S型SnO2/BiOBr異質結光催化還原CO2的電荷傳輸機理

關鍵詞：SnO2/BiOBr；光催化還原二氧化碳；S型異質結；fs-TA；原位XPS

1 Introduction

Promoting carbon peaking and carbon neutrality is recognizedas an inherent requirement for achieving high-qualitydevelopment 1–5. It is conducive to accelerating the achievementof the dual-carbon goal by converting CO2 into high-valueaddedfuels utilizing solar energy 6–10. Unfortunately， theconversion efficiency of photocatalysts is generally low forpractical application due to high dissociation energy （～750kJ·mol?1） of the C＝O bond of CO2 molecules 11–13. Therefore，designing an efficient photocatalyst system is a great challengeto boost the efficiency of CO2 photoreduction.

Generally， photocatalysts are improved by enhancing lightabsorption， accelerating charge transfer and promoting surfaceredox reactivity 14，15. It is well known that the construction ofsuitable heterojunction photocatalysts is an important means forimproving photocatalytic efficiency. In particular， the recentlyproposed S-scheme heterojunction plays a significant role inseparating carriers and retaining the strong redox capacity ofsemiconductor photocatalysts respectively 16–19. The explanationof the electron transfer mechanism of S-scheme heterojunction，as well as thermodynamics and kinetics， has attracted extensiveresearch. For example， S-scheme AgBr/BiOBr heterojunctionswith surface oxygen vacancies prepared by Miao et al. exhibitedexcellent CO yields up to 212.6 μmol?g?1?h?1， which is 9.2 timeshigher than that of pure BiOBr 20. The photocatalytic activity ofCeO2/BiOI S-scheme heterojunction designed by Xiao et al. wassignificantly improved 21. The FeNi2S4/ZnIn2S4 S-schemephotocatalyst reported by Wang et al. presented remarkablephotocatalytic H2 evolution rate with 7.7 mmol?g?1?h?1， which isabout 15.2 times higher than that of the pristine ZIS 22. TheIn2O3/Bi19Br3S27 S-scheme heterojunction synthesized by Bianet al. showed a CO yield 19 times higher than that of pristineIn2O3 23. A 3.12-fold increase in catalytic efficiency forphotocatalytic reduction of CO2 by CeO2/Bi2MoO6 S-schemeheterojunction was designed by Xu et al. 24. It has already beenproven that the development of S-scheme heterojunctionsprovides an excellent way to increase the redox capacity ofphotocatalyst. However， the specific transport mechanism ofphotogenerated carriers in S-scheme heterojunctions isambiguous， and probing the transport kinetics of electrons andholes is necessary.

Bismuth-based photocatalysts with unique anisotropicstructure， good optical， electrical properties and excellentphotovoltaic performance have attracted much attention in thefields of photocatalytic CO2 reduction and degradation ofpollutants 25，26. Among them， BiOBr has a suitable band gap aswell as unique interleaved and stacked [Bi2O2]2+ and Br– layers，facilitating the formation of an internal electric field andaccelerating the diffusion and separation of photo-inducedcarriers， which is promising in a broad range of applications 27–29.In addition， as a broadband semiconductor， SnO2 has attracted awide range of attention in the domain of photocatalysis due to itsenvironmental friendliness， low cost， non-toxicity， and stableoptical properties 30–32. Unfortunately， wide band gap of SnO2results in lower solar energy utilization， hindering its applicationunder visible light. Generally speaking， reduction photocatalystshave higher CB and Fermi energy levels 33. BiOBr has a narrowband gap， which could absorb visible light and exhibit strongreducing ability， thereby serving as a reduction semiconductor.Moreover， SnO2 has a low conduction band position and strongoxidizing ability. Therefore， we expect to improve thephotocatalytic efficiency by constructing SnO2/BiOBr S-schemeheterojunctions with strong reduction and oxidation capabilityand investigate the carrier kinetic law by femtosecond transientabsorption （fs-TA） spectroscopy measurements.

