Selective Oxidation of 5-Hydroxymethylfurfural to 2 ... - ACS Publications

Sep 28, 2017 - Chemistry Discipline, Queensland University of Technology, Brisbane, QLD 4001, .... between the reactants and the central metal sites o...
0 downloads 0 Views 1MB Size
Article pubs.acs.org/JACS

Cite This: J. Am. Chem. Soc. 2017, 139, 14775-14782

Selective Oxidation of 5‑Hydroxymethylfurfural to 2,5‑Furandicarboxylic Acid Using O2 and a Photocatalyst of Co‑thioporphyrazine Bonded to g‑C3N4 Shuai Xu,† Peng Zhou,† Zehui Zhang,*,† Changjun Yang,† Bingguang Zhang,† Kejian Deng,*,† Steven Bottle,‡ and Huaiyong Zhu‡ †

Key Laboratory of Catalysis and Materials Sciences of the Ministry of Education, South-Central University for Nationalities, Wuhan, 430074, People’s Republic of China ‡ Chemistry Discipline, Queensland University of Technology, Brisbane, QLD 4001, Australia S Supporting Information *

ABSTRACT: Selective oxidation of 5-hydroxymethylfurfural (HMF) to 2,5-furandicarboxylic acid (FDCA) is one of the key reactions for producing chemical commodities from biomass and their derivatives. The challenge for this reaction is to develop an efficient catalytic process that can be conducted under mild conditions (room temperature and atmospheric pressure, using oxygen molecules in air as the oxidant) and a recyclable catalyst. Herein we report a photocatalyst of cobalt thioporphyrazine (CoPz) dispersed on g-C3N4 (abbreviated as CoPz/g-C3N4), which exhibits excellent catalytic activity toward the selective oxidation of HMF into FDCA under simulated sunlight using oxygen molecules in air as a benign oxidant. For example, an FDCA yield of 96.1% in an aqueous solution at pH = 9.18 is achieved at ambient temperature and air pressure. At lower pH (4.01), the product generated is 2,5-diformylfuran. Hence, it is possible to control the reaction outcome by control of the pH of the reaction system. g-C3N4 itself is not a suitable catalyst for the selective oxidation because under the experimental conditions g-C3N4 generates hydroxyl radicals that initiate processes that oxidize HMF directly to CO2 and H2O. CoPz on the other hand activates O2 to give singlet oxygen (1O2), which more controllably oxidizes HMF to FDCA albeit at a more moderate yield (36.2%). The strong interaction between the CoPz and g-C3N4 in the CoPz/g-C3N4 catalyst is experimentally evidenced, which not only improves accessibility of the CoPz sites and makes the catalyst recyclable but also disables the hydroxyl radical generation by g-C3N4 and promotes 1O2 generation on the CoPz sites, significantly enhancing the catalytic performance. This study demonstrates the potential for efficient non-noble metal photocatalysts for organic transformations driven by sunlight.

1. INTRODUCTION

polyethylene furanoate (PEF). PEF has been considered as a promising alternative to the most widely used polymer for bottle production, polyethylene terephthalate.6,7 However, as illustrated in Scheme 1, the oxidation of HMF can generate several less valued products including 2,5-diformylfuran (DFF) and 5-hydroxymethyl-2-furancarboxylic acid (HFCA) in addition to FDCA.8 Thus, it is a challenge to achieve high FDCA yield from the HMF oxidation using molecular O2.

In recent years, the production of value-added chemicals from renewable resources has attracted increasing attention.1,2 5Hydroxymethylfurfural (HMF), produced via the dehydration of C6 carbohydrates, has been listed as one of the top-12 valueadded chemicals produced from sustainable biosubstances by the United States Department of Energy.3 A variety of highvalue chemicals derived from HMF or its derivates have been produced through various catalytic processes.4,5 Among these chemicals, 2,5-furandicarboxylic acid (FDCA) is regarded as an important monomer substitute for petrochemically derived terephthalic acid in the synthesis of polymers, especially © 2017 American Chemical Society

Received: August 19, 2017 Published: September 28, 2017 14775

DOI: 10.1021/jacs.7b08861 J. Am. Chem. Soc. 2017, 139, 14775−14782

Article

Journal of the American Chemical Society Scheme 1. Schematic Illustration of the Potential Oxidation Products from HMF

biomimetic catalysts that exhibit high redox activity.25 In our previous work, metal thioporphyrazines were observed to demonstrate some unique spectroscopic, electrochemical, and redox properties and were used for the photocatalytic degradation of organic pollutants.26,27 However, as shown in the present study, cobalt thioporphyrazine (CoPz) alone exhibits only moderate photocatalytic activity for the selection oxidation of HMF (40% conversion of HMF). We noted that CoPz molecules are hydrophobic, and in aqueous solution they are prone to aggregate,28 reducing the contact between the CoPz catalyst and reactants and thus lowering the catalyst’s performance. To decrease the aggregation, it is possible to disperse the metal thioporphyrazines on a support. The key issue for this approach is the interaction between the metal thioporphyrazine and the solid support. The strong interaction not only reduces aggregation, but also facilitates recovery of the photocatalyst from the reaction mixture after use. More importantly, the strong interaction involving chemical bonds between metal thioporphyrazine and the support will cause changes in the behavior of electron acceptance and donation between the reactants and the central metal sites of the thioporphyrazine. This behavior plays a key role in activating the reactant on the photocatalyst. Evidently the interaction between the metal thioporphyrazine and support varies with the property of the support. Therefore, the choice of a suitable support could be an effective means to improve the photocatalytic performance of biomimetic catalysts. Graphitic carbon nitride (g-C3N4), as a metal-free semiconductor, possesses advantages such as low toxicity, low cost, high electron-transfer ability, and high thermal stability, and so it has engendered increasing interest as a support to immobilize metal nanoparticles or metal complexes.29−31 Considering that g-C3N4 bears a π-conjugated system as well as an excellent electron-transition ability, in the present study, g-C3N4 was used as a support to immobilize cobalt thioporphyrazine (Figure 1) to generate a new hybrid photocatalyst for the selective oxidation of HMF into FDCA. The photocatalytic activity of the supported cobalt thioporphyrazine is significantly superior to that of the cobalt thioporphyrazine alone. We ascribe superior performance to the interaction between the cobalt thioporphyrazine and support.

