Aerobic oxidation of cyclohexane catalyzed by graphene oxide: Effects of surface structure and functionalization

Molecular Catalysis 431 (2017) 1–8

Contents lists available at ScienceDirect

Molecular Catalysis journal homepage: www.elsevier.com/locate/mcat

Aerobic oxidation of cyclohexane catalyzed by graphene oxide: Effects of surface structure and functionalization Yepeng Xiao, Jincheng Liu ∗ , Kaihong Xie, Weibin Wang, Yanxiong Fang ∗ School of Chemical Engineering and Light Industry, Guangdong University of Technology, Guangzhou 510006, China

a r t i c l e

i n f o

Article history: Received 27 June 2016 Received in revised form 12 January 2017 Accepted 13 January 2017 Keywords: Graphene oxide Nano graphene oxide Reduction Cyclohexane oxidation Surface functionalization

a b s t r a c t Despite extensive development efforts on new catalysts for cyclohexane oxidation, current commercial processes are still desirable to further decrease the economic and environmental costs in the oxidation procedures. In this study, the catalyst properties of metal-free functionalized graphene oxide (GO) in the selective oxidation of cyclohexane is investigated to understand the effects of structure, functional groups and oxygen-containing groups. Both GO sheets and nano graphene oxide (NGO) sheets possess high cyclohexane conversions of 16.7% and 17.7% comparable to that of metal catalysts in the aerobic oxidation of cyclohexane. The higher catalytic activity of NGO sheets is caused by the smaller size and the higher content of −COOH. Within the oxygen-containing groups, the carboxylic acid groups on the surfaces of GO are supposed as the active site for the generation of • O2 − radical, which is the active oxygen species in this catalytic oxidation process. © 2017 Elsevier B.V. All rights reserved.

1. Introduction Catalytic oxidation of cyclohexane (CyH) to a mixture of cyclohexanone and cyclohexanol (KA oil), which is one of the most challenging reactions among partial oxidation of the relatively inert C-H bond of alkanes [1,2], becomes increasingly important due to the great need of these products as precursors in the nylon-6 and nylon-6,6 polymers synthesis [3]. Recyclable catalysts with high stablity and activity are still highly desirable to further decrease the economic and environmental costs in the oxidation processes. The ideal catalytic oxidation process would use a recyclable and inexpensive catalyst in solvent-free or nontoxic solvents with good conversion and selectivity in the dry air or molecular oxygen as the only oxygen source. Recently, the transitional metal catalysts on the large surface supports or porphyrins are widely utilized in the catalytic cyclohexane oxidation [4–7]. However, drawbacks such as unsatisfactory activities, harsh reaction conditions, sensitivity to air or water, the high cost and the detrimental effects of toxic metal significantly limited its further applications with these metal-based catalysts [8,9]. Hence, it is highly desirable to design a recyclable catalyst that is cheap, metal- and solvent-free, easy-to-handle and highly efficient for the catalytic oxidation of cyclohexane.

∗ Corresponding authors. E-mail addresses: [email protected] (J. Liu), [email protected] (Y. Fang). http://dx.doi.org/10.1016/j.mcat.2017.01.020 2468-8231/© 2017 Elsevier B.V. All rights reserved.

Notably, carbon-based materials are considered as promising catalysts because of their low cost, environmental friendliness, high catalytic selectivity and good durability. Carbon or carbonbased materials such as graphene (G) [10,11], carbon nanotubes (CNTs) [12,13], carbon quantum dots (CQDs) [14] and graphitic carbon nitride (g-C3 N4 ) [8,15] have received increasing attention as metal-free heterogeneous catalysts owning to their environmental compatibility, outstanding thermal and chemical stability and large surface area. Moreover, carbon materials normally exhibit significant advantages in activity by structure modification [11] (e.g. heteroatom doping) and surface functionalization [16]. In particular, dopants in the carbon skeleton play important roles in tuning the electronic characteristics. The enhancement of oxidation of saturated C-H bonds catalytic activity by doping could be summarized that the dopants break the electroneutrality of sp2 carbon and induce changes of both atomic charge and spin density to create charged sites favorable for facilitating the adsorption of O2 and reactive intermediates no matter the dopants are electron-rich or electron-deficient [17]. On the other hand, it is well-known that the functionalization of carbon or carbon-based materials makes them excellent supports to anchor inorganic nanocrystals with good dispersion and stability [18]. Besides, surface functional groups play significant roles in some reactions. In gas-phase oxidative dehydrogenation (ODH) reactions, the catalytic cycle starts with the activation and dehydrogenation of hydrocarbons at quinone sites, and ends with the re-activation of carbon by oxygen [13]. Carboxylic acid groups on the surfaces of carbon catalysts are identified