In this work， SnO2/BiOBr S-scheme heterojunctions aresuccessfully synthesized through an in situ hydrothermal process.The photocatalytic performance for CO2 reduction ofSnO2/BiOBr is assessed under simulated sunlight illumination.It is demonstrated that photocatalytic efficiency of SnO2/BiOBrheterojunction is significantly improved up to 345.7μmol?g?1?h?1 for CO product in comparison to pristine BiOBr. Inparticular， fs-TA spectroscopy is employed to analyze theultrafast electron transfer at SnO2/BiOBr S-schemeheterojunction interface， demonstrating S-schemeheterojunction could suppress electron and hole complexation.In addition， it is evident from experimental and computationalcharacterization that the prepared SnO2/BiOBr has an S-schemestructure， realizing the effective separation of photogeneratedcarriers， retaining a strong redox capacity， and improving thephotocatalytic performance for CO2 reduction. It can beexpected that this work will contribute to the research of electrontransfer pathways for novel heterojunction photocatalysts.

2 Experimental

First， 5 mmol of Bi（NO3）3·5H2O was taken and dispersed into30 mL of deionized water with 15 min of sonication and 45 minof stirring， named as solution A. 5 mmol NaBr was dispersedinto 10 mL of deionized water， recorded as solution B. Thensolution B was gradually dropped to solution A and stirred for 30min. A quantity of SnCl4·5H2O was solubilized in a mixture of10 mL of anhydrous ethanol and 10 mL of NaOH， and theresulting solution with stirring was recorded as solution C.Solution C was added to the mixed solution of solution A andsolution B with magnetic stirring for 1 h. The final solution waspoured into a 100 mL polytetrafluoroethylene high-pressurereactor and hydrothermally heated at 110 °C for 12 h. Uponcooling to room temperature， the suspension was washed severaltimes with deionized water and alcohol， dried in an oven at 75 °C.By adjusting the amount of SnCl4·5H2O， SnO2/BiOBrheterojunctions with different molar ratios were obtained （SnO2 ：BiOBr = 0.35 ： 1， 0.4 ： 1， 0.45 ： 1）， which were recorded as 35%SnO2/BiOBr， 40% SnO2/BiOBr， 45% SnO2/BiOBr. Pure BiOBrand pristine SnO2 were prepared with the same methods.

The detailed measurement method of CO2 photoreduction andmaterials characterization are shown in the SupportingInformation.

3 Results and discussion

3.1 Structure composition of heterojunctions

Fig. 1a reveals the XRD patterns for pristine BiOBr， SnO2 andSnO2/BiOBr heterojunctions with various molar fractions.Diffraction peaks of as-prepared BiOBr and different ratios ofSnO2/BiOBr heterojunctions are consistent with the peaks oftetragonal BiOBr （JCPDS No. 09-0393） 34. The as-preparedSnO2 belongs to the tetragonal phase SnO2 with rutile structure（JCPDS No. 01-0625） 35. The diffraction peaks are relativelywide， indicating that the samples are smaller in size. No otherstray peaks are observed in the plots. The specific surface areaof 40%SnO2/BiOBr is the largest compared to other catalysts，which is more favorable for adsorption of CO2 and enhancementof photocatalytic reduction efficiency （Fig. 1b）. As illustrated bythe CO2 adsorption isotherms of the samples in Fig. S1， theadsorption of 40% SnO2/BiOBr is greater than that of pureBiOBr and pure SnO2. Fig. S2a–c show the SEM images ofpristine BiOBr， pristine SnO2， and 40%SnO2/BiOBr，respectively. It is evident that pure BiOBr consists of flakes withthe thickness of 25–30 nm. Meanwhile， SnO2 exhibits anirregular block structure. Fig. 1c，d present the TEM and HRTEMimages of 40%SnO2/BiOBr， in which distinct two-phase grainboundaries are readily viewed with lattice spacing of 0.27 nmand 0.32 nm belongs to the （110） facet of BiOBr and the （110）facet of SnO2， respectively. The elemental mapping （Fig. 1e–h）suggests that SnO2/BiOBr is comprised of Sn， Bi， O， Br withuniform distribution. All the above experimental data providefurther evidence of the satisfactory synthesis of heterojunctionwith high-quality exposed （110） crystal faces.