Currently the selective oxidation of HMF to FDCA is performed over noble metal catalysts at elevated temperatures (80−150 °C).9 The drawbacks of these thermal catalytic oxidation processes include high cost of the noble metal catalysts, high energy consumption (because of the high reaction temperatures), high oxygen pressures, and the requirement for excessive base, which impedes industrial applications. Therefore, developing routes for the selective oxidation of HMF into FDCA under mild conditions is an important challenge. In recent years, photocatalysis driven by light and using O2 as the oxidant has attracted growing attention in the field of oxidative transformations of organic compounds. As photocatalysis processes usually proceed under mild conditions, they have potential to successfully address this challenge.10−15 For example, many semiconductor photocatalysts have been studied for the oxidation of alcohols to give various carbonyl compounds.16,17 Nonetheless, the photocatalytic oxidation of alcohols with oxygen molecules still suffers from low efficiency and poor selectivity. Efficiently utilizing the photogenerated charges (electrons and holes) is difficult. When species with strong oxidizing power are generated in the photocatalytic system, nonselective processes predominate and overoxidation and/or mineralization often takes place, resulting in poor product selectivity. For example, the photocatalytic oxidation of HMF to DFF in water using TiO2 nanoparticles as catalyst under ultraviolet (UV) irradiation affords a 50% HMF conversion and a low DFF selectivity of 22%.18 This photocatalytic system generates highly active oxidative species, hydroxyl radicals, which destroy the furan ring. To achieve efficient transformation of HMF into FDCA, we have to devise catalytic systems that can generate oxidative species from molecular O2 under ambient temperature and pressure, with appropriate oxidation power to oxidize HMF into furanic compounds such as DFF and FDCA rather than overoxidize to CO2 and H2O. Recently photocatalytic selective oxidation of alcohols using molecular O2 and visible light under mild conditions has been reported,19,20 but noble metal nanoparticles were still required as photocatalysts in these processes. Hence, designing effective photocatalysts without noble metals for the selective oxidation of HMF to FDCA at ambient temperature and pressure using molecular O2 remains a valuable, unrealized goal. Natural enzymes21,22 with high redox activity are able to promote many types of reactions in organisms while exhibiting both high activity and reaction selectivity. The direct utilization of enzymes in chemical reactions is limited by the often short process-life of the enzymes, as these are often unstable under a range of environmental and industrial conditions. Nonetheless, the efficacy of enzyme catalysis has inspired a strategy of developing process catalysts for selective oxidation through mimicking the enzyme active center.23,24 Metalloporphyrins and metallophthalocyanines are two kinds of the most studied

Figure 1. Molecular structure of CoPz. 14776

DOI: 10.1021/jacs.7b08861 J. Am. Chem. Soc. 2017, 139, 14775−14782

Article

Journal of the American Chemical Society

2. RESULTS AND DISCUSSION Strong Interaction between CoPz and g-C3N4 in the Catalyst. The SEM-energy-dispersive X-ray (EDX) elemental mapping analysis of the CoPz/g-C3N4 catalyst is shown in Figure S1. The two characteristic elements of Co and S are evenly dispersed on the surface of the g-C3N4. The sulfur and cobalt weight percentages were 1.99% and 0.45%, respectively. Thioporphyrazine exhibits strong absorption in the UV−vis region due to its conjugated structure. As shown in Figure 2a,