2

Y. Xiao et al. / Molecular Catalysis 431 (2017) 1–8

as the active sites for the wet air oxidation of phenol [19]. However, so far, the mechanistic reason of carbon catalysis in most liquidphase reactions has not been rationalized in liquid-phase oxidative catalysis clearly. In this work, the effect of the oxygen-containing groups of graphene oxide (GO) in the catalytic oxidation of cyclohexane has been discovered during the investigation on the catalytic activity of GO and GO with various surface modifications. Nano-GO (NGO) sheet behaves surpassing activity than that from the big GO sheets. An increased conversion of cyclohexane has been obtained with a higher oxygen-containing GO, and reduced GO (r-GO) has inferior activity. Moreover, the carboxylic acid functionalization seems to benefit the reaction, but a sulfonic acid functionalized GO (GO-SO3 H) shows not much different compare to the counterpart of r-GO. It is implied that a certain mount and a certain type of oxygen-containing of GO plays a indispensable role in the catalytic oxidation of cyclohexane. 2. Experimental 2.1. Materials Natural graphite was purchased from Bay Carbon Company, USA. Sodium nitrate (NaNO3 , 99%), potassium permanganate (KMnO4 , 99%), hydrogen peroxide (H2 O2 , 35%), concentrated sulfuric acid (98%), acetone, ClCH2 COONa, ClCH2 CH2 SO3 Na, triphenylphosphine (pph3 , 99%) and benzoyl peroxide (BP) were purchased from Aladdin, China. All reagents were used without further purification. 2.2. Synthesis of GO, NGO, GO-COOH, GO-SO3 H and r-GO GO was synthesized by a modified Hummers method [20]. The GO-COOH and GO-SO3 H was produced by a bath sonication process [21,22]. Briefly, 20 mg of GO, 1000 mg of ClCH2 COONa (or BrCH2 CH2 SO3 Na) and a certain amount of NaOH were added into 50 mL of DI water, and the mixed solution was sonicated for 0.5, 1, 2, 5 h, respectively. The as-prepared GO-COOH (or GO-SO3 H) was neutralized by HNO3 and purified by DI water twice. The r-GO was produced by a simple bath sonication process. Typically, 20 mg of GO with 100, 200 and 400 mg of NaOH were added into 50 mL of DI water, and the mixed solution was sonicated for 0.5, 1, 2 and 5 h (marked as r-GO-100 mg-0.5 h etc.), respectively. The as-prepared r-GO was neutralized by HNO3 and purified by DI water twice. Meanwhile, NGO sheets prepared by refluxing 40 mg of GO in 5 mol/L solution of HNO3 for 72 h was used for comparison. 2.3. Characterization Transmission electron microscopy (TEM) images were obtained using a JEM-2100 microscope. Atomic force microscopy (AFM) was carried out using a non-contact mode on a PSIA XE-100 scanning probe microscope. X-ray powder diffraction (XRD) patterns were operated on a Ultima III using Cu K␣ irradiation. Fourier transform infrared spectroscopy (FTIR) spectra were recorded on a Perkin Elmer GX FT-IR system. X-ray photoelectron spectroscopy (XPS) measurements were carried out by using a Kratos Axis Ultra Spectrometer. The UV–vis absorption spectra were recorded by using an Evolution 300 spectrophotometer. 2.4. Typical oxidation procedures The cyclohexane oxidation reaction was performed in a Teflonlined 100 mL stainless-steel autoclave equipped with a magnetic stirrer. 43.5 mmol of cyclohexane and 20 mg of catalyst which was

Table 1 Effect of residual impurities on CyH oxidation. Catalyst.

