1 Introduction

Hydrogen is one of the most promising energy carriers for the future energy systems, but the low volumetric energy density of hydrogen (in both compressed gas and liquid forms) makes the storage of hydrogen a difficult problem [1]. Ammonia has been considered as one of the most promising hydrogen carriers for the high gravimetric (17.7 wt% H2) and volumetric (121 kg H2 m−3 in the liquid form) H2 density, high energy density, liquid state under mild conditions and billion tons of annual production in world-wide [2,3,4]. The decomposition of NH3, a critical step for production of H2 from NH3, is thermodynamically favorable under relatively low-temperature conditions. However, the sluggish kinetics of NH3 decomposition reaction greatly limits the production of H2 from decomposition of NH3. Thus, it is extremely expected to develop efficient catalysts exhibiting high catalytic activity and good stability for decomposition of NH3 at relatively low temperature [3, 5].

Among the different kinds of NH3 decomposition catalysts, supported Ru catalysts usually show high activity and stability for decomposition of NH3 [6,7,8,9]. It is generally accepted that the catalytic performances of supported Ru catalysts in NH3 decomposition are mainly determined by the active sites on the surface of Ru nanoparticles (NPs), which are closely correlated with the size, shape and interfacial structure between Ru NPs and oxide supports. As multicomponent catalyst systems, the catalytic performances of oxide supported Ru catalysts rely not only on the structure of Ru NPs but also on the properties of oxide supports. Usually, the catalytic performances of supported Ru catalysts in NH3 decomposition is strongly related to the properties of the oxide supports [7, 8, 10,11,12]. Previous studies show that oxides such as SiO2 [11, 13, 14], TiO2 [8], zeolite [14], Al2O3 [15, 16], MgO [8, 17,18,19], ZrO2 [20,21,22], La2O3 [23], barium hexaaluminate [24], Y2O3 [25], Pr2O3 [26] and Sm2O3 [27] are efficient supports for Ru-based catalysts.

Currently, development of new kind of oxide supported Ru catalysts is still highly desirable for expanding the scope of Ru-based NH3 decomposition catalysts. In recent years, neodymium oxide (Nd2O3) has attracted much attention due to its properties such as high chemical stability and high thermal conductivity. Also, the Nd2O3 materials have been widely applied as catalysts or supports in catalysis [28,29,30,31,32]. Nd2O3 is a promising support for the metal-based catalysts due to its ability to stabilize the active metal phase with high dispersion. Currently, Nd2O3 supported Co catalysts have been investigated in ammonia synthesis reaction [31, 32]. It has long been known that the function of oxide support is not only to disperse and stabilize metal particles but also to generate interface phenomena through metal–oxide interaction. Recently, the role of metal–support interaction in tuning the electronic properties and catalytic performances of oxide (e.g., FeOx, CeO2 and TiO2) supported metal NPs has been recognized in different catalytic reactions [33, 34]. However, little attention has been paid to the use of Nd2O3 for construction of efficient Nd2O3 supported Ru NPs catalyst with excellent activity and stability for NH3 decomposition. Meanwhile, the importance of the metal–support interaction in modulating the catalytic performances of Nd2O3 supported Ru catalysts in NH3 decomposition has not been recognized up to now.

Here, Ru/Nd2O3 analogues consisting of nearly identical Ru NPs and Nd2O3 by different methods (milling and precipitation method) were prepared and evaluated as catalysts for NH3 decomposition reaction. Compared with Ru/Nd2O3 obtained from milling method, Ru/Nd2O3 obtained from precipitation method has remarkably improved activity for NH3 decomposition under identical reaction conditions. A variety of characterization techniques have been applied to study the relationship between the catalytic performance and the structure of Ru/Nd2O3 catalysts obtained from different methods in detail, which reveal that the catalytic activity of Ru/Nd2O3 catalysts show high dependence on the interaction between the Ru NPs and Nd2O3 support.

2 Experimental

2.1 Preparation of Ru/Nd2O3 Catalysts

The support precursor of Nd(OH)3 was prepared by modifying established method [35]. First, Nd(NO3)3·6H2O and KOH were dissolved in deionized water (0.5 M), respectively. Then the Nd(NO3)3 solution was added into the KOH solution under stirring for 10 min. The mixture was transferred into a Teflon-lined stainless autoclave at 180 °C for 12 h. The Nd(OH)3 was obtained via filtration, washing and drying at 80 °C for 12 h.

