Mesoporous Carbon-Supported Pd Nanoparticles in the Metallic State-Catalyzed Acylation of Amides with Aryl Esters via C-O Activation.

Carbon, an abundant, inexpensive, and nonmetallic material, is an inimitable support in heterogeneous catalysis, and variable carbonaceous materials have been utilized to support metal nanoparticle catalysts. We developed an efficient and stable heterogeneous catalyst with highly dispersed metallic palladium nanoparticles embedded in an ordered pore channel of mesoporous carbon and first applied the catalyst to construct imides from amides using aryl esters as an acylation reagent via C-O activation. The catalyst represents excellent catalytic performance and could be reused and recycled five times without any significant decrease in activity. The heterogeneous nature of metallic state palladium was proven to be the active center in the acylation reaction.

Introduction

Imides are widely prevalent structural elements present in a variety of organic molecules that exhibit pharmaceutical and bioactive properties.[1,2] As a consequence, methods for installing imides structure into organic molecules are of great interest, and numerous strategies have been developed to construct those skeletons. In addition, the acylation of the amide with a range of activated acyl species derived from carboxylic acids (such as acid anhydrides and acyl halide) became a widely used method for the synthesis of the imide derivates, while stoichiometric amounts of salts were produced and the strong base was necessary for those reactions, resulting in the poor atom economy and environmental problems (Scheme a).[3−5] The development of homogeneous catalytic protocols carried the promise of the atom-economical and sustainable formation of imide derivates, such as the rearrangement of isocyanates, the oxidation of amides, and the oxidative decarboxylation of amino acids. Zhang et al. prepared imides from the oxidation of amides using CuBr as a catalyst and TBHP or TEMPO as an oxidant (Scheme b).[6] Moreover, aldehydes, benzyl alcohol, methyl arenes, and potassium acyltrifluoroborates could also be employed as acylation regents reacting with amide to afford the desired imides (Scheme c).[7−12] Notably, palladium-catalyzed aminocarbonylation of (hetero) aryl iodides with amides and CO has developed into an alternative reliable protocol for assembling various structurally diverse imides (Scheme d).[13] Although great progress has been realized in this realm, to the best of our knowledge, esters that are readily available and environmentally benign feedstock acting as acylation reagents to synthesize imide derivates are a formidable challenge.

Scheme 1

Preparation of Imide Derivates

Environmentally benign, operationally simple, and robust reactions, particularly those employing heterogeneous catalysts, are of significant interest to the chemical industry.[14−17] Inspired by our previous works which confirmed that supported palladium nanoparticles (PdNPs) could act as heterogeneous catalysts for the amidation reaction between aryl esters with formamides or amines,[18,19] we hypothesized whether they could serve as a catalyst to construct imide structures using an ester as an acylation reagent.

Carbonaceous materials possess a large specific surface area and good electrical and thermal conductivity and are widely used as supports for heterogeneous catalysts.[20−26] Metal supported on activated carbons (AC) is one of the most significantly considerable carbon-based catalysts which was studied and applied in chemical engineering.[27−29] Unfortunately, the mass-transfer process was limited by the microporous structure of activated carbons. The lack of anchor points in the active phase leads to the uneven distribution of the metal nanoparticles on the surface of carbon, easy agglomeration, and leaching. As a result, the synthesis of the carbon-supported metal catalyst with adjustable porosity and a high surface area was highly desired.[30]

Mesoporous carbon (MC) materials could improve the shortcomings of AC, such as insufficient anchor points and the aggregation and leaching of metal nanoparticles through the flexibly adjusted pore size. Also, the strong metal–support interactions triggered by the close contact of metal nanoparticles with the mesoporous material could improve the catalytic activity and stability.[31−35] Because of the unique physiochemical properties of MC, the mesoporous carbon-supported palladium nanoparticle (PdNP/MC) catalyst has been studied with respect to traditional C–C coupling and hydrogenation reactions. Regretfully, the wide application of the catalyst was hindered by the limited reaction types, poor catalytic performance, and lack of effective preparation methods for completely metallic state PdNPs, which are proven to be active species.[33,34]

Herein, we prepared various PdNPs/MC catalysts in the metallic state via the impregnation reduction method and first applied these heterogeneous catalysts to facilely construct imide structures using esters as acylation agents via C–O activation, which present a blueprint for the further application of PdNPs/MC in organic synthesis. Compared to commercially available heterogeneous Pd/AC, PdNPs/MC possesses a larger specific surface and exhibits better catalytic performance and stability. The XPS analysis of fresh and used catalyst indicates that the reaction may be performed by a Pd0/PdII catalytic cycle.