Conversion (%)

K/Ac

GO DI-GOa GO (0.5 wt%Mn)b

16.7 16.6 16.2

1.4 1.4 1.4

Selectivityd (%) KA

CHHP

AA

Others

45.2 44.0 44.1

3.0 2.2 2.4

46.3 48.0 47.7

5.5 5.8 5.8

a DI-GO: GO after an ion exchange processing in DI water at room temperature for 7 days. b As-prepared GO acetonic suspensions with 0.5 mg MnSO4 . c Molar ratio of cyclohexanone to cyclohexanol. d Selectivity of major products after 8 h reaction. KA: cyclohexanone and cyclohexanol, CHHP: cyclohexyl hydroperoxide, AA: adipic acid, the others included glutaric acid, succinic acid, hydroxyl hexylic acid and ␧-caprolactone and other undetected acids.

dispersed in 5 mL of acetone were introduced into the reactor without any initiator. The optimum temperature was set at 140 ◦ C with the air pressure 1.5 MPa initially. After 8 h reaction, the autoclave was removed from oil bath to an ice-water bath and maintained for about 1 h, which was necessary for the system to cool down completely. For analyzing the products of cyclohexane oxidation, FID gas chromatography (GC, Agilent-7890A, capillary column: HP5, 30 m × 0.25 mm, nitrogen as carrier gas with internal standard method, using cyclopentanol as the standard substance) was used for the qualitative analysis. The conversion was calculated based on the starting cyclohexane. Cyclohexyl hydroperoxide (CHHP) contents were determined by decomposition with triphenylphosphine (PPh3 ) and quantification of the additionally formed cyclohexanol by GC. Other products including acids and esters were determined by HPLC (Agilent 1260). Additionally, about 95% mass balance was obtained for the reaction based on the initial feeding amount of cyclohexane.

3. Results and discussions 3.1. Characterization The hydrophilic GO sheets produced by Hummers method have the average size of 1–2 ␮m (Fig. 1a). The as-prepared GO can be dispersed well in de-ionized (DI) water, forming a stable yellow-brown suspension (Fig. 1a, inset). Fig. 1b shows the UV–vis absorption spectra of GO, NGO, GO-COOH and GO-SO3 H. The absorption peak at 227 nm corresponding to ␲-␲* transitions of aromatic C C bonds [21]. However, the absorption peaks of GOCOOH and GO-SO3 H sheets show a slight red-shift due to the partial removal of oxygen functional groups during the functional processing [23]. As shown in Fig. 1c, the C, O and S photoelectron lines can be detected in the XPS survey spectra of GO. The TEM, UV–vis absorption and XPS analyses confirm the successful synthesis of high-quality GO, GO-COOH and GO-SO3 H sheets. One of the key concerns in identifying and describing metal-free catalysis is the suspected influence of metal impurities in the catalyst. In order to eliminate the catalytic influence of residual metal ions, all the materials have been suffered to a dialysis process for 1 week before the oxidation experiments. The possible impacts of Mn2+ or other impurities that may be introduced during the preparation of GO were investigated by a contrast experiment as Table 1. It can be seen that, the conversion of cyclohexane and the distribution of the reaction products have no evident difference among GO, DI-GO and GO with the addition of 0.5 wt% manganese (Mn) salt. This suggests that the residual Mn metals are not the active site or have negligible effect in the CyH oxidation.