possibly via the axial coordination of nitrogen atoms in g-C3N4 with the central Co2+ metal in CoPz.32 The deposition of a small amount of CoPz on the surface of g-C3N4 caused a remarkable change in the XRD pattern. As shown in Figure 2c, two distinct peaks at 27.5° and 13.1° were observed in the XRD patterns of g-C3N4, which are reflections from the (002) and (100) planes, respectively (JCPDS No. 871526). No reflection peaks are observed in the XRD patterns of CoPz, suggesting that CoPz had no crystalline structure. Nonetheless, the deposition of CoPz on the surface of g-C3N4 weakens the characteristic peak of g-C3N4 at 27.5°, and the other weak peak at 13.1° cannot even be observed from the pattern of the CoPz/g-C3N4 catalyst. The change in the XRD peaks of g-C3N4 reflects that CoPz is well dispersed on the surface of g-C3N4. Furthermore, such changes may also reveal significant structure change. Nitrogen sorption isotherms in Figure S2 show that the CoPz deposition resulted in a substantial decrease in nitrogen adsorption. The specific surface areas of g-C3N4 and CoPz/g-C3N4 were determined from the nitrogen adsorption data to be 74.8 and 14.8 m2/g, respectively. Given that a small amount of CoPz was loaded onto g-C3N4 at mild conditions (room temperature), the significant structure changes indicate a strong interaction between CoPz and gC3N4. The strong interaction between CoPz and g-C3N4 is also confirmed by comparison of the phosphorescence emission spectra of g-C3N4, CoPz, and CoPz/g-C3N4 samples (Figure 2d). A remarkable phosphorescence emission peak at 670 nm is observed for CoPz/g-C3N4 and CoPz. The phosphorescent intensity of CoPz/g-C3N4 is about 6 times stronger than CoPz. In contrast, no phosphorescence emission is observed for the gC3N4 at 670 nm. The phosphorescence emission phenomenon is due to the electron transition from the triplet state to the ground state in the molecules.33 The phosphorescence emission results suggest an electronic transfer interaction between CoPz and g-C3N4. It implies that there exists a strong electron interaction between g-C3N4 and CoPz, which involves the empty d orbital of Co2+ with the free electrons from the πconjugated molecule or the lone-pair electrons of the nitrogen atoms in g-C3N4.34 Photocatalytic Performance of CoPz/g-C3N4. Table 1 shows the results of exploratory experiments of the oxidation of HMF to FDCA with the different materials as the photocatalyst. Control experiments were also conducted to gain insight into the excellent performance of the CoPz/g-C3N4 catalyst in the photocatalytic process. First, the stability of HMF and FDCA under the reaction conditions was studied. The photolytic decomposition of HMF without CoPz/g-C3N4 was negligible (0.1%) after 14 h under the reaction conditions (Table 1, entry 1), and no photolytic decomposition of FDCA was observed either. These results indicate that both the substrate (HMF) and the product (FDCA) are stable under the reaction conditions without catalyst. Although a 99.6% HMF conversion was achieved on g-C3N4 catalyst after 14 h reaction, FDCA and other oxidation products of furan compounds such as DFF, FFCA, and HMFCA were not detected by HPLC (Table 1, entry 2). Total organic carbon (TOC) analysis revealed that the organic carbon in the reaction solution was less than 5%. Therefore, we infer that HMF mainly underwent photocatalytic mineralization to produce CO2 and H2O in the presence of g-C3N4. In contrast to g-C3N4, CoPz exhibited a high FDCA (Table 1, entry 3). Given that the experiments were conducted under identical conditions except for the

Figure 2. UV−vis DR spectra (a), Co 2p XPS spectra (b), XRD patterns (c), and phosphorescence emission spectra of CoPz, g-C3N4, and CoPz/g-C3N4 (d).

there are two characteristic bands peaked at 347 and 639 nm, respectively, in the UV−vis diffuse reflection (DR) spectrum of the CoPz/g-C3N4 catalyst, and the spectrum is similar to that of CoPz. They are ascribed to the electron transition between π and π*. For the latter, the band at short wavelength (the Bband) is ascribed to the electron transition between π and π* of the pyrrole rings in the CoPz molecule,26 while the band at long wavelength (the Q-band) is due to the electron transition between π and π* of CoPz molecules. In comparison to the bands of CoPz, the Q-band in the spectrum of the CoPz/gC3N4 catalyst is red-shifted, while a blue shift is observed for the B-band. Furthermore, the broad characteristic absorption peak at ∼370 nm in the spectrum of bulk g-C3N4 was not present in the spectrum of the CoPz/g-C3N4 catalyst. The changes in absorption bands are attributed to the strong interaction between CoPz and g-C3N4, suggesting that CoPz molecules were firmly bonded on the surface of g-C3N4. Comparison of the UV−vis DR spectra of CoPz, g-C3N4, and CoPz/g-C3N4 also shows that the absorption peaks of the CoPz/g-C3N4 catalyst are mainly from CoPz, and the intensities of the Q- and B-bands of the CoPz/g-C3N4 catalyst are higher than those of the corresponding bands of CoPz. The intense interaction between g-C3N4 and CoPz intensified the absorption of Q- and B-bands of CoPz. XPS spectra in Figure 2b show that the binding energies of Co 2p1/2 and 2p3/2 for CoPz are 795.3 and 780.4 eV, respectively, while they are 794.9 and 780.0 eV for the CoPz/gC3N4 catalyst. The slight changes of Co 2p binding energies can also be attributed to the interaction between CoPz and g-C3N4, 14777

DOI: 10.1021/jacs.7b08861 J. Am. Chem. Soc. 2017, 139, 14775−14782

Article

Journal of the American Chemical Society Table 1. Results of the Oxidation of HMF under Different Conditionsa entry 1 2 3 4 5 6 7 8 9 10

catalyst g-C3N4 CoPz CoPz/g-C3N4 CoPz/g-C3N4 CoPz g-C3N4 CoPz/g-C3N4 g-C3N4 CoPz/g-C3N4

incident light UV−visible UV−visible UV−visible UV−visible dark dark dark UV−visible UV−visible UV−visible

additive

HMF conc (%) 0.1 99.6 40.2 99.1 17.5 11.5

DMPO DMPO 4-chloro-2-nitrophenol

95.0 1.6 11.0

FDCA sel (%)

FDCA yield (%)

0.0

0.0

90.0 97.0 90.0 90.0

36.2 96.1 15.8 10.3

97.0

92.1

97.0

10.6

a

Reaction conditions: HMF (100.8 mg, 0.8 mmol), photocatalyst (5.0 mg), additive (5.0 mg), Na2B4O7 buffer solution (40 mL) at pH = 9.18, air flow rate at 20 mL/min at 1 bar, room temperature, reaction time 14 h, and light intensity of 0.5 W/cm2.