The Ru/Nd2O3-p samples with Ru mass loading of 1.5 wt% and 3 wt% were synthesized by precipitation method. Typically, 0.36 g of Nd(OH)3 and 0.09 g of KOH were dispersed in 30 mL of water. Subsequently, a designated amount of RuCl3 solution (0.007 M) was slowly added under vigorous stirring [36]. The mixture was stirred 2 h at room temperature. The solid was obtained by filtration, and then washed with deionized water. Finally, the product was dried at 80 °C for 12 h. The Ru/Nd2O3-p samples were obtained by reduction in 10% H2/N2 atmosphere at 500 °C for 2 h.

For comparison, 1.5 wt% and 3 wt% Ru/Nd2O3-m samples were prepared using milling method from Nd(OH)3 and RuO2 NPs. The preparation of RuO2 NPs is similar to the precipitation method for preparing Ru/Nd2O3-p catalyst. A designated amount of RuCl3 solution (0.007 M) was slowly added to the KOH solution (0.12 M) under vigorous stirring. The mixture was stirred 2 h at room temperature. The RuO2 NPs was obtained by filtration. In a typical synthesis, desired amount of Nd(OH)3 and RuO2 NPs were ground in a mortar for 20 min. The Ru/Nd2O3-m samples were obtained by reduction in 10% H2/N2 atmosphere at 500 °C for 2 h.

2.2 Characterization

X-ray diffraction (XRD) measurements were performed with Cu Kα radiation using an X’Pert Pro (PANAlytical) diffractometer.

The specific surface area of the catalysts was obtained by means of a nitrogen adsorption–desorption isotherm collected at − 196 °C using a Micromeritics ASAP 2020 analyzer. The samples were heated to 300 °C in a flow of N2 for 8 h prior to the experiment. The specific surface area was calculated by the Brunauer–Emmet–Teller (BET) method at the N2 relative pressure range of 0.05 < P/P0 < 0.30.

To determine the morphology and size of the catalysts, transmission electron microscopy (TEM) measurements were performed on a JEOL 2100X. The samples were prepared by ultrasonic dispersion of powders in ethanol and a droplet of the dispersion was then placed onto a copper grid. The dispersion of Ru particles (DRu) was calculated using the equation by Borodziński and Bonarowska [36]:

$$\text{D}_{\text{Ru}}=\sqrt[1.23]{\frac{\text{d}_{\text{at}}\times 3.32}{\text{d}_{\text{TEM}}}}\quad {\text{for}}\; 0.2 \le D_{Ru} \le 0.92.$$
$${\text{D}}_{\text{Ru}}=\frac{\text{d}_{\text{at}}\times 5.01}{\text{d}_{\text{TEM}}} \quad \text{for} D_{Ru}<0.2.$$

where dat is the atomic diameter of Ru (0.269 nm).

The exposed Ru dispersion of the catalyst was measured by CO pulse chemisorption using an AutoChem 2910. About 100 mg of the catalyst was reduced at 400 °C for 1 h in 10% H2/Ar (30 mL min−1) prior to the pulse CO chemisorption measurement, and then cooled down to room temperature in He (30 mL min−1). The CO uptake was measured using a gas chromatograph equipped with a TCD detector. The Ru dispersion was calculated by assuming a CO:Ru stoichiometry of 1:1.

Temperature programmed reduction (TPR) was performed with a fixed-bed reactor equipped with a gas chromatograph. TPR experiments were carried out from 30 to 800 °C (10 °C min−1) under 5% H2/Ar with a flow rate of 30 mL min−1.

CO adsorption diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) studies were performed to monitor the gas–solid interface using a Bruker TENSOR II equipped with a mercury cadmium telluride detector and operated at a resolution of 4 cm−1 for 16 scans. Previous to the CO adsorption experiment, the IR cell with catalyst was purged in Ar at room temperature for 0.5 h, and the spectrum was recorded as the background. Then a 5% CO/Ar mixture flow was switched to the IR cell for 0.5 h. Afterward, the chamber was purged with Ar (30 mL min−1) for 0.5 h to remove gaseous CO before the spectrum featuring CO adsorption was recorded.

X-ray photoelectron spectroscopy (XPS) measurement was performed on an ESCALAB MK-II spectrometer with Al Ka as the excitation source. Carbonaceous C 1s line (284.8 eV) was used as the reference to calibrate the binding energy.

2.3 Catalytic Reaction Test

Catalytic tests were carried out on a continuous fixed-bed flow quartz reactor. The dried catalysts (20–40 mesh) were reduced in an 10% H2/Ar (30 mL min−1) flow at 500 °C for 2 h. Then pure NH3 was inputted with a flow rate of 25 mL min−1. All the activity tests of catalysts were conducted under atmospheric pressure, and pure NH3 was used as the only reactant. The reaction temperature was in the range of 300–550 °C. The activity tests of catalysts with different weight hourly space velocity (WHSV) values in NH3 stream were conducted at 450 °C. Product gas composition was analyzed by on-line gas chromatograph (GC-7890A, Agilent) equipped with a thermal conductivity detector and Porapak N column. The tested reaction temperature was in the range of 300–500 °C.

