Selective hydrogenation of cinnamaldehyde catalyzed by Co-doped Pt clusters: a density functional theoretical study

Abstract

Using Pt₆ clusters as model catalysts in this paper, by means of the B3LYP functional in DFT at the 6-31+G(d) level (the LanL2DZ extra basis set was used for Pt atoms) we studied separately the microreaction mechanism of the selective hydrogenation of cinnamaldehyde catalyzed by pure Pt clusters and co-doped Pt clusters. The rationality of the transition state can be proved by vibration frequency analysis and IRC computations. Moreover, atoms-in-molecules (AIM) theory and natural bond orbital (NBO) theory were applied for discussing the interaction among orbitals and the bonding characteristics. The calculating results indicate that Pt₆ clusters are favored for catalyzing the activation and hydrogenation of the C=O bond in cinnamaldehyde molecules, eventually producing cinnamyl alcohol, which proved that Pt₆ clusters have a strong reaction selectivity for catalyzing the hydrogenation of cinnamaldehyde. Compared with Co-doped Pt clusters, Pt₆ clusters are more likely to catalyze the activation and hydrogenation of the C=O bond. Using Co-doped Pt clusters to catalyze the selective hydrogenation reaction of cinnamaldehyde to produce cinnamyl alcohol, the activation energy is lower than that of the reaction catalyzed by pure Pt clusters. The doped catalyst has a synergetic catalytic effect. Our findings have explained the mechanism of action of the doped catalyst and the experimental phenomena.

---

1. Introduction

The selective hydrogenation reactions of cinnamaldehyde play important roles in the fine chemicals and pharmaceutical industries.^1,2 Many experimental studies have reported that metals such as Pd, Pt and Au can be used in catalytic hydrogenation reactions. Different kinds of catalysts exhibit differences in catalytic activity and selectivity. In recent years, much work has shown that bimetallic or multi-metal catalysts possess higher catalytic activity and selectivity than single-metal catalysts.^3–6 Li et al.⁷ added Cr, Mn, Fe, Co, Ni and Cu, respectively, to a Pt catalyst, investigating the effect of the six types of bimetallic catalyst in the selective hydrogenation reaction of cinnamaldehyde. The result showed that the Pt–Ni catalyst had good catalytic activity and high hydrogenation selectivity for the C=C bond. The selectivity and conversion rate for producing 3-phenylpropionaldehyde reached 97.0% and 68.4%, respectively. The Pt–Co catalyst was favored for activating the C=O bond. The selectivity and conversion rate for producing cinnamyl alcohol reached 88.2% and 91.3%, respectively. Moreover, the Pt–Fe and Pt–Cu catalysts also displayed high hydrogenation selectivity for the C=O bond, but their conversion rates for producing cinnamyl alcohol were inferior to that of Pt–Co. Merlo et al.⁸ prepared a Sn–Pt bimetallic catalyst, which was used in the selective hydrogenation reaction of cinnamaldehyde. The result indicated that it not only enhanced the selectivity for cinnamyl alcohol significantly but also increased the reaction rate. Durndell et al.⁹ described the conversion of a batch reaction into continuous flow, affording tunable hydrogenation of C=O versus C=C bonds over a Pt/SiO₂ catalyst, which resulted in high steady-state activity and single-pass yields in the selective hydrogenation of cinnamaldehyde to cinnamyl alcohol under mild conditions. Sun et al.¹⁰ reported the selective hydrogenation of cinnamaldehyde, which was investigated using Pt nanoparticles deposited on reduced graphene oxide. Compared with carbon nanotubes or activated carbon as a support, reduced graphene oxide displayed the best activity and selectivity for the hydrogenation of the C=O bond. Jennings and Johnston ⁹ used density functional theory to discuss the influence on Pt clusters of doping with the transition metals Ti and V. The study found stable structures of Pt_x−yM_y (M = Ti, V; x = 2–6; y = 1, 2) clusters with different spin multiplicities. The result demonstrated that changing the spin multiplicity had a large influence on pure Pt clusters and a small influence on Pt clusters doped with metals. Lu et al.¹² used density functional theory to calculate the stable structures and relative properties of Au_nCs (n = 1–10) bimetallic clusters and pure Au_n (n ≤ 11) clusters. The result showed that the spatial configurations of metal clusters in which n = 4 and n = 6–10 were the most stable. Electrons were transmitted from Cs to Au in Au_nCs bimetallic clusters. Antolini et al.¹³ used Pt_1−xM_x (M = Fe, Co, Ni) to study the stability and alloy composition of doped metals. Xiong and Manthiram^14,15 reported that an alloy that had an ordered structure could further increase the catalytic activity. Xiang et al.¹⁶ synthesized hydrotalcite-supported platinum nanocrystals (Pt NCs) by a facile solution chemistry method and the catalysts displayed high stability with a marginal decrease in activity and selectivity after repeated use. Wang et al.¹⁷ reported that Pt nanocatalysts synthesized by reduction by an alcohol also exhibited high activity in the hydrogenation of cinnamaldehyde. Our group has studied the mechanism of a CuCl₂-catalyzed chlorocyclization reaction and explained the result reasonably.¹⁸ An investigation of the mechanism of the selective hydrogenation reaction of cinnamaldehyde catalyzed by Co-doped Pt₆ clusters has not been reported at present. In order to study the function of the doped metal, in this paper we used density functional theory to discuss the mechanism of the selective hydrogenation reaction of cinnamaldehyde catalyzed by pure Pt clusters and Co-doped Pt clusters. We hoped to understand the influences and characteristics of the catalytic hydrogenation reaction of an aromatic α,β-unsaturated aldehyde catalyzed by a doped metal.

