Ran-ran FengAn-an LiuShuo LiuJiao-jian ShiYi LiuZe-feng Ren

a.International Center for Quantum Materials(ICQM)and School of Physics,Peking University, Beijing 100871,China

b.Collaborative Innovation Center of Quantum Matter,Beijing 100871,China

Methanol Adsorption on TiO2Film Studied by Sum Frequency Generation Vibrational Spectroscopy

Ran-ran Fenga,b?,An-an Liua,b?,Shuo Liua,b,Jiao-jian Shia,b,Yi Liua,b,Ze-feng Rena,b?

a.International Center for Quantum Materials(ICQM)and School of Physics,Peking University, Beijing 100871,China

b.Collaborative Innovation Center of Quantum Matter,Beijing 100871,China

A broadband infrared surface sum frequency generation vibrational spectroscopy(SFG-VS) and anin situUV excitation setup devoted to studying surface photocatalysis have been constructed.With a home-made compact high vacuum cell,organic contaminants on TiO2thin flm surface prepared by RF magnetron sputtering werein situremoved under 266 nm irradiation in 10 kPa O2atmosphere.We obtained the methanol spectrum in the CH3stretching vibration region on TiO2surface with changing the methanol pressure at room temperature.Features of both molecular and dissociative methanol,methoxy,adsorbed on this surface were resolved.The CH3symmetric stretching vibration frequency and Fermi resonance of molecular methanol is red-shifted by about 6?8 cm?1from low to high coverage. Moreover,the recombination of dissociative methanol and H on surfaces in vacuum was also observed.Our results suggest two equilibria exist∶between molecular methanol in the gas phase and that on surfaces,and between molecular methanol and dissociative methanol on surfaces.

Surface sum frequency generation vibrational spectroscopy,Surface photocatalysis,TiO2

I.INTRODUCTION

Since the frst report of UV-induced redox chemistry on TiO2[1],TiO2has attracted extensive attention due to its wide applications in the areas of air purifcation, waste water treatment,self-cleaning glass and so on, especially in the last decade.Recently,the understanding of adsorption structure and photochemistry of organic molecules on TiO2has made much progress on the molecular level[2-7].

Methanol,as the simplest alcohol,has been thoroughly studied on TiO2(110)surfaces as a model molecule of photooxidation of organic contaminants. Henderson and co-workers conducted a detailed study of thermal reaction of CH3OH on TiO2(110)using variant surface chemistry techniques,and concluded that the majority of the CH3OH molecules were adsorbed on 5-coordinated Ti sites,and CH3OH adsorbed on bridge-bonded oxygen vacancy sites dissociated to methoxy and H on bridged-bonded oxygen sites(BBOH)[8],which was imaged by atomically resolved scanning tunneling microscopy(STM)[9]. They identifed that the methoxy group was the active species in the photochemical hole scavenging reaction of methanol on TiO2(110)surfaces[10].An excited electronic state on CH3OH/TiO2(110)was found with two-photon photoemission spectroscopy(2PPE) [11,12].Yang and co-workers employed the combination of time-dependent 2PPE and STM,and found site-specifc photoinduced dissociation of CH3OH at 5-coordinated Ti sites[12].They used a temperature dependent desorption(TPD)with laser excitation and observed a stepwise photocatalytic dissociation of CD3OH adsorbed on 5-coordinated Ti sites to methoxy,CD3O, and BBOH,and further to formaldehyde,CD2O,and BBOD under 400 nm light irradiation[13].Later,further oxidation of methanol to methyl formate was also found under UV light irradiation[14-16].

Theoretically,Yang and co-workers attributed the surface excited electronic state to a photocatalytic dissociated state of CH3OH on TiO2(110)[12],which was diferent from the assignment of wet electron states of Petek and co-workers[17].Due to the limited ability of density functional theory(DFT)on excited electronic state of surface,it’s difcult to know which species and what happen under excitation on the surface from the 2PPE data[18,19].However,the DFT based on the ground electronic state could interpret the stepwise dissociation of methanol on TiO2[13].

The TPD method not only provides us the adsorption information of methanol on TiO2[8,20],but also

In the present work,we used a newly setup,broadband infrared SFG vibrational spectrometer,toin situstudy methanol adsorption on TiO2thin flm prepared by RF magnetron sputtering.We obtained the SFG spectra of CH3OH/TiO2at diferent methanol pressure, and analyzed the SFG signal of molecular and dissociative methanol relating to the methanol pressure with global ftting.