be the strong interaction between CoPz and g-C3N4, which is confirmed by X-ray photoelectron spectroscopy (XPS), UVDRS, and phosphorescence emission. The control experiments in entries 5−7 reveal that CoPz is catalytically active for the oxidation of HMF into FDCA in the dark and thus the active sites in the CoPz/g-C3N4. The light irradiation can significantly enhance the catalyst performance. Effect of the Reaction System pH on Photocatalytic Performance of CoPz/g-C3N4. The thermal-catalytic oxidation of HMF into FDCA is generally performed using noble metal catalysts and an excess of base under demanding conditions such as high oxygen pressure and high temperatures (80−150 °C).9 In comparison, the photocatalytic oxidation of HMF over the CoPz/g-C3N4 catalyst can achieve excellent FDCA yield in a weakly alkaline solution (pH = 9.18). As illustrated in Figure 3(a), HMF conversion and selectivity to FDCA gradually increased during the reaction process and reached 99.1% and 97.0%, respectively, after 14 h. There have been a few reports of the photocatalytic oxidation of HMF into DFF, but low yields were achieved.18,35,36 To the best of our knowledge, the results presented here are the first example of the selective photocatalytic oxidation of HMF into FDCA with an excellent yield under such mild conditions. We also found that the reaction system pH has a critical influence on the product selectivity. At higher pH, the selectivity to FDCA is higher, and the trend for the photocatalytic system is similar to that observed in the oxidation of HMF over noble metal catalysts under heating.9 The high pH is beneficial for the cleavage of carbon−hydrogen bond (C−H) either in the −CH2OH or −CHO groups during the oxidation of HMF into FDCA.9 The results of Figure 3 indicate that (1) DFF is the probable intermediate in the oxidation of HMF to give FDCA and (2) it is possible to select the product by control of the pH of the reaction system. At low pH (4.01) the main product is DFF (the yield was above 80% after 14 h reaction), while at pH = 6.86 or 9.18 the reaction selectively produced FDCA. It was also noted that the reactions at pH = 4.01 and 6.86 were mainly promoted by light irradiation (Figure S3(a and b)). Stability of Heterogeneous Catalysts. It is of great importance, especially for practical applications, that the catalyst is recyclable. As shown in Figure 4, both HMF conversion and FDCA yield were almost the same in the first and second runs. There was a slight decrease in HMF conversions observed in the subsequent runs, but FDCA selectivity remained unchanged. The cobalt content in the liquid phase of the reaction system was below the detection

catalysts used, the striking difference in the product selectivity reveals the possibility that the oxidative species generated on gC3N4 is different from that on CoPz during the oxidation process. The most interesting finding is that the deposition of CoPz on the surface of g-C3N4 greatly enhanced the photocatalytic activity of CoPz. Excellent HMF conversion (99.1%) and selectivity to FDCA (97.0%) were achieved under light irradiation (Table 1, entry 4). In contrast, the HMF conversion in the dark over the CoPz/g-C3N4 catalyst is low (Table 1, entries 5 vs 4; see also the comparison of catalytic activities under light irradiation and dark water, Figure S3c). As mentioned above, the accessibility of the catalytic active sites of CoPz on CoPz/g-C3N4 catalyst for the reactant molecules is much better than that on CoPz due to the prevention of aggregation. The hydrophilicity of g-C3N4 reduces the aggregation of the CoPz/g-C3N4 catalyst in aqueous solution, and the loaded CoPz molecules are dispersed on g-C3N4 and are less aggregated than CoPz without support in the catalytic system. Thus, there should be more CoPz sites accessible to the polar substrate of HMF and the intermediate of DFF in the CoPz/g-C3N4 catalyst than in CoPz itself. The argument is suppored by the results shown in Table S1 that CoPz/g-C3N4 exhibited a stronger absorption for both HMF and the intermediate DFF than CoPz. The effect of the CoPz loading on the catalytic performance of CoPz/g-C3N4 was also investigated. As shown in Figure S4, FDCA yield increased with increasing CoPz loading, reached a maximum, and then decreased. The optimal weight ratio of CoPz to g-C3N4 was 1:10. The dependence of the FDCA yield on the CoPz loading further confirms that CoPz was the active site for the selective oxidation of HMF into FDCA under light irradiation. However, the excessive loading of CoPz resulted in a slight decrease of FDCA yield; CoPz may not be well dispersed on the surface of g-C3N4 when the loading is high. The performance of a metalfree Pz/g-C3N4 sample for the same reaction was also studied; a low HMF conversion of 34.5% and a low selectivity of 52.7% were achieved under the same reaction conditions as listed in entry 4. Hence, the central cobalt atom in CoPz has important contribution to the excellent activity of the CoPz/g-C3N4, possibly by increasing the interaction of CoPz with g-C3N4 or the oxygen molecules as nitrogen or oxygen can interact with its empty d orbitals. Nonetheless, the contribution due to the improved accessibility of the catalytic active site is not sufficient to explain the striking change in the catalytic performance between CoPz and CoPz/g-C3N4. The prevailing reason for the excellent performance of the CoPz/g-C3N4 catalyst should 14778

DOI: 10.1021/jacs.7b08861 J. Am. Chem. Soc. 2017, 139, 14775−14782

Article

Journal of the American Chemical Society

Figure 4. Recycling study for the CoPz/g-C3N4 catalyst. Reaction conditions: HMF (100.8 mg, 0.8 mmol), photocatalyst (5.0 mg), Na2B4O7 buffer solution (40 mL) at pH = 9.18, air flow rate at 20 mL/ min at 1 bar, room temperature, reaction time 14 h, and light intensity of 0.5 W/cm2.