2. Computational details

In this paper, the mechanism of the selective hydrogenation reaction of cinnamaldehyde catalyzed by Co-doped Pt₆ clusters has been studied by the density functional theory. Geometry optimizations, as well as frequency calculations for all the stationary points, were performed at the density functional level of theory using the hybrid B3LYP functional.^19–23 For C, N, O, and H atoms, the 6-31+G(d) basis set^24,25 was used. The LanL2DZ extra basis set was used for Pt atoms.^26,27 For each optimized stationary point, vibrational analysis was used to determine its characteristics (minimum or saddle point) and to calculate the zero-point vibrational energy (ZPVE). For all transition states, intrinsic reaction coordinate (IRC) calculations were performed in both directions to connect the corresponding intermediates.^28,29 To obtain further insight into the nature of the bonding interactions, AIM analysis was performed on some intermediates and transition states using AIM2000.³⁰ The nature of the bonding between atoms can be characterized by the value of the electron density at the bond critical point (BCP) and ring critical point (RCP). The charge distribution and charge transfer direction were analyzed by natural bond orbital (NBO) theory.³¹ NBO analysis provided the second-order perturbation energy of the interaction of orbitals with each other. Referring to the research results of Jennings and others, Pt₆ clusters were selected as the object of the doping study. All calculations were carried out with the Gaussian 09 suite of programs.³²

3. Conclusions and discussions

3.1 Hydrogenation reaction of cinnamaldehyde catalyzed by Pt₆ clusters

Density functional theory (DFT) calculations at the B3LYP/6-31G(d) level of theory predict that the C=C and C=O bond lengths of cinnamaldehyde are 0.1353 nm and 0.1222 nm, respectively. These results are in accordance with reported studies,³³ which show that the bond lengths of C=C and C=O are 0.1384 nm and 0.1225 nm, respectively. These results indicate that the computational method that we chose is credible. The configuration of cinnamaldehyde, which we chose as the reactant, is from the literature as reported by Egawa et al.³⁴

The selective hydrogenation reaction of cinnamaldehyde catalyzed by Pt₆ clusters may give rise to three different products, which are P1 (3-phenylpropionaldehyde), P2 (3-phenylallyl alcohol) and P3 (cinnamyl alcohol), respectively. Each product is obtained via two different reaction channels (Fig. 1). Pt₆ clusters take part in a compound reaction with C=O and C=C bonds individually and generate intermediates M1 and M2 and then these are combined with H₂ to produce intermediates M3 and M4, respectively. In reaction pathway I_A, intermediate M3 can generate P3 by the route of T1-3 → M5-3 → T2-6 → M6-6. In reaction pathway I_B, intermediate M3 can generate M6-5 by the route of T1-4 → M5-4 → T2-5 and can eventually produce P3. In pathway I_C, intermediate M3 can generate P2 by the reaction route of T1-4 → M5-4 → T2-4 → M6-4. In reaction pathway I_D, intermediate M4 can generate P2 by the route of T1-1 → M5-1 → T2-3 → M6-3. In pathway I_E, intermediate M4 generates P1 via the route of T1-1 → M5-1 → T2-2 → M6-2. In pathway I_F, intermediate M4 generates P1 via the route of T1-2 → M5-2 → T2-1 → M6-1. The optimized configurations and structural parameters of all reactants in this reaction process are shown in Fig. 2. The energy values that are related to this reaction are listed in Table 1. A diagram of the reaction energy levels can be seen in Fig. 3.

Fig. 1 Reaction mechanism of hydrogenation of cinnamaldehyde catalyzed by Pt₆ clusters.

Fig. 2 Geometric parameters of the compounds in the selective hydrogenation reaction of cinnamaldehyde catalyzed by Pt₆ clusters (outside the brackets are the bond lengths (nm) and inside the brackets are the electric charge densities of the bond saddle points (a.u.)).

Table 1 Energy E (a.u.), relative energy E_rel (kcal mol⁻¹), and vibrational frequency v (cm⁻¹) of the compounds

Species		E (a.u.)	Erel (kcal mol^−1)	v (cm^−1)	Species		E (a.u.)	Erel (kcal mol^−1)	v (cm^−1)
IA	R(Re + H2 + Pt6)	−1138.937674	0.00		ID	M2 + H2	−1138.995805	−36.46
	M1 + H2	−1138.971194	−21.03			M4	−1139.035945	−61.64
	M3	−1139.011696	−46.43			T1-1	−1138.985911	−30.26	697.7
	T1-3	−1138.980434	−26.82	460.3		M5-1	−1139.006431	−43.13
	M5-3	−1138.995931	−36.54			T2-3	−1138.974656	−23.20	1027.7
	T2-6	−1138.941541	−2.43	984.6		M6-3	−1139.028722	−57.11
	M6-6	−1138.967848	−18.93		IE	T2-2	−1138.990520	−33.15	860.2
	P3 + Pt6	−1138.944614	−4.35			M6-2	−1138.999914	−39.04
IB	T1-4	−1138.995284	−36.14	1044.2		P1 + Pt6	−1138.965815	−17.65
	M5-4	−1139.020981	−52.26		IF	T1-2	−1138.987451	−31.22	542.5
	T2-5	−1138.986006	−30.32	868.2		M5-2	−1139.009184	−44.86
	M6-5	−1139.015095	−48.56			T2-1	−1138.956000	−11.50	1184.9
IC	T2-4	−1138.970759	−20.75	616.1		M6-1	−1139.012962	−47.23
	M6-4	−1139.002199	−40.48
	P2 + Pt6	−1138.946554	−5.57

---

Fig. 3 Energy diagram of the selective hydrogenation reaction of cinnamaldehyde catalyzed by Pt₆ clusters.

