The non-precious metal approach for production of chemicals from biomass resources is attracting immense research interest globally and is very important from the perspectives of green chemistry and sustainable development [1-6]. For large-scale biomass conversions, precious metal catalysts and basic additives are generally used to achieve high performance. A typical case is the oxidative esterification of furfural, which is derived from lignocellulosic biomass and recognized as a promising renewable resource [7-11], to produce methylfuroate. It is an attractive route because of the applications of methylfuroate as an anti-explosive additive in petrol and an anti-tumor drug. In most cases, homogeneous bases are added to promote the reaction and facilitate the abstraction of active hydrogen. For example, Au/FexOy-hydroxyapatite catalyst affords a 92% yield of methylfuroate with addition of K2CO3 after reaction at 140 ℃ for 4 h [12]. Christensen et al. [11] reported that addition of CH3ONa remarkably improved the catalytic performance of Au/TiO2 in the oxidative esterification of furfural. Recently, CoxOy-N@C showed high potential in catalysis, and gave a 95% yield of methylfuroate from furfural and methanol with addition of K2CO3 [13]. However, base additives lead to high energy consumption, leaching problems, and side reactions. In future, it is necessary to develop an inexpensive and environmentally friendly catalyst that provides efficient oxidative esterification of furfural without the requirement for a basic additive.
Metal-containing N-doped carbon catalysts can be used as an alternative to precious metal catalysts, and are inexpensive and show unique properties in electrocatalysis [14-16] and organic reactions [17-20]. To date, various cobalt-nitrogen-doped carbon catalysts have been synthesized by pyrolysis of nitrogen-rich metal complexes or mixtures of metal salts and N-containing ligands on supports under an inert atmosphere. Research has focused on the structures and factors influencing the catalytic activity. There is evidence that a single Co atom bonded to N atoms within graphitic sheets can act as an active center in aerobic oxidative cross-coupling and hydrogenative coupling reactions [21, 22]. Basic additives, such as LiOH or NaOH, are indispensable for Co-N-C catalysts in these reactions, and it is not known if Co-N-C catalysts can achieve high efficiency without the assistance of basic additives.
Previous studies performed in our laboratory showed that the basic support Mg(OH)2 remarkably increased the performance of gold catalyst in oxidative conversions [23]. Further research indicated that pyridinic N-doped carbon species might behave as Lewis bases and greatly promoted the catalysis of CoOx-N/C to achieve high efficiency for the oxidative esterification of diformylfuran in a neutral medium [24]. These findings suggested that cobalt-nitrogen-doped carbon supported on basic carrier could realize high efficiency oxidation of furfural without the requirement for a basic additive. In the present study, a Co-N-C/MgO catalyst was synthesized and used for production of high yields of methylfuroate under mild conditions without any basic additives. This method could be used for the development of high-performance non-precious metal catalysts for oxidative esterification conversions of biomass and biomass-derived compounds.
All reagents and chemicals used were of analytical grade unless otherwise specified. Furfural, MgO, ethanol, activated carbon, and 1, 10-phenanthroline were obtained from Aladdin Chemistry Co. Ltd. (Shanghai, China). Methanol was obtained from Kermel Chemical Reagent Development Center (Tianjin, China). Cobalt acetate tetrahydrate was from Sinopharm Chemical Reagent Co. Ltd. NaX and NaY were from Nanhua Catalyst Co. Ltd (Tianjin, China). CaO was from Damao Chemical Reagent Factory (Tianjin, China).
The Co-N-C/MgO catalyst was synthesized according to a reported method with some modifications [18]. 1, 10-Phenanthroline (182 mg, 1.0 mmol) and cobalt acetate tetrahydrate (125.8 mg, 0.5 mmol) were dissolved in 50 mL of ethanol and stirred for 30 min at room temperature. Then, 692.2 mg of MgO was added to the solution, and the mixture was stirred for 4 h at 60 ℃. The ethanol was removed under vacuum and the mixture was cooled to room temperature. The sample was then dried at 80 ℃ for 12 h. Then, the sample was placed in an oven, which was heated to 800 ℃ at 25 ℃ min-1 and maintained at this temperature for 2 h under a nitrogen atmosphere. For comparison, cobalt-based catalysts were prepared by a similar method using activated carbon and typical basic supports (NaX, NaY, and CaO). The resulting samples were denoted as CoOx-N/C, CoNC/NaX, CoNC/NaY, and CoNC/CaO.