species for the photocatalytic oxidation of HMF with the CoPz/g-C3N4 catalyst. In contrast, the addition of DMPO to the catalytic system with g-C3N4 catalyst resulted in a very low HMF conversion (Table 1, entry 9), validating that HO• radicals are the oxidant in this process. When a scavenger for singlet oxygen (1O2), 4-chloro-2-nitrophenol,39 was added in the reaction system with the CoPz/g-C3N4 catalyst, HMF conversion decreased greatly to 11% (Table 1, entry 10), indicating that 1O2 is the possible oxidant for the selective oxidation of HMF. Electron paramagnetic resonance (EPR) spin-trapping was then used to directly detect either 1O2 species in the reaction system with CoPz/g-C3N4 catalyst or HO• species in the system with g-C3N4 catalyst. 2,2,6,6-Tetramethylpiperidine (TMP)40 was employed as the spin-trapping reagent. The spectrum of the nitroxide produced through the reaction of 1O2 and TMP illuminated with a pulsed laser at pH = 9.18 shows a characteristic 1:1:1 triplet EPR signal (Figure 5). The hydroxyl

Figure 3. Influence of pH of the reaction mixture on the photocatalytic performance: (a) pH = 9.18; (b) pH = 6.86; (c) pH = 4.01. Reaction conditions: HMF (100.8 mg, 0.8 mmol), CoPz/gC3N4 (5.0 mg), buffer solution (40 mL), air flow rate at 20 mL/min at 1 bar, room temperature, reaction time 14 h, and a light intensity of 0.5 W/cm2.

limit of ICP-AES, suggesting that leaching of the CoPz from the CoPz/g-C3N4 catalyst was negligible. The slight decrease of the catalytic activity of the CoPz/g-C3N4 catalyst is possibly due to accumulation of trace insoluble byproducts,37 which blocked the active sites of the catalyst. Identifying Oxidative Species. The photocatalytic oxidation of HMF under different gas atmospheres (N2, O2, and air) was examined (see Table S2), and this indicates that O2 is the oxidant for the selective oxidation of HMF into FDCA. To identify the reactive oxygen species for the selective HMF oxidation with the CoPz/g-C3N4 catalyst, we added dimethylpyridine N-oxide (DMPO), a reasonable scavenger for hydroxyl radical (HO•) in photocatalytic systems.38 However, this scavenger did not influence either the HMF conversion or the FDCA selectivity (Table 1, entries 4 vs 8). Therefore, the HO• radicals are not likely to be involved as the reactive oxygen

Figure 5. EPR signals of (a) the DMPO-HO• adduct in water in the presence of bulk g-C3N4 and (b) the TMP−1O2 adduct in water in the presence of CoPz/g-C3N4 in the photocatalysis process.

radical HO•, on the other hand, was trapped by 5, 5-Dimethyl1-pyrroline (DMPO)to form the adduct of DMPO-OH• in water in the presence of bulk g-C3N4, and a characteristic quartet signal with 1:2:2:1 is observed (Figure 5), demonstrating that hydroxyl radicals were generated on g-C3N4. In addition, the addition of DMPO in the oxidation of HMF using g-C3N4 as a photocatalyst also confirmed that HO• was produced in the presence of g-C3N4 (Table 1, entries 2 vs 9). 14779

DOI: 10.1021/jacs.7b08861 J. Am. Chem. Soc. 2017, 139, 14775−14782

Article

Journal of the American Chemical Society Scheme 2. Possible Mechanism of the Photocatalytic Oxidation of HMF into FDCA with the CoPz/g-C3N4 Catalyst

conversion was observed under 595 ± 5 nm. These results indicated that light absorption of the Q-band is ineffective to induce the reaction, and the photons of higher energy are much more efficient to drive the photocatalytic selective oxidation process than those of lower energy, although the absorption of 400 nm light by the CoPz/g-C3N4 catalyst is less intensive than that at 595 nm. The excitation of the pyrrole rings in the CoPz molecule is critical: the excited CoPz* can activate 3O2 adsorbed on it to 1O2, which is able to selectively oxidize HMF adsorbed on the catalyst into FDCA by the release of energy, and excited CoPz* comes back to its ground state CoPz. In the absence of CoPz, the photogenerated charges by g-C3N4 under light irradiation would directly transform 3O2 to highly reactive HO• radicals that oxidize HMF into CO2 and H2O.

