Probing the methanol heterogeneous photochemistry processes by operando NMR – The role of bulk water

Probing the methanol heterogeneous photochemistry processes by operando NMR – The role of bulk water

Journal of Catalysis 378 (2019) 36–41 Contents lists available at ScienceDirect Journal of Catalysis journal homepage: www.elsevier.com/locate/jcat ...

1MB Sizes 0 Downloads 8 Views

Journal of Catalysis 378 (2019) 36–41

Contents lists available at ScienceDirect

Journal of Catalysis journal homepage: www.elsevier.com/locate/jcat

Probing the methanol heterogeneous photochemistry processes by operando NMR – The role of bulk water Man Ye a,1, Bei-Bei Xu a,1, Ran Zhang a,1, Yi-Ning Yang a, Ling-Yun Yang b, Xue Lu Wang a,⇑, Ye-Feng Yao a,⇑ a Physics Department & Shanghai Key Laboratory of Magnetic Resonance, School of Physics and Materials Science, East China Normal University, North Zhongshan Road 3663, Shanghai 200062, China b iHuman Institute, ShanghaiTech University, Shanghai 201210, China

a r t i c l e

i n f o

Article history: Received 21 May 2019 Revised 9 August 2019 Accepted 13 August 2019 Available online 4 October 2019 Keywords: Hetegeneous catalysis Bulk water Methanol reforming Operando NMR

a b s t r a c t Solid-liquid heterogeneous catalysis reactions play key roles in industrial catalytic reactions, throughout pollution control, fuel cell, methanol reforming and so on. But, few operando techniques have been performed to track the complex reaction processes, especially those in bulk liquid environment containing two or more liquid reactants, thus leaving a gap in our understanding of the reaction mechanisms deep into the working conditions. Here, we demonstrate an approach, using operando nuclear magnetic resonance (NMR) to probe the dynamics and kinetics of the methanol heterogeneous photochemistry processes in liquid condition with and without bulk water (H2O). It shows that not only can bulk H2O affect the types of methanol reforming products, but also its participation way can vary from H2O molecule to H-atom. Furthermore, the productivities of the main products - methylene glycol (HOCH2OH) and hemiacetal (CH3OCH2OH) – show different dependency on the H2O concentration, which is further supported by the density functional theory calculations (DFT). The operando NMR method opens up exciting opportunities for the mechanism studies of heterogeneous catalysis reaction in bulk liquid. Ó 2019 Elsevier Inc. All rights reserved.

1. Introduction Heterogeneous catalysis plays a key role in many industrial production processes, such as pollution control, fuel cell and artificial photosynthesis, with over 80% of catalytic processes using heterogeneous catalysts to direct a chemical reaction.[1] Hence, understanding the catalytic reaction processes deep into the solid/gas or solid/liquid interface is one of the most significant tasks in designing and optimizing effective heterogeneous catalysis.[2] However, after long term development, due to the limitation of the analytic techniques, catalysts and reaction intermediates are often characterized before and after the reaction in a nonworking condition. Thus, the detailed mechanisms in many important heterogeneous catalysis reactions especially those in bulk liquid environment captaining two or more liquid reactants are still not clear, which strongly limits the practical applications. The methanol (CH3OH) reforming reaction under the liquid phases reaction with a solid catalyst has gained a lot of research interests in the past decades [3–13]. The importance of this ⇑ Corresponding authors. E-mail addresses: [email protected] (X.L. Wang), [email protected] (Y.-F. Yao). 1 These authors contributed equally to this work. https://doi.org/10.1016/j.jcat.2019.08.016 0021-9517/Ó 2019 Elsevier Inc. All rights reserved.

