Chinese Chemical Letters  2018, Vol. 29 Issue (6): 752-756   PDF    
Thermal-, photo-and electron-induced reactivity of hydrogen species on rutile TiO2(110) surface: Role of oxygen vacancy
Zongfang Wu, Feng Xiong, Zhengming Wang, Weixin Huang    
Hefei National Laboratory for Physical Sciences at the Microscale, CAS Key Laboratory of Materials for Energy Conversion and Department of Chemical Physics, University of Science and Technology of China, Hefei 230026, China
Abstract: Interaction of hydrogen with TiO2 plays a vital role in TiO2-based photocatalysis and thermal catalysis. In this work, we compared thermal-, photo-, and electron-induced reactivity of various types of hydrogen species on a rutile TiO2(110) surface formed by atomic H exposure at 320 and 115 K by means of thermal desorption spectroscopy, X-ray photoelectron spectroscopy and low energy electron diffraction. Atomic H interaction with rutile TiO2(110) at 115 K forms surface Ti—H hydride, surface hydroxyl group, and chemisorbed water. Upon heating, surface Ti—H hydride reacts to produce H2 while surface hydroxyl groups react to form both water and H2. Atomic H interaction with rutile TiO2(110) at 320 K strongly reduces TiO2 due to the continuous formation and desorption of water and forms surface hydroxyl groups and likely subsurface/bulk hydrogen species. Upon heating, hydrogen forms as the only gas-phase product and its desorption activation energy decreases with the subsurface/bulk reduction extent of rutile TiO2(110). Surface Ti—H hydride exhibits photo-induced reactivity while both surface Ti—H hydride and surface hydroxyl group exhibit electro-induced reactivity. These results have important implications for understanding the hydrogen-involved thermal and photo reactions on TiO2-based catalysts.
Key words: Surface chemistry     Atomic H     Surface Ti hydride     Surface hydroxyl group     Subsurface/bulk hydrogen species    

TiO2-based materials show promising application prospects in heterogeneous catalysis, solar cells, electronic devices, and so on [1-6]. The reactivity of hydrogen species on TiO2 surfaces is a very important issue in both TiO2 thermal catalysis and photocatalysis, however, the relevant fundamental understanding is quite limited, and even the type of hydrogen species on TiO2 surfaces is not fully established [5-9].

Single crystal surfaces with well-defined surface structures have been commonly employed to model corresponding powder catalysts for fundamental investigations. Rutile TiO2(110) surface is the most frequently used surface to model TiO2 catalysts [4, 7, 10]. Molecular H2 generally does not react with TiO2 surface under UHV conditions. Thermally cracked energetic atomic hydrogen has been demonstrated as an effective method to prepare hydrogen species on oxide surfaces under UHV conditions [11-20]. Interaction of atomic H with rutile TiO2(110) surface at room temperature (r.t.) has been previously reported [11, 12, 21-26]. Surface hydroxyl groups were commonly observed, but dispute remains on the formation of hydride-type Ti—H species as well as subsurface hydroxyl groups. Meanwhile, the reactivity of various hydrogen species on rutile TiO2(110) surface was also reported to be complex and ambiguous [11, 12, 22, 25, 27, 28].

We previously reported the interaction of atomic H with rutile TiO2(110) surface at 115 K [13]. Formations of adsorbed water, surface hydroxyl groups and Ti—H hydride were observed and their thermos and photo reactivity were identified, particularly the surface Ti—H hydride was firstly found photo-reactive. In this paper, we employed thermal desorption spectroscopy (TDS), photo-simulated desorption (PSD) spectroscopy, X-ray photoelectron spectroscopy (XPS) and low energy electron diffraction (LEED) to study the interaction of atomic H with rutile TiO2(110) surface at 320 K and compared the results with those at 115 K. The results show that both the formed hydrogen species and their reactivity are significantly influenced by the reduction extent of rutile TiO2(110) sample.

