal of the templates (PMMA opal and Brij 58) produced hierarchically ordered porous Ni with well-dened macropores and mesopores. In addition, the HER catalytic properties of the prepared 3DOM/m Ni in aqueous alkaline medium are evaluated in detail.

Results and discussion
Structural and morphological characterization   magnication image of a (100) plane for 3DOM Ni, again indicating a highly periodic structure.
More detailed structural information of the 3DOM/m Ni was obtained by transmission electron microscopy (TEM) and highresolution TEM (HR-TEM). Fig. 2a, c, e and g show the TEM images of as-prepared 3DOM/m Ni in the (100), (111), (110) and (211) directions, respectively. The ordered face-centered cubic (fcc) macropores are uniform and the pore size is around 330 nm, which is consistent with the observation from the SEM images. The corresponding HR-TEM images (Fig. 2b, d, f and h) reveal that well-dened mesopores are located in the macropore walls, indicating the hierarchically ordered macro/mesoporous structure of the 3DOM/m Ni. The mesopore diameter is in the 3-5 nm range and the average size of the Ni particles is approximately 4 nm, which is smaller than that for the 3DOM Ni The nitrogen adsorption/desorption isotherm of as-prepared 3DOM/m Ni is shown in Fig. 3. The isotherm can be classied as type IV, indicative of the presence of a mesoporous structure according to the IUPAC nomenclature. 46 The Brunauer-Emmett-Teller (BET) surface area and pore volume of the 3DOM/m Ni are 116 m 2 g À1 and 0.31 cm 3 g À1 , respectively. The inset of Fig. 3 shows the pore size distribution of the 3DOM/m Ni calculated by the Barrett-Joyner-Halenda (BJH) method using the desorption isotherm. The average pore diameter of the mesoporous Ni is 3.9 nm. In addition, the mesopore size distribution is in the range of 3-5 nm, and such a narrow distribution implies substantial homogeneity of the mesopores for the Ni macropore walls. The BJH derived pore size distribution corresponds closely to the HR-TEM results.
Inductively coupled plasma-atomic emission spectrometry (ICP) analysis demonstrated that the 3DOM/m Ni, 3DOM Ni, and Ni NPs contain $6 wt% boron. Fig. 4a and S3 † show the wide-angle XRD patterns of these Ni samples. In each XRD pattern, a broad diffraction peak at around 2q ¼ 45 is observed, suggesting that the incorporation of the B phase in these asprepared Ni catalysts results in an amorphous-like state. 47 The selected-area electron diffraction (SAED) pattern of the 3DOM/ m Ni (Fig. S4 †) exhibits halo rings rather than distinct points, in agreement with the XRD analysis. No Ni oxide diffraction is observed from the SAED pattern, indicating that Ni is present in the metallic state in the sample, as is consistent with that reported for mesostructured Ni-based alloys. 22 The small-angle XRD (SAXRD) pattern ( Fig. 4b) of the 3DOM/m Ni exhibits a peak at around 2q ¼ 1.3 , which can be assigned to the (100) diffraction of the hexagonal mesoscale structure with a dspacing of 6.7 nm. The presence of this SAXRD peak indicates the formation of an ordered mesostructure in the 3DOM/m Ni. The distance between the mesopores, estimated from the (100) diffraction peak, is 7.7 nm ð6:7 nm Â 2= ffiffiffi 3 p Þ, which coincides with the HR-TEM observations.