Firstly, we confirm the catalytic configuration of Pt₆ clusters, as shown in Fig. 2. The bond length of Pt(1)–Pt(2) in Pt₆ clusters is 0.2704 nm. The first step is the complexation of Pt₆ clusters with C=O and C=C bonds in cinnamaldehyde to produce intermediates M1 and M2, respectively, which leads to the bond lengths of Pt(1)–Pt(2) and Pt(2)–Pt(3) increasing by 0.0135 nm and 0.0063 nm, respectively. Compared with cinnamaldehyde, the bond length of C(4)=O in intermediate M1 increases to 0.1310 nm and the electric charge density of the BCP decreases to 0.3334 a.u. The bond length of C(2)=C(3) in intermediate M2 increases to 0.1434 nm and the electric charge density of the BCP decreases to 0.2892 a.u. In intermediate M1, the bond lengths of Pt(1)–O and Pt(1)–C(4) are 0.2084 nm and 0.2192 nm, respectively. The bond angle of O–Pt(1)–Pt(2) is 145.6° and that of C(4)–Pt(1)–Pt(2) is 125.1°. The Milliken charge of Pt(1) is 0.5226 a.u. and that of Pt(2) is −0.0532 a.u. The above configuration data shows that after Pt₆ clusters are combined with cinnamaldehyde, both C(4)=O and C(2)=C(3) double bonds are weakened. The energy analysis data in Table 1 shows that in the course of Re + Pt₆ → M1 the energy is reduced by 21.03 kcal mol⁻¹ and in the course of Re + Pt₆ → M2 the energy is reduced by 36.46 kcal mol⁻¹. This shows that intermediates M1 and M2 are easy to form and can exist stably. By NBO computational analysis, the E(2) values of LP*(6)Pt(1) → BD*(2)C(4)–O and BD*(2)C(4)–O → LP*(6)Pt(1) in intermediate M1 are 7.15 kcal mol⁻¹ and 48.47 kcal mol⁻¹, respectively, and are 62.96 kcal mol⁻¹ and 13.65 kcal mol⁻¹, respectively, for LP(5)Pt(1) → BD*(2)C(2)–C(3) and BD(1)C(2)–C(3) → LP*(7)Pt(1) in M2, which shows the strong interaction between Pt(1) and C(4)–O and between Pt(1) and C(2)–C(3). Then M1 and M2 are combined with H₂ separately, which contributes to the formation of intermediates M3 and M4. In M3, the bond lengths of Pt(2)–H(4) and C(4)=O are 0.1569 nm and 0.1306 nm, respectively. In M4, the bond lengths of Pt(2)–H(4) and C(2)=C(3) are 0.1572 nm and 0.1432 nm, respectively. Table 1 indicates that after M1 and M2 are combined with H₂, the energy of the system further decreases, which shows that intermediates M1 and M2 are easy to form and can exist stably.

In reaction pathway I_A, intermediate M3 can generate P3 by the route of T1-3 → M5-3 → T2-6 → M6-6. Pt(1)–Pt(2)–H(4)–C(4)–O in transition state T1-3 can form a five-ring structure, and the electric charge density of the RCP is 0.0196 a.u. The bond lengths of Pt(2)–H(4) and C(4)–O increase to 0.1730 nm and 0.1323 nm, respectively. The electric charge densities of the BCP decrease to 0.0964 a.u. and 0.3213 a.u., respectively. The bond length of C(4)–H(4) is 0.1377 nm, and the electric charge density of the BCP is 0.1273 a.u. In intermediate M5-3, the bond lengths of C(4)–O and C(4)–H(4) are 0.1432 nm and 0.1099 nm, respectively. The electric charge densities of the BCP are 0.2474 a.u. and 0.2763 a.u., respectively. The bond length of Pt(2)–H(5) is 0.1573 nm. From the above configuration data we can conclude that the Pt(2)–H(4) bond is being broken and the C(4)–H(4) bond is being formed in the course of M3 → T1-3 → M5-3, which completes the first step of hydrogenation. In transition state T2-6, the bond lengths of O–H(5) and C(4)–O are 0.1468 nm and 0.1441 nm, respectively. The bond length of Pt(2)–H(5) increases by 0.0138 nm compared with that in intermediate M5-3. The bond length of O–H(5) in M6-6 is 0.0972 nm, and the electric charge density of the BCP is 0.3408 a.u. The bond length of C(4)–O is 0.1471 nm and the electric charge density of the BCP is 0.2224 a.u. As is shown by the above data, we can conclude that the Pt(2)–H(5) bond is breaking gradually and the O–H(5) bond is being formed gradually in the course of M5-3 → T2-6 → M6-6. From Table 1 we can see that the activation energy needed for this reaction, which is a rate-determining step of reaction pathway I_A, is 34.12 kcal mol⁻¹. The product of the hydrogenation of the C=O bond finally dissociates from the Pt₆ clusters to produce P3.