reaction partially lies on the so-called ‘methanol economy’, in which CH3OH has been advocated that it could be catalytically reformed to important chemical raw materials like methylamine, ethylene and chloromethane [14,15]. In order to explore the reaction mechanism on molecular level, many in-situ and ex-situ analytic techniques have been developed, such as X-ray absorption spectroscopy (XAS), Raman, Infrared (IR) spectroscopy, transmission electron microscope (TEM), temperature-programmed desorption (TPD), scanning tunneling microscopy (STM), two photon photoemission (2PPE) and density functional theory calculations (DFT) [3–9,16–21]. By using these techniques, fruitful results have been obtained. For example, Zhou et al. successfully developed a time-dependent TPD-2PPE technique combined with STM to propose a two-step mechanism for the photooxidation of CH3OH to formaldehyde (HCHO) on TiO2 [18]. Phillips et al. further demonstrate that two consecutive photo-oxidation steps lead to methyl formate (HCOOCH3), by using mass spectrometry (MS) and STM [22]. However, in most of these studies, the reaction environment suitable for the above mentioned in-situ techniques is the solidgas/solid-liquid interface under ultrahigh vacuum (UHV) condition, where trace CH3OH molecules were adsorbed on the solid catalysis surface in form of gas molecule or covered in form of sub-monolayer liquid molecule. This type of reaction environment actually is far away from the practical situation in the

M. Ye et al. / Journal of Catalysis 378 (2019) 36–41

heterogeneous catalysis. Meanwhile, the vast majority of industrial heterogeneous catalysis reactions are performed in bulk liquid environments containing not only one liquid reagent and the catalyst is immersed in the mixed bulk reactants [2]. For instance, many studies have shown the mass transport limitation imposed by liquid water on the proton exchange membrane (PEM) fuel cell, that is to say, the presence of indelible H2O has the effects on the final performance [23,24]. Therefore, it’s still a quite tough task to take into account of the complexities of practical bulk liquid conditions such as solvent effect, proton exchange, and hydrogen-bonding effect. Particularly, when the catalyst is in a mixture of two or more liquid reactants of a catalytic reaction, the processes and consequence of the heterogeneous catalysis become more complicated. Yamakata et al. observed that H2O can act as an electron remover from catalysts and enhances the activity of photocatalytic CH3OH oxidation [25]. Yang et al. found that the effect of multilayer H2O on the CH3OH heterogeneous catalysis has quite different effects from monolayer H2O [26]. Meanwhile, Henderson et al. found that co-adsorbed H2O on the Ti5c sites, and physisorbed H2O molecules on the bridge-bonded oxygen (BBO) sites via hydrogen bonding have different influences on the CH3OH photochemistry process. It thus can be concluded that H2O plays a complicated role in the CH3OH heterogeneous catalysis processes, and it is necessary to clarify the role of H2O in bulk liquid environment containing both CH3OH and H2O. Yet, to the best of our knowledge, detailed information about the influence and the role of the bulk H2O on CH3OH photochemistry processes is still lacking, and few experimental evidence has been reported for quantitative insight into the mechanism of CH3OH heterogeneous photochemistry process in real bulk liquid condition. Here we demonstrate an approach, using the operando NMR spectroscopy, which allows to quantify the varied heterogeneous catalysis products desorbed from catalyst in the bulk liquid condition. The role of bulk H2O in CH3OH photochemistry process on rutile-TiO2 deep into the real bulk liquid environment was explored, revealing that not only can H2O affect the types of CH3OH reforming products, but the ways of its participation vary from H2O molecule to H-atom for the different products. It has been found that the productivities of the main products, HOCH2OH and CH3OCH2OH, have the different dependency on the H2O concentration. This can be related to the different reaction active sites and generation mechanisms. The methodology used in this work offers the promising possibilities to deeply study the mechanisms of heterogeneous catalysis reactions in the bulk liquid environment. 2. Experimental 2.1. Samples preparation The rutile-TiO2 used in the reactions was purchased from Sigma-Aldrich directly. Methanol (CH3OH, 99.9%) was obtained from Sigma-Aldrich. Methanol-D4 (D, 99.8%) with 0.03% V/V TMS solvents was obtained from chemicals Cambridge Isotope Lab. Deuterium oxide (D2O, 99.9% atom D, containing 0.05% wt. 3-(tri methylsilyl)propionic-2,2,3,3-d4 acid) was obtained from Aldrich. Before being added into the NMR tube, the TiO2 catalysts were grinded carefully and dried in a vacuum oven for 8 h at 80 °C. 2.2. Material characterization The structures of the samples were characterized by X-ray diffraction (XRD, Rigaku Ultima IV X-ray Diffractometer with CuKa radiation (V = 35 kV, I = 25 mA, l = 1.5418 Å)), the angular range

