Abstract

Transition metal phosphating Pd@NiP nanoparticles supported by reduced graphene oxide were successfully synthesized by a one-step in situ synthesis method. The Pd@NiP/rGO nanoparticles were characterized by powder X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), transmission electron microscopy (TEM), confocal Raman spectroscopy (Raman), Fourier transform infrared spectroscopy (FT-IR), and BET. Compared with Pd/rGO and Pd@Ni/rGO, Pd1@Ni12.5P40/rGO exhibited stronger catalytic activity, with TOF value reaching 133.33 min-1 and activation energy (Ea) of 29.31 kJ mol-1. This excellent catalytic performance may be related to the good dispersion of nanoparticles and the synergistic electron interactions between transition metal phosphating compounds.

1. Introduction

Hydrogen energy is one kind of new energy promoted by various countries at present. Compared with traditional energy, hydrogen energy has the advantages of very high energy value, wide application range, and clean and pollution-free [13]. Hence, interest in developing suitable hydrogen storage materials has grown rapidly over the past few decades. However, the efficiency and safety of hydrogen storage still exist as the biggest challenges for the future development of the hydrogen economy. So far, some good hydrogen storage materials such as metal hydrides [4, 5], adsorption materials [6, 7], and chemical hydrides have been extensively studied [810]. Chemical hydrides containing boron and nitrogen have come into notice for the high hydrogen content and good hydrogen release kinetics. Ammonia borane (NH3BH3, AB) has a very high hydrogen content (19.6 wt%) and high aqueous stability, which is considered as a promising portable and attractive hydrogen storage material [11, 12]. As can be seen from equation (1), 3 mol of H2 can be released for 1 mol of AB hydrolysis with an appropriate catalyst [13]. It is well known that noble metals like Ru, Pd, Pt, and Rh have the highest catalytic performance towards the hydrolysis of AB; however, their rarity and high price have seriously hindered their popularity and application [1417]. Therefore, the development of efficient, economical, and stable AB hydrolysis catalysts is essential but still a great challenge.

As we all know that the catalytic activity of metal catalysts largely relies on the dispersity of active metals. In order to obtain high-performance catalysts, surfactants and high surface area carrier materials have been used to prepare various catalysts [18]. We added graphene oxide as a carrier in the synthesis process, so that the catalyst particles may have better dispersion and more active sites. Graphene is a single-atom thick carbon material with excellent thermochemical stability, high specific surface area, and electrical conductivity, making it an ideal carrier for the fixation and growth of nanoparticles [19]. In addition, transition metal phosphides (TMPs) have unique charged properties (transition metal positively charged and phosphorus negatively charged) and have attracted particular attention in hydrogen evolution reactions (HER) and hydrodesulfurization (HDS), possibly due to its catalytic mechanism that is similar with hydrogenase resulting in high catalytic activity [2024]. In recent years, many scholars have also studied transition metal phosphating as a kind of efficient catalyst for hydrogen storage materials. For example, Du et al. reported the synthesis of amorphous NiP/rGO hybrids of transition metal phosphide using one-pot coreduction, and the catalytic property of NiP/rGO NPs towards the hydrolysis of AB is also studied. This excellent performance may be related to the synergistic electron interaction between nickel and phosphorus [25]. Peng accounted the synthesis of nanostructure Ni2P using Ni(OH)2 powder and NaH2PO2 in argon at 543 K. The mechanism study showed that the binding of catalyst surface with the substrate molecules is the key factor to improve catalytic activity [26].

In this study, using MeAB as reducing agent and PdCl2, NiCl2, and NaH2PO2 as precursors, the reduced graphene oxide-supported Pd@NiP core-shell nanoparticles were synthesized by one-step method and used as an efficient and stable ammonia borane catalyst for hydrogen production. Firstly, Pd@NiP/rGO core-shell nanoparticles were synthesized and characterized. Thanks to the effective and efficient doping of P and the synergistic electron effect between Ni and P, the synthesized Pd@NiP/rGO nanoparticles exhibit stronger catalytic activity than the phosphorus-free nanoparticles for the hydrolysis of AB. The catalyst has excellent catalytic activity, the TOF value is 133.33 min-1, and the activation energy (Ea) value is 29.31 kJ·mol-1. This simple and fast synthesis way can also be applied to the synthesis of other graphene-supported transition metal phosphating catalysts.