The samples were analyzed by X-ray powder diffraction (XRD) on a Rigaku D/Max 2500/PC powder diffractometer with Cu Kα radiation (λ = 0.15418 nm) at 40 kV and 200 mA at a scanning rate of 5° min-1. Transmission electron microscopy (TEM) images were obtained using a JEOL JEM-2000EX electron microscope with samples deposited on a carbon polymer supported copper grid. X-ray photoelectron spectroscopy (XPS) measurements were performed on a Thermo ESCALAB 250Xi using a Mg Kα (1253.6 eV) radiation source and a chamber pressure lower than 5 × 10-8 Pa. Inductively coupled plasma optical emission spectrometry (ICP-OES) using a Perkin Elmer ICP-OES 7300DV was used to determine the cobalt content. A carbon/sulfur analyzer (EMIA-8100, HORIBA) and oxygen/nitrogen/hydrogen analyzer (EMGA-930, HORIBA) were used to determine the carbon and nitrogen contents, respectively.
Oxidative esterification of furfural was performed in a 10-mL Teflon lined stainless steel autoclave equipped with a magnetic stirrer, a pressure gauge, and automatic temperature control apparatus. Typically, 0.5 mmol of furfural, 80 mg of Co-N-C/MgO (3.9% mass fraction, 0.05 mmol of Co), and 5 mL of methanol were loaded into the autoclave. The autoclave was purged with oxygen and then pressurized to the required value. During the reaction, the pressure was maintained by supplying oxygen. After the reaction was completed, the autoclave was cooled to room temperature. The liquid reaction mixture was analyzed on an Agilent gas chromatograph (GC7890D) equipped with a flame ionization detector (FID) and Agilent GC/MS 6890-5973. The conversion and selectivity toward methylfuroate were evaluated by the internal standard method using n-dodecane as the internal standard.
First, the Co-N-C/MgO catalyst was analyzed by XRD in the range 5°-80°. The Co-N-C/MgO only displayed diffraction peaks for MgO (Fig. 1(e)), and no cobalt nanoparticles were observed on the surface of the MgO in the TEM image (Fig. 1(b)). After acid treatment and microwave digestion, the Co-N-C/MgO was analyzed by ICP-OES. The results confirmed the formation of cobalt, nitrogen, and carbon after pyrolysis under N2. The cobalt was highly dispersed on the catalyst surface, and cobalt-nitrogen-doped carbon species formed. To verify this, the Co-N-C/MgO was treated with 2 mol L-1 HCl to remove MgO, which was confirmed by the disappearance of its characteristic peaks in the XRD pattern (Fig. 1(f)). Additionally, a diffraction peak appeared at 26°, indicating that the carbon formed (carbon mass fraction = 71.8% by ICP). In contrast to Co-N-C/MgO, the TEM image (Fig. 1(c)) of self-supported Co-N-C(HCl) showed obvious MgO-shaped cavities, which would have been occupied by MgO before the acid treatment. Moreover, the ICP-OES results showed that the Co-N-C(HCl) contained cobalt and nitrogen. These observations were consistent with previous reports that the moderate interaction of MgO with Co species may be beneficial for generation of cobalt-nitrogen-doped carbon species by pyrolysis of a cobalt(Ⅱ) phenanthroline complex on MgO [22].
XPS was used to characterize the N and Co. The N 1s spectrum of Co-N-C/MgO displayed three peaks at 399.0, 400.8, and 402.0-403.2 eV, which could be assigned to pyridinic N bonded to cobalt, graphitic N, and pyridine N-oxide, respectively (Fig. 2) [21, 25]. Meanwhile, in agreement with the N 1s spectrum and XRD pattern, the binding energy of the Co 2p3/2 peak at 780.5 eV and its satellite peak corresponded to Co(Ⅱ), which may be coordinated with N atoms in the carbon species [26, 27]. These results revealed that cobalt-nitrogen-doped carbon species and carbon were formed by pyrolysis of a cobalt(Ⅱ) phenanthroline complex on MgO at 800 ℃ under a N2 atmosphere, which agrees with previous research [15, 21, 22, 28].