37

was 5°–90° with a scanning rate of 10° min 1. Scanning electron microscope (SEM, Hitachi S-4800) and transmission electron microscopy (TEM, JEM 2100, 200 kV) were used to characterize the morphology of the samples.

2.3. NMR experiments 13 C and 1H solution nuclear magnetic resonance (NMR) spectra were obtained with Bruker AV-800 spectrometer (18.8 T), Bruker AV-500 instrument (11.7 T) or Varian 700 M spectrometer (16.4 T). 2D 1H–13C Heteronuclear Singular Quantum Correlation (HSQC) spectrum was recorded on Bruker AV-800 instrument (18.8 T), and 13 C-decoupling during acquisition. 2D 1H NMR exchange (EXSY) spectroscopy was recorded on Bruker AV-500 instrument (500 MHz) with 800 ms mixing time. All NMR experiments were referenced to 3-trimethylsilylpropionic acid (TMSP) and performed at 276 K. Data processing of NMR spectra were performed using the Bruker software TopSpin. The prepared samples were irradiated with 300 W Xe lamp with a 400 nm filter. (Beijing Perfectlight Science and Technology Co., Ltd).

3. Results and discussion The NMR apparatus for the study of the solid-liquid heterogeneous photocatalytic reaction (Fig. S1) in this work has been described previously in detail.[27] The samples (a suspension containing TiO2 photocatalyst together with the bulk CH3OH or CH3OH-H2O liquid mixtures) were prepared in NMR tubes at room temperature and ordinary pressure, allowing the liquid phase reactions with a solid catalyst of the systems to be directly studied inside the NMR probe. In order to disperse TiO2 powders uniformly in the solvent, a ultrasonication process (5 min) was applied on the NMR tubes. The obtained TiO2 suspension is quite stable during the NMR measuring time (see Fig. S1 0). The sample was standing for another 5 min before being transferred into the NMR magnet for the experiments. A 300 W Xe lamp was used as the light source, and the reactions were performed under 400 nm light irradiation by using the optical filter. The incident photon flux was estimated by measuring the incident light intensity at different NMR tube depth (The detailed information can be found in Fig. S2). The experimental processes are illustrated in Scheme 1. Titanium dioxide (TiO2), which is widely used in the photocatalysis, is served as a model catalyst (The detailed characterization of TiO2 can be found in Figs. S3–S5) [28]. NMR spectra of some benchmark experiments were measured first to test the signal sensitivity and ascribe the signals (the detailed descriptions can be found in Figs. S6–S10). To explore the role of bulk H2O in the photocatalytic CH3OH reforming process in the real bulk liquid environment, two solidliquid reaction systems, namely System 1 (TiO2/CH3OH) and System 2 (TiO2/CH3OH/H2O), were deliberately designed. The main distinction of the two systems lies on the addition of bulk H2O. In System 1, the solid-liquid heterogeneous photocatalytic reaction suspension environment only contains TiO2 and bulk pure CH3OH. Whereas in System 2, the TiO2, CH3OH and bulk H2O form the similar heterogeneous photocatalytic reaction suspension. Comparing the types and generation tendency of the CH3OH reforming products in System 1 with System 2 provides a way to probe the role of H2O in the CH3OH photochemistry process. The gas chromatography–mass spectrometry (GC–MS), is also used to probe the reaction products. However, no clear CH3OH reforming products can be found in GC–MS analysis, albeit with exquisite resolution. The high injection temperature of GC–MS, product separation, along with the low productivity could be regarded as the key factor resulting the invalid detection (details can be found in Fig. S11).

