Due to the distinct advantages including stability, high hydrogen storage capacity, hydrolysis controllability and high hydrogen purity, borohydrides form a kind of very promising hydrogen storage materials [1-5]. A typical representative of them is sodium borohydride (NaBH₄) [6, 7]. The hydrolysis of 1 g NaBH₄ can release 0.21 g hydrogen with high purity. However, the self-hydrolysis of NaBH₄ without a catalyst is extremely slow [7, 8]. Therefore, great efforts have been made to develop suitable catalysts which could endow NaBH₄ hydrolysis decent rates [9-11].

Varieties of metal catalysts, such as Ru [12-19], Pd [13, 20], Co [21-32], Ni [32-36] and complexes of them [37-46], were extensively studied in the last years. Plenty of researches suggest that Ru-based catalysts have excellent activity in catalyzing NaBH₄ hydrolysis [8, 12-15, 40, 41]. Nevertheless, although monometallic Ru catalysts have high activity for catalyzing NaBH₄ hydrolysis under neutral conditions, their performance is dramatically compromised under alkaline conditions [19, 38]. Yet, NaBH₄ solutions are usually stabilized in alkaline environment to prevent the uncontrolled self-hydrolysis. On the contrary, when Ru iscombined with other metals, the formed complex catalysts could maintain the high catalytic activity for NaBH₄ hydrolysis under alkaline conditions [15, 41]. Furthermore, considering the high cost and scarcity of Ru metal, it is very important to develop catalysts with special composition and nano/microstructures to improve the Ru efficiency. For this propose, Ru species in the catalysts must be dispersed well enough. It means that the catalysts composed of Ru nanoparticles, clusters, or even single atoms are highly demanded. However, freeRunanoparticles, clusters, or singleatomsarenot stable. They tend to aggregate to reduce the surface energy. Thus, varieties of substrates have been employed to support Ru species to improve their stability [12-15, 40-43]. Nevertheless, Ru species in these catalysts is normally in the form of nanoparticles, leaving space for further reducing the size to improve the catalytic efficiency.

Herein, we report a novel Ru-based catalyst (Ru/Co-TiO₂), which was fabricated through coprecipitation of Ru³⁺ and Co²⁺ on TiO₂ nanoparticles in aqua regia. Ru and Co species in this catalyst were well dispersed. NaBH₄ hydrolysis experiments in alkaline solutions showed that the Ru/Co-TiO₂ catalyst had superior catalytic activity and durability.

Scheme 1 shows an illustration of the fabrication of the Ru/Co-TiO₂ catalyst. Briefly, RuCl₃·3H₂O and Co(NO₃)₂·6H₂O were firstly dissolved in aqua regia to form a homogenous solution, into which TiO₂ nanoparticles (~40 nm) were added. The mixture was stirred in 140 ℃ oil bath until it was dry. The obtained solid was then reduced in aqueous solution of ammonia borane, washed with water and dried at 60 ℃ in vacuum overnight to obtain Ru/Co-TiO₂ catalyst. Ru-TiO₂ and Co-TiO₂ were fabricated for control experiments with the same method, but without Co(NO₃)₂·6H₂O or RuCl₃·3H₂O in the original aqua regia solutions, respectively.

After the successful fabrication, the loading density of Ru and Co in the catalysts was measured with inductively coupled plasmaoptical emission spectroscopy (ICP-OES). The results demonstrated that Ru/Co-TiO₂ catalyst had 0.06 wt% Ru and 0.61 wt% Co, while Ru-TiO₂ and Co-TiO₂ had 0.09 wt% Ru and 0.77 wt% Co, respectively. The Brunauer-Emmett-Teller (BET) surface measurement indicated that TiO₂ nanoparticles used as the substrate had a specific surface area of 64.3 m²/g. Considering the low density of Ru and Co, the high surface area of TiO₂ could provide sufficient space for them to anchor.

Transmission electron microscopy (TEM) and high-angle annular dark field-scanning transmission electron microscopy (HAADF-STEM) images in Fig. 1 have shown the structures of the fabricated Ru/Co-TiO₂ catalyst. The lattice fringes highlighted in Fig. 1b show the planes with interplanar spacing of 3.5 Å, corresponding to the (101) planes of TiO₂. However, due to their far lower density than Ti, no distinct Co and Ru atoms could be observed with HAADF-STEM, which is consistent with the results previously reported by Tao et al. [47]. This also verified the good distribution of Co and Ru species in the sample. As Co and Ru species could not be observed directly from HAADF-STEM images, elementalmappingwas carriedout toverify theirexistence (Fig. 1c). It could be clearly seen that the Co and Ru species sparsely dispersed on TiO₂ nanoparticles, and that the signal of Co was distinctly stronger than Ru, which was consistent with their different loading density.