2. Experimental

2.1. GO Preparation

GO was prepared by following the modified Hummers method [27, 28], adding 360 mL H2SO4 and 40 mL H3PO4 (9 : 1) into the round-bottom flask in an ice bath, and add 18 g KMnO4 powder in batches with slow stirring, then heat to 50°C, and maintain the stirring reaction for 12 h. After the reaction, cool to room temperature (RT), pour the solution above into 400 mL deionized water comprising 3 mL 30% H2O2 in an ice-bath, and then add H2O2 under stirring until the solution becomes a constant yellow. The products were obtained by centrifuging and washing with 30% diluted HCl, deionized water, and ethanol for three times, respectively. GO powder was obtained by vacuum drying at RT for 12 h.

2.2. Preparation of MeAB

The synthesis method of MeAB is the same as in the reference [29], at room temperature, and 3.783 g NaBH4 and 6.752 g CH3NH3Cl were placed into a 500 mL two-port round-bottom flask. Nitrogen was inserted into one mouth of the round-bottom flask, and 200 mL tetrahydrofuran was added under agitation. After 12 hours of reaction at RT, the initial product was obtained by extracting and evaporating the solvent. Then, the initial product is added to the ether under an ice bath and stirred for 2 hours. Finally, highly purified MeAB was also obtained by filtration and evaporation.

2.3. In Situ Synthesis of Pd@NiP/rGO Catalysts

For Pd1@Ni12.5P40/rGO NPs, 10 mg of GO powder was placed in a two-necked flask, and 5 mL of deionized water was added to disperse the mixture by ultrasound. 0.5 mL PdCl2 solution (0.01 mol/L), 0.625 mL NiCl2 solution (0.1 mol/L), and 2 mL NaH2PO2 solution (0.1 mol/L) were added to the flask. MeAB solution was added by a constant pressure drip funnel on one neck, and the other neck is connected to a tube of gas quantity tube, which is used to detect gas generation and measure the volume of produced gas. Use a syringe to add 1 mL MeAB solution (1 mol/L) to a constant pressure drop funnel and cover with a glass stopper. The reaction begins when MeAB solution drops into the solution in the flask. The generated hydrogen is put into the gas burette by the catheter. When no bubbles are generated in the gas tube, the reaction is complete, and the precursor has been reduced completely. The catalyst was obtained by filtrating, washing, and drying.

In order to optimize the catalyst for higher catalytic activity to the hydrolysis of AB, Pd@NiP NPs with different composition are synthesized by a similar method, while keeping the molar ratio of Pd to AB () at 0.005. The mole ratios of Ni/Pd are 0, 2.5, 5, 7.5, 10, 12.5, and 15, and the mole ratios of P/Pd are 0, 10, 20, 30, 40, and 50.

2.4. Materials Characterization

X-ray diffraction (XRD) was used to characterize the crystal structure of the material. The XRD pattern was obtained by Empyrean X-ray diffractometer with a Cu Kα as the radiation source (λ = 0.154178 nm), the scan velocity was 8°/min, and the angle of 2θ ranged from 5 to 80 degrees. The microstructure was obtained by JEM-2100F and JEM-ARM200F field-emission HETEM with EDAX Elite T energy spectrometer at 200 kV. Before testing, a small amount of powder samples was dispersed into ethanol by ultrasonic wave to prepare TEM samples, and then, a drop of suspended liquid was dropped onto a 300-mesh copper net for measurement. FT-IR spectra were obtained at RT using a Nicolet 670 spectrometer (Thermo Fisher Scientific, USA) in the wave number range of 4000 to 400 cm-1. Raman spectra in the range of 800 to 2000 cm-1 were obtained using a Renishaw confocal micro-Raman spectrometer excited by a laser beam at 514.5 nm. XPS spectra were surveyed by ESCALAB 250Xi X-ray photoelectron spectrometer. The element content was verified by ICP-OES (Prodigy 7). BET was scanned with the ASAP 2020 to test specific surface area and pore volume.