38

M. Ye et al. / Journal of Catalysis 378 (2019) 36–41

Scheme 1. Schematic diagram for operando NMR studies. Two suspensions (TiO2/ CH3OH and TiO2/CH3OH/H2O) containing TiO2 photocatalyst and the bulk liquid solution (CH3OH vs CH3OH-H2O) are investigated directly in NMR tubes at room temperature and ordinary pressure under the light irradiation.

Fig. 1a shows the 1H NMR spectra of system 1 (CH3OH/TiO2) before and after the irradiation. No CH3OH reforming product signal is observed in the spectrum of the sample without irradiation. The peak at 5.04 ppm can be assigned to the peak of tiny HDO which is the trace residue in the CH3OH solution (99.9%) [27]. The peak at 4.80 ppm is tentatively assigned to the signal of –OH of CH3OH/CD3OH (the detailed signal assignment can be found in Fig. S12). Upon irradiation for 12 h, a resonance signal (singlet, 4.64 ppm) is clearly observed in the spectrum. This signal was assigned to CH2 of CH3OCH2OH. The detailed signal assignment can be found in Figs. S13-S14, S16. The formation of CH3OCH2OH can be assigned to the reaction between HCHO and CH3OH on the surface of TiO2: For the TiO2 catalysts immersed in the CH3OH solvent, it is highly possible that the TiO2 surfaces will be fully solvated by

Fig. 1. 1H NMR spectra of different systems. (a) 1H NMR spectra of system 1 (CH3OH/CD3OD/catalyst) before and after the irradiation. The spectra were acquired from samples containing CD3OD (500 ll), CH3OH (20 ll) and TiO2 (1 mg). (b) 1H NMR spectra of system 2 (CH3OH/D2O/catalyst) before and after the irradiation. The spectra were acquired from samples containing D2O (500 ll), CH3OH (20 ll) and TiO2 (1 mg) at 276 K.

