Highly Efficient Heterogeneous Pd@POPs Catalyst for the N-Formylation of Amine and CO₂

Abstract

Utilization of CO₂ for the production of fine chemicals has become a research hotspot for a long time. In order to make use of CO₂, we developed a highly efficient heterogeneous catalyst (denoted as Pd@POPs) for the N-formylation reaction of amine and CO₂ under mild conditions. The Pd catalyst was based on a porous organic polymer derived from the solvothermal polymerization of vinyl-functionalized PPh₃. A series of characterizations and comparative experiments demonstrated that the Pd@POPs catalyst has high BET (Brunauer-Emmett-Teller) surface areas, hierarchical pore structure, and uniform dispersion of Pd active sites resulting from the formation of strong coordination bonds between Pd species and P atoms in the porous organic polymer (POP) support. In addition to the excellent activity, the Pd@POPs catalyst shows good stability for the N-formylation reaction of amine and CO₂.

1. Introduction

Carbon dioxide (CO₂) is an abundant, low-cost, sustainable, and nontoxic C1 raw material. The transformation of CO₂ into value-added chemicals has attracted wide attention in academia and the industry. Additionally, the utilization of CO₂ as a carbon source for fine chemical synthesis can contribute to the reduction of CO₂ in the atmosphere [1,2,3]. Great efforts have been made by researchers for the conversion of CO₂ into value-added chemicals, such as formic acid [4,5], cyclic carbonates [6,7,8], and formamides [9]. Formamides are often used as intermediates for the synthesis of fine chemicals [10,11], Vilsmeier–Haack reaction, and solvents [12,13], and their production is one of the potential ways for the fixation of CO₂. For example, N,N-dimethylformamide (DMF) is produced by an NaOOCH-catalyzed reaction of dimethylamine with CO in industrial production [14]. However, the toxic CO hampers its wider use. The carbonylation reaction using cheap and abundant CO₂ is a safer way to synthesize organic chemicals. Using CO₂ as C1 block and H₂ as formylating reductant is a substitute green way for the N-formylation of amines.

In recent years, various homogeneous catalysts have been developed to improve the efficiency of N-formylation reaction under mild conditions. The most active homogeneous catalyst with a TON (Turnover Number) of up to 1,940,000 for the N-formylation of morpholine was developed by Ding [15], which is a pincer catalyst based on Ru coordinated with P and N atoms in the form of tridentate chelating ligand. Milstein et al. reported another similar chelating coordination catalyst based on Co, and a yield of up to 99% was achieved within 36 h at 150 °C, P_CO2 = P_H2 = 30 bar [16]. Homogeneous catalysts possess highly catalytic activity due to their well-defined and uniform single active sites. Nevertheless, homogeneous catalysts suffer the problem of catalyst recyclability. Thus, heterogeneous catalysts are desirable for industrial applications because they allow easy separations and catalyst recovery and recycle. A large number of heterogeneous catalysts supported on inorganic materials (including PAL [17], TiO₂ [18], and so on [19]) have been developed for N-formylation reaction. For instance, Kaneda et al. reported that the Au-based TiO₂ catalyst exhibited excellent activity and reusability for the catalysis of the selective N-formylation of functionalized amines [18]. Recently, few metal catalysts immobilized on porous organic polymers (POPs) have been explored for N-formylation reaction. Compared with conventional inorganic supports, POPs stand out for their permanent porosity, high surface areas, good thermal stability, and structure diversity [20]. For example, Liu et al. synthesized pyridine-functionalized porous organic polymers (CarPy-CMP) and CarPy-CMP@Ru. Morpholine in this catalytic system can get a conversion of 97% and a yield of 94% at P_CO2 = P_H2 = 4 MPa at 130 °C within 24 h [21].

Our group has been committed to the preparation and application of triphenylphosphine-based porous organic polymers. Rh/POL-PPh₃ [22], Rh/CPOL-BP&PPh₃ [23], and Rh/CPOL-BP&P(OPh)₃ [24] have been applied in the hydroformylation of ethylene, propylene, and butene, respectively. These catalysts exhibited exciting catalytic activity and stability. We have also designed and synthesized another two kinds of highly active catalysts (PPh₃-ILX@POPs [8] and Mg-por/pho@POPs [25] for CO₂ conversion). From our previous research, we know that the phosphine-rich backbone has a strong adsorption capacity for CO₂ and can fix metals, thus preventing metal loss [8]. Inspired by that, we envisioned a promising application of heterogeneous Pd@POPs as catalysts for the N-formylation of amine and CO₂. Here, in order to expand the application of POP materials easily obtained in nearly 100% yield, a general and highly efficient Pd-based heterogeneous catalyst synthesized by the solvothermal polymerization of 3v-PPh₃ for N-formylation reaction was developed. A conversion of 93% was obtained under mild conditions, such as 100 °C, P_CO2 = P_H2 = 3 MPa, within 24 h. Besides its excellent performance, the Pd@POPs catalyst is more applicable to secondary amines than primary amines and more stable than other common Pd catalysts supported on an inorganic carrier due to the formation of Pd–P coordination bonds.

