Formation of gold oxide on Au(111) induced by transition metal oxides

Formation of gold oxide on Au(111)

induced by transition metal oxides

requirements for the degree of MASTER OFSCIENCE

in PHYSICS

Author : Sabine Wenzel

Student ID : s1578499

Supervisor : Dr. Irene Groot

2ndcorrector : Dr. ir. Sense Jan van der Molen Leiden, The Netherlands, October 17, 2016

Formation of gold oxide on Au(111)

induced by transition metal oxides

Sabine Wenzel

Huygens-Kamerlingh Onnes Laboratory, Leiden University P.O. Box 9504, 2300 RA Leiden, The Netherlands

October 17, 2016

Abstract

Catalysts using gold single crystals or nanoparticles, especially for oxidation reactions, are currently receiving increasing attention in fundamental

research as well as for commercial applications. Still there is no clear understanding of the way these catalysts function. A major point under

debate is the existence and relevance of oxidized gold species on gold/transition metal oxide catalysts and the suitability of the detection

methods used so far is in doubt. We present the first in situ scanning tunnelling microscopy measurements that show the formation of a gold surface oxide on a Au(111) single crystal from molecular oxygen induced by

the presence of small amounts of rhenium or tungsten oxide on the surface. The observation of different non-stable as well as stable oxide phases present

at room temperature and formed in different oxygen pressure regimes is evidence for the potential relevance of gold oxide as the active phase of

Introduction

Although traditionally known as an inert element gold has received significant at-tention in catalysis research for the last 30 years, amongst others in the context of (selective) oxidation, hydrogenation and dehydrogenation reactions, the water-gas shift reaction, the treatment of substances harmful to the environment and chemical processing [1–4]. Currently there is siderable effort in making catalysts con-taining gold suitable for a number of ap-plications [1, 4–7] and there is a first suc-cess in its commercial use for the produc-tion of specialty chemicals [1].

Since the early days of gold catal-ysis, using combinations of gold with transition metal oxides has been espe-cially promising [8, 9]. Apart from gold nanoparticles on oxide supports [6, 10, 11] many inversed model catalysts based on Au(111) have been investigated in recent years [12–19]. It has been demonstrated that these systems can exceed the effi-ciency of catalysts currently used for ox-idation reactions [12], showing their po-tential relevance for a wide range of useful industrial applications for oxidation catal-ysis [20].

Despite the large amount of research there is currently no full understanding of the mechanisms by which these cat-alysts function. Particularly the role of the interface between gold and oxide re-mains unclear and there is an ongoing de-bate about the active species on such cat-alysts [21]. Some research has identified cationic gold species and has related them to the catalytic activity [22–26]. This in-cludes recent new evidence that the activ-ity of gold nanoparticles on a titanium ox-ide support is higher for a Auδ+ _species

than for Au0[27]. Other groups could not confirm the relevance or existence of oxi-dized species on active catalysts [28, 29].

Furthermore, the detection of gold ox-ide with x-ray based methods is under doubt as it has been shown that the forma-tion of oxidized gold can be induced by x-rays [30].

It is known that molecular oxygen does not chemisorb on clean gold at room temperature [31] but it does have a signif-icantly higher dissociation probability on oxidized gold [32]. New insights into the formation of a gold oxide at room tem-perature in molecular oxygen will thus shed light on the mechanism by which the dissociation of molecular oxygen is facili-tated on gold/transition metal oxide cata-lysts during oxidation reactions.

Here, we report the first observation of gold oxide formation from molecular oxygen on the Au(111) surface induced by small amounts of two different transi-tion metal oxides. The use of an inverted catalyst, prepared by depositing the tran-sition metal oxide onto a Au(111) sin-gle crystal, provides a larger, visible gold surface for oxidation compared to gold nanoparticles, thus enabling the observa-tion of the gold oxide by scanning tun-nelling microscopy (STM) instead of x-ray based methods. Using tungsten and rhe-nium oxide respectively, transition met-als which are more reducible compared to the more frequently used cerium and tita-nium oxides when deposited on Au(111), the gold oxidation is enhanced and large coverages can be achieved facilitating the investigation of the gold oxide.

Results and Discussion

An amorphous as well as a crystalline gold oxide phase were observed at room temperature on Au(111) after deposition of a transition metal oxide. The amor-phous gold oxide forms in a low oxygen pressure of 10−5 mbar, whereas the

crys-2

a) b)

c) d)

Figure 1: The amorphous as well as the crystalline phase observed here in comparison with the gold oxides imaged by Min et al. a) 50 nm x 50 nm STM image of the Au(111) surface after tungsten oxide was deposited at room temperature. During the deposition the surface was exposed to 10−5_{mbar of oxygen for 60 minutes. The gold oxide appears lighter than the}

substrate, darker areas are identified with gold vacancy islands. The image was taken in UHV at a bias voltage of -1 V and a tunnelling current of 100 pA. b) Amorphous gold oxide observed by Min et al. after exposing Au(111) to ozone [33]. c) 10 nm x 10 nm image of the boundary area between the Au(111) substrate (left) and the crystallites (right) formed under high pres-sure, taken at a bias voltage of 0.3 V and a tunnelling current of 140 pA. In order to improve the visibility of the atoms a differential background subtraction was used for this image. d) Crystalline phase imaged by Min et al. after annealing the surface with amorphous gold oxide to 400 K [33].