the CH3OH molecules. Therefore, after its formation on the surface, HCHO is immediately surrounded by the CH3OH molecules on the surface. It is thus very likely that in such a local environment HCHO will have a further reaction with the neighbored CH3OH molecule. We thus have considered this mechanism as the dominant way for the formation of CH3OCH2OH. The previous reports show that CH3OCH2OH has a relatively low binding energy with the reaction centers (i.e., the Ti5c sites) [8,30]. This results in the fast desorption process of CH3OCH2OH on the catalyst surface and in turn the appearance of CH3OCH2OH in the solution. It is worthy of noting that no clear signals of HCOOH, HCOOCH3 or other CH3OH reforming products can be found in the spectrum of system 1. This means that in pure CH3OH environment the HCHO molecules are prone to react with the surrounded CH3OH molecules rather than to undergo a recombination with HBBO to form methoxy [31]. This result is not consistent with the previous studies, which reaction environment is the solid-gas/solid-liquid interface under ultrahigh vacuum (UHV) condition, where trace CH3OH molecules were adsorbed on the solid catalysis surface in form of gas molecule or covered in form of sub-monolayer liquid molecule [8,9,29]. Note that, in the solution NMR experiments in this work, we can only probe the signals of the chemicals which have been desorbed from the catalyst surface and entered into the solution, whereas the signals of the chemicals remaining on the catalyst surface are silent for the NMR detection. This could also result in the difference in the reaction products observed in this work and in the other experiments. Fig. 1b shows the 1H NMR spectra of system 2 (CH3OH/H2O/ TiO2) before and after the irradiation (The 13C NMR spectrum of system 2 after the irradiation can be found in Fig. S15). Similarly, only CH3OH (D) (3.34 ppm) and HDO (5.04 ppm) peaks are observed in the dark. After 1 h of irradiation, three singlets are found in the spectrum, which can be assigned to CH3OCH2OH (4.74 ppm), HOCH2OH (4.82 ppm) and HCOOH (8.45 ppm), respectively. Among these reaction products, CH3OCH2OH is likely formed through the direct coupling of HCHO and CH3OH on the TiO2 surface. HOCH2OH seems to have two types of formation routes. It can be formed through the direct coupling of HCHO and H2O on the TiO2 surface, or formed through the coupling of the desorpted HCHO and H2O in the solvent, after HCHO molecules have diffused into the solvent.[32] Although both formation ways of HOCH2OH are possible, we prefer the latter one (in the solvent) to be the dominate way. Because the adsorption energy of HOCH2OH reaches the high value of 1.19 eV, thus it is difficult for it to desorb from the TiO2 surface after its formation. The detailed mechanisms of the reactions will be discussed in detail later. To confirm that the reactions reflected by the above NMR study are indeed driven by the photocatalysis, a quantitative study of the changes of products for System 1 and 2 as a function of irradiation time was carried out. Fig. 2a shows the 1H NMR spectrum of System 1 acquired under light irradiation during the first hour. In the spectrum, the CH3OCH2OH signal peak (4.64 ppm) shows a clear liner growth tendency upon increasing the irradiation time. It is observed that the signal intensity of CH3OCH2OH in the NMR spectrum monotonically increases with the CH3OH concentration. Fitting the data yields the CH3OCH2OH production rate of 0.69 nmol/min at the first one hour. Note that the H/D isotope effect has been realized in literature [33,34]. However, in this study it does not affect the CH3OCH2OH quantity (Fig. S17). A similar quantitative kinetic study of the changes of products for System 2 has been also performed (see Fig. 2c and d). In the spectra, two NMR signal peaks (CH3OCH2OH, 4.74 ppm, and HOCH2OH, 4.82 ppm) both increased with the irradiation time, and meanwhile both HOCH2OH and CH3OCH2OH products show the similar linear increasing tendency as that in system 1 (Fig. 2d). The fitting yields 11.41 nmol/min for HOCH2OH and

M. Ye et al. / Journal of Catalysis 378 (2019) 36–41

39

Fig. 2. (a) Time dependent quantitative 1H NMR spectrum of TiO2 at 276 K in CD3OD/CH3OH mixtures under the 400 nm irradiation; (b) Yield of CH3OCH2OH plotted against the irradiation time. The NMR experiments were acquired form samples containing CD3OD (500 ll, 0.03% TMS as the internal reference), CH3OH (20 ll) and TiO2 (1 mg) at 276 K; (c) Time dependent quantitative 1H NMR spectrum of TiO2 at 276 K in D2O/CH3OH mixtures under the 400 nm irradiation; (d) Yield of HOCH2OH and CH3OCH2OH plotted against the irradiation time. The NMR were acquired form samples containing D2O (500 ll, 0.05% TMSP as the internal reference), CH3OH (20 ll) and TiO2 (1 mg) at 276 K.