3.2 Property of CO2 photoreduction

Photocatalytic CO2 reduction performance of photocatalystsis evaluated under 300 W xenon lamp irradiation. Fig. 2aillustrates the photocatalytic CO and CH4 yield of BiOBr， SnO2and SnO2/BiOBr with different molar ratios under lightirradiation. The yields of CO and CH4 for pristine SnO2 are 49.2and 1.8 μmol?g?1?h?1. The yields of CO and CH4 for pristineBiOBr are 61.8 and 2.6 μmol?g?1?h?1， respectively. The CO andCH4 yields of 40%SnO2/BiOBr are significantly increased，reaching 345.7 and 6.7 μmol?g?1?h?1， respectively， which are 5.6and 3.7 times higher than those of pristine BiOBr. The CO2conversion rate of this work is more excellent than other studies（Table S1）. Fig. 2b indicates the CO conversion of40%SnO2/BiOBr under different reaction conditions. Withoutadding catalysts， without giving light or switching the CO2atmosphere to N2 atmosphere， a small amount of CO can bedetected， indicating that the CO originated from CO2， andsuggesting that CO2， photocatalyst and light in the reactionsystem are all necessary for photocatalytic reaction. Isotopetracer13CO2 experiments are carried out to investigate the originof the products for CO2 reduction （Fig. S3）. The peaks of the ionfragments at m/z = 17 and 29 correspond to 13CH4 and 13CO，respectively. Cycling experiment reveals that there is still nosignificant attenuation of CO yield after five cycles，demonstrating the stability of the heterojunction photocatalysts（Fig. 2c）. Fig. 2d illustrates that CO yield enhances with lighttime increasing， which is basically a linear relationship.

3.3 Energy band structure， electron transfermechanism on photocatalysts

XPS is utilized to discuss the electronic states of BiOBr， SnO2and 40%SnO2/BiOBr. In survey XPS spectra of40%SnO2/BiOBr （Fig. S4）， elements Bi， O， Br， and Sn arepresent in the sample. High-resolution XPS spectra of thecatalysts are as following. It can be observed that Bi 4f XPSspectra have four characteristic peaks located at 159.8， 161.2，165.2， and 166.4 eV （Fig. 3a）， which are ascribed to the Bi（3?x）+and Bi3+ of BiOBr 36，37. As presented in Fig. 3b， the O 1s XPSspectra show the simultaneous presence of lattice oxygens，oxygen vacancies （Ovs）， and surface hydroxyls in BiOBr at530.5， 531.9 and 533.8 eV， respectively 38–40. Notably， thecharacteristic peaks observed at 66.8 and 67.8 eV are caused byBr 3d5/2， Br 3d3/2 （Fig. 3c） 41. The peaks at 486.7， 495.2 eV inFig. 3d are typical peaks for Sn 3d5/2， Sn 3d3/2 of Sn4+ 42.Significantly， peaks of Bi， O， and Br in heterojunctions shift tohigher binding energies with respect to the original BiOBr.Meanwhile， the peak of Sn exhibits a negative displacement. Thevariation on binding energy indicates the spontaneous transfer ofelectrons from BiOBr to SnO2 due to the formation ofheterojunction， leading to energy band bending and generationof an internal electric field （IEF） at the interface. Furthermore，peaks of Bi， O， Br shift negatively while peak of Sn shiftspositively under light. It is a distinct evidence for light-inducedelectron transport from SnO2 to BiOBr， conforming to the Sschemephotocatalytic mechanism.