3 Results and Discussion

3.1 Structure Characterization of Catalysts

In this study, RuO2 NPs were firstly synthesized using precipitation method. Figure S1 shows TEM images of RuO2 NPs in different scale. It is obvious that RuO2 colloid NPs have a spherical-like morphology with a uniform particle size with an average particle size of 1.9 ± 0.2 nm (Fig. S1). Well-crystallized Nd2O3 were used as the support precursor to disperse RuO2 NPs. Precipitation and solid milling were applied to disperse highly dispersed RuO2 NPs onto the Nd2O3 support (Scheme 1). The samples were reduced at 500 °C, and the obtained Ru/Nd2O3 catalysts were denoted as 1.5% Ru/Nd2O3-m, 3% Ru/Nd2O3-m, 1.5% Ru/Nd2O3-p and 3% Ru/Nd2O3-p samples, respectively.

Scheme 1
scheme 1

Schematic illustration of formation of RuO2 NPs (a) and deposition of highly dispersed Ru NPs on Nd2O3 support by precipitation method (b) and solid milling method (c)

Full size image

To investigate the structure of the catalysts, the XRD patterns of Ru/Nd2O3 samples obtained from different methods and Nd2O3 support are shown in Fig. 1. Similar with that of the Nd2O3 (JCPDS 00-121-0579), it was seen from the pattern that the diffraction peaks of Nd2O3 support were relatively wide, which was related to the low degree of crystallinity of the support. After loading of Ru NPs on the Nd2O3 support, no characteristic diffraction peak of Ru (JCPDS 011253) can be found in the XRD patterns of Ru/Nd2O3 samples obtained from different methods. This is related to the high dispersion of Ru NPs on the support surface. The major characteristic peak intensity of Ru/Nd2O3-m samples was similar with that of the Nd2O3 support, indicating the well preservation of Nd2O3 in Ru/Nd2O3-m samples. By contrast, the diffraction peak intensity of the 1.5% Ru/Nd2O3-p and 3% Ru/Nd2O3-p catalysts were much stronger than that of the Nd2O3 support, which indicates that the crystallinity of the Nd2O3 in the Ru/Nd2O3-p catalysts was improved.

Fig. 1
figure 1

XRD patterns of Nd2O3, 1.5% Ru/Nd2O3-m, 3% Ru/Nd2O3-m, 1.5% Ru/Nd2O3-p and 3% Ru/Nd2O3-p obtained after reaction. Standard pattern of Ru and Nd2O3 were included as references

Full size image

Table S1 shows the specific surface area and average Ru particle size (estimated by TEM) of the Ru/Nd2O3 catalysts. The BET surface areas of Ru/Nd2O3 samples and Nd2O3 support are summarized in Table S1. The Nd2O3 support has a high surface area of 70.6 m2 g−1. It is obvious that the BET surface areas of Ru/Nd2O3-m samples decrease obviously after the deposition of Ru NPs. In comparison with that of the Ru/Nd2O3-m samples, Ru/Nd2O3-p samples possess relatively large surface areas. The influence of preparation method on the particle size of Ru NPs was studied by TEM. Based on the TEM and HRTEM images shown in Fig. 2, it can be seen that loading of Ru by precipitation method and milling method leads to the formation of highly dispersed Ru NPs on the surface of Nd2O3 support. The Ru NPs were uniformly dispersed on the surface of the support, which is similar with that of the pristine RuO2 NPs precursor (Fig. S1). Based on the Ru NPs in the images of different Ru/Nd2O3 samples, the average particle sizes of the Ru NPs can be preliminary estimated. The average particle sizes of the Ru NPs for the 3% Ru/Nd2O3-m and 3% Ru/Nd2O3-p from statistical analysis are about 2.6 and 2.5 nm, respectively. Meanwhile, the average particle size of Ru NPs in 1.5% Ru/ZrO2-m and 1.5% Ru/ZrO2-p are 2.3 and 2.4 nm, respectively, similar with that of the 3% Ru/Nd2O3-m and 3% Ru/Nd2O3-p samples. Based on the average particle size from TEM results, the Ru dispersion values of the Ru/Nd2O3 samples estimated are determined to be about 41–46%. Therefore, loading of Ru on the Nd2O3 support by impregnation and precipitation methods show no obvious influence on the Ru dispersion in Ru/Nd2O3-m and Ru/Nd2O3-p catalysts. The metal dispersion and average Ru particle size on Ru/Nd2O3 catalysts were also estimated from the CO chemisorption and the corresponding results are listed in Table S1. The metal dispersion for the 1.5% Ru/Nd2O3-p and 3% Ru/Nd2O3-p catalyst reaches 39.5% and 44.8%. The size of the Ru NPs from CO chemisorption is about 3.3 and 3.0 nm for the 1.5% Ru/Nd2O3-p and 3% Ru/Nd2O3-p samples, which shows a good agreement with the TEM images. Therefore, the Ru NPs in Ru/Nd2O3-p samples show similar particle size with that of the Ru/Nd2O3-m samples with similar Ru contents.