4.75 nmol/min for CH3OCH2OH during the first hour. By comparison, the production rate of CH3OCH2OH is 1.4 times lower than that of HOCH2OH, suggesting that the two products might undergo different generation paths and mechanisms. For the formation of CH3OCH2OH, it can be formed via either HOCH2OH and CH3OH, or HCHO and CH3OH. However, the simple mixture of HOCH2OH, HCHO and CH3OH in the dark hardly produce any CH3OCH2OH (Fig. S18). It thus can be concluded that the formation of CH3OCH2OH occurs on the TiO2 surface under the illumination. The process of forming CH3OCH2OH involving the H or OH intermediates from the dissociation of H2O on TiO2 surface could be the molecular mechanism of the formation of CH3OCH2OH [35,36]. The bulk H2O plays an important role in this mechanism, which will be discussed in detail later. For a better understanding of the role of H2O and the molecular mechanisms of HOCH2OH and CH3OCH2OH, systematic DFT calculations concerning the generation of HOCH2OH and CH3OCH2OH on the rutile-TiO2 (1 1 0) surface have been conducted. The calculated results provide a possibility to understand the mechanism to some extent. As one can see in Fig. 3, CH3OCH2OH can be formed by the transfer of a H- to CH3OCH2O– (black line). Note that in System 1 the CH3OCH2OH are formed through the reaction between HCHO and CH3OH, indicating that the H- in System 2 is mainly from the dissociation of H2O but not CH3OH.[37] The red line in Fig. 3 gives the other possible way to generate the HOCH2OH on the surface of TiO2. Similar to the formation of CH3OCH2OH, the generation of HOCH2OH includes a two-step processes. In contrast, on the TiO2 surface, it is more favorable to generate HOCH2OH ( 0.06 eV) than CH3OCH2OH ( 0.28 eV). However, as seen from the calculated

adsorption energy (Fig. S19) for CH3OCH2OH and HOCH2OH, it is difficult for HOCH2OH to desorb from the TiO2 surface because its adsorption energy reaches the high value of 1.19 eV. In addition, because of its weak adsorption on the surface, the HCHO generated from the CH3OH may desorb easily from the TiO2 surface in System 2. Combination of the NMR observations and the DFT results thus strongly indicates that most of HOCH2OH observed are from the coupling reaction of desorption HCHO and H2O molecule in the bulk liquid environment, which principally will never be present in the vacuum systems. Note that there might still exit a gap between over-idealized theoretical calculations and the realistic reactions. To bridge this gap, more consideration in the calculation models are required, which is out of the scope of this study. The CH3OH reforming process in System 2 was further investigated in the samples having the different CH3OH concentrations (0.04 M, 0.08 M and 0.12 M); the amount of H2O is kept dominant in all of the samples.). Fig. 4 shows the kinetic evolution of HOCH2OH (Fig. 4a) and CH3OCH2OH (Fig. 4b). It is observed that the amount of HOCH2OH increases linearly against the irradiation time in all of the three samples. Interestingly, the growth rates of HOCH2OH are almost constant with the CH3OH concentrations. This is completely different with the observation in Fig. 4b, where the growth rates of CH3OCH2OH increase with the increasing of CH3OH concentration. The differences between HOCH2OH and CH3OCH2OH have been related to the different formation mechanisms and the participation ways of bulk H2O. As discussed above, the HOCH2OH molecules are mainly from the hydration of the desorbed HCHO molecules which are initially formed on the surface of TiO2. However, with increasing the CH3-

40

M. Ye et al. / Journal of Catalysis 378 (2019) 36–41

Fig. 3. Calculate formation energy of (a) CH3OCH3OH; (b) HOCH2OH on the TiO2 (1 1 0) surface. The optimized structures of CH3OCH2OH precursors are CH3O–, HCHO and –H, whereas the optimized precursor structures of HOCH2OH are HCHO and –OH. The Ti atoms are in light grey and O in red, while the C atoms in dark grey and H in white. This notation is used throughout this paper. The calculations are carried out using the Vienna ab initio simulation package (VASP). More details can be found in SI.

Fig. 4. Quantitative 1H NMR spectra of photocatalytic products in different methanol concentration. Yields of (a) HOCH2OH and (b) CH3OCH2OH produced from heterogeneous catalysis in three different methanol concentration (0.12 M, 0.08 M, 0.04 M) as a function of irradiation time.