Electrochemical measurements for characterizing the HER
The HER polarization curves (j-V plots) of the samples measured in 1.0 M NaOH solution at a scan rate of 2 mV s À1 and a rotating speed of 2000 rpm are shown in Fig. 5a for the different samples. As expected, the commercial Pt/C catalyst used for base-line comparison showed excellent HER activity, providing a near-zero overpotential. The 3DOM/m Ni provides a small onset overpotential of $63 mV, beyond which the cathodic current rises rapidly under more negative potentials. In sharp contrast, 3DOM Ni and Ni NPs show much larger onset overpotentials of $101 mV and $112 mV, respectively, and a bare GC electrode exhibits almost no catalytic activity for the HER. To drive a cathodic current density of À10 mA cm À2 , the 3DOM/m Ni only requires an overpotential (h) of 171 mV, which  is much smaller than those exhibited by the 3DOM Ni (270 mV) and Ni NPs (331 mV) electrodes. Also, this overpotential (171 mV) of the 3DOM/m Ni at À10 mA cm À2 is comparable to or smaller than that of other reported non-Pt HER electrocatalysts operating in alkaline aqueous solutions, including the Ni inverse opal ($240 mV), 48 bulk Mo 2 C ($200 mV) and MoB ($220 mV), 49 Ni 2 P nanoparticles ($180 mV), 50 porous Mo 2 C nanorods ($150 mV), 51 Ni 3 S 2 /carbon nanotube composite ($350 mV), 52 and Ni x Co 10Àx /C nanoakes ($370 mV), 53 suggesting that the 3DOM/m Ni is a relatively super catalyst for the HER.
The Tafel plots of these samples (Fig. 5b), obtained by plotting the overpotential versus the log(cathodic current density), t well with the Tafel equation (h ¼ b log(j) + a, where j is the current density and b is the Tafel slope). The Tafel slope is an inherent property of the catalyst, and a smaller one means a faster increase of HER rate with increasing potential. The commercial Pt/C exhibits a Tafel slope of 31 mV per decade, which is consistent with reported values. 50 The 3DOM/m Ni, 3DOM Ni, and Ni NPs show the Tafel slopes of 52, 81, and 92 mV per decade, respectively, indicating that 3DOM/m Ni has the highest HER activity among the three samples. This Tafel slope of 52 mV per decade is lower than or comparable to those observed for the other nanostructured non-precious metal HER catalysts, such as Ni 2 P nanoparticles 18  It is known that the Tafel slope is determined by the ratelimiting step of the HER. 61 According to the two-electron-reaction models, the HER in aqueous alkaline solution proceeds in the following three possible principle steps: 62,63 Discharge step -Volmer reaction Desorption -Heyrovsky reaction Recombination -Tafel reaction In the above equations, M represents a free active site on the metal surface, MH ads is the metal surface involving an adsorbed H intermediate, R is the ideal gas constant, T is the absolute temperature, a $ 0.5 is the symmetry coefficient, and F is the Faraday constant. If the discharge process (Volmer reaction) is the rate determining step, a slope of 120 mV per decade should be observed. On the other hand, a Tafel slope of 40 or 30 mV per decade would be expected when the Heyrovsky or Tafel reaction acts as the rate limiting step, respectively. The observed Tafel slope of 52 mV per decade for the 3DOM/m Ni suggests that the Volmer-Heyrovsky mechanism with the electrochemical  desorption as the rate-determining step is operative in the 3DOM/m Ni catalyzed HER. The exchange current density (j 0 ), calculated by the Butler-Volmer equation, 64 is considered a reasonable measure of the catalytic efficiency for the HER. Table 1 lists the j 0 values of the samples. The j 0 for the 3DOM/m Ni is determined as 0.11 mA cm À2 , which is within the same order of magnitude as that of the commercial Pt/C catalyst (0.73 mA cm À2 ), and much higher than those of some other reported HER catalysts summarized in Table S1. † Electrochemical impedance spectroscopy (EIS) measurements were conducted to study the electrode kinetics under HER operating conditions. Fig. 6a displays the Nyquist plots collected at an overpotential of h ¼ 100 mV for the various electrodes. The EIS data were tted to an equivalent circuit according to the model of the simple Randles cell (inset of Fig. 6a). The observed one semicircle in the Nyquist plot suggests that the charge-transfer process controls the kinetics at the electrode interface. It is well accepted that the diameter of the semicircle reects the charge-transfer resistance (R ct ) of the porous electrode. 65 The 3DOM/m Ni electrode exhibits a R ct of 33 U cm 2 , which is much lower than that of the 3DOM Ni (320 U cm 2 ) and Ni NPs (772 U cm 2 ), demonstrating the higher charge transport efficiency of the 3DOM/m Ni. The Nyquist plot for the 3DOM/m Ni at the selected overpotentials ranging from 0 to 200 mV, which is almost located in the linear portion of the Tafel plots, is presented in Fig. 6b. The diameter of the semicircle reduces with increase in overpotential, indicating faster charge transfer kinetics for the HER at higher overpotentials.
The highly active electrocatalytic performance of the 3DOM/ m Ni for the HER is likely related to the following factors: (1) the ordered mesopores with high surface area provide more catalytically active sites for the HER; 60 (2) the 3D ordered interconnected macro/mesoporous structure improves the utilization efficiency of active sites by facilitating the mass transport of uids and enhancing the access of reactant molecules to the active sites; [66][67][68] and (3) the 3D interconnected framework with ordered macropores and mesopores can promote electron transport and facilitate ion diffusion, and thus increase the conductivity, leading to improved HER kinetics. 69,70 The durability is also an essential criterion for a practical electrocatalyst. In this study, the durability tests of the 3DOM/m Ni, 3DOM Ni, and Ni NPs catalysts were carried out by cycling the electrode potential between À0.3 and +0.2 V vs. RHE at a scan rate of 100 mV s À1 for 2000 cycles. As seen in Fig. 7a, the polarization curve of the 3DOM/m Ni was little affected by cycling. The inset of Fig. 7a shows the time-dependent catalytic overpotential curve for the 3DOM/m Ni under a static current density of À10 mA cm À2 . Aer 25 hours of continuous operation, the catalytic overpotential shows a negligible increase, suggesting that the 3DOM/m Ni catalyst has superior stability in the long-term electrochemical process compared with some other recently reported HER catalysts, such as nanoscale NiO/Ni  heterostructures, 71 CoP nanoparticles, 72 and FeP nanorod arrays. 73 Furthermore, from the potential values at a current density of À10 mA cm À2 before and aer the test (Fig. 7b), the 3DOM/m Ni exhibits high durability with only a 6.9% increase of the overpotential, while 3DOM Ni shows a larger overpotential increase of 34.8% and Ni NPs produce even a greater overpotential increase of 48.5%. These results demonstrate that the 3DOM/m Ni catalyst has higher stability than the 3DOM Ni and Ni NPs under alkaline conditions. The high durability of the 3DOM/m Ni catalyst may be attributed to its unique 3D interconnected macro-and mesoporous framework structure, which has higher stability against growth and agglomeration of Ni particles, as well as the nanoparticle separation from the electrode that can occur for aggregated Ni nanoparticles. [74][75][76][77] Conclusions In summary, a 3D ordered macro-/mesoporous (3DOM/m) Ni catalyst has been successfully synthesized by employing a dual templating approach, using the PMMA opal as a macropore template and lyotropic liquid crystals (LLC) of the nonionic surfactant Brij 58 as a mesopore template. The macropore walls of the prepared 3DOM/m Ni exhibit a well-dened mesoporous structure with a narrow pore size distribution of 3-5 nm. As a result of the high surface area of the mesopores, efficient mass transport and increased accessible surface area from the macropores, and improved conductivity from the 3D interconnected network structure, the 3DOM/m Ni shows greatly enhanced electrocatalytic performance for the hydrogen reduction reaction (HER) compared to the 3D ordered macroporous (3DOM) Ni and Ni nanoparticles (NPs). The as-prepared 3DOM/m Ni electrocatalyst exhibited a low onset overpotential of 63 mV (vs. RHE), a low Tafel slope of 52 mV per decade, a high exchange current density of 0.11 mA cm À2 and a low charge transfer resistance of 33 U cm 2 at an overpotential of 100 mV, as well as a good durability with a negligible activity loss aer 2000 CV cycles in alkaline medium. The excellent catalytic activity and stability demonstrate that the 3DOM/m Ni could be a promising alternative to Pt-based HER electrocatalysts. Also, this novel 3D hierarchical ordered porous Ni may potentially be used in other applications, such as magnetics, batteries, adsorption, separation, and catalytic hydrogenation.