Results and Discussion

Pd nanoparticles immobilized on mesoporous carbon (3 wt % PdNPs/MC) was analyzed by X-ray diffraction (XRD). Clear reflections of fresh and used catalysts (after the fifth recycle) can be observed at 40.0, 46.2, and 68.0° and 39.9, 46.4, and 67.9°, respectively, which correspond well with the (111), (200), and (220) lattice planes of metallic palladium. The diffraction peak intensities of used PdNPs/MC decreased, which may indicate that the particle size of used PdNPs became smaller than that of the fresh catalyst (Figure ). In addition, the fresh and used catalysts were characterized by the transmission electron microscope (TEM), which showed that the PdNPs were distributed uniformly on the pore channels of MC (Figure a,b). The Pd particle size was measured and found to be 5.18 and 4.60 nm for fresh and used catalysts, respectively (Figure c,d). A slightly smaller Pd particle size of the used catalyst was observed, in line with the results of XRD analysis.

Figure 1

XRD analysis of fresh and used 3 wt % PdNPs/MC.

Figure 2

(a) TEM image of fresh 3 wt % PdNPs/MC. (b) TEM image of used 3 wt % PdNPs/MC. (c, d) PdNP size distributions of fresh and used catalysts.

The PdNPs on the support of fresh and used catalysts exist in the metallic state, which were confirmed by the X-ray photoelectron spectroscopy (XPS) signal that appeared at binding energies of 335.62 eV for Pd 3d5/2 and 340.73 eV for Pd 3d3/2 over a fresh catalyst and 335.68 eV for Pd 3d5/2 and 340.50 eV for Pd 3d3/2 over a used catalyst (Figure ). The binding energies of Pd0 in the literature are 335.20 eV fpr Pd 3d5/2 and 340.50 eV for Pd 3d3/2.[36] This result confirmed the hypothesis that the PdNPs/MC-catalyzed C–O activation reaction was performed via a catalytic cycle that began with Pd0.

Figure 3

(a) XPS survey of the fresh catalyst. (b) XPS analysis of fresh and used 3 wt % PdNPs/MC.

The N2 gas adsorption/desorption analysis of MC, fresh, and used catalysts all exhibit typical type IV adsorption–desorption isotherms. The H2-type hysteresis loops with typical columnar channels at relative pressures of 0.5–0.9 confirmed that fresh and used catalysts both have mesoporous structural features (Figure a). The fresh and used catalysts have uniform, small, and arranged mesopores of about 4 nm (Figure ). The Brunauer–Emmett–Teller (BET) analysis is shown in Table . Compared to fresh 3 wt % PdNPs/MC, the smaller specific surface and pore volume of used 3 wt % PdNPs/MC may be attributed to the blockage of the pores by input reaction substrates, which is supported by a comparison of the IR analysis of fresh and used catalysts (Figure ).

Figure 4

(a) N2 adsorption/desorption isotherms of MC and fresh 3 wt % PdNPs/MC. (b) Pore size distribution of MC and fresh 3 wt % PdNPs/MC.

Table 1

N2 Gas Adsorption/Desorption Analysis of Catalysts

sample	specific surface (m2/g)	pore volume (m2/g)	pore size (nm)
MC	736	0.72	3.9
fresh (3 wt % PdNPs/MC)	647	0.66	4.0
used (3 wt % PdNPs/MC)	481	0.50	4.2

Figure 5

IR analysis of fresh and used catalysts.