Fig. 2
figure 2

TEM and HRTEM images of 1.5% Ru/Nd2O3-m (a and e), 3% Ru/Nd2O3-m (b and f), 1.5% Ru/Nd2O3-p (c and g) and 3% Ru/Nd2O3-p (d and h) obtained after reaction

Full size image

XPS analyses were performed to investigate the chemical state of the Ru NPs in different Ru/Nd2O3 catalysts (Fig. 3). Due to the overlapped of Ru 3d3/2 and C 1s peak at c.a. 284.8 eV, Ru state was analyzed by the peak of Ru 3d5/2 peaks. The Ru 3d5/2 peak of 1.5% Ru/Nd2O3-m and 3% Ru/Nd2O3-m was found to be about 280.1 and 280.3 eV, consistent with the Ru 3d5/2 peak observed for the metallic Ru and highly dispersed Ru metal NPs on the oxide supports (279.6–280.2 eV) [37], indicating the dominant metallic state of Ru NPs in the Ru/Nd2O3-m samples. Similar with Ru/Nd2O3-m samples, Ru 3d5/2 peak centered at 280.0 eV can also be observed in the XPS spectra of 1.5% Ru/Nd2O3-p and 3% Ru/Nd2O3-p samples, demonstrating that the similar chemical state of metallic Ru NPs in the two Ru/Nd2O3-p samples. Besides the peak associated with the metallic Ru species at 280.0 eV, a broad peak at about 280.9 eV can also be observed for the 1.5% Ru/Nd2O3-p and 3% Ru/Nd2O3-p samples. As the Ru NPs in all the samples have been fully reduced, the much different binding energy value of Ru 3d5/2 peak at about 280.2 and 280.9 eV can be attributed to the metallic Ru species in different electronic environment states. In particular, the Ru 3d5/2 peak at about 280.9 eV may be associated with the Ruδ+ species located at the interface between the Ru NPs and Nd2O3 support. This result demonstrates that the electronic environment of Ru NPs in the Ru/Nd2O3-p samples is much different from that of the Ru NPs in the Ru/Nd2O3-m samples. Based on the Ru 3d XPS spectra (Fig. 3), the concentration of Ruδ+ species in total surface Ru species on the 1.5% Ru/Nd2O3-p sample is estimated to be about 45.0%. Compared to the 1.5% Ru/Nd2O3-p, 3% Ru/Nd2O3-p shows a relatively higher surface Ruδ+ fraction of 67.6%. TEM results reveal that nearly identical reduced Ru NPs can be observed in the different samples. An obvious XPS peak at 531.6 eV, associated with surface –OH or oxygen vacancies, can be observed in the O 1s XPS spectra of 1.5% Ru/Nd2O3-m, 3% Ru/Nd2O3-m, 1.5% Ru/Nd2O3-p and 3% Ru/Nd2O3-p catalysts samples (Fig. S2), indicating the presence of high amount of surface oxygen vacancies. At the same time, nearly identical Nd 3d XPS spectra can be observed for the different Ru/Nd2O3 samples (Fig. S3). However, substantial difference of chemical structure can be observed over the two samples even with nearly identical metallic Ru NPs. Considering the similar Ru NPs on the Nd2O3 support in different 1.5% Ru/Nd2O3 samples, the appearance of peak at about 280.9 eV is supposed to be caused by the presence of Ruδ+ species located at the interface between Ru NPs and Nd2O3 support, which may be beneficial for the catalytic performances in NH3 decomposition.

Fig. 3
figure 3

Ru 3d core level XPS spectra of 1.5% Ru/Nd2O3-m, 3% Ru/Nd2O3-m, 1.5% Ru/Nd2O3-p and 3% Ru/Nd2O3-p catalysts obtained after reaction