Preparation of PMMA colloidal crystal (opal) monoliths
Monodispersed PMMA spheres were synthesized by emulsierfree emulsion polymerization of MMA for about two hours at 70 C using AAPH as an initiator, as reported by Stein et al. previously. 78 The resulting PMMA sphere suspension was then transferred to a wide mouth glass bottle and le for about 2 months at room temperature thereby allowing the spheres to completely precipitate. Aer that, the supernatant liquid was carefully removed using a dropper without disturbing the precipitate, and then the precipitate was kept in open air at room temperature ($10 days) for complete evaporation of the remaining water to form PMMA opal monoliths. The prepared PMMA opal monoliths were heated at 120 C for 15 minutes to strengthen contact between each of the spheres. The PMMA opals in this study are composed of $330 nm diameter PMMA spheres.

Synthesis of 3D ordered macro-/mesoporous (3DOM/m) Ni catalysts
The synthetic procedure for the 3DOM/m Ni is illustrated in Scheme 1. To make a templating precursor, 5.0 g Brij 58 was added to a solution containing 3.0 g NiCl 2 $6H 2 O and 2.8 g water, and the concentration of Brij 58 was adjusted to be 55 wt% (Brij 58/(water + Brij 58)). The obtained mixture was heated to 80 C and stirred until a homogeneous green color mixture was produced. A closed container was used to avoid evaporation of the water. Then the mixture was inltrated within the void spaces of the PMMA opal template by immersing the opal in the mixture for 5 hours at 80 C. Aer cooling to room temperature, the LLC was formed and the pores of the PMMA opal were lled with LLC containing Ni 2+ . Subsequently, the LLC-lled PMMA opal was carefully removed from the bulk LLC phase, and put in a closed vial with the reducing agent DMAB ($0.8 g) in a small dish. The vial was then kept at 25 C for 3 days, allowing the LLC-lled opal to be exposed to saturated DMAB vapor under autogenous pressure. During the process, the Ni 2+ was gradually reduced by the inltration of DMAB vapor into the LLC-Ni 2+ / opal composite. Aer the deposition of Ni, the sample was treated with THF, ethanol, and deionized water to remove the PMMA opal and surfactants to produce the freestanding 3DOM/ m Ni.