The catalytic properties of the 3 wt % PdNPs/MC were initially evaluated for the acylation of amides (1a) and esters (2a) in ethyl acetate at 120 °C (Table ). Initially, commercially available 5 wt % Pd/C was used as the catalyst for the reaction and gave 3a in 45% yield (entry 1). Pd/γ-Al2O3, an efficient catalyst in our previous work, was also examined and afforded 3a in 73% yield. Notably, when 3 wt % PdNPs/MC acted as the catalyst for this reaction, gains in yields were observed, which could afford the desired products in 87% yield (entry 3), presumably owing to the comprehensive larger specific surface area of 3 wt % PdNPs/MC. Employing Pd(OAc)2 led to a slight decrease in the yield of the reaction (entry 4). PdNPs supported on different carbon materials, such as porous carbon, multiwalled carbon nanotubes (MWCNT), N-doped mesoporous carbon materials (NMC), and porous carbon (PC), can also be applied to the reaction, albeit in lower yields (entries 5–7). Any fluctuation in the Pd content of MC from 3 wt % (1, 2, 4, and 5 wt %) resulted in a decrease in reaction yields (entries 8–11). A brief screening of solvents suggested that EtOAc is the optimal solvent (entries 12–14). Omitting Pd failed to afford even trace amounts of the desired product (entries 15). Changing the reaction temperature offered poor yields (entries 16 and 17).

Table 2

Optimization of Reaction Conditionsa

entry	catalyst	solvent	yield (%)b
1	5 wt % Pd/AC	EtOAc	45
2	3 wt % Pd/γ-Al2O3	EtOAc	73
3	3 wt % PdNPs/MC	EtOAc	87
4	Pd(OAc)2	EtOAc	78
5	3 wt %Pd/PC	EtOAc	35
6	3 wt %Pd/MWCNT	EtOAc	52
7	3 wt %Pd/NMC	EtOAc	52
8	1 wt % PdNPs/MC	EtOAc	31
9	2 wt % PdNPs/MC	EtOAc	48
10	4 wt % PdNPs/MC	EtOAc	65
11	5 wt % PdNPs/MC	EtOAc	48
12	3 wt % PdNPs/MC	PhCl	8
13	3 wt % PdNPs/MC	MeCN	23
14	3 wt % PdNPs/MC	EtOH	35
15	MC	EtOAc
16c	3 wt % PdNPs/MC	EtOAc	63
17d	3 wt % PdNPs/MC	EtOAc	42

Reaction conditions: 1a (0.1 mmol), 2a (3 equiv), catalyst (25 mg), EtOAc (0.5 mL), 24 h, 120 °C.

Isolated yields.

130 °C.

100 °C.

With optimal reaction conditions in hand, we explore the versatility of aryl esters. As illustrated in Table , a series of substituted aryl esters allow the reaction to proceed smoothly to afford the desired products in moderate to excellent yields. An aryl ester bearing an electron-withdrawing group or an electron-donating group at the para or meta position as well as multiple substituents gave the desired products in moderate to excellent yields (3aa–3na). Interestingly, a preference for the electron-withdrawing groups on the phenyl group of an aryl ester was observed for this reaction (3ea–3ga vs 3ca–3da). The strong electron-deficient aryl esters exhibited excellent reactivity and afforded the desired imides at up to 99% yield (3ia, 3ma, and 3na). However, substrates with ortho-substituted groups delivered the products in relatively lower yields, which might be attributed to the steric hindrance (3la) and thus substrate 1r with two −Cl substituents in the ortho position of the ester group failed to give any expected products. Fortunately, substrates containing a heteroaryl ring were suitable for this protocol, providing products in moderate yields (3pa and 3qa).

Table 3

Versatility of Aryl Estersa,b

Reaction conditions: 1b (0.1 mmol), 2 (3 equiv), 3 wt % PdNPs/MC (25 mg), EtOAc (0.5 mL), 24 h, 120 °C.

Isolated yields.

Subsequently, the scope and limitations of amide 2 were examined using 1a as the coupling partner (Table ). An aryl amide with electron-donating or electron-withdrawing groups (4-F, 4-OCH3, and 2-I) participated in the reaction to afford the desired product in moderate yields (3ab–3ae). A range of primary amides containing different alkyl groups display relatively better reaction efficiencies, furnishing imides in moderate to excellent yields (3af–3ak). Satisfactorily, an amide with terminal alkenes gave an imide in 47% yield, thereby offering an opportunity for further derivatization. In addition, the 1H-indene containing an internal alkene was also tolerated in this protocol and converted to the 3am product, albeit in a somewhat low yield. Notably, N-phenylformamide served as a suitable substrate for this reaction (3an, 35% yield). Imides are widespread in pharmaceuticals. We could construct aniracetam, a drug improving brain function, in 87% yield (3ao).