Full size image

3.2 Catalytic Performances of Ru/Nd2O3 Catalysts

Figure 4a shows the temperature dependence of NH3 conversion obtained over Ru/Nd2O3 catalysts prepared by precipitation and milling method. The Ru NPs alone show very poor catalytic performance under reaction conditions (Fig. S4). However, upon dispersing Ru NPs on the Nd2O3 support, the catalytic performance of Ru NPs can be greatly improved (Fig. 4a). It can be seen that the catalytic activity of the catalyst Ru loaded on the same Nd2O3 support were greatly influenced by the preparation method. Under identical conditions, the Ru/Nd2O3-p samples exhibit much higher catalytic activity than the Ru/Nd2O3-m at temperature ranging from 350 to 550 °C, implying the critical role of preparation method for the formation of efficient catalysts. The Ru/Nd2O3-p readily show apparent activity from 350 °C, while Ru/Nd2O3-m samples display negligible activity at temperature ranging from 350 to 450 °C. At the same temperature, the activities follow the order of 3% Ru/Nd2O3-p > 1.5% Ru/Nd2O3-p > 3% Ru/Nd2O3-m > 1.5% Ru/Nd2O3-m. For example, the NH3 conversion of 3% Ru/Nd2O3-p can reach 84.2% at 450 °C, which is much higher than that of the 1.5% Ru/Nd2O3-p (49.0%). By contrast, no obvious activity can be observed over the 3% Ru/Nd2O3-m and 1.5% Ru/Nd2O3-m samples at the temperature of 450 °C.

Fig. 4
figure 4

a NH3 conversion over 1.5% Ru/Nd2O3-m, 3% Ru/Nd2O3-m, 1.5% Ru/Nd2O3-p and 3% Ru/Nd2O3-p catalysts as a function of reaction temperature. b Effect of WHSV on the NH3 conversion and H2 formation rate of 3% Ru/Nd2O3-p catalyst for NH3 decomposition at 450 °C. c H2 formation rates of 3% Ru/Nd2O3-p and typical efficient Ru-based NH3 decomposition catalysts at 450 °C with a WHSV of 30,000 mL gcat−1 h−1. d Stability test of 3% Ru/Nd2O3-p catalyst for NH3 decomposition at 450 °C

Full size image

Additionally, the catalytic performance of 3% Ru/Nd2O3-p was further examined with a supply of pure NH3 at various WHSV values (5000–70,000 mL gcat−1 h−1). Figure 4b depicts effect of WHSV on the catalytic performance of 3% Ru/Nd2O3-p catalyst at 450 °C. It can be inferred that the higher NH3 conversion rate directly contributes to a higher H2 formation rate due to the absence of side reactions in NH3 decomposition reaction. It can be seen that NH3 conversion at 450 °C increases remarkably with decrease of WHSV because of the increase in the residence time of the reactant. At the temperature of 450 °C, nearly complete conversion (96.8%) can be obtained with a WHSV as high as 10,000 mL gcat−1 h−1. The H2 formation rate of catalyst based on the weight of the 3% Ru/Nd2O3-p catalyst was calculated at various WHSV values. With the WHSV value increases from 5000 to 70,000 mL gcat−1 h−1, the H2 formation rate of 3% Ru/Nd2O3-p catalyst shows a remarkable increase from 365.5 to 2250.0 mmol gcat−1 h−1 at 450 °C. A comparison of activity of Ru/Nd2O3-p catalysts with those of the other efficient Ru-based catalysts for NH3 decomposition reported in literatures are shown in Fig. 4c and Table 1. Under similar reaction conditions, not only the NH3 conversion but also the H2 formation rate of 3% Ru/Nd2O3-p is comparable with many of the efficient supported Ru catalysts in literatures. The H2 formation rate of 1.5% Ru/Nd2O3-p and 3% Ru/Nd2O3-p catalyst per gram Ru is up to 859.9 and 1002.1 mmol gRu−1 min−1 at 450 °C, which is about 1.7 and 2.0 times higher than one of typical efficient Ru/MgO-DP catalyst (495.2 mmol gRu−1 min−1) reported previously.

Table 1 NH3 conversion and H2 formation rate of typical supported Ru catalysts at 450 °C
Full size table

Above results clearly show that Nd2O3 is a good support for preparation of efficient supported Ru NPs-based catalysts. Apart from catalytic activity, stability is another crucial parameter for evaluation the catalytic performance of Ru/Nd2O3-p catalyst in NH3 decomposition reaction. Here, a stability test of the 3% Ru/Nd2O3-p catalyst was also carried out at 450 °C with a relatively high WHSV of 30,000 mL gcat−1 h−1. As shown in Fig. 4d, no obvious activity loss is observed and approximately 81.0% NH3 conversion is maintained over 120 h, demonstrating the excellent stability of 3% Ru/Nd2O3-p catalyst for NH3 decomposition reaction. XRD (Fig. S5) pattern of 3% Ru/Nd2O3-p obtained after the stability test shows that no characteristic XRD peaks associated with Ru metal phase can be observed, evidencing that no serious aggregation of Ru NPs occurred under harsh reduction conditions. TEM image (Fig. S6) further shows that the Ru NPs after 100 h reaction were uniformly distributed on the Nd2O3 support. The Ru NPs on the Nd2O3 support is the active component for NH3 decomposition reaction. Here, the well preservation of highly dispersed Ru NPs of 3% Ru/Nd2O3-p at 450 °C agrees well with its high stability under NH3 decomposition reaction conditions. Overall, the 3% Ru/Nd2O3-p catalyst not only gives high activity but also gives high stability, demonstrating the promise for potential application as efficient catalysts in NH3 decomposition reaction.