To further analyze the surface of the Ru/Co-TiO₂ catalyst, X-Ray Photoelectron Spectroscopy (XPS) investigations were carried out. Fig. S1a (Supporting information) shows the full survey XPS spectrum of the Ru/Co-TiO₂ catalyst. In the spectrum the peaks of Ti and O could be unambiguously observed, while the peaks of Ru and Co were not significantly visible. Due to the low content of Ru and the disturbance of Ti 2p peaks, the Ru 3p peaks could not be distinguished during fine scanning (Fig. S1b in Supporting information). Nevertheless, the Co 2p peaks in Fig. S1c (Supporting information) distinctly show that Co species in the catalyst had two oxidation states. The peaks at 786.0 eV and 801.7 eV corresponded to Co²⁺, and the peaks at 780.9 eV and 796.4 eV corresponded to Co³⁺. The analysis based on XPS investigation has further confirmed the low content of Ru and Co species in the Ru/CoTiO₂ catalyst. In addition, the X-ray Diffraction (XRD) pattern in Fig. S2 (Supporting information) shows apparent diffraction peaks of Ru/Co-TiO₂ catalyst, while all of them belong to the planes of TiO₂. No diffraction peaks associated to Ru or Co species could be observed, indicating the lack of related nanoparticles again.

The catalytic performance of the fabricated Ru/Co-TiO₂ catalyst was evaluated with catalyzed NaBH₄ hydrolysis under different conditions in a home-built setup [48]. In a typical experiment, 100 mg catalyst and 25 mg NaBH₄ were mixed in an aqueous medium with 10 wt% NaOH to form 10 g reaction mixture to start the catalyzed hydrolysis, and the NaBH₄ hydrolysis rate was reflected by the hydrogen generation (HG) rate. Various factors which influenced the hydrolysis rate including composition of catalysts, NaOH concentration, NaBH₄ concentration, catalyst dosage and temperature were investigated in detail.

Aqueous solutions of NaBH₄ are usually stabilized under alkaline conditions to prevent the unwanted self-hydrolysis. Therefore, a competent catalyst should have high activity for catalyzing NaBH₄ hydrolysis under alkaline conditions. In our study, we performed NaBH₄ hydrolysis experiments with different NaOH concentrations from 0 to 20 wt% (Fig. 2). It was found that NaBH₄ hydrolysis catalyzed by Ru/Co-TiO₂ had the highest HG rate of 10.34 mL/min when NaOH concentration was 10 wt%. Considering the loading density of Ru in the catalyst (0.06 wt%), the HG rate per gram Ru could be calculated to be 172 L min^-1 g_Ru^-1. This value is higher than the Ru efficiency of most recently reported Ru-based catalysts (Table S1 in Supporting information) [13-18, 39-41]. If the average HG rate was calculated based on the amount of Co in the catalyst, it was 17 L min^-1 g_Co^-1, also very high due to the contribution of Ru (Table S2 in Supporting information) [23-27].

When NaOH concentration was lower than 10 wt%, the existence of NaOH apparently accelerated the catalyzed NaBH₄ hydrolysis. However, when the concentration of NaOH was too high (higher than 10 wt%), the hydrolysis reaction became slower. It is widely accepted that the slowdown of the hydrolysis reaction when NaOH concentration was too high was caused by the increasing viscosity of the reaction mixture, but how low concentration of NaOH accelerated the catalyzed NaBH₄ hydrolysis is still unknown.

The Co-TiO₂ catalyst showed a similar trend as Ru/Co-TiO₂, providing highest catalytic activity when NaOH concentration was 12.5 wt%. This is consistent with the results reported in previous studies [30, 32, 46]. However, Ru-TiO₂ showed a different trend. The HG rate of the reactions catalyzed by Ru-TiO₂ (3.65 mL/min) was on the same level with Ru/Co-TiO₂ (5.47 mL/min) when NaOH was absent. But the existence of NaOH had distinct inhibiting effect for Ru-TiO₂ catalyst. With the increasing of NaOH concentration, the NaBH₄ hydrolysis rate decreased dramatically. When NaOH concentration was above 12.5 wt%, hydrogen could hardly be produced.