Aiming to further investigate the electron transfer mechanismof SnO2/BiOBr heterojunction， an attempt is made to investigatethe energy band structures over SnO2 and BiOBr by UV-Visdiffuse reflectance spectroscopy （DRS） and valence band X-rayphotoelectron spectrum （VBXPS）. Fig. 4a shows UV-Vis DRSof pristine BiOBr， SnO2， and SnO2/BiOBr with various molarproportions. It can be found that the absorption edge ofSnO2/BiOBr is red-shifted compared to pure SnO2. In Fig. 4b，the band gaps of BiOBr and SnO2 are 2.94 and 3.32 eVrespectively， according to the following equation： αhν = A （hν ?Eg） n/2， where α， hν， A and Eg represent the absorption coefficient，photon energy， constant and band gap， respectively 43，44. Flatband potentials for BiOBr and SnO2 with positive slopes areshown in the Mott Schottky curves （Fig. 4c）. The conductionbands of the n-type semiconductors BiOBr and SnO2 arecalculated as ?0.59 and 0.11 V， respectively 45. Based on theformula Eg = EVB ? ECB 46， the valence bands of BiOBr and SnO2can be calculated as 2.35 and 3.43 V， respectively. As determinedfrom VBXPS tests， the distance from the valence band to theFermi energy level is 2.40， 3.00 eV for BiOBr and SnO2，respectively， indicating that the Ef of SnO2 is lower than that ofBiOBr. Depending on the information obtained， the energy bandstructures of BiOBr， SnO2 are deduced in Fig. 4d.

Ultraviolet photoelectron spectroscopy （UPS） is applied forinvestigating the work function （Φ） of BiOBr and SnO2 （Fig.5a，b）. According to the formula Φ = hν ? （Ecutoff ? EVF） 47–50， thework functions of BiOBr and SnO2 are calculated to be 6.15 and6.60 eV， respectively. Therefore， when the two semiconductorsare in contact， the electrons in BiOBr will be transferred to SnO2spontaneously， which is in agreement with the XPS. Fig. 5creveals the charge density difference of SnO2/BiOBr by Densityfunctional theory （DFT） simulation. The blue and yellow regionsrepresent electron enrichment and electron depletion regions，respectively. Yellow regions on the surface of BiOBr indicatethe decrease of its electron density， whereas blue regions on thesurface of SnO2 indicate an increase of its charge density.Overall， BiOBr provides electrons while SnO2 gains electrons，generating an electron transfer path from BiOBr to SnO2， whichis consistent with experimental findings. In accordance withprevious experimental and computational results， a mechanismfor photogenerated carrier separation and transfer overphotocatalysts is proposed （Fig. 5d–f）. Before contact， the Efposition of BiOBr is higher than that of SnO2. Whenheterojunction catalyst is formed， free electrons in BiOBr arespontaneously transferred to SnO2 until the Ef bends. SnO2 gainselectrons to form an electron-rich region， and BiOBr loseselectrons to form an electron-depleted region. Thereby， theenergy bands at the interface of SnO2 and BiOBr are bent.Simultaneously， an internal electric field pointing from BiOBr toSnO2 is created. Given full-band illumination， electronspositioned in the VB of SnO2 and BiOBr are excited into CB，respectively， driven by energy band bending and IEF， electronson SnO2 CB migrate to BiOBr VB and recombine with itsphotogenerated holes. Ultimately， CO2 reduction takes place inthe CB of BiOBr. This process not only achieves effectiveseparation of photogenerated carriers， but also retains a strongredox capacity， which is thermodynamically and kineticallyfavorable for photocatalytic reduction of CO2.