Y. Xiao et al. / Molecular Catalysis 431 (2017) 1–8

3

3.2. The effect of oxygen content of GO on the catalytic activity The crystalline structures of GO, r-GO-100mg–1 h, r-GO100mg–5 h and r-GO-400mg–5 h are evaluated by XRD. As shown in Fig. 2a, the diffraction peak position at 11◦ can be attributed to the (001) interlayer spacing of 7.43 Å in the solid GO stacks. However, as the consequence of prolongation of sonication time as well as the weight increment of NaOH, diffraction peak at 22.9◦ becomes stronger and sharper, which corresponds to the diffraction of the (002) plane with an interlayer spacing of 3.36 Å due to the removal of oxygen functional groups [24,25]. This indicates that GO was partially deoxygenated under 100 mg of NaOH conditions with 5 h sonication time, and a deeper reduction degree could be obtained by enhancing the mass of NaOH. The FT-IR spectra are obtained in order to characterize the oxygen-containing functional groups on the GO after reduction. As shown in Fig. 2b, a broad peak at 3300 cm−1 indicates the presence of −OH group. The C O and C O stretching vibration bands of −COOH groups are at 1712 and 1041 cm−1 , and the peak centered at 1345 cm−1 is corresponding to the in-plane −OH bending mode. After sonication, a weak shoulder peak at 1580 cm−1 was present, which is caused by the C C aromatic ring stretching [26]. Furthermore, as shown in Fig. 2c, the oxygen group contents at the surface of all modified GO sheets can be determined by the Boehm titration method [27]. Obviously, the contents of three surface oxygen groups (−COOH, −O− and −OH) were gradually decreased. Both the FT-IR spectrum analysis and the Boehm titration results indicate that the efficient reduction and deoxygenation of GO have occurred during the sonication process in the NaOH solutions. The effect of oxygen contents in GO sheets on cyclohexane oxidation is shown in Fig. 2d. A clear decline in the conversion of cyclohexane could be observed ranging from 15.2% to 7.4% due to the reduction of GO sheets, the distribution of products shows a slight change. With a lower oxygen content and the ensued restoration of the ␲-conjugation net work in GO sheets, less AA and other by-products have been generated, and lower level of CHHP could be detected, which may demonstrates the superior capability to decompose the peroxide by more intact sp2 system. This phenomenon also had been observed by Peng’s group using CNTs as the catalyst for the cyclohexane oxidation [13]. In a conclusion, GO with higher content of oxygen-containing groups facilitates the enhancement of cyclohexane conversion, and more oxidation products would be produced as a consequence. 3.3. The effect of types of oxygen-containing groups on the catalytic activity

Fig. 1. (a) TEM image of GO; (b) UV–vis absorption spectra of GO, NGO, GO-COOH and GO-SO3 H; (c) XPS survey spectrum of GO, NGO, GO-COOH and GO-SO3 H.

The oxygen contents in GO sheets have important impact on the cyclohexane oxidation as we discussed earlier. However, there are several types of oxygen-containing groups such as hydroxyl groups (−OH), carboxylic acid groups (−COOH) and oxygen epoxide groups (−O−) on the surface of GO sheets. So the question is which oxygencontaining groups plays the most important roles in the catalytic oxidation of cyclohexane? In order to address this question clearly, the carboxylated GO sheets have been applied in the aerobic oxidation of cyclohexane. Line shape analysis results of the XPS C1s spectrum for the GO-COOH sheets after carboxylation treatment for 0.5 and 5 h are shown in Fig. 3a and b. It is clearly that the lower total amount of oxygen-containing groups and higher ratio of carboxylic acid groups to oxygen-containing groups in GO-COOH–5 h can be obtained than that of GO-COOH-0.5 h. IR spectra of GO-COOH with various carboxylation times are depicted in Fig. 3c. The carbonyl peak at 1712 cm−1 and a broad peak of hydroxyl groups at 3300 cm−1 can be seen in the FT-IR spectrum of GO-COOH. The Boehm titration further confirms the above inference. As shown in Fig. 3d, longer carboxylation time leads to a decreased total

4

Y. Xiao et al. / Molecular Catalysis 431 (2017) 1–8

Fig. 2. (a) XRD patterns of GO, r-GO-100mg–1 h, r-GO-100mg–5 h and r-GO-400mg–5 h; (b) IR spectra of GO, r-GO-100mg–1 h, r-GO-100mg–5 h and r-GO-400mg–5 h; (c) Boehm titration results of GO, r-GO-100mg–1 h, r-GO-100mg–5 h and r-GO-400mg–5 h; (d) The effect of oxygen content of GO on CyH oxidation.