Table 4

Scope of Amidesa,b

Reaction conditions: 1b (0.1 mmol), 2 (3 equiv), 3 wt % PdNPs/MC (25 mg), EtOAc (0.5 mL), 24 h, 120 °C.

Isolated yields.

The ester bearing different aryl groups was tested (Scheme ). Changing pentafluorophenyl to an ester pyridyl could also afford desired product 3ea. Furthermore, the electron-deficient aryl picolinates are also compatible in the reaction.

Scheme 2

Scope of Other Esters

The possibility of the recyclability of the 3 wt % PdNPs/MC in this reaction was also investigated using substrates 1a and 2a under optimal reaction conditions. As illustrated in Figure , the heterogeneous catalyst could be recovered by centrifugation and reused five times without any dramatic decrease in catalytic efficiency, and a slight decrease in yield was observed for the fifth cycle.

Figure 6

Recyclability of the catalyst.

This protocol appeared to operate mainly by heterogeneous pathways, as demonstrated by a hot filtration test. After filtering the 3 wt % PdNPs/MC in the middle of the reaction, no further reaction was observed, precluding the possibility of homogeneous catalysis, which was confirmed by the TOF value of 1a, as shown in Table . This result was also proven by a Hg(0) poisoning test (ESI). The ICP-MS analysis indicated 6.08 ppm Pd leaching into the reaction solution, thus further supporting the stability of the heterogeneous catalyst.

Table 5

Results of the Hot Filtration Test

time (h)	conversion (%)a	TOF
0	0	0
2	35.4	9.61
4 (after hot filtration)	36.1	9.67

GC data.

On the basis of previous reports,[19,37,38] we proposed a plausible mechanism for the acylation of amide (Scheme ). First, the process would begin with the oxidative addition of Pd0 and the C(acyl)–O bond of aryl ester 1, generating PdII intermediate A. Subsequently, intermediate A would coordinate with amide to give intermediate B. Then the alkoxide attacked the hydrogen of the amide, resulting in phenol and generating intermediate C. The reductive elimination of intermediate C would generate the imide products and the Pd0 catalyst.

Scheme 3

Proposed Reaction Mechanism

Conclusions

We have prepared an MC-supported metallic state heterogeneous Pd catalyst which would be applied successfully to construct imides from aryl esters and amides via C–O bond activation. This catalyst represents excellent catalytic activity and can be reused and recycled five times without any significant decrease in activity. The morphology of fresh and used catalysts was investigated by XRD, XPS, TEM, and N2 gas adsorption/desorption. The XPS analysis of fresh and used catalyst suggested that Pd0 was the active species. These findings provide an avenue for the wide application of the PdNPs/MC catalysts in undeveloped areas, including C–O, C–N, and C–H activation reactions.

Experimental Section

Mesoporous Carbon Preparation

Resorcinol (30 mmol) and F127 (0.4 mmol) were dissolved in a mixed solution of absolute ethanol and ultrapure water (40 mL, VH/VEtOH = 1:1), and then concentrated HCl (mass fraction 37%, 4 mmol) was added to the mixture, which was stirred at room temperature for 1 h. Formaldehyde solution (mass fraction 37%, 60 mmol) was added dropwise to the solution and stirred vigorously for another hour at room temperature. The resulting mixed solution was transferred to the reactor that was placed in an oven at 100 °C for 3 days. The rough product was cooled to room temperature, washed via centrifugation three times with water, and dried overnight at 80 °C. The obtained solid was carbonized for 3 h at 600 °C under a nitrogen atmosphere in a tube furnace at a heating rate of 2 °C/min. The product was cooled to room temperature and ground to obtain a black powder.[39]