2.5. Catalytic Activity Test

The synthesized Pd@NiP/rGO nanoparticle catalyst was placed in a two-necked flask, one port of the flask was used to add ammonia borane aqueous solution (1 mmol ammonia borane dissolved in 1 mL water), another port is connected to a eudiometer to monitor gas production, and the reaction stops when no H2 is produced. Keep the amount of AB at 1 mmol and calculate the catalytic rate of this catalyst by testing the catalysts with different concentrations () at RT and determine whether the reaction is a first-order reaction. According to the requirements of the Arrhenius equation, the ratio of the molar amount of the catalyst to the molar amount of ammonia borane () was kept constant; the catalytic reaction was determined by the water bath of the magnetic stirrer temperatures during the process (298 K, 303 K, 308 K, and 313 K), and we can calculate the Ea value of the reaction. In order to measure the cyclic durability of the catalyst for catalyzing the hydrolysis of AB, the same amount of AB was added to continue the reaction after the last catalytic reaction under the same conditions, and the reaction was repeated 5 times. To calculate the conversion frequency (TOF) value of the hydrolysis of AB, the following formula is used: where is the volume of hydrogen, is the atmospheric pressure, is the temperature of the reaction, is the molar gas constant, is the molar mass of Pd, and is the reaction time.

3. Results and Discussion

3.1. Synthesis and Characterization

During the synthesis of reduced graphene oxide-supported Pd@NiP core-shell nanoparticles, the MeAB solution was added to the aqueous solution which contained PdCl2, NiCl2, NaH2PO2, and GO. During the synthesis, due to the low reduction potentials of Ni2+ and H2PO2- ( vs. SHE and vs. SHE), MeAB with weak reducibility cannot reduce them, but it can reduce the Pd2+ which has high reduction potential ( vs. SHE) to form Pd core. The Pd-H bond reduces Ni2+ and H2PO2- [30], which makes it continuously grow on the surface of the Pd core to deposit a NiP shell, and finally, Pd@NiP core-shell nanoparticles are obtained.

To verify the morphological characteristics of the catalyst, the microstructure of Pd1@Ni12.5P40/rGO nanoparticles was characterized by TEM. In Figures 1(a)1(c), the catalyst nanoparticles synthesized in situ by one-step method are uniformly dispersed on the graphene oxide carrier, and the light-colored folds on the bottom layer are graphene oxide carriers, which can be clearly seen; seeing its edge profile shows that graphene oxide can effectively prevent the aggregation of particles. In Figure 1(d), at a larger magnification, the obvious contrast between the core and the shell can be clearly observed; the black core is Pd, and the light gray shell is NiP, indicating that Pd is first reduced by MeAB, and then, as the seed of the core, it helps to promote the reduction of Ni2+ and H2PO2- to form a shell. The SAED result (Figure 1(e)) shows that Pd1@Ni12.5P40/rGO NPs are polycrystalline, and the lattice spacing was calculated to be 0.224 nm, which is consistent with the Pd (111) crystal face. In order to determine the elemental composition and distribution of the nanoparticles, the EDS mapping images of Pd1@Ni12.5P40/rGO are determined in Figures 1(f)1(j). The presence and uniform distribution of Pd, Ni, and P were confirmed by the EDS mapping images. In addition, the distribution area of Ni and P is bigger than that of Pd, which can further indicate the existence of a core-shell structure.