All experiments were performed in a UHV chamber with a base pressure of 5 ×10-11 mbar [10, 13, 29]. The chamber was equipped with a differentially-pumped quadrupole mass spectrometer, as well as facilities for low energy electron diffraction and X-ray photoelectron spectroscopy. A rutile TiO2(110) single crystal bought from MaTeck was mounted onto a Ta support plate (1 mm thick and of the same dimensions as the crystal) using a high temperature alumina-based inorganic adhesive (Aremco 503) and graphite powder (99.9995%, Alfa Aesar China Co., Ltd.). The Ta support was cooled and resistively heated by two Ta wires spotwelded to its backside. The sample temperature could be controlled between 100 K and 1273 K and was measured using a chromel-alumel thermocouple spot-welded to the backside of the sample. Prior to experiments, the rutile TiO2(110) sample was cleaned by repeated cycles of Ar ion sputtering, oxidation and annealing at 1000 K for 10 min until LEED gave a sharp (1 ×1) diffraction pattern and no contaminants could be detected by XPS. The reproducible preparation of the clean rutile TiO2(110) surface was verified by means of CO2 TDS experiments. D2O (D>99.9%, SIGMA-ALDRICH) and H2O (>18 MV) were purified by repeated freeze-pump-thaw cycles. D2 (>99.999%, Nanjing ShangYuan Industry Factory) and CO2 (>99.99%, Nanjing ShangYuan Industry Factory) were used as received. The purity of all reactants was checked by QMS prior to experiments. The exposure of atomic hydrogen (deuterium) was accomplished using a MGC75 thermal gas cracker with an Ir capillary. All exposures were reported in Langmuir (1 L = 1.0 ×10-6 Torr·s) without corrections for the gauge sensitivity. During the TDS measurements, the sample was positioned ~1 mm away from a collecting tube of a differentialpumped QMS and the heating rate was 2 K/s. XPS spectra were recorded using Mg Kα radiation ( = 1253.6 eV) with a pass energy of 20 eV.

UV irradiation was accomplished using a 100 W Hg arc lamp (Oriel 6281) which provided a pressure-broadened emission spectrum from gaseous Hg with significant intensity in the UVlight region. A water filter was used to remove the IR portion of the emission spectrum. The UV-light was focused onto the tip of a single strand, a 0.6 mm diameter fused silica fiber optic cable that directed the light through a UHV-compatible feedthrough onto the rutile TiO2(110) face without exposure to extraneous surfaces. Exposure of the TiO2(110) crystal at 115 K to the UV-light resulted in the increase of crystal temperature by no more than 3 K.

Fig. 1 compares TDS spectra after an exposure of 50 L atomic D on rutile TiO2(110) surface at 115 and 320 K. As reported previously [13], both D2/HD and D2O/HDO desorption traces were observed after the exposure at 115 K. A sharp D2 desorption feature at the onset edge of D2 desorption trace (at ~150 K) without the corresponding HD desorption trace in the HD desorption trace could be attributed to the desorption of molecularly-chemisorbed D2 on the Ta plate used to support the TiO2 sample. The major D2/ HD peak at 205 K and the shoulder peak at 165 K arise from surface reactions of surface Ti—H hydride species while the shoulder peak at 292 K arises from surface reaction of surface hydroxyl groups. The major water desorption at 315 K and the broad shoulder peak at 465 K respectively arise from the desorption of molecularlyadsorbed water and surface reaction of surface hydroxyl groups. These results demonstrate the formation of surface Ti—H hydride species, surface hydroxyl group and adsorbed water on rutile TiO2(110) surface upon atomic H exposure at 115 K. The formation of surface Ti—H hydride species is related with the reduction of rutile TiO2(110) surface upon the formation of surface hydroxyl group and water from atomic hydrogen [13].

Download:
Fig. 1. TDS spectra after 50 L atomic D exposures on rutile TiO2(110) surface at 320 K (black lines) and 115 K (red lines).