In reaction pathway I_B, intermediate M3 can generate M6-5 by the route of T1-4 → M5-4 → T2-5 and can eventually produce P3. In transition state T1-4, Pt(1)–Pt(2)–H(4)–O–C(4) can form a five-ring structure, and the electric charge density of the RCP is 0.0089 a.u. The bond lengths of Pt(2)–H(4), C(4)–O and O–H(4) are 0.2090 nm, 0.1470 nm and 0.1576 nm, respectively. In intermediate M5-4, the bond length of C(4)–O is 0.1370 nm and the electric charge density of the RCP is 0.2854 a.u. The bond length of O–H(4) is 0.0972 nm and the electric charge density of the RCP is 0.3418 a.u. The bond length of Pt(2)–H(5) is 0.1583 nm. According to the preceding data, intermediate M3 can generate intermediate M5-4 via transition state T1-4, in which the O–H(4) bond is being formed gradually and the Pt(2)–H(4) bond is breaking gradually with the weakening of the O=C(4) bond. In transition state T2-5, the bond lengths of Pt(2)–H(5) and C(4)–H(5) increase to 0.1704 nm and 0.1559 nm, respectively, whereas the electric charge density of the RCP is 0.0934 a.u. In intermediate M6-5, the bond lengths of C(4)–H(5) and C(4)–O are 0.1092 nm and 0.1412 nm, respectively, whereas the electric charge densities of the BCP are 0.2297 a.u. and 0.2623 a.u., respectively. The bond length of Pt(1)–C(4) increases to 0.3060 nm. According to the preceding date, we can deduce that the C(4)–H(5) bond is formed and the Pt(2)–H(5) bond is broken in the course of M5-4 → T2-5 → M6-5, which finally generates P3 after dissociation. From the energy statistics in Table 1 we can see that the activation energy needed in this process, which is the rate-determining step in reaction pathway I_B, is 21.94 kcal mol⁻¹.

In pathway I_C, the reaction process of M3 → T1-4 → M5-4 is similar to that in I_B. Intermediate M5-4 forms M6-4 via transition state T2-4, which eventually generates the product P2. In transition state T2-4, the bond length of C(2)–H(5) is 0.1721 nm and the electric charge density of the BCP is 0.0592 a.u. The bond lengths of Pt(2)–H(5) and C(2)–C(3) are 0.1630 nm and 0.1435 nm, respectively, and the electric charge densities of the BCP are 0.1226 a.u. and 0.0871 a.u., respectively, whereas the bond length of C(3)–C(4) decreases to 0.1448 nm. In intermediate M6-4, the bond lengths of C(3)–C(4), C(2)–C(3) and C(2)–H(5) are 0.1422 nm, 0.1525 nm and 0.1096 nm, respectively, whereas the corresponding electric charge densities of the BCP are 0.3019 a.u., 0.2445 a.u. and 0.2729 a.u., respectively. From the preceding configuration data we can conclude that in the course of M5-4 → T2-4 → M6-4, the Pt(2)–H(5) bond is breaking gradually and the C(4)–H(5) bond is being formed gradually. Meanwhile, the C(2)–C(3) bond is being formed. As can be seen in Table 1, the activation energy for the process of M5-4 → T2-4 → M6-4, which is the rate-determining step in I_C, is 31.50 kcal mol⁻¹.

In reaction pathway I_D, intermediate M4 can generate P2 by the route of T1-1 → M5-1 → T2-3 → M6-3. Pt(1)–Pt(2)–H(4)–C(2)–C(3) forms a five-ring structure in transition state T1-1, where the electric charge density of the BCP is 0.0183 a.u. The bond lengths of Pt(2)–H(4) and C(2)–H(4) are 0.1630 nm and 0.1650 nm, respectively, whereas the electric charge densities of the BCP are 0.1250 a.u. and 0.0680 a.u., respectively. Compared with M4, the bond length of C(2)–C(3) increases by 0.0017 nm, whereas the electric charge density of the BCP decreases correspondingly by 0.0070 a.u. In intermediate M5-1, the bond lengths of C(2)–H(4) and C(2)–C(3) are 0.1097 nm and 0.1537 nm, respectively, whereas the electric charge densities of the BCP are 0.2742 a.u. and 0.2488 a.u., respectively. The bond length of Pt(2)–H(5) is 0.1579 nm. The preceding configuration data suggest that the Pt(2)–H(4) bond is breaking gradually and the C(2)–H(4) bond is being formed step by step in the course of M4 → T1-1 → M5-1. The activation energy needed in this reaction process, which is the rate-determining step in pathway I_D, as shown in Table 1, is 31.38 kcal mol⁻¹. Intermediate M5-1 forms intermediate M6-3 via transition state T2-3. In transition state T2-3, the bond length of O–H(5) is 0.1359 nm and the electric charge density of the BCP is 0.1149 a.u. The bond lengths of C(4)–O and C(3)–C(4) are 0.1275 nm and 0.1448 nm, respectively, whereas the electric charge densities of the BCP are 0.3538 a.u. and 0.2899 a.u., respectively. Compared with intermediate M5-1, the bond length of Pt(2)–H(5) increases by 0.0130 nm. In intermediate M6-3, the bond length of O–H(5) is 0.0972 nm and the electric charge density of the BCP is 0.3389 a.u. The bond lengths of C(3)–C(4) and C(4)–O are 0.1430 nm and 0.1383 nm, respectively, whereas the electric charge densities of the BCP are 0.3477 a.u. and 0.2592 a.u., respectively. The bond length of C(2)–C(3) increases by 0.1532 nm, whereas the electric charge density of the BCP decreases by 0.2510 a.u. From the preceding configuration data we can conclude that the Pt(2)–H(5) bond is broken and the O–H(5) bond is formed in the course of M5-1 → T2-3 → M6-3, which indicates that the 1,4-hydrogenation process of cinnamaldehyde has been finished. Finally, the 1,4-hydrogenation product dissociates from the Pt₆ clusters to generate the product P2.