Figure 2 shows the XRD images of Pd1@Ni12.5P40/rGO and GO. It can be seen from the figure that there is a distinct diffraction peak at 43°. According to the comparison of the PDF# 88-2335 card, this peak is Pd (111) plane, which is consistent with the SAED result. No diffraction peaks of Ni appear in the images of Pd1@Ni12.5P40/rGO, which may be due to the amorphous nature of Ni in the composites. In addition, compared with the spectrum of GO, the peak of GO (001) at 8.4° disappeared, and the peak of rGO appeared at 23.5° in the spectrum of Pd1@Ni12.5P40/rGO, which indicated that GO has been completely reduced and rGO is formed. As can be seen from Figure 3, two distinct characteristic peaks can be observed in the Raman spectra of GO and Pd1@Ni12.5P40/rGO at both 1342 cm-1 and 1589 cm-1, which denote the D peak and G peak of the carbon material, respectively. The D-band is caused by GO disorder caused by defects such as vacancies, grain boundaries, and amorphous carbon, and the G-band is caused by E2g phonons of C sp2 in a two-dimensional hexagonal lattice [31, 32]. The intensity ratio of D peak and G peak () is used to describe the degree of defect and graphitization of carbon materials [33]. The of Pd1@Ni12.5P40/rGO and GO are 1.01 and 1.29, which means that GO was successfully reduced to rGO during the reaction. The Fourier transform infrared spectra of Pd1@Ni12.5P40/rGO and GO are shown in Figure 4. The strong wide band of GO at 3405 cm-1 is attributed to the moisture absorption of graphene oxide. The C=O stretching band of carbonyl or carboxyl group at 1736 cm-1 is weak. The peak value at 1619 cm-1 is related to the skeleton vibration of unoxidized graphite domains and the vibration of adsorbed water. The peaks at 1228 cm-1 and 1058 cm-1 are related to the C-OH vibration and the deformation of the C-O band in carboxylic acids, respectively [34]. In the FTIR result of Pd1@Ni12.5P40/rGO, the peaks of C=O, C-OH, and C-O are disappeared, which further verified the reducing of GO during the reaction [35].

The composition and chemical valence of transition metal phosphide nanoparticles supported by GO were detected by XPS analysis. Figure 5(a) shows the full spectrum of Pd1@Ni12.5P40/rGO nanoparticles. The signals of Pd, Ni, P, C, O, and other elements are clearly visible, indicating that Pd, Ni, and P elements have been successfully introduced into the composite. In Figure 5(b), the two obvious peaks at 335.1 eV and 340.4 eV represented Pd 3d5/2 and Pd 3d3/2 with zero valence state Pd, respectively. Figure 5(c) represents the peaks of Ni 2p, and the two peaks at 856.2 eV and 874.1 eV represent the oxidized Ni 2p3/2 and Ni 2P1/2, respectively [36]. These results indicate that Pd as the core is stable, so it is not oxidized, whereas the surface nickel is easily oxidized [37]. In Figure 5(c), there is a peak at 852.4 eV belonging to Niδ+, which shows a positive displacement of 0.2 eV compared with the peak of metal Ni at 852.2 eV, with a small positive charge (Niδ+, ) [38]. Figure 5(d) shows the XPS spectrum of P 2p. The peak at 129.7 eV is very close to the zero-valence P, indicating that P atoms in nanoparticles have marginal negative charge (Pδ-, ) [39]. The other peak at 133.5 eV belongs to the oxidation state P formed on the surface of the catalyst due to air exposure [40]. These results indicate that there is a strong electron transfer between Ni and P, which is the key to improve the catalytic activity of the catalyst [41]. According to existing literature reports, hydrogenase has high HER activity, and the active sites are the pendant bases near the metal center [42]. The Ni(δ+) and the pendant base P(δ-) of Pd1@Ni12.5P40/rGO nanoparticles have similar electronic structures to hydrogenase, so the high catalytic activity of Pd1@Ni12.5P40/rGO may be related to the hydrogenase-like catalytic mechanism [43]. In Figure 5(e), XPS result for C1s of rGO indicates that four peaks at 284.8, 286.1, 287.2, and 289 eV are observed and identified as sp2C, –C–O, –C=O, and –COO groups, respectively [44]. Figure 6 shows the nitrogen adsorption curve of Pd1@Ni12.5P40/rGO sample calculated using the BET model, indicating that the Pd1@Ni12.5P40 NPs occupy the orifice of the rGO. The surface area of Pd1@Ni12.5P40/rGO was calculated to be 73.74 m2g-1 using the Brunauer-Emmett-Teller (BET) model.

