Synthesis and hydrogenation application of Pt–Pd bimetallic nanocatalysts stabilized by macrocycle-modified dendrimer

Different generations of poly(propylene imine) (Gn-PPI) terminated with N-containing 15-membered triolefinic macrocycle (GnM) (n = 2, 3, 4, 5) were prepared. The bimetallic nanoparticle catalysts GnM-(Ptx/Pd10−x) (x = 0, 3, 5, 7, 10) were prepared by the synchronous ligand-exchange reaction between GnM and the complexes of Pt(PPh3)4 and Pd(PPh3)4. The structure and catalytic properties of GnM-(Ptx/Pd10−x) were characterized via Fourier transform infrared spectroscopy, 1H nuclear magnetic resonance spectroscopy, X-ray diffraction, X-ray photoelectron spectroscopy, high-resolution transmission electron microscopy, energy-dispersive spectroscopy and inductively coupled plasma atomic emission spectroscopy. The novel bimetallic Pd–Pt nanoparticle catalysts stabilized by dendrimers (DSNs) present higher catalytic activities for the hydrogenation of dimeric acid (DA) than that of nitrile butadiene rubber (NBR). It can be concluded that bimetallic Pd–Pt DSNs possess alloying and synergistic electronic effects on account of the hydrogenation degree (HD) of DA and NBR. Furthermore, the HD of DA and NBR shows a remarkable decrease with the incremental generations (n) of GnM-(Pt3/Pd7) (n = 2, 3, 4, 5).


Introduction
For the past few years, bimetallic nanoparticles have attracted widespread attention among the academic community owing to their applications in fine chemicals, nanomedicine and petrochemical technology [1][2][3][4].
As an oil chemical product with low toxicity, extensive sources and renewable materials, dimeric acid (DA) is widely applied to the synthesis of novel polymeric materials.
Hydrogenated dimeric acid (HDA) is produced by unsaturated DA through catalytic hydrogenation reaction. HDA has better thermostability and machinability than unsaturated DA, which is widely used in the synthesis of polyamides, polyesters, mineral oil additives, surface active agents and waterproofing agents [5][6][7].
Compared with nitrile butadiene rubber (NBR), hydrogenated nitrile butadiene rubber (HNBR) possesses physicochemical properties of thermostability, inoxidizability and high oil durability, due to its lower ratio of unsaturated carbon-carbon bonds than NBR [8]. Nowadays, the solution hydrogenation of NBR has been the main technical route for preparing HNBR hydrogenated by homogeneous or heterogeneous catalysts. To contrast the difference between the two types of catalysts, the former possesses higher catalytic activity and leads to higher degree of hydrogenation, which attracts extensive attention for the catalytic hydrogenation of NBR. However, their high cost as well as difficult catalyst removal hinders their application in HNBR production. Accordingly, extensive research and development work has been conducted on novel hydrogenation catalysts for NBR, aimed at obtaining high selectivity, efficient catalytic activity, cost-effectiveness and ease of removal [9].
In comparison to monometallic nanoparticles, bimetallic nanoparticles have unique properties, such as synergistic electric effect, high selectivity, efficient catalytic activity and cost-effectiveness [10][11][12][13]. Among these bimetallic nanoparticles, supported palladium-metal (Pd-M) bimetallic nanoparticle catalysts have attracted considerable attention because of their remarkable activity and high selectivity [14][15][16][17]. It is worth noting that the catalytic activity of Pd-M nanoparticle catalysts could be further enhanced by developing novel preparation strategies, selecting suitable support materials and introducing a second or even third ingredient to form bimetallic or multimetallic nanoparticles [15,18]. The methods of bimetallic nanoparticles synthesized in the template of dendrimer (DSNs) have been extensively studied towards their applications in catalytic hydrogenation. For example, Chung & Rhee [12] prepared Pd-Rh bimetallic nanoparticles in the presence of poly(amidoamine) dendrimers with surface hydroxyl groups for the selective partial hydrogenation of 1,3-cyclooctadiene.