In pathway I_E, the process of M4 → T1-1 → M5-1 is the same as that of I_D, whereas intermediate M5-1 can generate intermediate M6-2 via transition state T2-2. In transition state T2-2, the bond lengths of C(3)–H(5) and C(2)–C(3) are 0.1470 nm and 0.1548 nm, respectively, whereas the electric charge densities of the BCP are 0.1149 a.u. and 0.2346 a.u., respectively. The bond length of Pt(2)–H(5) increases to 0.1622 nm. In intermediate M6-2, the bond length of C(3)–H(5) is 0.1097 nm and the electric charge density of the BCP is 0.2708 a.u. The bond length of C(2)–C(3) is 0.1554 nm and the electric charge density of the BCP is 0.2374 a.u. From the preceding configuration data we can deduce that the Pt(2)–H(5) bond is breaking step by step and the O–H(5) bond is being formed gradually in the course of M5-1 → T2-2 → M6-2. The hydrogenation product of the C=C bond dissociates from the Pt₆ clusters and eventually produces P1. As is shown in Table 1, the activation energy needed in the process of M4 → T1-1 → M5-1, which is the rate-determining step in pathway I_E, is 31.38 kcal mol⁻¹.

In pathway I_F, intermediate M4 generates the product P2 via transition states T1-2 and T2-1 before and after intermediate M5-2. In transition state T1-2, Pt(1)–Pt(2)–H(4)–C(3)–C(2) form a 5-ring structure, whereas the electric charge density of the BCP is 0.0162 a.u. The bond length of Pt(2)–H(4) is 0.1640 nm and the electric charge density of the BCP is 0.1268 a.u. The bond lengths of C(3)–H(4) and C(2)–H(2) are 0.1604 nm and 0.1428 nm, respectively, whereas the electric charge densities of the BCP are 0.0769 a.u. and 0.2934 a.u., respectively. In intermediate M5-2, the bond lengths of C(3)–H(4) and C(2)–C(3) are 0.1099 nm and 0.1524 nm, respectively, whereas the electric charge densities of the BCP are 0.2700 a.u. and 0.2441 a.u., respectively. The bond length of Pt(2)–H(5) is 0.1579 nm. The Pt(2)–H(4) bond is breaking gradually and the C(3)–H(4) bond is being formed step by step in the course of M4 → T1-2 → M5-2. Intermediate M5-2 generates intermediate M6-1 via transition state T2-1. In transition state T2-1, the bond lengths of C(2)–H(5) and C(2)–C(3) are 0.1502 nm and 0.1525 nm, respectively, whereas the electric charge densities of the BCP are 0.1044 a.u. and 0.2464 a.u., respectively. The bond length of Pt(2)–H(5) increases to 0.2010 nm. In intermediate M6-1, the bond lengths of C(2)–H(5) and C(2)–C(3) are 0.1098 nm and 0.1552 nm, respectively, whereas the electric charge densities of the BCP are 0.2715 a.u. and 0.2315 a.u., respectively. The bond length of Pt(1)–C(2) increases to 0.3106 nm. The Pt(2)–H(5) bond is broken and the C(2)–H(5) bond is formed gradually in the course of M5-2 → T2-1 → M6-1. According to the data in Table 1, as a rate-determining step, the activation energy needed in this process is 33.36 kcal mol⁻¹.

As illustrated in Fig. 3, the C=O and C=C bonds in cinnamaldehyde can generate M1 and M2 by activated complexation with Pt₆ clusters separately and their energies decrease by 21.03 kcal mol⁻¹ and 36.46 kcal mol⁻¹, respectively, afterwards. Then, M1 and M2 can generate M3 and M4, respectively, after combining with H₂, in which the total energy in the system decreases further. From M3 to P3 via pathways I_A and I_B, the rate-determining steps are M5-3 → T2-6 → M6-6 and M5-4 → T2-5 → M6-5, respectively, and the activation energies are 34.12 kcal mol⁻¹ and 21.94 kcal mol⁻¹, accordingly, in the two pathways. By comparison, I_B is the optimal reaction pathway for generating P3. There are also two pathways for obtaining P2, namely, I_C and I_D. In pathway I_C, the process of M3 → T1-4 → M5-4 is the same as in I_B and the rate-determining step is M5-4 → T2-4 → M6-4, of which the activation energy is 31.50 kcal mol⁻¹. In pathway I_D, the rate-determining step is M4 → T1-1 → M5-1, of which the activation energy is 31.38 kcal mol⁻¹. I_D is the optimal reaction pathway for generating P2. Intermediate M4 can produce P1 by means of pathways I_E and I_F: the rate-determining step, which is the same as in I_D, is M4 → T1-1 → M6-2 in pathway I_E. The rate-determining step in pathway I_F is M5-2 → T2-1 → M6-1, of which the activation energy is 33.36 kcal mol⁻¹; by comparison, we can deduce that I_E is the optimal reaction pathway for generating P1. By comparing the activation energies of the different rate-determining steps of the various reaction pathways, we find that the activation energies of the rate-determining steps for obtaining P3, P2 and P1 by their optimal reaction pathways follow the order P1 > P2 > P3; that is to say, the activation energy for hydrogenating the C=C bond is the highest,^35–38 and it is therefore the most difficult to generate P1. The activation energy of P3 is the lowest,^39–51 which means that it is easiest to hydrogenate the C=O bond.

3.2 Hydrogenation reaction of cinnamaldehyde catalyzed by Co-doped Pt₅ clusters

The selective hydrogenation reaction of cinnamaldehyde catalyzed by Co–Pt₅ clusters can generate three different products, namely, P3, P2 and P1. Co–Pt₅ clusters take part in compound reactions with C=O and C=C bonds individually and generate intermediates M1′ and M2′ and then these are combined with H₂ to produce intermediates M3′ and M4′, respectively. The product P3 can be obtained via pathways I_A′ and I_B′. The product P2 can be produced via pathways I_C′ and I_D′. The product P1 can be obtained via pathways I_E′ and I_F′. The optimized configurations and structural parameters of all reactants in this reaction process are shown in Fig. 4. The energy values that are related to this reaction are listed in Table 2. A diagram of the reaction energy levels can be seen in Fig. 5.