a) b)

c) d)

Figure 2:80 nm x 80 nm STM images of the tungsten oxide/Au(111) surface during exposure to 0.5 bar of oxygen. The images were taken at room temperature using a bias voltage of 2 V and a tunnelling current of 100 pA after a) 50 minutes, b) 90 minutes and c) 2 hours in oxygen. d) All three orientations of the crystallites, taken one day after the end of the oxygen exposure in UHV at room temperature using a bias voltage of 1 V and a tunnelling current of 80 pA.

4

a)

b) c)

Figure 3:Model structure for the gold oxide phases in analogy to oxides investigated on plat-inum. Gold atoms are grey, oxygen atoms are red. a) Unit cell in agreement with the distances measured on the crystalline gold oxide, b) metastable oxide rows and c) crystallites (corre-sponding to one monolayer of oxygen).

talline phase forms in 0.5 bar of molecular oxygen. As shown in Figure 1 the STM measurements are in agreement with im-ages of the same phases taken by Min et al. after exposing the Au(111) surface to ozone [33]. The unit cell of the crystalline phase (Figure 1 c)) is rectangular and the shorter unit cell vector forms an angle of

(120.4±0.9)◦with the visible rows of gold atoms. Averaging over multiple STM im-ages and all three different orientations of the crystallites (see Figure 2 d)) the size of the unit cell was determined as a = (4.9±0.2) A and b˚ = (3.3±0.4) A.˚ Figure 3 a) shows how the observed unit cell size and orientation fit onto the fcc (111) surface.

Furthermore, metastable oxide stripes can form on the same surface from molec-ular oxygen. After annealing the transi-tion metal oxide/Au(111) surface to 500 K

in 10−5 mbar of oxygen for 30 minutes a large coverage of the stripe-like phase in three different orientations was observed (see Figure 4). This phase is stable in ultra-high vacuum (UHV) for less than 24 hours and can be regained by repeating the same annealing step.

The stripe phase as well as the crys-talline structure observed here can be de-scribed with models previously proposed for Au(111) and Pt(111). Comparing the different orientations of the stripes to the orientations of the crystalline phase in Figure 2 d) suggests that the crystallites consist of rows of gold atoms similar to the observed stripes. In this case the shape of the crystallites shows that these rows can more easily attach parallel to each other than growing longer at higher oxy-gen pressure. The surface oxide structure proposed by Baker et al. [34] for a

cov-a) b)

Figure 4: 150 nm x 150 nm STM images after annealing the surface to 500 K in an oxygen background of 10−5mbar for 30 minutes. The images were taken in UHV after cooling down to room temperature. a) Model catalyst with tungsten oxide, image taken at a bias voltage of 1 V and a tunnelling current of 50 pA. b) Model catalyst with rhenium oxide, image taken at a bias voltage of 2 V and a tunnelling current of 50 pA.

erage of one monolayer of oxygen atoms has the same unit cell as observed here. According to this model the oxide consists of alternating lifted and non-lifted rows of gold atoms and one oxygen atom per gold atom. Such oxides covering the whole surface as well as oxide stripes with a larger distance to each other have been ob-served and investigated more thoroughly on the fcc(111) surface of platinum. A model used for describing these [35, 36] is reproduced in Figure 3 b) and c) in or-der to visualise a possible structure for both phases. The model for the crystalline structure is similar to the one proposed for gold [34] while the oxide stripes described by Miller et al. [36] consist of rows of gold atoms in an additional layer on top of the surface layer with two oxygen atoms per gold atom. This is an indication of the low stability of the oxide stripes as observed here in UHV at room temperature.

All observed oxide phases cover large areas of the Au(111) surface with only a small amount of tungsten oxide or

rhe-nium oxide deposited, respectively. The STM studies were combined with addi-tional x-ray photoelectron spectroscopy (XPS) measurements in order to deter-mine the amount of transition metal oxide necessary to form gold oxide. A surface with amorphous gold oxide like the one in Figure 1 a) has a transition metal cover-age of less than 4 per cent. With the same coverage the other gold oxide phases can be formed yielding an amount as shown in Figures 2 d) and 4.