II.EXPERIMENTS

The schematic of SFG setup is shown in Fig.1. An optical parametric amplifer(TOPAS-USF,Light Conversion)pumped by 60%power of the amplifer Ti∶Sapphire laser system(Spitfre ACE,Newport,5 mJ, 800 nm,35 fs,1 kHz repetition rate)generates tunable near-IR pulses(signal and idler,1.16?2.6μm). The mid-IR(2.6?11μm)is subsequently generated by nonlinear diference frequency mixing(NCDFM1, Light Conversion)of signal and idler in AgGaS2.The femtosecond IR pulse at about 3.4μm wavelength is roughly 20 mW.We used a GaAs(100)surface to get the IR profle,whose apparent bandwidth is typically 240 cm?1(FWHM),smaller than the real one due to the limited diameter of the visible light.20%power from the amplifer is spectrally narrowed as the visible light (VIS)using a pulse shaper(1800 L/mm pulse compression grating,Spectrogon;cylindrical lens with 200 mm focal length),normally FWHM 7.5 cm?1and 14 mW in our experiment.The residual power is used to generate the third harmonic,UV 266 nm.Second harmonic, 400 nm,is generated in a type Iβ-barium borate.And a calcite plate compensates the delay between the fundamental and the second harmonic,which then pass a waveplate(λ/2=800 nm,λ=400 nm)and are mixed in another type I-barium borate to generate a third harmonic.The delay of 266 nm laser is controlled by a hybrid translation stage(M-511.HD,Physik Instrumente). Typically,30 mW UV light,tuned by combination of a waveplate and a thin flm polarizer(TFP),is used in our experiment.The diameter of 266 nm laser is about 1.5 mm on the surface,which is much larger than VIS and IR laser spots to easily insure overlap.

Both polarizations of the VIS and IR are controlled by true zero-order half-wave plates,and the SFG signal polarization is selected and controlled by the combination of an achromatic half-wave plate and a Glan polarizer.The SFG signal is collimated by a lens and isolated from the VIS and IR beams by optical irises and double 775 nm shortpass flters(Semrock)right in front of a 500 mm monochromator(SP2500,Princeton Instruments)entrance slit.The SFG signal is focused onto the entrance slit,dispersed by a 1200 L/mm grating and imaged onto 5 pixels(full width)vertically of a thermoelectrically cooled(?60?C)electron-multiplied CCD camera(Princeton Instruments,ProEM Excelon) containing a 1600×400,16μm2,pixel array.Here,all SFG spectra correspond to the ssp(SF,VIS,IR)polarization combination.Our SFG system is extremely stable benefting from the stable temperature(±0.2?C when running laser)in the home-design clean room.

FIG.1 A schematic of SFG light path,and a thin flm sample cell.

The CaF2substrates(25 mm in diameter,3 mm in thickness)were soaked and ultrasonically cleaned in toluene,rinsed in deionized water and soaked in methanol,rinsed in deionized water again,dried with nitrogen gas,and fnally cleaned with plasma(Harrick)for 10 min.CaF2substrates were stored in a Tefon box before using.We prepared the TiO2thin flm on CaF2substrates by RF magnetron sputtering.The working pressure of the working gas,argon,was 1 Pa,and the fow was 30 sccm.The 60 nm thick TiO2flm was prepared with RF power 60 W for 30 min and the substrate at room temperature.This method produced the mixed phase of anatase and rutile[26].The methanol (99.95%purity)was purifed further by several freezepump-thaw cycles before experiment.

The CaF2window with TiO2thin flm was mounted on a home-made cell,as shown in Fig.2.We used doubleo-ring sealing and diferential pumping methods to avoid leakage and to achieve high vacuum condition in the cell.The ultimate pressure is less than 5×10?8Torr.The absolute pressure of gas in the cell is monitored with a capacitance gauge(CPCA-330Z,Shanghai Zhentai Instruments).The whole cell is mounted on a mirror mount(Polaris,Thorlabs)and two-axis linear stages,with which SFG signal can be easily optimized. All the gas handling system is assembled by CF fanges and VCR connections,which can reduce the risk of leakage.

The SFG intensity is given by

FIG.2 The cut view of the cell and the schematic of SFG and THG light path on the TiO2thin flm.