content of oxygen-containing groups, but the amount of −COOH is improved from 0.7 mmol g−1 to 1.1 mmol g−1 with the increased carboxylation time from 0.5 to 5 h. The effect of carboxylic acid group content on the catalytic activity is shown in Fig. 3e. Although the increased reduction degree by longer ultrasonic treatment may decrease the catalytic activity of GO, the improvement of the carboxylation degree seems to reverse the decline. With the improved level of carboxylation in GO sheets, a dramatic increased activity is obtained. This indicates that −COOH groups play very important roles in the catalytic reaction. To further confirm this regulation, excess mount of oleylamine is added to the as-prepared GO sheets for the formation of oleylamine-carboxylic salt (R-NH3 + COO− -GO), which will shield the catalytic effect of carboxylic acid group (Supporting information Fig. S1) [28,29]. Due to the lack of −COOH groups, only 5.9% CyH conversion is obtained (Supporting information Table S1), which is significantly dropped down compared to 16.7% from GO sheets. All the above experiments point out the indispensable role of the −COOH in the catalytic oxidation of cyclohexane. Sulfonation of carbon materials have been reported for many oxidation reactions. Sulfonic acid groups (-SO3 H) on the surface of carbon materials can react with hydrogen peroxide to form the corresponding peroxy-sulfides which can further oxidize substrates to generate products [30–32]. We also investigate the effect of sulfonation on cyclohexane oxidation. GO-SO3 H is synthesized as the same way as the carboxylation of GO. Typically, 20 mg of GO, 100 mg of NaOH and 1000 mg of BrCH2 CH2 SO3 Na are dissolved in 50 mL DI water and sonicated for 0.5, 1, 2 and 5 h. However, GOSO3 H does not perform as good as our expectation. The GO-SO3 H sheets show the same low activity as the r-GO sheets (Supporting information Table S2), indicating that the −SO3 H groups have no effect in the catalytic oxidation of cyclohexane.

3.4. The effect of size of GO sheets on the catalytic activity AFM is used to verify the layer numbers and the size of synthesized GO and NGO. As shown in Fig. 4a, the thickness of the GO obtained is about 1.1 nm, matching well with the reported apparent thickness of GO reported by the literatures [33]. As a result of refluxing in 5 mol L−1 solution of HNO3 for 72 h, the size of GO sheets has decreased significantly from 1–2 ␮m to about 30–50 nm (Fig. 4b) [20]. According to the line shape analysis for the C1s XPS spectra of GO and NGO in Fig. 5a and b, the content of carboxylic acid in the surface states of GO was slightly changed after HNO3 refluxing process. Since −COOH groups are manily on the verge of GO sheets, the NGO sheets have more verges and an accordingly higher content of −COOH groups, which has a good agreement with the Boehm titration in Fig. 5c. Fig. 5d shows the catalytic activity of GO and NGO. It is interesting to find that NGO has a slightly advantage in the catalytic oxidation of cyclohexane comparing to the as-prepared large GO sheets. This phenomenon can be explained by the increased −COOH mounts at the open edges of NGO. The feature that distinguishes graphene from other sp2 carbon family members is the presence of open edges, especially when the size of the graphene is at the nanoscale. The edges have a different electronic state from the bulk due to a localized state and defects at the edges [34]. Below 600 ◦ C, these defect sites or edges of graphene have been reported to convert O2 molecules to electrophilic oxygen species, such as superoxide O2 − or peroxide O2 2− [35], which could attack the electron-rich C-H bonds in cyclohexane. Since GO sheets are further oxidized and broken into much smaller NGO sheets due to the HNO3 -refluxing process, the stronger localized state is and more created defects at the edges can be gained. As a result, the activation of cyclohexane is facilitated, and the improved catalytic activity could be achieved.

Y. Xiao et al. / Molecular Catalysis 431 (2017) 1–8

5

Fig. 3. Line shape analysis for the C1 s XPS spectra of (a) GO-COOH-0.5 h and (b) GO-COOH–5 h; (c) IR spectra of GO-COOH (GO with 500 mg NaOH and 1000 mg ClCH2 COONa in 50 mL DI water sonicates for 0.5, 1, 2 and 5 h, respectively); (d) Boehm titration results of GO-COOH-0.5 h, GO-COOH–2 h and GO-COOH-0.5 h; (e) The effect of carboxyl acid group content on CyH oxidation.