OH concentration, more CH3O– would appear on the TiO2 surface. This leads to a competition between the formation of the intermediate (H-HCHO-CH3O) and the desorption of HCHO on the surface of TiO2. If the formation of H-HCHO-CH3O dominates, less HCHO and the consequent product, HOCH2OH, are expected. This may have formed a dynamic balance, resulting in the almost constant production rate of HOCH2OH in the studied CH3OH concentrations. As to CH3OCH2OH, its formation is closely related to the formation of the intermediate, H-HCHO-CH3O. With the increase of CH3OH concentration, more CH3O– would adsorb on the TiO2 surface, thus resulting in more production of CH3OCH2OH. Because of the low desorb energy, the formed CH3OCH2OH molecules on the TiO2 surface can easily detach themselves from the surface and enter into the liquid environment. This provides a plausible expla-

nation why the productivity of CH3OCH2OH shows an apparent dependence on the CH3OH concentration.

4. Conclusion In summary, by using operando NMR spectroscopy, we have investigated the role of bulk H2O in methanol heterogeneous photochemistry processes on the rutile-TiO2 surface deep into the bulk liquid environment. The results reveal the presence of the complicated dynamic balance among the multiple surface reactions, where the bulk H2O does play critical roles. It shows that not only can bulk H2O affect the types of methanol reforming products, but also the way of its participation can vary from H2O molecule to H-

M. Ye et al. / Journal of Catalysis 378 (2019) 36–41

atom. The productivity of the main products, HOCH2OH and CH3OCH2OH, shows different dependency on the H2O concentration. The operando NMR method in this work opens up exciting opportunities for the mechanism study of the methanol reforming, and will find more applications on the study of heterogeneous catalysis reactions in bulk liquid environments in general. Acknowledgements This work was supported by National Natural Science Foundation of China (21603073 and 21574043). Appendix A. Supplementary material Supplementary data to this article can be found online at https://doi.org/10.1016/j.jcat.2019.08.016. References [1] J. Dou, Z.C. Sun, A.A. Opalade, N. Wang, W.S. Fu, F. Tao, Chem. Soc. Rev. 46 (2017) 2001–2027. [2] F. Tao, P.A. Crozier, Chem. Rev. 116 (2016) 3487–3539. [3] Q. Guo, C. Xu, W. Yang, Z. Ren, Z. Ma, D. Dai, T.K. Minton, X. Yang, J. Phys. Chem. C. 117 (2013) 5293–5300. [4] M. Setvin, X. Shi, J. Hulva, T. Simschitz, G.S. Parkinson, M. Schmid, C. Di Valentin, A. Selloni, U. Diebold, ACS Catal. 7 (2017) 7081–7091. [5] C. Xu, W. Yang, Q. Guo, D. Dai, M. Chen, X. Yang, J. Am. Chem. Soc. 135 (2013) 10206–10209. [6] C. Xu, W. Yang, Q. Guo, D. Dai, M. Chen, X. Yang, J. Am. Chem. Soc. 136 (2014) 602. [7] C. Xu, W. Yang, Q. Guo, D. Dai, T.K. Minton, X. Yang, J. Phys. Chem. Lett. 4 (2013) 2668–2673. [8] Q. Yuan, Z. Wu, Y. Jin, F. Xiong, W. Huang, J. Phys. Chem. C. 118 (2014) 20420– 20428. [9] Q. Yuan, Z. Wu, Y. Jin, L. Xu, F. Xiong, Y. Ma, W. Huang, J. Am. Chem. Soc. 135 (2013) 5212–5219. [10] T.J. Wang, Q.Q. Hao, Z.Q. Wang, X.C. Mao, Z.B. Ma, Z.F. Ren, D.X. Dai, C.Y. Zhou, X.M. Yang, J. Phys. Chem. C. 122 (2018) 26512–26518. [11] I. Yarulina, A.D. Chowdhury, F. Meirer, B.M. Weckhuysen, J. Gascon, Nat. Catal. 1 (2018) 398–411.