Following the exposure at 320 K, only a strong D2 desorption peak initiating at 550 K and centering at 825 K appears while no HD and water desorption traces were observed. At 320 K, the adsorption of H2O from the residual gas in the UHV chamber on rutile TiO2(110) surface is negligible and thus no H-D exchange reaction occurs. Meanwhile, as demonstrated in the TDS spectra following the exposure at 115 K, it is reasonable that the surface Ti—H species and molecularly-adsorbed D2O species are unstable at 320 K. However, it is interesting that the water desorption peak resulting from surface reaction of surface hydroxyl groups is not observed. Both HREELS and SPM results previously verified the formation of surface hydroxyl groups upon atomic H exposure to rutile TiO2(110) surface at RT [11, 25, 30]. Thus our TDS results suggest that surface hydroxyl groups formed on rutile TiO2(110) surface upon atomic H exposure at 320 K selectively react to produce hydrogen. However, surface hydroxyl groups formed on rutile TiO2(110) surface upon atomic H exposure at 115 K react to produce both water and hydrogen. Therefore, the water production pathway from surface hydroxyl groups formed on rutile TiO2(110) surface upon atomic H exposure at 320 K is completely suppressed. By investigating the reactivity of surface hydroxyl groups on iron oxide thin films and ceria thin films and nanocrystals [14-20], we established a concept of local oxygen vacancy-controlled reactivity of hydroxyl groups on oxide surfaces. Hydroxyl groups on stoichiometric oxide surfaces preferentially react to produce water and surface oxygen vacancy, but this reaction pathway is gradually poisoned and eventually suppressed with the increasing coverage of local oxygen vacancy on oxides, and other reaction pathways open, including the reaction of surface hydroxyl groups to produce hydrogen and the interfacial reaction of hydroxyl groups on oxide surface with CO adsorbed on metal surface to produce CO2. A recent DFT calculation study also reported that although surface hydroxyl groups prefer to produce H2O on stoichiometric rutile TiO2(110) surface, H2 production becomes more feasible than water production with the increasing coverage of surface oxygen vacancy, especially in the local areas where hydroxyl groups are close to oxygen vacancies [28]. Therefore, the suppress of water production pathway from surface hydroxyl groups formed on rutile TiO2(110) surface upon atomic H exposure at 320 K indicates a much more extensive oxygen vacancy formation on rutile TiO2(110) surface exposed to atomic H at 320 K than at 115 K. This is reasonable because the creation of oxygen vacancies of rutile TiO2(110) surface by atomic H exposure is self-terminated by the formation of adsorbed water at 115 K but it becomes continuous by the formation and subsequent desorption of water from the surface at 320 K.

Figs. 2A and B compare TDS spectra after repeated cycles of 50 L atomic D exposure on rutile TiO2(110) surface respectively at 115 K and 320 K. Both D2/HD and water desorption traces after the 1st and 2nd cycles of atomic D exposure at 115 K are almost identical. This indicates that the surface reduction of rutile TiO2(110) surface resulting from the water formation in the 1st cycle of TDS experiment should not affect much its reactivity towards atomic D in the subsequent 2nd cycle of TDS experiment. We consider two likely reasons: one is that the surface oxygen vacancies created by the formation and desorption of water are refilled by the migration of subsurface/bulk lattice oxygen to the surface at elevated temperatures and the oxygen vacancies are located in the subsurface/bulk region after the TDS measurement; the other is that the coverage of surface oxygen vacancies is quite low and thus the resulting oxygen vacancies located in the subsurface/bulk region does not affect the surface reactivity of rutile TiO2(110) surface.

Download:
Fig. 2. TDS spectra following repeated cycles of 50 L atomic D exposures on rutile TiO2(110) surfaces at (A) 115 K and (B) 320 K.