To elucidate the impact of NaBH₄ concentration on the rate of hydrolysis catalyzed by Ru/Co-TiO₂, experiments with different NaBH₄ concentrations were carried out. It could be concluded from Figs. 3a and b that the hydrolysis rate increased sharply when NaBH₄ concentration was increased from 0.10 wt% to 0.25 wt%, but deceased gradually when NaBH₄ concentration was further increased from 0.25 wt% to 10 wt%. As shown in Fig. 3a, when NaBH₄ concentration was higher than 0.25 wt%, hydrogen bubbles emerged as soon as the NaBH₄ solution was mixed with Ru/Co-TiO₂ catalyst. However, if NaBH₄ concentration decreased from 0.25 wt%, the initiation of the hydrolysis reaction gradually became more difficult, and therefore the volume of hydrogen generated at the very beginning was very little. Actually, in the catalyzed NaBH₄ hydrolysis experiments, we noticed that the colour of the Ru/Co-TiO₂ catalyst changed from grey to black during the reactions. When NaBH₄ concentration was low enough, it could be further observed that hydrogen bubbles could be generated in the solution only after the colour change of the catalyst had finished. This was consistent with the phenomenon reported by Demirci et al. in 2011 [49]. It indicates that the active form of the catalyst might not be the fresh prepared Ru/Co-TiO₂, but be some product from Ru/Co-TiO₂ reacted with NaBH₄ solution. This is a very important point for the investigation on the potential mechanisms of NaBH₄ hydrolysis catalyzed by Ru/Co-TiO₂. However, as emphasized by Demirci et al., the elucidation of the mechanisms will be a much tougher work to do [50]. Anyway, due to the lag phase at the beginning, the average hydrolysis rate of reactions decreased sharply as NaBH₄ concentration decreased from 0.25 wt% (Fig. 3b). However, when NaBH₄ concentration was increased from 0.25 wt%, the difference at the very beginning of the reactions became not distinguishable. On the contrary, with the increase of NaBH₄ concentration, the increase of solution viscosity and the formation of more NaBO₂ around the catalyst nanoparticles started to hinder mass exchange in the reaction [33, 43, 46]. As a result, the general hydrolysis rate gradually decreased (Fig. 3b). Nevertheless, although the increase of NaBH₄ concentration brought about acceleration or slowdown of hydrolysis reactions, the yield of hydrogen was not significantly affected. The final volume of hydrogen generated increased proportionally with the increasing of NaBH₄ concentration (Fig. S3 in Supporting information).

The change of dosage of the Ru/Co-TiO₂ catalyst dramatically affected NaBH₄ hydrolysis rate in another way. As shown in Figs. 3c and d, when the catalyst amount increased from 50 mg to 200 mg, the hydrolysis rate increased linearly from 4.60 mL/min to 20.06 mL/min, indicating that the catalyst was fully utilized in the hydrolysis reactions.

Temperature plays an important role in the NaBH₄ hydrolysis as it is an exothermic reaction. To investigate the temperature effects on our system, the reactions were carried out at 30 ℃, 40 ℃, 50 ℃ and 60 ℃. As shown in Figs. 4a and b, the NaBH₄ hydrolysis rate increased sharply with the increasing of temperature. The HG rate of the hydrolysis reactions at 30 ℃ was just 10.34 mL/min, while it jumped to 45.54 mL/min when it was 60 ℃, with 3.4 times increasing (Fig. 4b). Moreover, the activation energy (E_a) of the hydrolysis reactions catalyzed by Ru/Co-TiO₂ was calculated with Arrhenius equation (Eq. 1):

Where T is the reaction temperature of NaBH₄ hydrolysis in Kelvin (K), k is the rate constant of the hydrolysis reaction, R is the gas constant, and A is pre-exponential factor. Based on the NaBH₄ hydrolysis rates reflected by the HG rates at different temperatures, the E_a was calculated to be 38.74 kJ/mol (Fig. 4c), which is comparable with the E_a of NaBH₄ hydrolysis reactions catalyzed by other reported Ru-based catalysts (Table S1 in Supporting information) [13-17, 39-41], demonstrating good catalytic activity of Ru/Co-TiO₂.

To verify the practicability of Ru/Co-TiO₂ in practical applications, the durability tests were carried out by evaluating the activity of catalyst in cycle use. After each use, the used catalyst was collected from the reaction mixture by centrifuging, washed with water, and dried in vacuum at 60 ℃. And then it was used for catalyzing NaBH₄ hydrolysis again. As show in Fig. 4d, the NaBH₄ hydrolysis rate at the 55^th cycle was 97.1% of the rate at 1^st cycle. After the 5^th experiment, the used catalyst was further characterized with high resolutionTEM, XPS and XRD. The results (Fig. S4 in Supporting information) showed that no significant changes could be observed on the catalyst after being used for 5 cycles. It demonstrates that the Ru/Co-TiO₂ catalyst had excellent durability in catalyzing NaBH₄ hydrolysis.

In summary, a Ru/Co-TiO₂ catalyst composed of TiO₂ nanoparticles with Ru and Co species highly dispersed on the surface was developed by coprecipitation of Ru³⁺ and Co²⁺ on TiO₂ nanoparticles in aqua regia. The structures and chemical compositions of the Ru/Co-TiO₂ catalyst were verified with ICPOES, TEM, HAADF-STEM, XPS, and XRD investigations. The experimental results showed that the Ru/Co-TiO₂ catalyst had superior performance for catalyzing NaBH₄ hydrolysis in the alkaline solution with 10 wt% NaOH, providing a hydrogen generation rate per gram Ru at 30 ℃ as high as 172 L min^-1 g_Ru^-1, better than most of the previously reported Ru-based catalysts. Moreover, the Ru/Co-TiO₂ catalyst also showed excellent durability in cycle use, with just 2.9% of activity loss after being used for 5 times. These advantages made it a potential choice for catalyzing NaBH₄ hydrolysis to supply hydrogen in portable systems.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Supplementary material related to this article can befound, in the online version, at doi:https://doi.org/10.1016/j.cclet.2020.03.072.