Fig. 4 Geometric parameters of the compounds in the selective hydrogenation reaction of cinnamaldehyde catalyzed by Co–Pt₅ clusters (outside the brackets are the bond lengths (nm) and inside the brackets are the electric charge densities of the bond saddle points (a.u.)).

Table 2 Energy E (a.u.), relative energy E_rel (kcal mol⁻¹) and vibrational frequency v (cm⁻¹) of the compounds in the hydrogenation reaction of cinnamaldehyde catalyzed by Co–Pt₅ clusters

Species		E	Erel	v	Species		E	Erel	v
IA^′	R(Re + H2 + Co–Pt5)	−1164.912563	0.00		ID^′	M2′ + H2	−1164.987031	−46.71
	M1′ + H2	−1164.964975	−32.88			M4′	−1165.010320	−61.32
	M3′	−1164.987262	−46.86			T1-1′	−1164.936846	−15.23	744.9
	T1-3′	−1164.956073	−27.29	184.6		M5-1′	−1164.990795	−49.07
	M5-3′	−1164.965390	−33.14			T2-3′	−1164.958858	−29.04	822.1
	T2-6′	−1164.951813	−24.62	136.7		M6-3′	−1164.996187	−52.45
	M6-6′	−1164.974701	−38.98		IE^′	T2-2′	−1164.966387	−33.76	765.7
	P3 + Co–Pt5	−1164.919503	−4.35			M6-2′	−1164.994931	−51.67
IB^′	T1-4′	−1164.982455	−43.84	762.3		P1 + Co–Pt5	−1164.940704	−17.65
	M5-4′	−1164.996427	−25.61		IF^′	T1-2′	−1164.939165	−16.69	757.8
	T2-5′	−1164.948732	−22.67	819.9		M5-2′	−1164.992892	−50.39
	M6-5′	−1164.997042	−52.99			T2-1′	−1164.959341	−29.34	808.2
IC^′	T2-4′	−1164.942252	−18.62	697.5		M6-1′	−1164.994878	−51.63
	M6-4′	−1164.991701	−49.64
	P2 + Co–Pt5	−1164.921443	−5.57

---

Fig. 5 Energy diagram of selective hydrogenation reaction of cinnamaldehyde catalyzed by Co–Pt₅ clusters.

The first step is the complexation of Co–Pt₅ clusters with C=O and C=C bonds in cinnamaldehyde to produce intermediates M1′ and M2′, respectively. The bond lengths of Pt(1)–O and Pt(1)–C(4) are 0.2100 nm and 0.2203 nm, respectively. The bond angles of O–Pt(1)–Co and C(4)–Pt(1)–Co are 115.8° and 111.9°, respectively. The Milliken charge of Pt(1) is 0.11143 a.u. and that of Co is 2.6674 a.u. In M1′, both the Milliken charge and the bond angles of Pt(1) decrease, which shows that there are charge transfers between Co and Pt and the presence of a synergistic catalytic effect. In M2′, both the Milliken charge and the bond angles of Pt(1) decrease, which shows that there are also charge transfers between Co and Pt and the presence of a synergistic catalytic effect. The energies in the course of Re → Co–Pt₅ → M1′ and Re → Co–Pt₅ → M2′ decrease by 32.88 kcal mol⁻¹ and 46.71 kcal mol⁻¹, respectively, which shows that intermediates M1′ and M2′ are easy to form and can exist stably. From Table 2 we know that after M1′ and M2′ combine with H₂ the energy of the reaction system further decreases, which shows that intermediates M3′ and M4′ are easy to form and can exist stably.

In reaction pathway I_A′, intermediate M3′ can generate P3 by the route of T1-3′ → M5-3′ → T2-6′ → M6-6′. The activation energy needed in the process of M3′ → T1-3′ → M5-3′, which is the rate-determining step in reaction pathway I_A′, is 19.57 kcal mol⁻¹. In reaction pathway I_B′, intermediate M3′ can generate P3 by the route of T1-4′ → M5-4′ → T2-5′ → M6-5′. The activation energy in the process of M5-4′ → T2-5′ → M6-5′, which is the rate-determining step in pathway I_B′, is 29.92 kcal mol⁻¹.

In reaction pathway I_C′, the process of M3′ → T1-4′ → M5-4′ is the same as in I_B′. However, intermediate M5-4′ forms intermediate M6-4′ via transition state T2-4′ and then generates the product P2. The process of M5-4′ → T2-4′ → M6-4′ is the rate-determining step in pathway I_C′, of which the activation energy is 33.98 kcal mol⁻¹. In reaction pathway I_D′, the product P2 can be obtained via the route of M4′ → T1-1′ → T2-3′ → M6-3′. The activation energy needed in the process of M4′ → T1-1′ → M5-1′, which is the rate-determining step in pathway I_D′, is 46.09 kcal mol⁻¹.

In pathway I_E′, the process of M4′ → T1-1′ → M5-1′ is the same as in I_D′. Intermediate M5-1′ generates intermediate M6-2′ via transition state T2-2′. The activation energy needed in the process of M4′ → T1-1′ → M5-1′, which is the rate-determining step in pathway I_E′, is 46.09 kcal mol⁻¹. Finally, the hydrogenation product of the C=C bond dissociates from the Co–Pt₅ clusters to generate the product P1. In pathway I_F′, intermediate M4′ generates P1 via the route of T1-2′ → M5-2′ → T2-1′ → M6-1′. The activation energy needed in this reaction process, which is the rate-determining step in pathway I_F′, is 44.63 kcal mol⁻¹. Intermediate M5-2′ generates intermediate M6-1′ via transition state T2-1′, which dissociates from the catalyst to generate P1.