3.5. Recycling test The performance of recyclability is studied by repeated cyclohexane oxidations. Fig. 6a displays the stability and possible reusability for selective oxidation of cyclohexane based on asprepared GO and NGO. About 20% and 13% decrease of cyclohexane conversion are observed during the 5-cycling tests with GO and NGO, respectively. This indicates that the NGO sheets shows better recycling stability than that of GO sheets. In order to investigate the reason for the slight decrease of catalytic activity, the FTIR as well as UV–vis absorption spectra are involved to explore the reasons for the deactivation. From Fig. 6b, the characteristic peaks corresponding to C O C and C OH decreased obviously during the recycling tests, indicating that oxygen content of GO has decreased. Furthermore, UV–vis absorption also provides strong evidence of the reduction. Fig. 6c shows that aromatic C C bonds gradually red shift to 265 nm, which is indicative of the restoration of the

␲-conjugation net work within the graphene nanosheets. All the results confirm that reduction of GO has occurred during the recycling test. In addition, we use a GC-TCD to detect if there is any additional water formed in the case GO has been reduced. But compared to the blank experiment without catalysts, the variation of water content is not apparent in the catalytic process. 124.4 mg of water in the reaction system can be obtained for the blank experiment, and 95–135 mg of water is generated from the catalytic process (Table S3). So we can infer that GO and NGO is just slightly reduced in each cycle test, which explained the slight decrease of activity during each catalytic cycle.

3.6. Proposed mechanism study In this work, trapping experiments of the reactive species are carried out with three different radical scavengers of tert-butanol (a • OH radical scavenger), benzquinamide (BQ, a • O2 − radical

6

Y. Xiao et al. / Molecular Catalysis 431 (2017) 1–8

Fig. 4. AFM image of (a) GO and (b) NGO.

Fig. 5. Line shape analysis for the C1s XPS spectra of (a) GO; (b) NGO; (c) Boehm titration results of GO-COOH-0.5 h, GO-COOH–2 h and GO-COOH-0.5 h; (d) the effect of size of GO sheets on CyH oxidation.

scavenger) and disodium ethylenediaminetetraacetate (Na2 -EDTA, a hole scavenger) [14]. Table 2 gives the results after the addition of scavengers. It is shown that 2.3% conversion rate of cyclohexane could be obtained in the blank experiment. The presence of benzquinamide substantially suppressed the catalytic activity for the oxidation of cyclohexane from 16.7% to 1.5% after 8 h reaction, while the introduction of tert-butanol and Na2 -EDTA barely have influence on the catalytic reaction results. These experiments confirm that the superoxide radicals (• O2 − ) are the active oxygen species

in this catalytic oxidation process, which is similar to that of the cyclohexane oxidation with CNTs as the catalyst [12]. In order to investigate the important role of intact sp2 system in GO sheets in the catalytic oxidation of cyclohexane, a catalytic test of peroxide decomposition was carried out using benzoyl peroxide (BP) as a model compound. Fig. 7a shows the percentage BP consumed within 1 h at 90 ◦ C in the presence of GOs. Without any catalyst, only about 14% BP was decomposed in 1 h. With the addition of GO sheets, the BP consumption is significantly increased to

Y. Xiao et al. / Molecular Catalysis 431 (2017) 1–8

7

Table 2 Effects of GO, tert-butanol, BQ and Na2 -EDTA on the CyH conversion. Entry

GO

tert-butanol

Na2 -EDTA

BQ

Conversion (%)

1 2 3 4 5

×a    

× ×  × ×

× × ×  ×

× × × × 

2.3 16.7 16.9 16.6 1.5

a

: added; ×: not added.

Fig. 7. (a) Effect of GOs with different oxygen contents on the thermolysis of benzoyl peroxide; (b) A possible mechanism of the CyH oxidation with GOs as the catalyst.

Fig. 6. (a) Recycling test of GO in liquid-phase CyH oxidation; (b) FT-IR spectra of GO, GO after 3-cycle use and GO after 5-cycle use; (c) UV–vis spectra of GO, GO after 3-cycle use and GO after 5-cycle use.