3.3 Structure–Performance Relationship of Ru/Nd2O3-m and Ru/Nd2O3-p Catalysts

To clarify the origin of high activity of 4% Ru/Sm2O3‒p, the apparent activation energy (Ea) of different Ru/Nd2O3-m and Ru/Nd2O3-p catalysts was calculated based on NH3 conversion values far away from the equilibrium value (less than 20%). From Arrhenius-type plots (Fig. 5), it can be seen that no obvious difference in the Ea can be observed for the different Ru/Nd2O3-m and Ru/Nd2O3-p samples. The apparent activation energy of 3% Ru/Nd2O3-p is 142 kJ mol−1, which is little smaller than that of 3% Ru/Nd2O3-m (132 kJ mol−1). At the same time, similar Ea value can be obverted over the 1.5% Ru/Nd2O3-p (151 kJ mol−1) and 1.5% Ru/Nd2O3-m (152 kJ mol−1) samples. For the Ru/Nd2O3-m and Ru/Nd2O3-p samples with different Ru contents, it can be seen that the Ea values of both samples show a little decrease with increase of Ru content. Among the different Ru/Nd2O3 samples, the 3% Ru/Nd2O3-m sample gives a relatively small Ea of 132 kJ mol−1.

Fig. 5
figure 5

a Arrhenius plots of 1.5% Ru/Nd2O3-m, 3% Ru/Nd2O3-m, 1.5% Ru/Nd2O3-p and 3% Ru/Nd2O3-p catalysts. b TOFH2 values of 1.5% Ru/Nd2O3-m, 3% Ru/Nd2O3-m, 1.5% Ru/Nd2O3-p and 3% Ru/Nd2O3-p at 400 and 450 °C respectively

Full size image

Comparison the intrinsic activity of Ru NPs is useful for understanding the remarkable different activity of Ru/Nd2O3-m and Ru/Nd2O3-p catalysts. Usually, the H2 formation turnover frequency (TOFH2), which can be calculated by normalizing the observed H2 formation rate (mole formed, H2 gcat−1 s−1) to the number of exposed Ru surface atoms per gram of catalyst, is a standard parameter for the evaluation of intrinsic catalytic performance of different catalysts. Here, the TOFH2 of the Ru/Nd2O3-m and Ru/Nd2O3-p samples can be calculated by assuming that all surface Ru atoms act as active sites (Fig. S7). Correlating well with the catalytic activity results, a remarkable difference in the TOFH2 can be observed for Ru/Nd2O3-p catalysts in comparison with that of the Ru/Nd2O3-m catalysts under identical conditions. From the Fig. 5b and Fig. S2, it can be seen that the reaction rate of Ru/Nd2O3-p samples were significantly higher than that of Ru/Nd2O3-m samples for the decomposition of NH3 and formation of H2, revealing the excellent intrinsic activity of Ru NPs in Ru/Nd2O3-p samples. As shown in Fig. 5b, the 1.5% Ru/Nd2O3-p sample possesses a TOFH2 value of 6.04 s−1 at 500 °C, higher than that of the 3% Ru/Nd2O3-p (4.34 s−1) sample. The TOFNH3 value of 1.5% Ru/Nd2O3-p at 500 °C is 67.1 and 22.4 times higher than that of the 1.5% Ru/Nd2O3-m (0.09 s−1) and 3% Ru/Nd2O3-m (0.27 s−1) samples. Remarkably, the TOFH2 value of 3% Ru/Nd2O3-p (4.34 s−1) at 500 °C is 48.2 and 16.1 times higher than that of the 1.5% Ru/Nd2O3-m (0.09 s−1) and 3% Ru/Nd2O3-m (0.27 s−1) samples at the same temperature.