As illustrated in Fig. 5, the C=O and C=C bonds in cinnamaldehyde can generate M1′ and M2′ by activated complexation with Co–Pt₅ clusters separately and their energies decrease afterwards. Then M1′ and M2′ can generate M3′ and M4′, respectively, after combining with H₂, in which the total energy in the system decreases further. From M3′ to P3 via pathways I_A′ and I_B′, the rate-determining steps are M3′ → T1-3′ → M5-3′ and M5-4′ → T2-5′ → M6-5′, respectively, and the activation energies are 19.57 kcal mol⁻¹ and 29.92 kcal mol⁻¹, accordingly, in the two pathways. The activation energy of pathway I_A′ is lower than that of I_B′. Therefore, I_A′ is the optimal reaction pathway for generating P3. P2 can be generated via pathways I_C′ and I_D′. The rate-determining steps are M5-4′ → T2-4′ → M6-4′ and M4′ → T1-1′ → M5-1′, respectively, and the activation energies are 33.98 kcal mol⁻¹ and 46.09 kcal mol⁻¹, accordingly, in the two pathways. Hence I_C′ is the optimal reaction pathway for generating P2. Product P1 can be obtained via pathways I_E′ and I_F′. The rate-determining step of I_E′ is the same as that of I_D′ and the activation energy is 46.09 kcal mol⁻¹. The rate-determining step of I_F′ is M4′ → T1-2′ → M5-2′ and the activation energy is 44.63 kcal mol⁻¹. Hence I_F′ is the optimal reaction pathway for generating P1. In comparison with the result of the study of the selective hydrogenation reaction of cinnamaldehyde catalyzed by Pt₆, we find that the activation energy of the rate-determining step for generating P3 is somewhat lower (the activation energy of the rate-determining step in pathway I_B is 20.19 kcal mol⁻¹), whereas the activation energies of the rate-determining steps for generating P2 and P1 both increase. Products P1 and P2 are difficult to produce under the thermodynamic conditions, which shows that Co–Pt₅ clusters are beneficial for the selective hydrogenation reaction of the C=O bond in cinnamaldehyde. In the first step, the C=O and C=C bonds undergo complexation with Co–Pt₅ clusters to generate intermediates M1′ and M2′, respectively. In intermediate M1′, the bond angles of O–Pt(1)–Co and C(4)–Pt(1)–Co are 115.8° and 111.9°, respectively. The Milliken charges of Pt(1) and Co are 0.11143 a.u. and 2.6674 a.u., respectively. In M1, both the Milliken charge and the bond angle of Pt(1) decrease, which shows that there is charge transfer between Co and Pt in the complexion process of cinnamaldehyde with Co–Pt₅ clusters. All the evidence proves that Co–Pt₅ clusters have a synergistic catalytic effect in the selective hydrogenation reaction of cinnamaldehyde, which is in accordance with experimental studies.⁵²

4. Conclusions

In this paper, we studied the reaction mechanism of the selective hydrogenation of cinnamaldehyde catalyzed by Co-doped Pt clusters. The following are the main conclusions:

Firstly, the selective hydrogenation reaction of cinnamaldehyde catalyzed by Pt₆ clusters can produce three different products, which are P1 (3-phenylpropionaldehyde), P2 (3-phenylallyl alcohol) and P3 (cinnamyl alcohol), respectively. The activation energy needed in the rate-determining steps follows the order P1 > P2 > P3. Each product can be generated by two different pathways. Different clusters only change the catalytic activation and do not change the reaction mechanism. The activation energy needed to hydrogenate the C=C bond in cinnamaldehyde is the highest.^35–38 The activation energy needed to hydrogenate the C=O bond is the lowest.^39–51

Secondly, after being mixed with Co, Pt clusters are more favorable for catalyzing the hydrogenation reaction of the C=O bond. The pathway of the mechanism catalyzed by Co–Pt₅ clusters is the same as that for Pt clusters without Co. Compared with the catalytic effect of Pt₆ clusters, the activation energy needed to hydrogenate the C=O bond in cinnamaldehyde when catalyzed by Co–Pt₅ clusters is lower.^39–51 There is a synergistic catalytic effect between Co and Pt in the selective hydrogenation reaction of cinnamaldehyde catalyzed by Pt₅ clusters mixed with Co.

Acknowledgements

We are grateful for the financial support of this work by the Natural Science Foundation of Sichuan province (2014JY0099).