52%, indicating that the superior catalytic capacity of GO sheets in the peroxide decomposition. With the increased −COOH content in the GO-COOH sheets, the BP consumption is further improved to 71%, which could be attributed to the accelerated conjugated ␲-electrons transfer process in a intact sp2 system within slightly reduced GO-COOH sheets and the speeding up of peroxide decomposition reaction. According to the previous reports, carboxylic acid groups on the surfaces of carbon catalysts are identified as the active sites for many hydrocarbon oxygenation such as dehydration of alcohols, wet air oxidation (WAO) of phenol and oxidative dehydrogenation (ODH) of alkane [19,35]. Even an enhanced conversion of cyclohexane could be observed in the case of using acetic acid as the solvent [36]. Therefore, the conclusions can be summarized as: (1) The carboxylic groups could effectively improve the formation of the radicals, and the higher content of the carboxylic groups in the GO is in favor of producing • O2 − , (2) The intact sp2 system within GO sheets plays a very important role in the peroxide decomposition. Combing with the work we have done above, we propose a

8

Y. Xiao et al. / Molecular Catalysis 431 (2017) 1–8

possible reaction mechanism in the cyclohexane oxidation using GO as the catalyst. As shown in Fig. 7b, O2 in the cyclohexane adsorbs on the surface of GO, and then the carboxylic groups (COOH) on GO and O2 could produce • O2 − by the hydrogen band. The superoxide radicals can arouse radical chain reactions, leading to a hydrogen abstraction, which is a significant step for chain propagation in the cyclohexane oxidation. After that, the as-generated CHHP could be decomposed by intact sp2 system within GO sheets to produce the KA oil and AA. 4. Conclusion The modified and functionalized GO sheets show high catalytic activity in the oxidation of cyclohexane under the air atmosphere. GO with higher oxygen contents and smaller size presents higher activity in the cyclohexane oxidation. The carboxylic acid groups on the surface of GO sheets are supposed as the active site to convert O2 into • O2 − , which could abstract ·H radical from cyclohexane by hydrogen abstraction. The highest activity reaches 17.7% in the catalytic oxidation of cyclohexane catalyzed by the NGO sheets. The NGO sheets perform better stability than that of GO sheets with only 13% decrease of cyclohexane conversions after 5 cycling catalytic experiments. The performance of NGO sheets in our study points out the indispensable role of the oxygen-containing groups and carboxylic acid groups, which is interesting and helpful for the better design of metal-free catalysts for the oxidation of cyclohexane. Acknowledgments The authors would like to acknowledge National Natural Science Foundation of China (21547014), National Natural Science Foundation of China (21276052) and National Key Technology Support Program (2015BAK44B01) support for this work. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.mcat.2017.01. 020. References [1] D.J. Xiao, J. Oktawiec, P.J. Milner, J.R. Long, J. Am. Chem. Soc. 138 (2016) 14371–14379. [2] Y. Wang, X. Wen, C. Rong, S. Tang, W. Wu, C. Zhang, Y. Liu, Z. Fu, J. Mol. Catal. A Chem. 411 (2016) 103–109.