Metal–support interaction is widely present in oxide supported metal catalysts, and plays critical role in regulating the activity of catalysts. Here, TPR was performed to probe the metal‒support interaction of the Ru/Sm2O3 catalysts obtained from different methods. As shown in Fig. 6, two broad peaks centered at around 127 and 213 °C, corresponding to a two-step reduction process from RuO2 NPs to metallic Ru NPs, can be observed in the TPR profile of RuO2 NPs. For the 1.5% Ru/Nd2O3-m and 3% Ru/Nd2O3-m samples, two major broad reduction peaks centered at around 190 and 590 °C can be observed in the TPR profiles. The onset reduction temperature of 1.5% Ru/Nd2O3-m and 3% Ru/Nd2O3-m samples is also at about 100 °C, similar with that of the RuO2 NPs. About three broad peaks at ca. 187 °C can be observed for 1.5% Ru/Nd2O3-m and 3% Ru/Nd2O3-m samples. In combination with the TPR profile of Nd2O3, we can reasonably ascribe the reduction peaks at temperature below 240 °C to the reduction of the RuO2 NPs in Ru/Nd2O3‒m samples while reduction peaks at temperature above 240 °C can be attributed to reduction of surface oxygen on Nd2O3 support. As shown in Fig. 6, major part of RuO2 NPs in Ru/Nd2O3‒m samples can be reduced at temperature below 180 °C, similar with free standing RuO2 and Ru/Nd2O3-m samples. For the 1.5% Ru/Nd2O3-p and 3% Ru/Nd2O3-p sample, four broad peaks centered at 90, 120, 240 and 510 °C can be observed in the TPR profiles. The onset reduction temperature of 1.5% Ru/Nd2O3-p and 3% Ru/Nd2O3-p sample is about 53 °C, along with three reduction peaks centered at 84, 104, 122 and 240 °C, respectively. In comparison with that of the RuO2 NPs in Ru/Nd2O3-m sample, RuO2 NPs in Ru/Nd2O3-p samples give much different reduction behavior. Both the onset and peak reduction temperature of the Ru/Nd2O3-p catalysts shift towards low temperature, revealing the enhanced interaction between the RuO2 NPs and Nd2O3. Consider the similar and uniform dispersion of Ru NPs in different Ru/Nd2O3 samples, the difference of reduction peaks indicates that much enhanced interaction between RuO2 NPs and Nd2O3 support in Ru/Nd2O3-p samples. The enhanced interaction between the RuO2 NPs and Nd2O3 can not only facilitate the reduction of RuO2 NPs but also promote the reduction of surface oxygen of the Nd2O3 support. At the same time, the reduction peaks associated with the reduction of surface oxygen of the Nd2O3 support show an obvious decrease towards lower temperature.

Fig. 6
figure 6

TPR profiles of RuO2 colloid, Nd2O3, 1.5% Ru/Nd2O3-m, 3% Ru/Nd2O3-m, 1.5% Ru/Nd2O3-p and 3% Ru/Nd2O3-p

Full size image

DRIFTS of CO adsorption spectroscopy was applied to investigate the metal–support interaction of different catalysts. The DRIFTS spectra of adsorbed CO on the different catalysts, including 3% Ru/Nd2O3-m and 3% Ru/Nd2O3-p samples, are presented in Fig. 7. The 3% Ru/Nd2O3-p sample displays four bands, i.e., 2180, 2115 cm−1, 2039–2062 cm−1 and 1900–2025 cm−1, similar to those reported in previous literatures. Generally, the CO bands originated from CO adsorption on supported Ru catalysts can be divided into three kinds of groups: high frequency 1 (HF1) at 2156–2133 cm−1, high frequency 2 (HF2) at 2100–2060 cm−1 and low frequency (LF) at 2060–1970 cm−1. According to the literature results, the HF2 and LF bands at 2039 and 1965 cm−1 are ascribed to the linearly adsorbed CO and bridge bonded CO on the low-coordinated surface Ru atoms located on metallic Ru NPs [44,45,46,47], while the HF1 band at 2180 and 2115 cm−1 is usually assigned to the multi-carbonyl species on the partially oxidized Ru sites [Ruδ+(CO)x] on the oxide supports [48, 49]. The CO bands originated from CO adsorption on supported Ru NPs in Ru/Nd2O3 samples were significantly influenced by the preparation method. Different from that of the 3% Ru/Nd2O3-p sample, the CO DRIFTS spectra of 3% Ru/Nd2O3-m does not show any adsorption band in the wide region, which should have direct relation with the metal–support interaction in Ru/Nd2O3 catalysts. As the partially oxidized Ru sites [Ruδ+(CO)x] are usually located at the interface of Ru NPs–oxide support, we can reasonably conclude that the interaction between Ru NPs and oxide support can be enhanced in 3% Ru/Nd2O3-p samples.