The catalytic performance of Co-N-C/MgO, CoOx-N/C, CoNC/NaX, CoNC/NaY, and CoNC/CaO was compared. Exploratory experiments were performed in methanol in a 10-mL autoclave reactor (Fig. 3). Co-N-C/MgO exhibited excellent performance in the oxidative esterification of furfural without a basic additive, where it gave 89.3% conversion of furfural to methylfuroate with 90.1% selectivity at 100 ℃ for 6 h. When the reaction time was increased to 12 h, a satisfactory result (93.0% conversion with 98.5% selectivity toward methylfuroate) was obtained. In previous studies, cobalt oxide-nitrogen/catalyst (CoOx-N/C) gave a good yield for methylfuroate produced by oxidative esterification of furfural with K2CO3 as an additive [13, 26]. Here, CoOx-N/C only gave a 67.2% conversion of furfural and 84.4% selectivity toward methylfuroate without addition of a base. By contrast, the Co-based catalysts with the typical basic zeolites NaX and NaY as supports exhibited moderate catalytic activity with 83.6% and 85.4% selectivity toward methylfuroate, respectively. CoNC/CaO only achieved 22.0% conversion and 38.8% selectivity toward methylfuroate. The choice of support obviously affected both the catalytic activity and selectivity. As demonstrated by XRD, TEM, XPS, and previous reports, the use of MgO as a support gave good distribution and formation of cobalt-nitrogen-doped carbon species, which may be closely related to the remarkable catalytic activity of Co-N-C/MgO.
Considering that hydrogen abstraction is a key step in the oxidative esterification of an aldehyde with an alcohol, we compared the catalytic performance of Co-N-C/MgO with other Co-based catalysts in the dehydrogenative esterification of furfural with methanol under N2 (Table 1). The choice of support was critical for the selectivity to the dehydrogenative esterification product, which was consistent with the results for oxidative esterification. As expected, in the presence of Co-N-C/MgO, the reaction of furfural with methanol at 100 ℃ mainly converted the aldehyde into ester. Co-N-C/MgO exhibited the highest selectivity toward methylfuroate, whereas CoOx-N/C, CoNC/NaX, CoNC/NaY, and CoNC/CaO were not efficient for the generation of methylfuroate. These results imply that the excellent performance of Co-N-C/MgO could be attributed to its good catalytic ability in hydrogen abstraction.
To further investigate the catalytic performance of the cobalt-nitrogen-carbon species, Co-N-C(HCl) (0.05 mmol of Co) was synthesized by removing MgO with HCl, and used as a catalyst in the oxidative esterification of furfural in methanol under 0.5 MPa O2 at 100 ℃ for 5 h. Interestingly, in the absence of MgO, the majority of furfural was converted to the acetal (2-(dimethoxymethyl)furan, 81.0% conversion), and no methylfuroate was detected by GC analysis. To obtain more information, MgO was introduced (molar ratio of MgO/Co = 34.8) to the Co-N-C(HCl). The GC/MS chromatograms obtained before and after addition of MgO were distinct (Fig. 4). MgO dramatically increased the selectivity toward the ester, with the selectivity toward methylfuroate increasing from not detected to 84.4% (Table 2, entry 2). This showed the catalytic performance of the cobalt-nitrogen-doped carbon species was greatly affected by HCl treatment, but could be recovered to a great extent by addition of MgO. The condensation of furfural with methanol to acetal and the oxidative esterification of furfural to methylfuroate are competitive pathways, with the dominant process determining the main product. These differences indicated that the -CHO groups were mainly converted via the competing condensation pathway with Co-N-C(HCl), or via the oxidative esterification pathway with Co-N-C(HCl) and MgO. Other research has shown that specific anions, including Cl-, greatly affect the catalyst performance of M-N-C catalyst because they might coordinate with the active metal center [29]. In our case, although the MgO could be removed from the Co-N-C/MgO by washing with HCl, residual Cl- may affect the metal center and lead to an obvious loss of catalytic activity for the oxidative esterification.
Two experiments were carried out to confirm the negative role of Cl-. When NaCl (0.5 mmol) was introduced to the reaction with the Co-N-C/MgO catalyst, the conversion decreased from 93.0% to 56.3% and the selectivity toward methylfuroate decreased slightly (Table 2, entry 3). If Co-N-C/MgO was pretreated with NaCl solution, thoroughly washed with water, and then calcined at 800 ℃, both the conversion and selectivity decreased obviously (Table 2, entry 4). These changes confirmed that Cl- ions had a negative effect on the oxidative esterification reaction. Therefore, we believe that using MgO as a support or additive allows for high catalytic efficiency of the cobalt-nitrogen-doped carbon catalyst and greatly reduces the negative effects of Cl-. In addition, preliminary studies indicated that the use of basic additives such as K2CO3 could increase the catalytic activity and selectivity in oxidation and oxidative esterification reactions by promoting the elimination of active hydrogen. Thus, Co-N-C/MgO was compared with the representative Co-based catalyst CoOx-N/C in the oxidative esterification of furfural in methanol with the addition of 0.1 mmol of K2CO3. Both catalysts achieved excellent results. Using CoOx-N/C, the conversion increased from 67.2% to 92.6% and the selectivity toward methylfuroate increased from 84.4% to 95.0%. Meanwhile, Co-N-C/MgO gave 99.6% conversion with 95.2% selectivity toward methylfuroate. Therefore, the effect of MgO is similar to that of K2CO3, which implies that MgO may play a similar role to a base like K2CO3 in this reaction.