2. Results and Discussions

The synthesis method of the Pd@POPs catalyst is illustrated in Scheme 1. The vinyl-functionalized PPh₃ ligand was polymerized under solvothermal conditions (THF, 100 °C); then Pd(OAc)₂ was added to the reaction mixture.

The pore structure of the catalyst can be determined from the N₂ adsorption–desorption isotherm (Figure 1a) and pore size distribution (Figure 1b). The Pd@POPs exhibits high BET surface areas and pore volume (900.3 m²/g and 1.504 cm³/g). The pore size distribution is calculated by the nonlocal density functional theory (NLDFT) method. The pore sizes of Pd@POPs are mainly distributed in the region of micropores (<2 nm) and mesopores (2–10 nm). The existence of micropores can be determined by the region of P/P₀ = 0–0.01. The hysteresis loop in the N₂ adsorption–desorption isotherm suggests the existence of mesopores. The TEM image (Figure 1c) and SEM image (Figure 1d) for Pd@POPs also provide evidence for the hierarchical porosity that facilitates the diffusion of reactants and products during the reaction. Furthermore, the good thermal stability of POP and Pd@POPs is proved by thermal gravimetric analysis (TGA, Figure S1). The initial decomposition temperature is 400 °C, which can be well adapted to industrial requirements.

The TEM images of fresh and used Pd@POPs are shown in Figure 2. No obvious big metal particles or clusters are observed in both of the TEM images, which suggests that the Pd active species are uniformly dispersed on the POP support. In addition, SEM mapping images of used Pd@POPs reveal that functional elements (P and Pd) are highly dispersed. It means that these elements are well integrated in the used Pd@POPs catalyst.

To explore the oxidation states of Pd species and the coordination effect between Pd and P, X-ray photoelectron spectroscopy (XPS) analysis was performed for POP and Pd@POPs, and the results are listed in Figure 3. The P2p spectrum (Figure 3a) mainly shows the presence of P with BE (Bonding Energy) = 131.79 and 130.42 eV in the POP backbone. After Pd loading, the binding energy of P2p (Figure 3b) shifts forward to a high field to 132.1 and 130.6 eV. In the Pd3d XPS spectrum (Figure 3c), four peaks with binding energies at 342.79 and 338.15 eV, which can be ascribed to Pd²⁺, and 340.4 and 336.8 eV, which are assigned to Pd⁰, can be deconvoluted. Compared with Pd(OAc)₂ (343.8 and 338.6 eV) [26], the Pd binding energy shifted negatively. These results demonstrate the coordination of Pd species with POP, in accordance with the conclusion of ³¹P solid-state NMR. The presence of Pd⁰ could be due to the fact that POP support possesses reducibility.

The solid-state ¹³C NMR spectrum of Pd@POPs (Figure 3d) shows two strong broad peaks at 20–50 ppm (polymerized vinyl group) and 120–160 ppm (aromatic carbons), which suggests that the polymerization process was completed adequately. The solid-state ³¹P NMR spectrum of Pd@POPs is employed to auxiliary XPS, proving the existence of a P–Pd coordination bond. The peak at −6 ppm in spectrum (Figure 3e) is assigned to the uncoordinated P specie. The peak at 24 ppm can be ascribed to the P specie coordinated with Pd due to the fact that the peak of 24 ppm is enhanced after Pd loading, which is consistent with the literature [27].

Sequentially, the catalytic activity of a few kinds of metal catalysts for the N-formylation reaction of morpholine to N-formylmorpholine was assessed, and the results are summarized in

Table 1. Catalytic performance of metal catalysts for N-formylation reaction.

Entry	Catalyst	Conv.%
1 a	Pd(OAc)2	49
2	Pd(OAc)2	92
3	PdCl2	71
4	RhCl3	11
5	IrCl3	36
6	Co(NO3)2	20

Conditions: morpholine, 1 mmol; precursor, 2.3 mol% based on morpholine; K₃PO₄, 0.3 mmol; P_CO2 = P_H2 = 3 MPa; DMI, 4 mL; 100 °C; 24 h; ^a no K₃PO₄.