On the basis of the results (Table 1 and Figure 5), we infer that the 1O2 is the reactive species that possesses the appropriate oxidation capacity for the selective oxidation of HMF into FDCA over the CoPz/g-C3N4 catalyst, while the HO• radical has oxidizing power able to destroy the furan ring of HMF, converting HMF into CO2 and H2O (Table 1, entry 2). It has been reported that 1O2 generated from the activation of molecules’ oxygen by photocatalysts is capable of oxidizing alcohols to carbonyl compounds or oxidizing aldehydes to carboxylic acids with quantitative yields.41,42 For example, Cho and co-workers reported a method for the nearly quantitative oxidation of aldehydes to carboxylic acids by 1O2 generated by visible light in the presence of a Ru or Ir photocatalyst.41 Interestingly, combining the results of the catalytic performance and oxidative species, we learn that the interaction between the CoPz and g-C3N4 in the CoPz/g-C3N4 catalyst dampens the ability of g-C3N4 to generate hydroxyl radicals under light irradiation (comparing entry 2 with entry 4), while promoting the 1O2 generation ability of CoPz (comparing entry 3 with entry 4). Thus, the interaction significantly enhances the efficiency of the new photocatalyst. Tentative Mechanism of the Photocatalysis Oxidation Process with CoPz/g-C3N4. A tentative mechanism for the photocatalytic oxidation is depicted in Scheme 2 based on the results we acquired and literature knowledge. In this photocatalytic system, the composite CoPz/g-C3N4 catalyst is able to strongly absorb light (Figure 2a) and the CoPz bonded on gC3N4 converts triplet O2 (3O2) into 1O2 via the energy transit ion. 1O2 species possess suitable oxidation power that can selectively oxidize HMF to the desirable product FDCA. It has been recently reported that the activation of 3O2 to 1O2 is positively correlated with hybrid molecule electronic cloud density.42 As revealed by the results in Figure 2, there exists a strong interaction between CoPz and g-C3N4. The interaction between the empty d orbital of Co2+ with the free electrons from a π-conjugated molecule or the lone-pair electrons of the nitrogen atoms in g-C3N4 would result in a higher electronic cloud density than that in pure CoPz, and thus it could exhibit much higher catalytic activity with a much higher FDCA yield under the same reaction conditions. We also conducted the photocatalytic oxidation of HMF under light irradiation with two selected wavelengths, 400 ± 5 nm and 595 ± 5 nm. As shown in Table S3, a high HMF conversion of 83.2% and an excellent FDCA selectivity of 98.1% were achieved under the irradiation of 400 ± 5 nm wavelength. In contrast, the conversion of HMF was only 2.6% under the 595 ± 5 nm wavelength. In addition, a low HMF conversion of 18.6% with an FDCA selectivity of 90.9% was observed when the reaction is catalyzed by CoPz alone under irradiation of 400 ± 5 nm wavelength, while no detectable

3. CONCLUSIONS This study demonstrates that CoPz/g-C3N4 is a highly efficient photocatalyst for the selective oxidation of HMF to FDCA under simulated sunlight using oxygen molecules in air as a benign oxidant. It can achieve excellent product yield in aqueous solution at ambient temperatures and pressures in neutral and basic conditions. The catalytic system can also produce another selective oxidation product, DFF, in a good yield under slightly acidic conditions. The experimental results indicate there is strong interaction between the CoPz and gC3N4 in the CoPz/g-C3N4 catalyst, which is the key factor for the excellent performance. The interaction can promote 1O2 generation ability of CoPz under light irradiation while dampening the ability of g-C3N4 to generate hydroxyl radicals. 1 O2 species are expected to selectively oxidize HMF into FDCA. The harvested light energy enhances the generation of 1 O2 species from O2 molecules on the catalyst. Moreover, the CoPz/g-C3N4 catalyst can be recycled. This study reveals a new approach to develop efficient photocatalysts from biomimetic catalysts and may also inspire research on utilizing sunlight for production of value-added chemicals from biomass derivatives. 4. EXPERIMENTAL SECTION Materials. All solvents were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). All of the chemicals were purchased from Aladdin Chemicals Co. Ltd. (Beijing, China). Preparation of the CoPz/g-C3N4 Photocatalyst. Cobalt thioporphyrazine with 2,3-bis(buthylthio)maleonitrile (abbreviated as CoPz) was prepared and characterized as previously described.27 g-C3N4 was prepared via the slow pyrolysis of carbamide in air from 30 to 550 °C at a rate of 15 °C/min and kept at 550 °C for 3 h following a method reported in the literature.43 CoPz and g-C3N4 with a mass ratio of 1:10 were dispersed into 50 mL of CH2Cl2, and the mixture was stirred at room temperature for 48 h. Finally, the CoPz/gC3N4 catalyst was collected by the removal of the solvent under 14780