The case is very different at 320 K. Only a broad symmetric D2 desorption feature was observed during the TDS measurements following repeated cycles of atomic D exposure at 320 K, corresponding to a second-order recombinative reaction of two hydroxyl groups to produce hydrogen. With the cycle number of TDS measurement increasing, the shape of the D2 desorption peak does not change, but the desorption temperature keeps shifting to low temperatures from 825 K in the 1st cycle to 650 K in the 6th cycle, and meanwhile the peak intensity keeps decreasing and no D2 could be observed in the 8th cycle of TDS measurement. The D2 desorption trace following the 1st atomic D exposure can be restored by oxidation of the sample in 10-6 mbar O2 at 700 K followed by annealing at 1000 K. For a desorption peak with a second-order desorption kinetics, the desorption temperature generally shifts to high temperatures as the peak intensity decreases [31], thus the shifting of D2 desorption peak temperature to low temperatures with the peak intensity decreasing suggests a significant restructure of rutile TiO2(110) sample during the repeated cycles of TDS measurement of atomic D exposure at 320 K that decreases the desorption activation energy of chemisorbed D species. Such a restructure is reasonably associated with the extensive reduction of rutile TiO2(110) sample by atomic D exposure at 320 K.

The surface structure of rutile TiO2(110) sample after the 8th TDS cycle of 50 L atomic D exposure at 320 K was examined carefully by LEED pattern, Ti 2p and O 1s XPS spectra and CO2-TDS spectrum. The results shown in Fig. 3 demonstrate that its LEED pattern, Ti 2p and O 1s XPS spectra and CO2-TDS spectrum are almost identical to those of the starting clean rutile TiO2(110) surface. Particularly, it was previously reported that the CO2 adsorption on rutile TiO2(110) surface is sensitive to the surface structure and CO2 adsorbed at the surface oxygen vacancy site exhibits a desorption peak at a higher temperature than that from the Ti5c site [2]. Therefore, although experiencing a serious restructure, the rutile TiO2(110) surface after the repeated cycles of TDS measurement following 50 L atomic D exposure at 320 K should not change much. This agrees with the above results at 115 K to indicate that the surface oxygen vacancies created by the formation and desorption of water are healed by the migration of subsurface/bulk lattice oxygen to the surface at elevated temperatures and the oxygen vacancies are located in the subsurface/bulk region after the TDS measurement. However, different from the results at 115 K that the coverage of resulting oxygen vacancies located in the subsurface/bulk region is too small to affect the surface reactivity of rutile TiO2(110) surface, their coverage is large enough due to the extensive reduction at 320 K to significantly affect the reactivity of surface hydroxyl groups on rutile TiO2(110) surface. The accumulation of subsurface/bulk oxygen vacancies on rutile TiO2(110) sample results in not only the selective production of hydrogen from surface hydroxyl groups but also the decrease of the hydrogen formation activation energy from surface hydroxyl groups.

Download:
Fig. 3. (A) LEED patterns, (B) Ti 2p XPS spectra, (C) O 1s XPS spectra, and (D) CO2 TDS spectra of rutile TiO2(110) surfaces before and after the 8th TDS cycle of 50 L atomic D exposure at 320 K.

Similar surface structures between the starting rutile TiO2(110) sample and the sample after the repeated cycles of TDS measurements following 50 L atomic D exposure at 320 K also suggests that the formation of surface hydroxyl groups on rutile TiO2(110) surface among repeated cycles of TDS measurements at 320 K should be similar although the produced D2 keeps decreasing and eventually no D2 is produced. This indicates the existence of other pathways of surface hydroxyl groups that compete with the hydrogen formation pathway on rutile TiO2(110) surface with enough subsurface/bulk oxygen vacancies and do not form any gasphase products. Migration of H atoms in surface hydroxyl groups into subsurface/bulk regions was previously identified on a highly hydrogenated rutile TiO2(110) surface [11], and its activation energy was reported to be 1.1 eV, comparable to the activation energy of surface hydroxyl recombination reaction into H2 [21, 28]. Therefore, the migration of H atoms in surface hydroxyl groups into subsurface/bulk regions should account for the decrease and eventual disappearance of H2 production from surface hydroxyl groups on highly reduced rutile TiO2(110) samples. Moreover, the H diffusion pathway becomes more favorable as the subsurface/ bulk oxygen vacancies accumulate in the rutile TiO2(110) sample.