FIG.3 SFG-VS in the CH stretching vibration region of the organic contaminants on TiO2thin flm with UV cleaner andin situ266 nm irradiation in 10 kPa O2atmosphere.

whereANRis nonresonance amplitude;Aq,Φq,ωq, Γqare the amplitude,phase,resonant frequency,and damping constant(half width at half maximum)of theq-th mode,respectively;Nis density of adsorbates,〈β〉is the angular average of the molecular hyperpolarizability.The measured SFG signal was normalized to the apparent IR profle(∝IVisIIR)obtained at GaAs(100)surface located at the same position of the TiO2thin flm. Due to the limited resolution and signal-to-noise ratio and much more parameters than simple linear absorption spectra,the meaningful quantitative information is difcult to be obtained with the ftting of SFG spectra [27].We developed a global ftting program based on Matlab software,which can ft multiple resonant frequencies of multiple curves with any algebraic relationship among ftting parameters based on SFG(Eq.(1)). And 85 parameters were reduced to 53 with setting the same phase Φqand damping constant Γqfor 4 resonant terms of 5 curves.

III.RESULTS AND DISCUSSION

The newly prepared TiO2thin flm was contaminated with organic molecules under ambient conditions.Figure 3 shows the SFG spectra in the CH stretching vibration region of TiO2thin flm,and most contaminants were removed after UV cleaner(SAMCO’s UV-1,N2purged).To avoid contamination under ambient conditions,we used the third harmonic,266 nm,irradiation toin situremove the contaminants with flling 10 kPa O2in the cell by photodegradation reaction.The typical spectrum of residual contaminants on bare surface is shown in Fig.3.Routinely,the TiO2surface was refreshed with 30 mW 266 nm treatment for 1 h in 10 kPa O2atmosphere in the cell before measurements.

Figure 4 shows the SFG-VS of methanol on TiO2thin flm at diferent methanol pressures,which is similar to the results of Shultz and co-workers[24].Based on our global ftting program,we got four resonances located at 2826(Res 1),2848(Res 2),2928(Res 3) and 2958 cm?1(Res 4)for 30 Pa CH3OH,respectively. The ftting results are shown in the supplementary material.All the four features arise as the pressure of CH3OH,while the second and fourth features increase more obviously.We assigned the resonances at 2826 and 2928 cm?1to the symmetric stretch(νs)and symmetric Fermi resonance(νF)[28,29]of CH3group of methoxy,chemically adsorbed on TiO2,and the other two resonances at 2848 and 2958 cm?1to that of molecular methanol,physically adsorbed on TiO2[24].As increasing the methanol pressure to 4 kPa,the frst and third resonant frequencies almost keep invariant,whilethe second and forth are red-shifted by 6?8 cm?1.It is most likely that the interaction between the methanol molecules at the frst layer induces the shift of CH3vibrational frequency.The Fermi resonance at around 2910 cm?1at the vapor/methanol interface is not obtained in our SFG spectra[30].One possible reason could be that the feature is too small to be resolved.

The studies on single crystal rutile TiO2(110)and anatase TiO2(101)showed there were both molecular adsorption and dissociative adsorptionn of methoxy [8,31].Similar results were obtained on rutile TiO2nanoparticles[32].There are several diferent sites for adsorption of methoxy on single crystal rutile TiO2(110)and anatase TiO2(101)from TPD studies [8,31].Here,we cannot distinguish methoxy groups at diferent sites from the vibrational frequency of CH3. However,two types of methoxy groups were distinguished in our experiment,which had diferent photocatalytic reaction efciencies under both aerobic and anaerobic conditions.

When we evacuated the cell after flling 300 Pa CH3OH,the SFG peaks arising from molecular methanol largely decrease while the methoxy peaks are little afected,which is in accordance with the data of Shultz and co-workers[24].Moreover,we found that further pumping out the cell results in the continuous decrease of the SFG signal of chemically adsorbed methoxy,as shown in Fig.6.The TPD results on TiO2(110)show that the desorption temperature of methanol is around 295 K[8],so there is still some molecular methanol on the thin flm at room temperature.DFT calculation showed that the adsorption energy of both methanol and its dissociated products, methoxy and H on BBO sites were almost degenerate, and the barrier was very low[13,36,37].Thus,it is reasonable to assume that species including chemically adsorbed methoxy,molecular methanol,and the methanol in gas phase are dominated by the following equilibriums at room temperature∶

FIG.4 SFG spectra(dot)of methanol on TiO2thin flm as a function of methanol pressure and their global fttings (line).

FIG.5 Densities of methoxy and methanol on TiO2vs. methanol pressure in the cell.