3.2. Catalytic Performance for the Hydrolysis of AB by Pd@NiP/rGO NPs

Pd@Ni/rGO catalysts with different ratios of palladium and nickel were synthesized using MeAB as a reducing agent, is fixed at 0.005, and the catalytic performance of the catalysts for AB hydrolysis was investigated. As shown in Figure 7, in the absence of Pd, MeAB cannot reduce Ni2+, resulting in the inability to catalyze the hydrolysis and dehydrogenation of AB. When pure palladium particles are supported on graphene oxide, it takes about 42 minutes to fully release 3 equivalents of hydrogen by catalyzing the hydrolysis of AB. By changing the amount of Ni, Pd@Ni/rGO exhibited different catalytic activities. As the ratio of Ni:Pd increased from 2.5 to 12.5, the catalytic activity of the catalyst increased rapidly, and the catalytic activity decreased when the ratio continued to rise, which indicated that the best ratio of Pd to Ni in the Pd@Ni/rGO catalyst was 1 : 12.5.

To obtain catalysts with different phosphorus content, we control the phosphorus content by adding different doses of NaH2PO2 solution. Figure 8 shows the effect of different phosphorus contents on catalytic hydrolysis performance of the Pd@NiP/rGO catalysts. From the figure, it can be observed that the catalyst activity is greatly improved after adding phosphorus. In addition, the catalytic activity of the catalysts enhances with increasing the content of phosphorus and increasing the ratio of P:Pd from 0 to 40. The best ratio of Pd:P is 1 : 40, namely, Pd1@Ni12.5P40/rGO. By calculation, the TOF value of the catalyst Pd1@Ni12.5P40/rGO is 133.33 min-1, which is higher than that of many Pd-based nanoparticles without P (as shown in Table 1). The mass loading of Pd@NiP nanoparticles and on reduced graphene oxide is 19.12 wt% which is calculated from the results of ICP-OES.

Figure 9 shows the curves of AB hydrolysis catalyzed by Pd@NiP/rGO nanoparticles gained by using different reducing agents (MeAB, AB, and NaBH4). The activity of Pd@NiP/rGO NPs reduced by MeAB is significantly higher than the catalysts obtained by the reduction of AB and NaBH4, and their catalytic activity is inversely proportional to the reducibility, further indicating that the reducibility of the reducing agent plays an important role in the synthesis of the catalyst, thus affecting the catalytic activity.

Figure 10 shows the hydrolysis and dehydrogenation curves of AB (1 mmol) catalyzed by Pd1@Ni12.5P40/rGO nanoparticle catalysts with different concentrations at room temperature. The ratios of are 0.005, 0.006, 0.007, and 0.008, respectively. The dehydrogenation rate is calculated by taking the linear part of the first segment with the most stable slope of each dehydrogenation curve. Each dehydrogenation rate corresponds to a point in the logarithmic graph, and the four points in the logarithmic graph are fitted to a straight line, which can be calculated when a slope of 1.0323 was obtained, which indicated that the Pd1@Ni12.5P40/rGO nanoparticles catalyzed a first-order reaction of the AB hydrolysis process relative to the catalyst concentration. In order to get the activation energy (Ea) of the hydrolysis reaction catalyzed by Pd1@Ni12.5P40/rGO nanoparticles, the reaction was carried out in at different temperature ranged from 298 K to 313 K. The value of the rate constant is calculated from the dehydrogenation curve in Figure 11(a). Figure 11(b) plots the Arrhenius curve of the of the catalyst as a function of . It can be seen from the image that Ea for catalyzing the hydrolysis of AB is 29.31 kJ·mol-1, which is lower than many reported catalysts shown in Table 1, indicating that the synthesized Pd1@Ni12.5P40/rGO nanoparticle catalyst has preeminent catalytic performance towards the hydrolysis of AB.