The above results demonstrate that the subsurface/bulk H species on the highly reduced rutile TiO2(110) sample is more stable than the surface hydroxyl groups. Since the employed gasphase atomic H is very energetic, it is likely that the subsurface/ bulk H species forms upon gas-phase atomic H exposure on the highly reduced rutile TiO2(110) sample at 320 K.

The photo-reactivity of various hydrogen species formed by gas-phase atomic D exposures at 115 K and 320 K was further examined. Fig. 4 shows the TDS spectra of rutile TiO2(110) surfaces exposed to 50 L D at 115 K and 320 K without and with a 150 s UV light illumination. As reported previously [13], the surface Ti—H hydride species is photoreactive while the surface hydroxyl group is not. The UV light illumination at 115 K induced the desorption of Ti—H hydride species, resulting in the weakening of D2/HD desorption peaks at 205 K and 165 K arising from surface reactions of surface Ti—H hydride species, while the D2/HD desorption feature at 292 K arising from surface reaction of surface hydroxyl groups does not change upon the UV light illumination. The UV light illumination at 320 K does not change the D2 TDS feature, in consistence of the absence of surface Ti—H hydride species and the presence of surface hydroxyl groups. It is noteworthy that the TDS experiment without UV light irradiation shown in Fig. 4B corresponds to the 6th TDS cycle of 50 L D exposure at 320 K while the subsequent TDS experiment with UV light irradiation corresponds to the 7th TDS cycle. As shown in Fig. 2B, the D2 desorption peak in the 7th TDS cycle without UV light illumination should be weaker than that in the 6th TDS cycle. However, the D2 desorption peak in the 7th TDS cycle with UV light illumination is almost identical to that the 6th TDS cycle without UV light illumination (Fig. 4B). This suggests that the UV light illumination should increase the ratio of surface hydroxyl groups that react to produce hydrogen upon the subsequent heating rather than that diffuse into the subsurface/bulk region. We consider two likely reasons: One is that the UV light illumination can drive the migration of subsurface/bulk H species to the surface to form surface hydroxyl groups with a large barrier of 2 eV [12]; the other is that the charge carriers within rutile TiO2(110) sample generated upon UV light illumination might stabilize surface hydroxyl groups. Our previous results show that the charge carriers within rutile TiO2(110) sample generated upon UV light illumination can survive after the shut-off of the UV light [13].

Download:
Fig. 4. D2 and HD TDS spectra following 50 L atomic D exposures on rutile TiO2(110) surface at 115 K (A) and 320 K (B) without and with a 150 s UV light irradiation.