References

1. B. B. Bachiller, R. I. RódrıGuez and R. A. Guerrero, Appl. Catal., A, 2001, 205, 227–237.
2. J. M. Nhut, R. Vieira, L. Pesant, J. P. Tessonnier, N. Keller, G. Ehret, P. H. Cuong and M. J. Ledoux, Catal. Today, 2002, 76, 11–32.
3. L. E. Murillo and J. G. Chen, Surf. Sci., 2008, 602, 919–931.
4. Z. Rong, Z. Sun, Y. Wang, J. Lv and Y. Wang, Catal. Lett., 2014, 144, 980–986.
5. A. Fortunelli and A. M. Velasco, J. Mol. Struct.: THEOCHEM, 2002, 586, 17–27.
6. W. J. Yan, Z. Guo, X. L. Jia, V. Kariwala, T. Chen and Y. H. Yang, Chem. Eng. Sci., 2012, 76, 26–36.
7. Y. Li, R. X. Zhou and G. H. Lai, React. Kinet. Catal. Lett., 2006, 88, 105–110.
8. A. B. Merlo, B. F. Machado, V. Vetere, J. L. Faria and M. L. Casella, Appl. Catal., A, 2010, 383, 43–49.
9. L. J. Durndell, K. Wilson and A. F. Lee, RSC Adv., 2015, 5, 80022–80026.
10. Z. Sun, Z. Rong, Y. Wang, Y. Xia, W. Du and Y. Wang, RSC Adv., 2013, 4, 1874–1878.
11. P. C. Jennings and R. L. Johnston, Int. J. Comput. Theor. Chem., 2013, 1021, 91–100.
12. C. Lu, X. Y. Kuang, Z. W. Lu, A. J. Mao and Y. M. Ma, J. Phys. Chem. A, 2011, 115, 9273–9281.
13. E. Antolini, J. R. C. Salgado and E. R. Gonzalez, J. Power Sources, 2006, 160, 957–968.
14. L. Xiong and A. Manthiram, J. Electrochem. Soc., 2005, 152, A697–A703.
15. L. Xiong and A. Manthiram, J. Mater. Chem., 2004, 14, 1454–1460.
16. X. Xiang, W. He and L. Xie, Catal. Sci. Technol., 2013, 3, 2819–2827.
17. Y. Wang, W. He and L. Wang, Chem.–Asian J., 2015, 10, 1561–1570.
18. G. Wang, W. F. Cai, L. C. Li and A. M. Tian, Sci. China: Chem., 2013, 2, 185–197.
19. A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652.
20. C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 37, 785–789.
21. B. Miehlich, A. Savin and H. Stoll, Chem. Phys. Lett., 1989, 157, 200–206.
22. C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 7, 785–789.
23. P. J. Stephens, F. J. Devlin, M. J. Frisch and C. F. Chabalowski, J. Phys. Chem., 1994, 98, 11623–11627.
24. J. P. Bhakthi and P. G. Chathurika, Can. J. Chem., 2014, 92, 809–813.
25. M. Muzomwe, G. Maes and O. E. Kasende, Nat. Sci., 2012, 4, 286–297.
26. P. J. Hay and W. R. Wadt, J. Chem. Phys., 1985, 82, 270–283.
27. H. Teramae, Y. Ito and U. Nagashima, J. Comput. Chem., Jpn., 2013, 12, 133–137.
28. C. Gonzalez and H. B. Schlegel, J. Chem. Phys., 1989, 90, 2154–2161.
29. C. Gonzalez and H. B. Schlegel, J. Phys. Chem., 1990, 94, 5523–5527.
30. R. W. F. Bader, Atoms in Molecules: A Quantum Theory, Oxford University Press, 1990.
31. A. E. Reed, L. A. Curtiss and F. Weinhold, Chem. Rev., 1988, 88, 899–926.
32. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. J. A. Montgomery, J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, Ö. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox, Gaussian 09, Revision A.02, Gaussian, Inc., Wallingford CT, 2009.
33. H. Noller and W. M. Lin, J. Catal., 1984, 85, 25–30.
34. T. Egawa, M. Rui and D. Yamamoto, J. Mol. Struct., 2008, 892, 158–162.
35. P. Gallezot and D. Richard, Catal. Rev.: Sci. Eng., 1998, 40, 81–126.
36. P. Kluson and L. Cerveny, Appl. Catal., A, 1995, 128, 13–31.
37. V. Ponec, Appl. Catal., A, 1997, 149, 27–48.
38. U. K. Singh and M. A. Vannice, Appl. Catal., A, 2001, 213, 1–24.
39. N. Mahata, F. Gonçalves, M. F. R. Pereira and J. L. Figueiredo, Appl. Catal., A, 2008, 339, 159–168.
40. G. Neri, I. Arrigo, F. Corigliano, L. D. Luca and A. Donato, Catal. Lett., 2011, 141, 1590–1597.
41. K. J. You, C. T. Chang, B. J. Liaw, C. T. Huang and Y. Z. Chen, Appl. Catal., A, 2009, 361, 65–71.
42. H. Rojasa, G. Diazb, J. J. Martinez, C. Castãneda, A. G. Cortes and J. A. Alatorre, J. Mol. Catal. A: Chem., 2012, 363, 122–128.
43. P. Gao, A. Q. Wang, X. D. Wang and T. Zhang, Catal. Lett., 2008, 125, 289–295.
44. X. J. Ni, B. S. Zhang, C. Li, M. Pang, D. S. Su, C. T. Williams and C. H. Liang, Catal. Commun., 2012, 24, 65–69.
45. K. J. A. Raj, M. G. Prakash, T. Elangovan and B. Viswanathan, Catal. Lett., 2012, 142, 87–94.
46. B. Zhang, X. B. Zhang, L. Y. Xu, Y. J. Zhang, Y. H. Qin and C. F. Liang, React. Kinet. Catal. Lett., 2013, 110, 207–214.
47. V. Gutierrez, M. Alvarez and M. A. Volpe, Appl. Catal., A, 2012, 413–414, 358–365.
48. C. Milone, C. Crisafulli, R. Ingoglia, L. Schipilliti and S. Galvagno, Catal. Today, 2007, 122, 341–351.
49. J. S. Qiu, H. Z. Zhang, X. N. Wang and C. H. Liang, React. Kinet. Catal. Lett., 2006, 88, 269–276.
50. K. Abdur-Rashid, A. J. Lough and R. H. Morris, Organometallics, 2001, 20, 1047–1049.
51. E. Farnetti, M. Pesce, J. Kašpar, R. Spogliarich and M. Graziani, J. Chem. Soc., Chem. Commun., 1986, 10, 746–747.
52. Z. Tian, X. Xiang, L. Xie and F. Li, Ind. Eng. Chem. Res., 2013, 52, 288–296.

---