. The results show that the homogeneous catalyst of Pd(OAc)₂ gave a conversion of 49% (entry 1) in the absence of K₃PO₄. After adding K₃PO₄ to the reaction system, the conversion increased to 92% (entry 2). It means that alkali is beneficial to the formation of N-formylmorpholine, which is consistent with the literature [23,28]. Other precursors of Pd were researched, such as PdCl₂ (entry 3), showing a conversion of 71%. Some other metal catalysts were also investigated, and the results are listed in Table 1 (entry 4–6). RhCl₃ (entry 4) is almost not active for the N-formylation of morpholine with CO₂ in a homogeneous system, and the conversion is only 11%. Both IrCl₃ (entry 5) and Co(NO₃)₂ (entry 6) have low activity for N-formylation and get conversions of 36% and 20%, respectively. It can be concluded that the metal catalysts of palladium are more active for N-formylation than RhCl₃, IrCl₃, and Co(NO₃)₂. Therefore, we used Pd(OAc)₂ as the metal precursor to investigate the effect of supports.

A comparison of catalytic performances for Pd@supports is listed in

Table 2. Comparison of the catalytic performance of Pd@supports.

Catalyst	CO2/H2/MPa	T/°C	Conv./%
Pd@POPs	3/3	100	92/91 a
Pd@SBA-15	3/3	100	92/82 a
Pd@Al2O3	3/3	100	61
Pd@TiO2	3/3	100	93/71 a
Pd@NC [29]	3/4	130	93
Pd/LDH [30]	3/3	140	91.7

Conditions: morpholine, 1 mmol; Pd, 2.3 mol% based on morpholine (Pd loading is 0.5 wt%); K₃PO₄, 0.3 mmol; P_CO2 = P_H2 = 3 MPa; DMI, 4 mL; 100 °C; 24 h. ^a The data were obtained with a reused catalyst for the second run.

, employing morpholine as a substrate. When POP was used as support (Pd@POPs), the catalyst obtained conversions of 92% (first run) and 91% (second run). By comparison, the conversions for the first run over Pd@SBA-15, Pd@Al₂O₃, and Pd@TiO₂ were 92%, 61%, and 93%, respectively, while the conversions for the second run of using Pd@SBA-15 and Pd@TiO₂ decreased to 82% and 71%, respectively. The decrease of catalytic activity may be attributed to the fact that the interaction between Pd and supports is not as strong as the coordination bond between Pd and P species in POP support. These results show that POP is more active and stable for the N-formylation of morpholine and CO₂ than catalysts with other supports we selected in the same reaction condition. The high BET surface areas, hierarchical pore structure, and Pd–P coordination bonds of the Pd@POPs catalyst are responsible for its improved catalytic performance. Compared with the Pd-based heterogeneous catalysts (Pd@NC, Pd/LDH) reported in the literature, Pd@POPs exhibits a similar activity under milder conditions.

Inspired by the above results, Pd@POPs was explored as a catalyst for the N-formylation of other secondary or primary amines (

Table 3. Substrate prolongation experiments of the Pd@POPs catalyst.

Entry	Substrate	Product	Conv.%
1	1a	1b	93
2	2a	2b	80
3	3a	3b	17
4	4a	4b	22
5	5a	5b	47
6	6a	6b	33
7	7a	7b	30
8	8a	8b	61
9	9a	9b	24

Conditions: morpholine, 1 mmol; Pd, 2.3 mol% based on morpholine (Pd loading is 2.5 wt%); K₃PO₄, 0.33 mmol; P_CO2 = PH₂ = 3 MPa; DMI, 4 mL; 100 °C; 24 h.

). The catalyst was successful for secondary amines, and the N-formylation products of cyclic secondary amines (1a-2a and 1b-2b) were synthesized with yields of 80–93% (entries 1–2). When aliphatic N-methylpentylamine was the substrate, a 61% yield of the corresponding formamide was obtained (entry 8). The Pd@POPs catalyst exhibited good applicability for secondary amines but was limited for primary amines. In Table 3, when primary amines, such as 4-methylbenzylamine (5a), were used as the raw material, the highest yield of 47% was achieved (entry 5). For other primary amines, including cyclohexamine (3a), benzylamine (4a), β-phenylethylamine (6a), 1-heptanamine (7a), and n-pentylamine (9a), the yields of the desired formamides were only 17–33% (entries 3, 4, 6, 7, 9).