A dendrimer has a controlled molecular weight, well-defined structure, high symmetry of geometric structure, nanoscale of molecular size and surface modification [13]. Therefore, it can act as an ideal template to control the morphology and size of dendrimer-encapsulated metal nanoparticle catalysts. Moreover, the DSNs can be prepared when transferring the catalytic active centres to the surface of the end-functionalized dendrimer [19].
The 15-membered triolefinic macrocycle (MAC), possessing excellent oxidation-resistant property and insolubility of alcohols, is a suitable electron donor for transition metals and easily removed from the reaction system [20]. Therefore, bimetallic DSNs with high stability, easy-removal and favourable catalytic activity and selectivity can be obtained [21], which will be a promising research field for the catalytic hydrogenation of unsaturated compounds.
In this paper, G n M-(Pt x /Pd 10−x ) (n = 2, 3, 4, 5) DSN catalysts were prepared by the synchronous ligand-exchange reaction between precursors (Pt(PPh 3 ) 4 , Pd(PPh 3 ) 4 ) and MAC-terminated G n -PPI (G n M) (n = 2, 3, 4, 5) as the template. The hydrogenation of unsaturated compounds was carried out to evaluate the catalytic ability and selectivity of bimetallic DSN catalysts. It can be concluded that bimetallic Pd-Pt DSNs possess alloying and synergistic electronic effects. Furthermore, the HD of DA and NBR shows a remarkable reduction with the incremental generations (n) of G n M-(Pt 3 /Pd 7 ). the literature [13]. DA (98 wt%, M w = 300.2) was purchased from Fujian Liancheng Baixin Science and Technology Co. Ltd. NBR (N31, ACN = 33.5 wt%, M w = 350 000) was purchased from ZEON Corporation in Japan. All other reagents and solvents were of analytical grade and used without further purification.

Hydrogenation of the dimeric acid and nitrile-butadiene rubber
The hydrogenation of DA was carried out in the presence of bimetallic catalysts in xylene at 180°C under hydrogen (scheme 3). Xylene (30 ml), 1.8 g DA and a certain amount of G 4 M-(Pt x /Pd 10−x ) (catalyst/DA = 0.45 wt%) were added to a 50 ml high-pressure reactor. The hydrogen gas was introduced into the reactor and maintained at 2.0 MPa after the reactor was degassed using hydrogen for six times at room temperature. The reaction system was maintained for 8 h at 180°C with an agitating speed of 600 r.p.m. The system was cooled down. The mixture was evaporated to obtain viscous liquid; then, G n M-(Pt X /Pd 10−X ) was precipitated by adding n-pentane (60 ml). Finally, the precipitate was separated and the filtrate was dried to obtain HDA. The hydrogenation of NBR was carried out in the presence of bimetallic catalysts in xylene at 140°C under hydrogen (scheme 4). Xylene (30 ml), 1.5 g HNBR and a certain amount of G n M-(Pt x /Pd 10−x ) (catalyst/DA = 0.45 wt%) were added to a 50 ml high-pressure reactor. The hydrogen gas was introduced into the reactor and maintained at 4.0 MPa after the mixture was degassed using hydrogen for six times at room temperature. The reaction system was maintained for 8 h at 140°C with an agitating speed of 600 r.p.m. After the completion of the reaction, the system was cooled down. Then HNBR was obtained by unloading the crude product, adding methanol to flocculate the rubber and drying flocculate for 6 h at 90°C.

Characterization
MAC, G n -PPI, G n M and G n M-(Pt x /Pd 10−x ) (n = 2, 3, 4, 5) were characterized by 1 H nuclear magnetic resonance spectroscopy ( 1 H NMR, Bruker Avance III HD, 600 MHz, CDCl 3 ) and Fourier transform   infrared spectroscopy (FTIR, Bruker Vertex 70). The degree of hydrogenation of HDA and NBR was recorded via FTIR. X-ray diffraction (XRD) data were measured using an X-ray diffractometer (D8 Advance, anode material: Cu, step size: 0.013°). Transmission electron microscopy (TEM) was carried out using a JEM-2100F (Jeol, Japan) instrument equipped with an energy-dispersive spectroscopy (EDS) detector. X-ray photoelectron spectroscopy (XPS) was conducted using a Kratos Axis Ultra DLD of British Kratos company. The inductively coupled plasma atomic emission spectroscopy (ICP-OES) data were obtained by a PE Optima8300. The hydrogenation degree (HD) of DA was calculated by iodine number (w) referring to ASTM D 4607-94 [22]: where w is the iodine number of DA and w 0 is the iodine number of HDA. The HD of NBR was obtained from [23]: The peaks at 970, 917 and 2236 cm −1 are ascribed to trans-1,4-ethenyl group (1,4-BD) out-of-plane bending vibration, 1,2-ethenyl group (1,2-BD) out-of-plane bending vibration and nitrile group (C≡N) stretching vibration, respectively, as shown in scheme 4. A is the peak area of the correlative characteristic peaks for HNBR in the FTIR absorption spectra, while A is the peak area of the correlative characteristic peaks for NBR in the FTIR absorption spectra. HD calculated by FTIR is based on the internal standard peak of nitrile group (C≡N; scheme 4).