Next, the reaction conditions for the oxidative esterification of furfural with Co-N-C/MgO catalyst were optimized. The catalytic performance of Co-N-C/MgO was investigated at 100 ℃ for 6 h with oxygen pressures of 0.3, 0.5, 0.7, and 1.0 MPa (Table 3, entries 1-3 and Fig. 3). The oxygen pressure had a negligible effect on the oxidative esterification of furfural with Co-N-C/MgO in this pressure range. Considering the conversion of furfural may reach or be close to maximum values at 6 h, we shortened the reaction time to 1.5 h to further investigate the effect of the oxygen pressure. As expected, even at moderate conversion, there was also no apparent change in the conversion or selectivity toward methylfuroate with changes in the oxygen pressure. Additionally, the influence of temperature was examined at 60, 80, 100, and 120 ℃ (Table 3, entries 8-10 and Fig. 3). The conversion of furfural gradually increased as the reaction temperature increased, and 100 ℃ and 120 ℃ were the optimum temperatures for this reaction.
Recycling of the Co-N-C/MgO catalyst was investigated by evaluating the catalytic activity in repeat use for the oxidation esterification of furfural at 100 ℃ under 0.5 MPa O2. The Co-N-C/MgO catalyst retained its activity for at least four recycling runs. In the fourth cycle, the catalyst gave 84.7% conversion of furfural and 87.5% selectivity toward methylfuroate. These results show that the Co-N-C/MgO catalyst is stable and can be reused in the oxidation of furfural to methylfuroate.
The oxidative esterification of furfural involves oxidative esterification and condensation of -CHO with -OH (Scheme 1). In methanol, furfural is converted into the intermediate hemiacetal, and subsequently undergoes dehydrogenation to form the target product methylfuroate. A competitive reaction of hemiacetal condensation with methanol can occur to generate the side product acetal in the presence of specific catalytic sites or in the blank reaction (without any catalyst or additive). In our case, with Co-N-C/MgO as the catalyst, the main product was methylfuroate. This suggests that the dehydrogenation of hemiacetal to methylfuroate is dominant. By contrast, with Co-N-C(HCl) as the catalyst, the main reaction was the condensation of -CHO with -OH generating the acetal as the main product. To exclude the possibility of producing methylfuroate via esterification of the hydroxyl and carboxyl groups, an oxidative esterification experiment using furancarboxylic acid as the starting substrate was performed in the presence of Co-N-C/MgO under the same conditions. No methylfuroate was detected in the GC chromatogram (data not shown). This clearly proved that the oxidative esterification did not proceed along the esterification pathway. The details for the mechanism of the Co-N-C/MgO catalyst are being clarified in further investigations.
A non-precious metal catalyst of cobalt-nitrogen-doped carbon supported on MgO was prepared via impregnation and pyrolysis of a cobalt(Ⅱ) phenanthroline complex on MgO at 800 ℃ in N2. Investigation of various supports (MgO, activated carbon, NaX, NaY, and CaO) showed the choice of support greatly affected the catalytic activity and selectivity toward the target ester product. The Co-N-C/MgO catalyst exhibited excellent performance for the oxidative esterification of furfural to methylfuroate (93.0% conversion and 98.5% selectivity) without the requirement for a basic additive under 0.5 MPa O2 at 100 ℃ for 12 h. However, use of Co-N-C(HCl) as the catalyst produced mainly an acetal as a condensation product. Chloride ions had a negative effect on the oxidative esterification. The inclusion of MgO as a support increases the catalytic efficiency of the cobalt-nitrogen-doped carbon species, and as additive greatly reduces the negative effect of Cl-. Additionally, the oxygen pressure (0.3-1.0 MPa) has a negligible effect on the oxidative esterification of furfural with Co-N-C/MgO. This knowledge could be used to develop non-precious metal catalysts for the oxidative esterification of biomass-derived compounds.