3.4 Ultrafast electron transfer at the SnO2/BiOBrheterojunction interface

To further understand the charge transfer dynamics within theheterojunction， femtosecond transient absorption （fs-TA）spectroscopy measurements are measured on SnO2 andSnO2/BiOBr. As shown in pseudocolor fs-TA plots （Fig. 6a） andfs-TA spectra （Fig. 6c） of SnO2， a broad negative peak at ～500nm can be found， corresponding to the superposition of groundstate bleach （GSB） signal and stimulated emission （SE）signal 51–53. Since similar fs-TA signals were also detected inthe SnO2/BiOBr composites （Fig. 6b，d）， the fs-TA signals ofSnO2/BiOBr are attributed to the GSB and SE signals of SnO2component. Subsequently， the normalized fs-TA decay kineticsof pure SnO2 probed at 500 nm within 200 ps are fitted a doubleexponential function （Fig. 6e）. For SnO2/BiOBr， the fs-TA decaycurves probed at 500 nm within 200 ps are fitted with a triexponentialfunction （Fig. 6f）. As exhibited in Fig. 6g，photogenerated electron relaxation within pure SnO2 isassociated with two pathways （I） electron relaxation in the SnO2CB （τ1 = 1.51 ps） and （II） electron trapping at shallow trap states（τ2 = 17.44 ps） 54–56. Two lifetime components with similar timevalues （τ1 = 2.57 ps and τ2 = 12.36 ps） can also be obtained in theheterojunction， and they can be attributed to the same electronquenching dynamics as in pure SnO2. Upon forming theSnO2/BiOBr heterojunction， an ultrafast channel forphotoelectrons in SnO2 CB to transfer to the BiOBr VB appears.Therefore， the newly fitted lifetime （τ3 = 34.61 ps） could beplausibly attributed to S-scheme heterojunction interfacialelectron transfer （Fig. 6h）.

3.5 DFT calculations of photocatalysts

Partial density of states （PDOS） of pristine BiOBr， pristineSnO2 and SnO2/BiOBr are obtained by DFT calculations （Fig.7a–c）. The PDOS of pure BiOBr shows the valence bandmaximum （VBM） is composed of Br 4p and O 2p orbitals whilethe conduction band minimum （CBM） is made up of Bi 6porbitals （Fig. 7a）. Fig. 7b reveals that the VBM of pristine SnO2is mainly determined by the O 2p orbitals and the CBM is formedof O 2p and Sn 5s states. With the formation of the SnO2/BiOBrheterojunction， the VBM is made up of O 2p orbitals， and theCBM is occupied by Bi 6p， O 2p， Sn 5s orbitals （Fig. 7c）. Thetotal density of states （Fig. S5） indicate the shift in the positionof the DOS peak for the heterojunction， illuminating the orbitalhybridization occurs. Fig. 7d–f is band structure of pure BiOBr，pristine SnO2 and SnO2/BiOBr. It can be found that the CBM andVBM of the semiconductor are at different high symmetry points，indicating that both of SnO2 and BiOBr are indirectsemiconductors. SnO2/BiOBr exhibits a denser energy bandstructure compared to pristine BiOBr and pristine SnO2，contributing to higher electron mobility rates 57–59.

3.6 Electrochemical measurement of photocatalysts

The carrier separation and transfer efficiency of the catalystsare investigated by transient photocurrent， electrochemicalimpedance， photoluminescence emission spectroscopy， andtime-resolved PL decay. As seen in Fig. 8a， photocurrent densityof 40%SnO2/BiOBr is obviously stronger compared to pureBiOBr and pure SnO2， suggesting that the formation ofheterojunction results in an improvement in carrier separationefficiency. Fig. 8b is the electrochemical impedance curve of thepristine BiOBr， pristine SnO2 and 40%SnO2/BiOBr.40%SnO2/BiOBr shows the smallest radius， which suggests thesmallest carrier transfer resistance. As depicted in Fig. 8c， the PLintensity of 40%SnO2/BiOBr is significantly lower than that ofpure SnO2 and pure BiOBr， implying that the formation ofheterojunctions could effectively inhibit the recombination of h+and e?. In Fig. 8d， the average fluorescence lifetime is extendedfrom BiOBr （8.25 ns） and SnO2 （7.04 ns） to SnO2/BiOBr （8.93ns） in the time-resolved PL decay curves 60. The results ofphotoelectrochemical tests can indicate that the formation ofSnO2/BiOBr S-scheme heterojunction is capable of acceleratingthe charge transfer rate and effectively inhibiting thephotogenerated carrier complexes.