[3] S. Khare, P. Shrivastava, J. Mol. Catal. A Chem. 411 (2016) 279–289. [4] W.-J. Zhou, R. Wischert, K. Xue, Y.-T. Zheng, B. Albela, L. Bonneviot, J.-M. Clacens, F. De Campo, M. Pera-Titus, P. Wu, ACS Catal. 4 (2014) 53–62. [5] X. Fang, Z. Yin, H. Wang, J. Li, X. Liang, J. Kang, B. He, J. Catal. 329 (2015) 187–194. [6] J.-C. Wang, F.-W. Ding, J.-P. Ma, Q.-K. Liu, J.-Y. Cheng, Y.-B. Dong, Inorg. Chem. 54 (2015) 10865–10872. [7] T. Wang, Y. She, H. Fu, H. Li, Catal. Today 264 (2016) 185–190. [8] X.-H. Li, J.-S. Chen, X. Wang, J. Sun, M. Antonietti, J. Am. Chem. Soc. 133 (2011) 8074–8077. [9] D.-H. Lan, L. Chen, C.-T. Au, S.-F. Yin, Carbon 93 (2015) 22–31. [10] L. Qu, Y. Liu, J.-B. Baek, L. Dai, ACS Nano 4 (2010) 1321–1326. [11] S. Dou, A. Shen, Z. Ma, J. Wu, L. Tao, S. Wang, J. Electroanal. Chem. 753 (2015) 21–27. [12] H. Yu, F. Peng, J. Tan, X. Hu, H. Wang, J. Yang, W. Zheng, Angew. Chem. Int. Ed. 50 (2011) 3978–3982. [13] Y. Cao, X. Luo, H. Yu, F. Peng, H. Wang, G. Ning, Catal. Sci. Technol. 3 (2013) 2654–2660. [14] R. Liu, H. Huang, H. Li, Y. Liu, J. Zhong, Y. Li, S. Zhang, Z. Kang, ACS Catal. 4 (2014) 328–336. [15] J. Li, Y. Zhang, X. Zhang, J. Han, Y. Wang, L. Gu, Z. Zhang, X. Wang, J. Jian, P. Xu, B. Song, ACS Appl. Mater. Interfaces 7 (2015) 19626–19634. [16] J. Liu, H. Jeong, J. Liu, K. Lee, J.-Y. Park, Y.H. Ahn, S. Lee, Carbon 48 (2010) 2282–2289. [17] D. Yu, Q. Zhang, L. Dai, J. Am. Chem. Soc. 132 (2010) 15127–15129. [18] J. Liu, H. Bai, Y. Wang, Z. Liu, X. Zhang, D.D. Sun, Adv. Funct. Mater. 20 (2010) 4175–4181. [19] S. Yang, X. Li, W. Zhu, J. Wang, C. Descorme, Carbon 46 (2008) 445–452. [20] J. Liu, S. Xu, L. Liu, D.D. Sun, Carbon 60 (2013) 445–452. [21] X. Sun, Z. Liu, K. Welsher, J.T. Robinson, A. Goodwin, S. Zaric, H. Dai, Nano Res. 1 (2008) 203–212. [22] T. Zhang, J. Liu, D.D. Sun, RSC Adv. 2 (2012) 5134–5137. [23] Y. Zhou, Q. Bao, L.A.L. Tang, Y. Zhong, K.P. Loh, Chem. Mater. 21 (2009) 2950–2956. [24] D.W. Boukhvalov, M.I. Katsnelson, J. Am. Chem. Soc. 130 (2008) 10697–10701. [25] H. Wang, Y.H. Hu, Ind. Eng. Chem. Res. 50 (2011) 6132–6137. [26] K.-W. Park, J.H. Jung, J. Power Sources 199 (2012) 379–385. [27] H. Yan, H. Wu, K. Li, Y. Wang, X. Tao, H. Yang, A. Li, R. Cheng, ACS Appl. Mater. Interfaces 7 (2015) 6690–6697. [28] Y. Matsuo, T. Miyabe, T. Fukutsuka, Y. Sugie, Carbon 45 (2007) 1005–1012. [29] J. Liu, O. Margeat, W. Dachraoui, X. Liu, M. Fahlman, J. Ackermann, Adv. Funct. Mater. 24 (2014) 6029–6037. [30] A. Zali, A. Shokrolahi, M.H. Keshavarz, M.A. Zarei, Acta Chim. Slov. 55 (2008) 257–260. [31] A. Shokrolahi, A. Zali, M.H. Keshavarz, Chin. J. Catal. 31 (2010) 1427–1432. [32] L. Zhou, B. Dong, S. Tang, H. Ma, C. Chen, X. Yang, J. Xu, J. Energy Chem. 22 (2013) 659–664. [33] X. Fan, W. Peng, Y. Li, X. Li, S. Wang, G. Zhang, F. Zhang, Adv. Mater. 20 (2008) 4490–4493. [34] Y. Gao, D. Ma, C. Wang, J. Guan, X. Bao, Chem. Commun. 47 (2011) 2432–2434. [35] J. Zhang, X. Liu, R. Blume, A. Zhang, R. Schloegl, D.S. Su, Science 322 (2008) 73–77. [36] X. Liu, J. He, L. Yang, Y. Wang, S. Zhang, W. Wang, J. Wang, Catal. Commun. 11 (2010) 710–714.