The cycle durability of the catalyst is essential for practical application. In the catalytic hydrolysis process of AB, the cyclic stability of Pd1@Ni12.5P40/rGO is shown in Figure 12. After five cycles of continuous catalytic hydrolysis of ammonia borane, Pd1@Ni12.5P40/rGO nanoparticles still maintain 52.98% of the incipient catalytic activity. The catalytic activity and conversion rate remained good after the cycle, indicating that Pd1@Ni12.5P40/rGO nanoparticles can be recycled for at least five times.

In order to further research the mechanism of hydrolysis AB catalyzed by Pd1@Ni12.5P40/rGO, the isotope experiment was conducted by using D2O as the solvent instead of H2O. From Figure 13, the hydrolysis rate of AB in D2O is much lower than that of H2O as the solvent. The high kinetic isotope effect (KIE) of D2O () stated that the division of the O-H bond of water is the rate-determining step of the hydrolysis reaction, which was also consistent with previous research [55]. The interaction between H with positive charge in water and Pδ- can speed up the rate-determining step to obtain faster reaction kinetics. In this context, we can infer the possible mechanism of Pd1@Ni12.5P40/rGO catalytic hydrolysis of AB when considering the report published earlier. At the start of AB hydrolysis, H2O may offend AB to form [H3NBH2H] H-OH species [12] and interact with Pd@NiP NPs sequentially. The B-H bond breaks subsequently, and the H atom shifts to the surface of Pd@NiP which leads to an increase in the oxidation capacity of the B-H bond; the synergistic effect makes this process easier. The H of B-H in ammonia borane with negative charge will have electric charge effect with Niδ+[56]. The synergistic catalytic effects of the Ni and P may be the key to enhance the catalytic activity. At the same time, the water molecule splits into H+ and OH -, and one of H atom in AB is replaced by OH in water. Therefore, the two active H atoms provided by AB and H2O combine to release H2 molecules, and with that comes the creation of BH(OH)2NH3 and B(OH)3 as the by-products.

4. Conclusion

Reduced graphene oxide-loaded Pd@NiP core-shell nanoparticles were synthesized by a one-step method as the catalyst for hydrogen production from hydrolysis of ammonia borane. The Pd@NiP/rGO core-shell nanoparticles were characterized by TEM, XRD, FTIR, Raman, XPS, and BET. The catalytic performance test shows that the synthesized Pd@NiP/rGO nanoparticles exhibit stronger catalytic activity than the phosphorus-free nanoparticles, which may be caused by the effective doping of phosphorus and the synergistic electron effect between Ni and P. The as-synthesized catalyst puts up splendid catalytic activity, with a TOF value of 133.33 min-1 and activation energy (Ea) of 29.31 kJ mol-1. MeAB with weak reducibility has better control capability in reducing metal ions than AB and NaBH4, so it can obtain catalyst with better catalytic activity. In addition, the results of the isotope experiment manifested that the division of the O-H bond of water was the rate-determining step of the hydrolysis reaction, and a detailed mechanism for the hydrolysis of AB catalyzed by nanoparticles was also put forward. Furthermore, the simple one-step synthesis method can be expanded to other graphene-supported transition metal phosphating core-shell nanoparticles.

Data Availability

The manuscript does not contain the original data.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Acknowledgments

This research was supported by the National Natural Science Foundation of China (Grant No. 21805217) and the Fundamental Research Funds for the Central Universities (WUT: 2019IVB014, 2021IVA014).