DOI: 10.1021/jacs.7b08861 J. Am. Chem. Soc. 2017, 139, 14775−14782

Journal of the American Chemical Society



reduced pressure. Other types of CoPz/g-C3N4 catalysts with different weight ratios were also prepared in the same way. Characterization of Catalysts. Nitrogen physisorption measurements were conducted at 77 K on a Quantachrome Autosorb-1-C-MS instrument. Specific surface areas of the samples were estimated by using the adsorption data and the BET method. The molecular emission spectra were recorded on a Hitachi Ltd. F-7000 instrument. X-ray powder diffraction (XRD) measurements in a 2θ range from 5° to 70° were conducted on a Bruker Advanced D8 powder diffractometer using Cu Kα at a scanning rate of 0.016°/s. XPS experiments were carried out on a Thermo VG Scientific ESCA MultiLab-2000 spectrometer with a monochromatized Al Kα source (1486.6 eV) at a constant analyzer pass energy of 25 eV. UV−visible absorption and UV−vis DR spectra and the molecular absorption spectra were recorded on a Shimadzu ultraviolet and visible spectrophotometer (Kyoto, Japan). The morphology of the catalyst was characterized by a field emission scanning electron microscope (S4800, Hitachi, Japan) with a voltage of 10 kV. EPR spectra were recorded on a Bruker EMX spectrometer. The EPR experiments were conducted with a center field of 3507.815 G and a frequency of 9.83 GHz by using an Elexsys probe head with 15 mg of the sample placed in a 4 mm tube. Photocatalytic Reactions. The photocatalytic reaction was performed in the bespoke reactor made in our laboratory (Figure S5). In a typical run, HMF (0.8 mmol, 100.8 mg) and 40 mL of Na2B4O7 buffer solution with a pH = 9.18 were charged into a 50 mL quartz flask with a magnetic stirring at the speed of 1000 rpm. The CoPz/g-C3N4 catalyst (5.0 mg) was then dispersed into the reactant solution. Air was bubbled from the bottom of the reactor at a flow rate of 20 mL/min. The photocatalytic reaction commenced when the reactor was irradiated using a Xe light (from Nelson, emitting wavelengths over the range 300−1000 nm). The light intensity was measured to be 0.5 W/cm2. The temperature of the reaction system was carefully controlled at room temperature by circulating cold water. At given irradiation time intervals, 0.5 mL aliquots were collected and then filtered through a Millipore filter (pore size 0.45 μm) to remove the catalyst particulates. The liquid-phase products were analyzed by HPLC. Control experiments were conducted in the dark in a water bath under the same conditions. To avoid the exposure of the reaction mixture to light, the reactor was wrapped with aluminum foil. Potassium biphthalate (C8H5KO4) and Na2B4O7·10H2O buffer solutions were used to maintain the solution pH at 4.01 and 9.18, respectively. Product analysis was conducted on a DIONEX 3000 HPLC instrument. Furan compounds including HMF, DFF, and FDCA were successfully separated by a C18 column (200 × 4.6 mm). The furan compounds were detected by a UV detector at 270 nm. The mobile phase was composed of acetonitrile and water with a volume ratio of 55:45, and the flow rate was set at 1.0 mL/min. The column oven temperature was kept at 25 °C. The retention times of FDCA, HMF, and DFF were 2.2, 3.1, and 3.9 min, respectively. The content of furan compounds was quantified by the external standard method. Action Spectrum Experiments. Light emission diode lamps (Perfect Light, Beijing, China) with two wavelengths (400 ± 5 and 590 ± 5 nm) were used as the light sources. Other experimental procedures were the same as described above. Recycling of Catalyst. After reaction, the catalyst was collected by centrifugation at 1000 rpm to avoid loss and washed by water several times until no FDCA was detected. The spent catalyst was dried in a vacuum oven at 50 °C and reused for the next run under the same conditions. Other cycles were repeated with the same procedure.



Article

AUTHOR INFORMATION

Corresponding Authors

*[email protected] (Z. Zhang) *[email protected] (K. Deng) ORCID

Peng Zhou: 0000-0002-3317-3747 Zehui Zhang: 0000-0003-1711-2191 Kejian Deng: 0000-0003-4070-9076 Steven Bottle: 0000-0003-0436-2044 Huaiyong Zhu: 0000-0002-1790-1599 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We gratefully acknowledge financial support from National Natural Science Foundation of China (Grant Nos. 21203252 and 21272281).



REFERENCES

(1) Besson, M.; Gallezot, P.; Pinel, C. Chem. Rev. 2014, 114, 1827. (2) You, B.; Liu, X.; Jiang, N.; Sun, Y. J. J. Am. Chem. Soc. 2016, 138, 13639. (3) Zhou, P.; Zhang, Z. H. Catal. Sci. Technol. 2016, 6, 3694. (4) Binder, J. B.; Raines, R. T. J. Am. Chem. Soc. 2009, 131, 1979. (5) Teixeira, I. F.; Lo, B. T. W.; Kostetskyy, P.; Stamatakis, M.; Ye, L.; Tang, C. C.; Mpourmpakis, G.; Tsang, S. C. E. Angew. Chem., Int. Ed. 2016, 55, 13061. (6) Sousa, A. F.; Coelho, J. F. J.; Silvestre, A. J. D. Polymer 2016, 98, 129. (7) Kanetaka, Y.; Yamazaki, S.; Kimura, K. Macromolecules 2016, 49, 1252. (8) Dijkman, W. P.; Groothuis, D. E.; Fraaije, M. W. Angew. Chem., Int. Ed. 2014, 53, 6515. (9) Zhang, Z. H.; Deng, K. J. ACS Catal. 2015, 5, 6529. (10) Fava, E.; Millet, A.; Nakajima, M.; Loescher, S.; Rueping, M. Angew. Chem., Int. Ed. 2016, 55, 6776. (11) Senaweera, S.; Weaver, J. D. J. Am. Chem. Soc. 2016, 138, 2520. (12) Hu, X. Q.; Lan, Y.; Xiao, W. J. Nat. Commun. 2016, 7, 11188. (13) Chambers, M. B.; Wang, X.; Ellezam, L.; Ersen, O.; Fontecave, M.; Sanchez, C.; Draznieks, C. J. Am. Chem. Soc. 2017, 139, 8222. (14) Li, H.; Qin, F.; Yang, Z. P.; Cui, X. M.; Wang, J. F.; Zhang, L. Z. J. Am. Chem. Soc. 2017, 139, 3513. (15) Chen, Y. Z.; Wang, Z. U.; Wang, H. W.; Lu, J. L.; Yu, S. H.; Jiang, H. L. J. Am. Chem. Soc. 2017, 139, 2035. (16) Lang, X. J.; Ma, W. H.; Chen, C. C.; Ji, H. W.; Zhao, J. C. Acc. Chem. Res. 2014, 47, 355. (17) Liang, S. J.; Wen, L. R.; Lin, S.; Bi, J. H.; Feng, P. Y.; Fu, X. Z.; Wu, L. Angew. Chem., Int. Ed. 2014, 53, 2951. (18) Yurdakal, S.; Tek, B. S.; Alagoz, O.; Augugliaro, V.; Loddo, V.; Palmisano, G.; Palmisano, L. ACS Sustainable Chem. Eng. 2013, 1, 456. (19) Xiao, Q.; Liu, Z.; Bo, A.; Zavahir, S.; Sarina, S.; Bottle, S.; Riches, J. D.; Zhu, H. J. Am. Chem. Soc. 2015, 137, 1956. (20) Tana, T.; Gao, X. W.; Xiao, Q.; Huang, Y. M.; Sarina, S.; Christopher, P.; Jia, J. F.; Zhu, H. Y. Chem. Commun. 2016, 52, 11567. (21) Mahanta, N.; Zhang, Z. A.; Hudson, G. A.; van der Donk, W. A.; Mitchell, D. A. J. Am. Chem. Soc. 2017, 139, 4310. (22) Wolf, F.; Bauer, J. S.; Bendel, T. M.; Kulik, A.; Kalinowski, J.; Gross, H.; Kaysser, L. Angew. Chem., Int. Ed. 2017, 56, 6665. (23) Park, S. Y.; Hwang, I. S.; Lee, H. J.; Song, C. E. Nat. Commun. 2017, 8, 14877. (24) Wang, L.; Zhang, J. Y.; Kim, B.; Peng, J. J.; Berry, S. N.; Ni, Y.; Su, D. D.; Lee, J.; Yuan, L.; Chang, Y. T. J. Am. Chem. Soc. 2016, 138, 10394. (25) Shanmugam, S.; Xu, J. T.; Boyer, C. J. Am. Chem. Soc. 2015, 137, 9174.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.7b08861. Supplementary Tables S1−S3, Figures S1−S7 (PDF) 14781