3.7 Reaction mechanism of CO2 photoreduction

To acquire a deeper explanation of the photocatalytic CO2reduction process， in situ FTIR of 40%SnO2/BiOBr is employedto test the reaction intermediates （Fig. 9a）. The absorption peakslocated at 1370.9， 1437.3， and 1738.9 cm?1 can be assigned tobidentate adsorption of CO2 （b-CO2）， monodentate adsorption ofCO2 （m-CO2）， and bridging adsorption of CO2 （c-CO2） 61–63.Vibration bands at 3593.6， 3638.1， 3702.1， 3725.8 cm?1 arecaused by OH in H2O （ν（OH）） 64–66. The peaks at 1103.7， and1152.2 cm?1 are originated from CHO* and CH3O* 67，68. Theinfrared peaks at 1232.2 and 2985.3 cm?1 are typical of *COOHgroup， which is regarded a critical intermediate in photocatalyticconversion from CO2 to CO 69，70. What?s more significant is theabsorption signal of CO （2075.5 cm?1） is observed in the infraredspectra， providing evidence for the eventual conversion of theintermediates to CO to some extent 71. Summarizing the aboveresults， a potential pathway for photocatalytic conversion of CO2to CO can be proposed （Fig. 9b）. The conversion pathway forCO2 on the surface of SnO2/BiOBr S-scheme heterojunction isrepresented via a schematic diagram. Hydrogenation of CO2 toCOOH* is an accessible process. Subsequently， COOH*intermediates undergo two different transformation processes.On the one hand， the converted COOH* combines with the H+/e?pair forming CO* and releasing CO. On the other hand，successive hydrogenations result in the sequential conversion ofCOOH* to CHO*， CH2O*， CH3O*， and eventually CH4.

4 Conclusion

In summary， tight SnO2/BiOBr S-scheme heterojunctionphotocatalysts are prepared by in situ hydrothermal method. Thecatalytic performance of SnO2/BiOBr is evaluated under fullbandlight conditions. 40%SnO2/BiOBr shows optimalperformance with CO yield up to 345.7 μmol?g?1?h?1， which is5.6 times higher than that of the pristine BiOBr. Fs-TAS isutilized to prove ultrafast charge transfer on the SnO2/BiOBrinterface， concerning the complexation of photogeneratedelectrons in SnO2 CB and holes in BiOBr VB. In addition， thephotocatalytic mechanism of SnO2/BiOBr S-schemeheterojunctions is speculated according to the energy bandstructure of the photocatalysts， in situ irradiation XPS， fs-TA andDFT calculations. This work contributes to accelerating theunderstanding of ultrafast charge transfer utilizing fs-TAanalysis.

Author Contribution： Conceptualization， Y.A. and Y.Z.;Methodology， Y.A. and Y.Z.; Software， Y.A.; Validation， Y.A.and W.L.; Formal Analysis， Y.A.， W.L. and J.Z.; InvestigationY.A.; Resources， Y.A.， W.L. and J.Z.; Data Curation， Y.A.;Writing-Original Draft Preparation， Y.A.; Writing-Review amp;Editing， Y.Z.， Z.L. and J.Z.; Visualization， Y.A. and W.L.;Supervision， Y.Z.， Z.L. and J.Z.; Project Administration， Y.Z.，Z.L. and J.Z.; Funding Acquisition， Y.Z.， Z.L. and J.Z.

Supporting Information： available free of charge via theinternet at https：//www.whxb.pku.edu.cn.