Fig. 7
figure 7

DRIFTS spectra of CO adsorption on Nd2O3, 3% Ru/Nd2O3-m and 3% Ru/Nd2O3-p

Full size image

NH3 decomposition reaction over Ru-based catalysts is known to be a structure–sensitive reaction. Due to the change of concentration of active B5 sites on the surface of Ru NPs, the size of Ru NPs play crucial role in tuning the catalytic performances of the Ru-based catalysts [50,51,52]. It worthy of mentioning that NH3 synthesis, a rise in TOF over promoted Ru/Al2O3 and AC-supported Ru catalysts has been observed with increase of size of Ru particles [50, 51]. It has been found that NH3 decomposition on supported Ru catalysts is also highly structure sensitive, with TOFH2 values increasing by almost 2 orders of magnitude as the particle size increases from 0.8 to > 7 nm [52]. Here, the different Ru/Nd2O3 catalysts were obtained from the same Nd(OH)3 support precursor and RuO2 colloid NPs. At the same time, TEM images and CO chemisorption results show that Ru NPs were distributed uniformly on the Nd2O3 support with similar particle size in different Ru/Nd2O3 samples. However, it could be seen that the H2 formation rate and TOFNH3 value of Ru/Nd2O3-p samples were significantly higher than that of Ru/Nd2O3-m samples with similar Ru NPs under identical conditions. Thus, the particle size of Ru NPs may not be the key factor in causing the difference of the catalytic performances of Ru/Nd2O3-m and Ru/Nd2O3-p catalysts studied here. The remarkable increase in the catalytic activity and intrinsic activity of Ru/Nd2O3-p catalysts should be highly related with other important key structure factor rather than particle size of Ru NPs.

For oxide supported Ru NPs catalysts, structural factors such as the dispersion, size and morphology of Ru NPs, interaction between Ru NPs and support, and surface property of support can influence the surface structure and chemical state of Ru NPs, leading to a great difference in the catalytic activity. Based on the similar particle size and metal dispersion of Ru NPs and similar composition in the different Ru/Nd2O3 samples, we propose that the remarkable difference of intrinsic activity in Ru/Nd2O3-m and Ru/Nd2O3-p catalysts may arise from the difference in the metal–oxide interaction in different catalysts. In combination with the TEM, XPS, TPR and CO DRIFTS results, we can correlate the observed remarkable different activity of Ru/Nd2O3 catalyst to the difference of metal–support interaction. TEM results reveal that similar Ru NPs can be observed in the different samples. However, substantial difference of chemical structure can be observed over the two samples even with similar Ru NPs. For the Ru/Nd2O3-p with relatively strong metal–oxide interaction, the relative high binding energy value of Ru 3d associated with the presence of Ruδ+ species located at the interface between Ru NPs and Nd2O3 support can be detected. Thus, it is worthwhile to emphasize the critical role of metal–support interaction in modulating the electronic structure and intrinsic activity of Ru NPs in different Ru/Nd2O3 catalysts. Suitable metal–support interaction is believed to be beneficial for improving the catalytic performance of Ru/Nd2O3 catalysts for NH3 decomposition, and much enhanced TOFH2 can be observed even the sample has similar particle size of Ru NPs. We believe that this phenomenon is highly related to the change in the adsorption strength of the reactants and products on the similar Ru NPs in different Ru/Nd2O3 catalysts. It has been generally considered that the associative desorption of nitrogen atoms from the catalyst surface is the rate-determining step in NH3 decomposition reaction. Considering that B5 site on the Ru NPs is the active site for NH3 decomposition, above results confirm that the modulated electron structure of Ru NPs can eventually change activation and dissociation of adsorbed NH3 on the surface of Ru/Nd2O3-p catalysts during NH3 decomposition. At the same time, it can be expected that the interaction between Ru NPs and Nd2O3 support can change in the electronic property of the metallic Ru NPs and weaken strong adsorption of nitrogen atoms and further facilitate associative desorption of nitrogen atoms from the B5 active sites on the Ru NPs in Ru/Nd2O3-p catalysts. The easy desorption of dissociated nitrogen atoms from the B5 active sites on the Ru NPs made the residence time of reactants at active sites shorter and provide more active sites for the decomposition of NH3 molecules, leading to a faster reaction rate in NH3 decomposition reaction.

4 Conclusions

In summary, we have explored the potential of Nd2O3 supported Ru NPs as a new efficient catalyst for NH3 decomposition using Ru/Nd2O3 prepared via precipitation and milling method. We found that the Ru/Nd2O3-p catalysts from precipitation method give much higher activity than that of the Ru/Nd2O3-m analogues from milling method with nearly identical Ru NP in NH3 decomposition. A detailed analysis based on the XRD, TEM, XPS, TPR and CO DRIFTS characterization results reveal that enhancing metal–support interaction can modulate the electronic structure of Ru NPs and efficiently enhance the intrinsic activity of Ru NPs in the Ru/Nd2O3-p catalysts. In view of the excellent activity and robust stability achieved by the Ru/Nd2O3-p catalysts, this work offers promise to improve the catalytic performance of Ru-based catalysts by delicately modulating metal–support interaction.