DOI: 10.1021/jacs.7b08861 J. Am. Chem. Soc. 2017, 139, 14775−14782

Article

Journal of the American Chemical Society (26) Zhang, Z. H.; Zhang, M. J.; Deng, J.; Deng, K. J.; Zhang, B. G.; Lv, K. L.; Sun, J.; Chen, L. Q. Appl. Catal., B 2013, 132, 90. (27) Zhou, Q.; Yang, C. J.; Deng, K. J. Appl. Catal., B 2016, 192, 108. (28) Cui, L. Y.; Tokarz, D.; Cisek, R.; Chen, J.; Barzda, V.; Zheng, G. Angew. Chem., Int. Ed. 2015, 54, 13928. (29) Gao, G. P.; Jiao, Y.; Waclawik, E. R.; Du, A. J. J. Am. Chem. Soc. 2016, 138, 6292. (30) Kuriki, R.; Yamamoto, M.; Higuchi, K.; Yamamoto, Y.; Akatsuka, M.; Lu, D.; Yagi, S.; Yoshida, T.; Ishitani, O.; Maeda, K. Angew. Chem., Int. Ed. 2017, 56, 4867. (31) Ju, E. G.; Dong, K.; Chen, Z. W.; Liu, Z.; Liu, C. Q.; Huang, Y. Y.; Wang, Z. Z.; Pu, F.; Ren, J. S.; Qu, X. G. Angew. Chem., Int. Ed. 2016, 55, 11467. (32) Cao, R.; Thapa, R.; Kim, H.; Xu, X.; Kim, M. G.; Li, Q.; Park, N.; Liu, M. L.; Cho, J. Nat. Commun. 2013, 4, 2076. (33) Franz1, K. A.; Kehr1, W. G.; Siggel, A.; Wieczoreck, J.; Adam, W. Ullmann’s Encyclopedia of Industrial Chemistry; Wiley-VCH, 2000. (34) Ong, W. J.; Tan, L. L.; Ng, Y. H.; Yong, S. K.; Chai, S. P. Chem. Rev. 2016, 116, 7159. (35) Krivtsov, I.; Garcia-Lopez, E. I.; Marci, G.; Palmisano, L.; Amghouz, Z.; Garcia, J. R.; Ordonez, S.; Diaz, E. Appl. Catal., B 2017, 204, 41. (36) Zhang, H. L.; Wu, Q.; Guo, C.; Wu, Y.; Wu, T. H. ACS Sustainable Chem. Eng. 2017, 5, 3517. (37) Maiti, N. C.; Mazumdar, S.; Periasamy, N. J. Phys. Chem. B 1998, 102, 1528. (38) Ning, X. F.; Meng, S. G.; Fu, X. L.; Ye, X. J.; Chen, S. F. Green Chem. 2016, 18, 3628. (39) Wan, Y. X.; Xie, Y. B.; Sun, H. Q.; Xiao, J. D.; Cao, H. B.; Wang, S. B. Catal. Sci. Technol. 2016, 6, 2918. (40) Zheng, Y. Z.; Zhang, D. K.; Shah, S. N. A.; Li, H. F.; Lin, J. M. Chem. Commun. 2017, 53, 5657. (41) Iqbal, N.; Choi, S.; You, Y.; Cho, E. J. Tetrahedron Lett. 2013, 54, 6222. (42) Chen, Y. Z.; Wang, Z. U.; Wang, H. W.; Lu, J. L.; Yu, S. H.; Jiang, H. L. J. Am. Chem. Soc. 2017, 139, 2035. (43) Martin, D. J.; Qiu, K. P.; Shevlin, S. A.; Handoko, A. D.; Chen, X. W.; Guo, Z. X.; Tang, J. W. Angew. Chem., Int. Ed. 2014, 53, 9240.

14782

DOI: 10.1021/jacs.7b08861 J. Am. Chem. Soc. 2017, 139, 14775−14782