FIG.6 The methoxy signal as a function of pumping time.300 Pa CH3OH is flled in the cell before pumping. (a)Evacuation for 3 min,＜4×10-6Torr,(b)evacuation for 8 min,＜8×10-7Torr,and(c)evacuation for 24 min,＜3×10-7Torr.

where the dissociation of methanol to methoxy equilibrates with the formation of methanol on surfaces,not directly with the methanol in the gas phase.It mayinterpret why the methoxy quantity does not obey the Langmuir isothermal model.According to the reaction (2),the methoxy and H on the surface can recombine to form CH3OH and desorb subsequently,which is responsible for the decrease in SFG signal of chemically adsorbed methoxy as longer pumping out.The data show that this recombination rate is small at room temperature.Although methoxy on 5-coordianted Ti sites can recombine with the H on bridge-bonded oxygen sites on rutile TiO2(100)surface at 250 K,much lower than 300 K[14],and calculation also shows that the barrier of both dissociation and recombination of methanol 5-coordianted Ti sites are very small,about 0.2 eV[13,36],methoxy groups adsorbed at variable sites on TiO2thin flm,like step,kink,and point defect sites,may have high adsorption energy,which raises the recombination barrier and lowers the recombination rate.

IV.CONCLUSION

We describe a surface specifc broadband infrared sum frequency generation vibrational spectrometer with third harmonic,266 nm,excitation that has been newly set up in our laboratory for studying surface photocatalysis on surfaces.Thein situcleanup of surface contaminants was realized by photodegradation reactions in O2atmosphere.We measured the SFG vibrational spectrum of methanol on TiO2thin flm in a diferential pumping cell,and ft with a self-developed global ftting program.Both the molecular methanol and methoxy adsorption were identifed with changing the methanol pressure,and the CH3symmetric stretching vibration frequency and Fermi resonance of molecular methanol is red-shifted by about 6?8 cm?1from low to high coverage due to its interaction with the TiO2surface at the frst layer.Obvious recombination between methoxy and H on TiO2surface was observed,which rate is small at the room temperature. Thus,we propose that two equilibriums exist∶molecular methanol between in the gas phase and on the TiO2surface,and between molecular methanol and dissociative methanol on surfaces.The thin flm surface is closer to the real applied material,but it is always too complicated to obtain more information at the atomic and molecular level due to its disorder and complexity. Therefore,the single crystal surface,which is well defned,is a good model to illuminate the adsorption and photocatalytic reaction details of organic molecules on TiO2surfaces.

Supplementary material∶The ftting results of SFG spectra dependent on the methanol pressure are given.

V.ACKNOWLEDGMENTS

This work was supported by the National Basic Research Program of China(No.2013CB834600) and the National Natural Science Foundation of China(No.1127002/B030403,No.11290162/A040106, and No.21322310/B030402).

[1]A.Fujishima and K.Honda,Nature238,37(1972).

[2]M.A.Henderson,Surf.Sci.Rep.66,185(2011).

[3]C.L.Pang,R.Lindsay,and G.Thornton,Chem.Rev.113,3887(2013).

[4]C.L.Pang,R.Lindsay,and G.Thornton,Chem.Soc. Rev.37,2328(2008).

[5]X.Lang,W.Ma,C.Chen,H.Ji,and J.Zhao,Acc. Chem.Res.47,355(2013).

[6]H.Chen,C.E.Nanayakkara,and V.H.Grassian, Chem.Rev.112,5919(2012).

[7]Z.Dohn′alek,I.Lyubinetsky,and R.Rousseau,Prog. Surf.Sci.85,161(2010).

[8]M.A.Henderson,S.Otero-Tapia,and M.E.Castro, Faraday Discuss.114,313(1999).

[9]Z.R.Zhang,O.Bondarchuk,J.M.White,B.D.Kay, and Z.Dohn′alek,J.Am.Chem.Soc.128,4198(2006).

[10]M.Shen and M.A.Henderson,J.Phys.Chem.L2, 2707(2011).

[11]K.Onda,B.Li,J.Zhao,and H.Petek,Surf.Sci.593, 32(2005).

[12]C.Y.Zhou,Z.F.Ren,S.J.Tan,Z.B.Ma,X.C.Mao, D.X.Dai,H.J.Fan,X.M.Yang,J.LaRue,R.Cooper, A.M.Wodtke,Z.Wang,Z.Y.Li,B.Wang,J.L.Yang, and J.G.Hou,Chem.Sci.1,575(2010).

[13]Q.Guo,C.Xu,Z.Ren,W.Yang,Z.Ma,D.Dai,H. Fan,T.K.Minton,and X.Yang,J.Am.Chem.Soc.134,13366(2012).

[14]Q.Guo,C.Xu,W.Yang,Z.Ren,Z.Ma,D.Dai,T. K.Minton,and X.Yang,J.Phys.Chem.C117,5293 (2013).