The Ti—H hydride species and surface hydroxyl groups on rutile TiO2(110) surface were previously reported to exhibit electron stimulated desorption reactivity [25-27], which were also examined employing the secondary electrons generated during XPS measurements. Fig. 5 compares the TDS spectra following 20 L atomic D exposure at 115 K and 50 L atomic D exposure at 320 K on rutile TiO2(110) surface without and with XPS measurements. TDS experiment without XPS measurement shown in Fig. 5B corresponds to the 2nd TDS cycle of 50 L D exposure at 320 K. After the XPS measurement at 115 K, the HD/D2 desorption peaks arising from surface reactions of surface Ti—H hydride species significantly weaken, and the D2/HD desorption feature arising from surface reaction of surface hydroxyl groups also obviously weakens. Meanwhile, the desorption temperatures of all D2/HD desorption features shift toward high temperatures, in consistence with their second-order desorption kinetics. However, the desorption features of molecularly-adsorbed HDO/D2O does not change much. After the XPS measurement at 320 K, the D2 desorption peak arising from surface reaction of surface hydroxyl groups completely disappears. These observations indicate the occurrence of electron stimulated desorption of Ti—H hydride species and surface hydroxyl groups during the XPS measurements, but not of molecularly-adsorbed water. It can be seen that the XPS measurement induces completely electron stimulated desorption of surface hydroxyl groups at 320 K but only partially at 115 K. There are two possible reasons: one is a negative effect of coadsorbed water on the electron stimulated desorption reactivity of surface hydroxyl groups; the other is a positive effect of subsurface/bulk reduction of rutile TiO2(110) sample on the electron stimulated desorption reactivity of surface hydroxyl groups.

Download:
Fig. 5. TDS spectra following 20 L atomic D exposure at 115 K (A) and 50 L atomic D exposure at 320 K (B) on rutile TiO2(110) surface without and with XPS measurements.

In summary, we have successfully investigated thermal-, photo-, and electron-induced reactivity of various types of hydrogen species on a rutile TiO2(110) surface formed by atomic H exposure. Both the formed hydrogen species and their reactivity are significantly influenced by the reduction extent of rutile TiO2(110) sample. Atomic H exposure at 115 K leads to a slight reduction of rutile TiO2(110) surface with the formation of surface Ti—H hydride, surface hydroxyl group and chemisorbed water, while atomic H exposure at 320 K leads to a serious reduction of rutile TiO2(110) surface with the formation of surface hydroxyl groups and likely subsurface/bulk H species. Upon heating, surface hydroxyl groups react to produce both hydrogen and water on the slightly-reduced surface while they react to selectively produce hydrogen or migrate into the subsurface/bulk region on the intensely-reduced surface, and the H2 desorption activation energy decreases with the reduction extent. Surface Ti—H hydride exhibits photo-induced reactivity while both surface Ti—H hydride and surface hydroxyl group exhibit electro-induced reactivity. These results greatly broaden the fundamental understanding of the hydrogen-involved reactivity on oxide surfaces.

Acknowledgments

This work was financially supported by the National Natural Science Foundation of China (Nos. 21525313, 21761132005), Chinese Academy of Sciences (No. KJZD-EW-M03), MOE Fundamental Research Funds for the Central Universities (No. WK2060030017) and Collaborative Innovation Center of Suzhou Nano Science and Technology.