41

[12] S. Sengodan, R. Lan, J. Humphreys, D.W. Du, W. Xu, H.T. Wang, S.W. Tao, Renew. Sust. Energ. Rev. 82 (2018) 761–780. [13] G.N. Nomikos, P. Panagiotopoulou, D.I. Kondarides, X.E. Verykios, Appl. Catal. B-environ. 146 (2014) 249–257. [14] M. Nielsen, E. Alberico, W. Baumann, H.J. Drexler, H. Junge, S. Gladiali, M. Beller, Nature 495 (2013) 85. [15] G.A. Olah, Chem. Eng. News. 81 (2003) 5. [16] J. Singh, C. Lamberti, J.A. van Bokhoven, Chem. Soc. Rev. 39 (2010) 4754–4766. [17] M.A. Bañares, G. Mestl, Adv. Catal. 52 (2009) 43–128. [18] C. Zhou, Z. Ren, S. Tan, Z. Ma, X. Mao, D. Dai, H. Fan, X. Yang, J. LaRue, R. Cooper, A.M. Wodtke, Z. Wang, Z. Li, B. Wang, J. Yang, J. Hou, Chem. Sci. 1 (2010) 575– 580. [19] A. Vimont, F. Thibault-Starzyk, M. Daturi, Chem. Soc. Rev. 39 (2010) 4928– 4950. [20] E.R. White, M. Mecklenburg, B. Shevitski, S.B. Singer, B.C. Regan, Langmuir 28 (2012) 3695. [21] J.E. Evans, K.L. Jungjohann, P.C. Wong, P.L. Chiu, G.H. Dutrow, I. Arslan, N.D. Browning, Micron 43 (2012) 1085–1090. [22] K.R. Phillips, S.C. Jensen, M. Baron, S.C. Li, C.M. Friend, J. Am. Chem. Soc. 135 (2013) 574–577. [23] H. Li, Y. Tang, Z. Wang, Z. Shi, S. Wu, D. Song, J. Zhang, K. Fatih, J. Zhang, H. Wang, J. Power Sources 178 (2008) 103–117. [24] G. Avgouropoulos, J. Papavasiliou, M.K. Daletou, J.K. Kallitsis, T. Ioannides, S. Neophytides, Appl. Catal. B-environ. 90 (2009) 628–632. [25] A. Yamakata, A. Takaaki Ishibashi, H. Onishi, J. Phys. Chem. C. 107 (2003) 9820–9823. [26] W. Yang, Z. Geng, Q. Guo, D. Dai, X. Yang, J. Phys. Chem. C. 121 (2017) 17244– 17250. [27] X.L. Wang, W. Liu, Y.Y. Yu, Y. Song, W.Q. Fang, D. Wei, X.Q. Gong, Y.F. Yao, H.G. Yang, Nat. Commun. 7 (2016) 11918. [28] M.A. Henderson, Surf. Sci. Rep. 66 (2011) 185–297. [29] M.A. Henderson, S. Oterotapia, M.E. Castro, Faraday Discuss. 114 (1999) 399– 405. [30] Q. Guo, C. Xu, Z. Ren, W. Yang, Z. Ma, D. Dai, H. Fan, T.K. Minton, X. Yang, J. Am. Chem. Soc. 134 (2012) 13366–13373. [31] Q. Guo, Chin. J. Catal. 36 (2015) 1649–1655. [32] J. F Walker, RE Krieger: Malabar, FL, 1964. [33] L.I. Krishtalik, Electrochim. Acta. 46 (2001) 2949–2960. [34] H.C. Urey, F.G. Brickwedde, G.M. Murphy, Phy. Rev. 40 (1932) 1–15. [35] A.L. Goodman, G.M. Underwood, V.H. Grassian, J. Phys. Chem. A 103 (1999) 7217–7223. [36] S. Benkoula, O. Sublemontier, M. Patanen, C. Nicolas, F. Sirotti, A. Naitabdi, F. Gaie-Levrel, E. Antonsson, D. Aureau, F. Ouf, S. Wada, A. Etcheberry, K. Ueda, C. Miron, Sci. Rep. 5 (2015) 15088. [37] H. Liu, E. Iglesia, J. Phys. Chem. B 109 (2005) 2155–2163.