[15]K.R.Phillips,S.C.Jensen,M.Baron,S.C.Li,and C. M.Friend,J.Am.Chem.Soc.135,574(2012).

[16]Q.Yuan,Z.Wu,Y.Jin,L.Xu,F.Xiong,Y.Ma,and W.Huang,J.Am.Chem.Soc.135,5212(2013).

[17]B.Li,J.Zhao,K.Onda,K.D.Jordan,J.Yang,and H. Petek,Science311,1436(2006).

[18]A.Migani,D.J.Mowbray,A.Iacomino,J.Zhao,H. Petek,and A.Rubio,J.Am.Chem.Soc.135,11429 (2013).

[19]A.Migani,D.J.Mowbray,J.Zhao,H.Petek,and A. Rubio,J.Chem.Theory Comput.10,2103(2014).

[20]M.A.Henderson,S.Otero-Tapia,and M.E.Castro, Surf.Sci.412/413,252(1998).

[21]M.Xu,Y.Gao,Y.Wang,and C.Woell,Phys.Chem. Chem.Phys.12,3649(2010).

[22]M.Xu,Y.Gao,E.M.Moreno,M.Kunst,M.Muhler, Y.Wang,H.Idriss,and C.Woell,Phys.Rev.Lett.106, (2011).

[23]G.A.Kimmel,M.Baer,N.G.Petrik,J.VandeVondele, R.Rousseau,and C.J.Mundy,J.Phys.Chem.L3,778 (2012).

[24]C.Y.Wang,H.Groenzin,and M.J.Shultz,J.Phys. Chem.B108,265(2004).

[25]C.Y.Wang,H.Groenzin,and M.J.Shultz,Langmuir19,7330(2003).

[26]L.Miao,P.Jin,K.Kaneko,A.Terai,N.Nabatova-Gabain,and S.Tanemura,Appl.Surf.Sci.212/213, 255(2003).

[27]L.Velarde,X.Y.Zhang,Z.Lu,A.G.Joly,Z.M.Wang, and H.F.Wang,J.Chem.Phys.135,241102-1(2011).

[28]R.Lu,W.Gan,B.H.Wu,Z.Zhang,Y.Guo,and H. F.Wang,J.Phys.Chem.B109,14118(2005).

[29]Y.Yu,Y.Wang,K.Lin,N.Hu,X.Zhou,and S.Liu, J.Phys.Chem.A117,4377(2013).

[30]R.Superfne,J.Y.Huang,and Y.R.Shen,Phys.Rev. Lett.66,1066(1991).

[31]G.S.Herman,Z.Dohn′alek,N.Ruzycki,and U. Diebold,J.Phys.Chem.B107,2788(2003).

[32]D.A.Panayotov,S.P.Burrows,and J.R.Morris,J. Phys.Chem.C116,6623(2012).

[33]R.G.Snyder and J.R.Scherer,J.Chem.Phys.71, 3221(1979).

[34]R.G.Snyder,H.L.Strauss,and C.A.Elliger,J.Phys. Chem86,5145(1982).

[35]R.G.Snyder,A.L.Aljibury,H.L.Strauss,H.L.Casal, K.M.Gough,and W.F.Murphy,J.Chem.Phys.81, 5352(1984).

[36]R.Sanchez de Armas,J.Oviedo,M.A.San Miguel, and J.F.Sanz,J.Phys.Chem.C111,10023(2007).

[37]J.Zhao,J.Yang,and H.Petek,Phys.Rev.B80, 235416(2009).

10.1063/1674-0068/28/cjcp1409154ofers rich evidence of its photocatalytic reaction in the molecular level[13-16].However,TPD only probes the desorption species from the heating surface,and we cannot exclude any thermal reaction when heating up the sample.The high resolution electron energy loss spectroscopy(HREELS)and low energy electron difraction (LEED)can providein situmeasurements,but Henderson and co-workers showed that electrons can induce decomposition of methanol on TiO2(110)[8,20].IR absorption spectroscopy is a powerful tool for identifying the species on oxide surfaces.Only a few experiments have been done on TiO2single crystal due to extremely low signal[21-23]while there are a large set of studies on TiO2powders and thin flms[2].Instead of linear IR spectroscopy,Shultz and co-workers,for the frst time,employed a second-order surface sum frequency generation vibrational spectroscopy(SFG-VS), which is surface specifc,to characterize the methanol and methoxy adsorption structures on anatase TiO2nanoparticle thin flms[24,25].

?These authors contributed equally to this work.

?Author to whom correspondence should be addressed.E-mail：zfren@pku.edu.cn

(Dated:Received on September 16,2014;Accepted on October 20,2014)