References
[1]
A. Fujishima, K. Honda, Nature 238 (1972) 37-38. DOI:10.1038/238037a0
[2]
M.A. Henderson, Surf. Sci. Rep. 66 (2011) 185-297. DOI:10.1016/j.surfrep.2011.01.001
[3]
X. Chen, L. Liu, P.Y. Yu, S.S. Mao, Science 331 (2011) 746-750. DOI:10.1126/science.1200448
[4]
J.T. Yates, Surf. Sci. 603 (2009) 1605-1612. DOI:10.1016/j.susc.2008.11.052
[5]
D.V. Bavykin, J.M. Friedrich, F.C. Walsh, Adv. Mater. 18 (2006) 2807-2824. DOI:10.1002/(ISSN)1521-4095
[6]
M.R. Hoffmann, S.T. Martin, W. Choi, D.W. Bahnemann, Chem. Rev. 95 (1995) 69-96. DOI:10.1021/cr00033a004
[7]
C.L. Pang, R. Lindsay, G. Thornton, Chem. Soc. Rev. 37 (2008) 2328-2353. DOI:10.1039/b719085a
[8]
L. B. Mo, Y. Wang, Y. Bai, et al., Sci. Rep. 5 (2015) 17634. DOI:10.1038/srep17634
[9]
S. Chen, D. Li, Y. Liu, W. Huang, J. Catal. 341 (2016) 126-135. DOI:10.1016/j.jcat.2016.06.016
[10]
Q. Yuan, Z. Wu, Y. Jin, et al., J. Am. Chem. Soc. 135 (2013) 5212-5219. DOI:10.1021/ja400978r
[11]
X.L. Yin, M. Calatayud, H. Qiu, et al., ChemPhysChem 9 (2008) 253-256. DOI:10.1002/(ISSN)1439-7641
[12]
G.H. Enevoldsen, H.P. Pinto, A.S. Foster, et al., Phys. Rev. Lett. 102 (2009) 136103. DOI:10.1103/PhysRevLett.102.136103
[13]
Z. Wu, W. Zhang, F. Xiong, et al., Phys. Chem. Chem. Phys. 16 (2014) 7051-7057. DOI:10.1039/c4cp00697f
[14]
B. Chen, Y. Ma, L. Ding, et al., J. Phys. Chem. C 117 (2013) 5800-5810. DOI:10.1021/jp312406f
[15]
L. Xu, W. Zhang, Y. Zhang, et al., J. Phys. Chem. C 115 (2011) 6815-6824. DOI:10.1021/jp200423j
[16]
L. Xu, Y. Ma, Y. Zhang, Z. Jiang, W. Huang, J. Am. Chem. Soc. 131 (2009) 16366-16367. DOI:10.1021/ja908081s
[17]
W. Huang, W. Ranke, R. Schlögl, J. Phys. Chem. C 111 (2007) 2198-2204. DOI:10.1021/jp066584u
[18]
W. Huang, W. Ranke, Surf. Sci. 600 (2006) 793-802. DOI:10.1016/j.susc.2005.11.026
[19]
W. Huang, W. Ranke, R. Schlögl, J. Phys. Chem. B 109 (2005) 9202-9204. DOI:10.1021/jp0511505
[20]
L. Xu, Z. Wu, Y. Zhang, et al., J. Phys. Chem. C 115 (2011) 14290-14299. DOI:10.1021/jp204106u
[21]
Y. Du, N.G. Petrik, N.A. Deskins, et al., Phys. Chem. Chem. Phys. 14 (2012) 3066-3074. DOI:10.1039/C1CP22515D
[22]
M. Kunat, U. Burghaus, C. Woll, Phys. Chem. Chem. Phys. 6 (2004) 4203-4207. DOI:10.1039/B404629C
[23]
J. Leconte, A. Markovits, M.K. Skalli, C. Minot, A. Belmajdoub, Surf. Sci. 497 (2002) 194-204. DOI:10.1016/S0039-6028(01)01477-7
[24]
J.M. Pan, B.L. Maschhoff, U. Diebold, T.E. Madey, J. Vac. Sci. Technol. A 10 (1992) 2470-2476. DOI:10.1116/1.577986
[25]
S. Suzuki, K. Fukui, H. Onishi, Y. Iwasawa, Phys. Rev. Lett. 84 (2000) 2156-2159. DOI:10.1103/PhysRevLett.84.2156
[26]
M.L. Knotek, Surf. Sci. 91 (1980) L17-L22.
[27]
Z. Zhang, J.T. Yates, Surf. Sci. 652 (2016) 195-199. DOI:10.1016/j.susc.2016.04.015
[28]
R. Wang, H. Fan, Catal. Sci. Technol. 7 (2017) 251-264. DOI:10.1039/C6CY02007K
[29]
Z.F. Wu, Y.K. Jin, F. Xiong, G.H. Sun, W.X. Huang, Sci. China Chem. 59 (2016) 9.
[30]
J. Tao, Q. Cuan, X.Q. Gong, M. Batzill, J. Phys. Chem. C 116 (2012) 20438-20446. DOI:10.1021/jp3064678
[31]
P.A. Redhead, Vacuum 12 (1962) 203-211. DOI:10.1016/0042-207X(62)90978-8