Additionally, the reusability of the Pd@POPs catalyst was studied. The results (Figure 4) suggest that the catalyst could be recycled at least five times without obvious loss of catalytic activity. The good stability is attributed to high P ligand concentrations, high surface areas, and stable coordination bonds between Pd species and P atoms in the POP support.

As we know that the N-formylation reaction mechanism (Figure 5) is similar in most literatures, CO₂ is reduced to formic acid in a hydrogen atmosphere, and then reacts with amine to form amides [17,29,31]. We also demonstrated the reaction mechanism in a previous work [32].

3. Conclusions

In summary, we successfully synthesized a Pd@POPs catalyst by the solvothermal synthetic method for the N-formylation reaction of amine and CO₂. The Pd@POPs catalyst can immobilize Pd active species, which simplifies the recovery and reuse for the N-formylation reaction system. Especially, the heterogeneous catalyst has a good practical prospect because it not only is easy to synthesize but also has high catalytic efficiency and excellent stability. Characterization results indicate that the Pd@POPs catalyst has high BET surface areas, hierarchical pore structure, and uniform dispersion of Pd active species due to the formation of strong Pd–P coordination bonds. With the above-mentioned advantages, this method may be expected to replace the current ways for N-formylation reaction.

4. Materials and Methods

4.1. Material

All solvents and other chemicals were commercially available. Anhydrous THF (Tetrahydrofuran) is prepared by distillation from sodium benzophenone ketyl. The DMI (≥99.5%) was purchased from the Macklin Biochemical Co., Ltd. (Shanghai, China).

4.2. Synthesis of PPh₃-POP

PPh₃-POP was synthesized by solvothermal polymerization. Typically, 100 mL THF was added into a 250 mL round-bottom flask loaded with a mixture of 3v-PPh₃ (tris-(4-vinylphenyl)-phosphine, 10 g), AIBN (2,2’-azobis(isobutyronitrile), 0.25 g), and a magnetic rotor. After stirring for 1 h to make it completely dispersed, the solution was transferred into an autoclave under Ar atmosphere. After sealing, the solution was kept at 100 °C for 24 h. After the system was cooled to room temperature, the polymer was collected by vacuum filtration and washed by THF. Then, the polymer was dried at 60 °C under vacuum for 12 h. Eventually, white powder was got and labeled as PPh₃-POP.

4.3. Synthesis of Pd@POPs

Pd(OAc)₂ (0.0108 g) was dispersed in 50 mL THF under Ar atmosphere. Afterwards, POP (1.0103 g) was introduced. The resulting mixture was stirred at room temperature for 24 h. The precipitate was collected by filtration and washed with THF. After drying at 60 °C under vacuum for 12 h, a pale-yellow powder was acquired and labeled as Pd@POPs.

4.4. Synthesis of Pd@SBA-15, Pd@Al₂O₃, and Pd@TiO₂

The Pd@SBA-15, Pd@Al₂O₃, and Pd@TiO₂ catalysts were prepared by the same method as described above except that POP was replaced with SBA-15, Al₂O₃, and TiO₂.

4.5. A Typical Procedure for N-Formylation Reaction

Pd@POPs (Pd loading was 2.3 mol% based on morpholine), morpholine (1 mmol), K₃PO₄ (0.3 mmol, 0.063 g), and DMI (1,3-dimethyl-2-imidazolidinone, 4 mL) were successively added into a stainless steel autoclave reactor (30 mL inner volume). The autoclave was purged by mixed gas three times and charged with the mixed gas (CO₂:H₂ = 1:1) up to 6 MPa at room temperature. Then the system was heated by an electric heating jacket to 100 °C and stirred for 24 h. The products were analyzed by Agilent 7890A (Santa Clara, CA, USA) gas chromatography (GC) with a capillary column (HP-5, 30 m × 0.32 μm diameter) using a flame ionization detector. Gas chromatography analysis used toluene as an internal standard. ¹H NMR spectra were acquired on Bruker AVANCE III NMR spectrometer (Kloten, Zürich, Switzerland) at 400 MHz using trimethylsilane (TMS) as an internal standard.

4.6. Recycling Stability of Pd@POPs Catalyst

The used catalyst was obtained by centrifugation after every cycle. The above catalyst was thoroughly washed with ethanol and THF. After drying at 60 °C under vacuum for 12 h, the used catalyst was reused for the next reaction cycle.