Porous Carbon Preparation

p-Phenylenediamine (0.01 mol) was added to DMF (70 mL) at room temperature under magnetic stirring. After the p-phenylenediamine completely dissolved, 3,3′,4,4′-benzophenone tetracarboxylic dianhydride (BTDA) (0.01 mol) and ethylenediamine (0.01 mol) were added to the mixed solution in succession and stirred at room temperature for 12 h. The mixture was transferred to an autoclave to react at 150 °C for 12 h, cooled to room temperature, centrifuged at 3900 rpm for 15 min, washed sequentially with ethanol and DMF, and then dried at 60 °C in vacuum for 24 h. Finally, the solid was cooled to room temperature and calcined in a quartz tube furnace at 350 °C for 1 h under a N2 atmosphere at a 2 °C/min heating rate to complete the activation. The solid was then pyrolyzed at 900 °C for 2 h under a N2 atmosphere at a heating rate of 2 °C/min to form a porous carbon material.[40]

Doped Mesoporous Carbon Preparation

Hexamethylenetetramine (4.2 g) was added to a solution of resorcinol (30 mmol) and F127 (0.4 mmol) in a mixed solution of dry ethanol and ultrapure water (V(ultrapure water)/V(dry ethanol) = 1:1, 40 mL), and then concentrated hydrochloric acid (37 wt %, 4 mmol) was added to the mixture and stirred at room temperature for 1 h. Formaldehyde solution (37 wt %, 60 mmol) was added dropwise to the solution and stirred vigorously for 1 h under the same conditions. The resulting mixture was transferred to the reaction kettle, reacted in an oven at 100 °C for 3 days, cooled to room temperature, and centrifuged, and the solid was washed with water three times and dried overnight at 80 °C. The obtained solid was carbonized for 3 h in a tube furnace at 600 °C under nitrogen (heating rate of 2 °C/min), cooled to room temperature, and ground to obtain a black powder.[39]

Heterogenous Catalyst Preparation

The Pd NPs on MC and other supports such as γ-Al2O3 and multiwalled carbon nanotubes (MWCNTs) were prepared by the modified impregnation–reduction method. A PdNPs/MC heterogeneous catalyst (3 wt %) was prepared by the following procedure: a solid support (0.97 g) was suspended in distilled water (50 mL), followed by adding PdCl2 aqueous solution (0.01 M, 28.2 mL) and stirring at room temperature. Then an l-lysine aqueous solution (0.03 M, 1 mL) was added to this mixture. Subsequently, a NaOH aqueous solution (0.1 M) was added to the mixture to adjust the pH to 7 before being treated with a NaBH4 aqueous solution (0.35 M, 4.5 mL) dropwise within 10 min with vigorous stirring. The mixture was aged for 24 h and centrifuged to afford a solid, which was washed four times with distilled water and once with ethanol before being dried at 80 °C. The dried solid was used directly as the catalyst.

Activity Test

The acylation of 4-cyanobenzoic acid pentafluorophenol ester 1a with caprolactam 2a was used as a model reaction. In a typical reaction, 3 wt % PdNPs/MC (25 mg), 1a (31.3 mg, 0.10 mmol), 2a (33.9 mg 0.3 mmol), and EtOAc (0.5 mL) were added to a 25 mL oven-dried reaction tube. The resulting solution was stirred at 120 °C for 24 h via a magnetic stirrer. The crude product was concentrated under reduced pressure and purified by column chromatography to afford desired product 3.

Catalyst Recycling Experiment

After each reaction cycle, the solvent, substrate, and products were removed by centrifugation; the separated 3 wt % PdNPs/MC was washed thoroughly with 0.1 M NaOH ethanol solution (twice) and distilled water (four times) and then washed twice with ethanol, followed by centrifugal separation and drying at 80 °C for 12 h. The recovered catalyst was used for the next cycle.

Hot Filtration Test

PdNPs/MC (25 mg, 3 wt %), 4-cyanobenzoic acid pentafluorophenol ester (31.3 mg, 0.10 mmol), caprolactam (33.9 mg 0.3 mmol), and EtOAc (0.5 mL) were added to the reaction tube. The resulting solution was stirred at 120 °C for 2 h in a magnetic stirrer, and the reaction mixture was filtered through a preheated Celite pad after the reaction for 2 h. The filtrate was detected by GC to obtain the conversion of 1a and TOF. Then the filtered reaction solution continued to react for 2 h under normal conditions and was detected by GC again.