Synthesis of 3D ordered macroporous (3DOM) Ni
The PMMA opal monolith was rst immersed in a Ni solution containing 3.0 g NiCl 2 $6H 2 O, 3.0 g water, and 5.0 g ethanol for 5 hours at room temperature. Then, the chemical reduction deposition of Ni was conducted in the DMAB bath following the Scheme 1 Schematic of the preparation of the 3D ordered macro-/ mesoporous (3DOM/m) Ni. procedure described above for preparing the 3DOM/m Ni. Subsequent removal of the PMMA opal template by THF produced freestanding 3DOM Ni.

Synthesis of Ni nanoparticles (NPs)
3.0 g NiCl 2 $6H 2 O was dissolved in 2.8 g water to prepare the Ni solution used for Ni nanoparticle deposition. The chemical reduction of Ni 2+ was performed by adding 0.6 g DMAB into the Ni solution and then allowing reaction at 25 C for 1 day. The resulting precipitate was centrifuged, washed with deionized water and ethanol, and then dried at 25 C in a vacuum overnight to obtain Ni NPs.

Characterization
Scanning electron microscopy (SEM) images were obtained on a Hitachi S-4700 FEG scanning electron microscope. Transmission electron microscopy (TEM) and selected-area electron diffraction (SAED) were carried out on a JEOL JEM-3010 transmission electron microscope operating at 200 kV. Powder X-ray diffraction (XRD) data were collected on a Rigaku D/max 2500 VB2+/PC diffractometer with Cu-Ka radiation (l ¼ 1.5418Å). Nitrogen adsorption and desorption isotherms were measured at 77 K using Micromeritics ASAP 2010 apparatus. The Ni and B contents were determined by inductively coupled plasmaatomic emission spectrometry (ICP-AES, IRIS Intrepid II XSP, Thermo Elemental).

Electrochemical tests
Electrochemical measurements were performed in a threeelectrode electrochemical cell on a potentiostat/galvanostat (Reference 600, Gamry Instruments) with a 5 mm diameter glassy carbon (GC) rotating disk electrode (Pine Instruments), a double junction Ag/AgCl (saturated KCl) electrode and a Pt foil as the working, reference and counter electrodes, respectively. All electrode potentials are converted to the reversible hydrogen electrode (RHE) scale according to the Nernst equation: E RHE ¼ E Ag/AgCl + 0.059pH + 0.197. The working electrode was prepared as follows. The GC electrode was polished to a mirror nish using aqueous alumina suspensions and cleaned by ultrasonication in deionized water. 5 mg of catalyst (i.e., 3DOM/m Ni, 3DOM Ni, Ni NPs, or commercial Pt/C) was ultrasonically dispersed in 1 mL of 3 : 1 (v/v) water/isopropanol mixed solvent and 40 mL of Naon solution (5 wt%) for at least 30 minutes to generate a homogeneous ink. A volume of 10 mL of the catalyst ink was then pipetted onto the GC substrate (catalyst loading: $0.25 mg cm À2 ) and completely dried.
All electrochemical experiments were conducted in an argonpurged 1.0 M NaOH solution. The linear sweep voltammetry (LSV), beginning at 0.1 V and ending at À0.8 V vs. RHE at a scan rate of 2 mV s À1 , was conducted with a rotation speed of 2000 rpm aer repeating a few times to reach the steady state. Electrochemical impedance spectroscopy (EIS) was performed at an overpotential of 100 mV vs. RHE while sweeping the frequency from 5 MHz to 10 mHz at the amplitude of the sinusoidal voltage of 5 mV, and the charge transfer resistances were estimated from the Nyquist plots. The accelerated degradation studies were carried out by cycling the electrode 2000 times and recording the polarization curve aer every 200 cyclic voltammetric (CV) sweeps; each cycle started at À0.3 V and ended at +0.2 V vs. RHE with a scan rate of 100 mV s À1 and a rotational speed of 2000 rpm. The time dependency of catalytic overpotential during electrolysis for the 3DOM/m Ni was tested at a static current density of À10 mA cm À2 aer equilibrium for 25 hours.