XPS was used to confirm the coordination of Pd and Pt with G 4 M. Figure 6 shows the XPS spectra which were adjusted referring to the criterion of the most intense carbon (C1s; binding energy = 284.6 eV). As shown in figure 6a, the Pd(3d 3/2 ) and Pd(3d 5/2 ) peaks lie at 340.72 and 337.82 eV which are different from those peaks (340.63 and 336.15 eV) in G 4 -OH-(Pd) [27]. Meanwhile, figure 6b shows the Pt(4f) regions of G 4 M-(Pt 3 /Pd 7 ). The Pt(4f 5/2 ) and Pt(4f 7   which were higher than those peaks (72.86 and 76.22 eV) in the G 6 -OH-(Pt 147 ) [28]. These differences could also be indicative that the Pt-Pd species of G n M-(Pt 3 /Pd 7 ) are in a bimetallic alloy state. Table 3 shows the actual quantitative analysis of Pt/Pd atomic ratios for G 2 M-(Pt 3 /Pd 7 ) and G 4 M-(Pt 3 /Pd 7 ) carried out via ICP-OES. As shown in table 3, the actual Pt/Pd atomic ratios are 0.408 and 0.418 which are close to the experimental ratios (0.43) of G 2 M-(Pt 3 /Pd 7 ) and G 4 M-(Pt 3 /Pd 7 ), respectively. The decrease of Pt/Pd ratios can be due to the reason that the synchronous ligand-exchange reaction of the precursor Pt is much harder than that of the precursor Pd. The representative TEM micrographs of G n -(Pt 3 /Pd 7 ) (n = 2, 3, 4, 5) and G 4 -(Pt 5 /Pd 5 ) are shown in figure 7. The TEM images demonstrate that the particle size is almost uniform and nearly spherical. As shown in figure 7c,c 1 and figure 7d,d 1 , the average diameters of G 4 M-(Pt 5 /Pd 5 ) and G 4 M-(Pt 3 /Pd 7 ) deduced by Gaussian distribution are 4.04 ± 1.15 nm and 3.76 ± 1.30 nm, respectively. Moreover, electron diffraction patterns of G 4 M-(Pt 3 /Pd 7 ) in figure 8a show that the lattice spaces of 0.224 and 0.139 nm are assigned to Pd(1 1 1) and Pd(2 2 0) combined with the analysis of XRD data [25], respectively. While the lattice spaces of 0.195 and 0.118 nm are attributed to Pt(2 0 0) and Pt(3 1 1) combined with the results of XRD analysis [26], respectively. Using XRD and electron diffraction patterns, it is demonstrated that Pt-Pd species in G 4 M-(Pt 3 /Pd 7 ) have an alloy structure. As shown in figure 8b, the characteristic X-ray peaks at 2.846 and 9.401 keV are assignable to Pd(La) and Pt(La) [29] in G 4 M-(Pt 3 /Pd 7 ), respectively. It is also confirmed that G 4 M-(Pt 3 /Pd 7 ) was synthesized successfully.
G n -(Pt 3 /Pd 7 ) (n = 2, 3, 5) are also analysed by TEM in figure 7. The Gaussian-fitting average diameters for G 2 M-(Pt 3 /Pd 7 ), G 3 M-(Pt 3 /Pd 7 ), G 4 M-(Pt 3 /Pd 7 ) and G 5 M-(Pt 3 /Pd 7 ) are 1.74 ± 0.46, 1.96 ± 0.46, 3.76 ± 1.30, and 7.38 ± 1.44 nm, respectively. As shown in figure 8d, it can be indicated that the average sizes of G n -(Pt 3 /Pd 7 ) nanoparticles are positively correlated with the generations of G n -PPI (n = 2, 3, 4, 5). The sizes of G n -(Pt 3 /Pd 7 ) are increased with increasing generations of G n M. G 5 M has 64 terminated-MACs which is twice as many as G 4 -M and quadruple as many as G 3 M and possesses a lot of coordination sites which can coordinate with much more Pt-Pd metal; therefore, G 5 M-(Pt 3 /Pd 7 ) has the largest particle sizes.

Hydrogenation of DA and NBR
To evaluate the catalytic activity and selectivity, G n -(Pt x /Pd 10−x ) catalysts were applied in the hydrogenation of DA. Figure 9 shows the HD of HDA hydrogenated via G 4 -(Pt x /Pd 10−x ) (x = 0, 3, 5, 7, 10) and the physical mixture of G 4 M-Pt and G 4 M-Pd. HDA catalysed by bimetallic DSN catalysts has a higher HD than that catalysed by the physical mixture of monometallic DSN catalysts with equal components of Pd and Pt. The maximum HD is 59.7% catalysed by G 4 -(Pt 3 /Pd 7 ) which is much higher than that of 18.7% catalysed by G 4 M-Pd only. The improved catalytic activity of bimetallic DSNs indicates that there are alloying and bimetallic synergistic electronic effects [30] in the G n -(Pt 3 /Pd 7 ) nanoparticles. A synergetic effect can be caused by a structure effect and chemical effect [31]. Structure synergism indicates the increasing number of new active sites in bimetallic DSNs. It may lead to a greater dispersion of the two metals than in the monometallic catalysts; furthermore, it may have mutual effects between two metals. Chemical synergism means that the bimetallic catalyst contains new sites not present in the monometallic catalysts, which will lead to different activities of the reaction intermediates over the monometallic and bimetallic catalysts.  Figure 9 also shows catalytic hydrogenation effects for DA and NBR using G 4 -(Pt x /Pd 10−x ) (x = 0, 3, 5, 7, 10) catalysts. From the hydrogenated curves in figure 9 for HDA and HNBR, it can be seen that the HD for HDA is higher than that of HNBR with the equal composition of G n -(Pt x /Pd 10−x ). The maximum HD for HDA and HNBR were 59.7% and 48.4%, respectively. As shown in figure 10, the HD for HDA is also higher than that of HNBR catalysed by the same generation of G n -(Pt 3 /Pd 7 ). The maximum HD for HDA and HNBR were 72.4% and 66.8%, respectively. It may be due to the significant difference of molecular mass of DA and NBR: DA is classified as an oligomer (M w = 300.2), while NBR is classified as macromolecular (M w ≈ 350 000). The viscosity of NBR solution is much higher than that of DA. Therefore, it is the reason that the HD of HNBR is lower than that of HDA. Figure 10 presents the HD of NBR and HDA catalysed by G n -(Pt 3 /Pd 7 ) (n = 2, 3, 4, 5). It can be seen that the HD of G n -(Pt 3 /Pd 7 ) can be fitted with a remarkable negative correlation with the incremental generations of G n -(Pt 3 /Pd 7 ). The maximum HD for HDA and HNBR catalysed by G 2 -(Pt 3 /Pd 7 ) are 72.4% and 66.8%, respectively, while the minimum HD for HDA and HNBR catalysed by G 5 -(Pt 3 /Pd 7 ) are 44.2% and 34.2%, respectively. The average particle sizes of G n -(Pt x /Pd 10−x ) were increased with increasing generations of G n M. Nano-catalysts with smaller sizes, having a higher specific surface area, can enhance the dispersity of active sites, which is beneficial to absorb reactants and H 2 on the surface of catalysts [32]. As a consequence, G n -(Pt x /Pd 10−x ) with smallest sizes have the highest catalytic activity for unsaturated compounds.