We propose that tungsten oxide and rhenium oxide function as a catalyst for the gold oxidation by providing atomic oxygen. In an extensive DFT study [37] Shi et al. investigated the stability of a number of possible crystal structures for a Au(111) surface oxide and compared the most stable one to bulk oxide. Al-though the 4 x 4 unit cell structure used for these calculations is not applicable to our findings, the resulting stability curve can serve as an estimate for the stability of the surface oxide observed here. As

6

Figure 5:XPS 4f signals of rhenium metal (blue squares) and rhenium oxide (orange triangles) in comparison. The measurements were done in UHV at room temperature and rescaled for comparability.

such it is in agreement with the existence of a thermodynamically stable gold oxide at room temperature, even in UHV. How-ever, no oxidation takes place on the clean gold surface as Au(111) is not able to dis-sociate molecular oxygen since there is a high dissociation barrier of 2.23 eV [38]. From XPS measurements it is clear that the deposited compound is an oxide of the transition metals, a shift of 4 eV was mea-sured compared to the XPS signal after de-positing a metal (see Figure 5). Tungsten oxide is a known oxidation catalyst and has the ability to dissociate molecular oxy-gen [39, 40]. The transfer of atomic oxyoxy-gen from tungsten and rhenium to gold is pos-sible because these transition metal ox-ides are easily reducible when deposited on Au(111). Annealing the tungsten ox-ide/gold surface to 500 K in UHV is suffi-cient to reduce the tungsten oxide within 15 minutes. Subsequently taken STM im-ages (Figure 6) show that a gold-tungsten surface alloy is formed after the reduction. This is evidence for a high formation

en-ergy of the transition metal oxide when deposited on Au(111), which allows for the transfer of oxygen to gold atoms. Sub-sequently, the transition metal oxide can be re-oxidized with molecular oxygen as long as the surface is in an oxygen pres-sure of 10−5 mbar or more. Additionally, it is possible that oxidized gold promotes the dissociation of molecular oxygen for subsequent gold oxidation.

Under a different chemical potential the transition metal oxide catalysts pro-mote the reverse reaction reducing the gold oxide formed previously on the Au(111) surface. The amorphous phase is stable in UHV at room temperature but not visible anymore after annealing the surface to 400 K for 15 minutes. This is a reversible process as the amorphous gold oxide reforms within 30 minutes un-der subsequent exposure to 10−5 mbar of molecular oxygen. In contrast to these measurements, Min et al. [33] still detect a gold oxide on the Au(111) surface ox-idized using ozone after annealing it to

Figure 6:80 nm x 80 nm image of the tungsten oxide/Au(111) surface after annealing to 500 K for 15 minutes. Darker areas in between the herringbone are associated with a gold-tungsten surface alloy. The image was taken in UHV at room temperature using a bias voltage of -1 V and a tunnelling current of 100 pA.

400 K. In their case no other metal was present on the surface to promote the re-duction of the gold oxide.

It is likely that the crystallization of the gold oxide phase formed under high pressure is caused by a different oxidation state compared to the amorphous gold ox-ide phase formed in a lower oxygen pres-sure. The oxidation states of the differ-ent gold phases could not be measured with XPS due to considerable beam dam-age to the gold oxide, which was observed in STM after the x-ray exposure. When ex-posing the surface with amorphous gold oxide (see Figure 1 a)) to a high oxygen pressure no change is visible. When there was however no prior oxidation of the gold, the crystalline phase (Figure 2 d)) was observed to form in high pressure on the transition metal oxide/Au(111) sur-face. It is thus clear that the high pres-sure environment does not only induce the mobility of the gold oxide but has additional effects, which are responsible for the difference between the amorphous and the crystalline phase. As can be seen

in Figure 2 the crystalline phase is pre-ceded by an intermediate mobile species, which forms under exposure to 0.5 bar of molecular oxygen within 3 hours. Af-ter reducing the pressure to UHV a high mobility was still observed (not shown). Within another 15 hours in UHV the high-pressure structure crystallizes completely (see Figure 2 d)) and becomes stable. It follows that the mobility of the intermedi-ate stintermedi-ate does not stem from contaminants in the high-pressure environment such as water and that the crystalline phase does not have a higher oxidation state than the mobile intermediate. The most likely ex-planation for the mobility, which enables the crystallization, is thus that the mobile predecessor of the crystalline phase has a higher oxidation state than the amor-phous gold oxide. In this case the dis-crepancy with these observations and Min et al. [33] inducing the crystalline struc-ture by annealing the amorphous gold ox-ide might be explained by the presence of additional oxygen atoms on the sur-face after ozone exposure or by a residual

8 ozone background during the annealing

step combined with mobility induced by surface temperature.

Conclusion

We presented the first observation of the formation of gold oxide from molecular oxygen on a Au(111) surface promoted by a small amount of transition metal oxide. At room temperature an amorphous gold oxide forms in low oxygen pressure and a crystalline phase at a higher pressure. Annealing the catalyst to 500 K in an oxy-gen background yields a metastable phase of stripes. We propose that the transi-tion metal oxide functransi-tions as an oxidatransi-tion catalyst for the crucial step of dissociat-ing oxygen molecules and thus providdissociat-ing atomic oxygen to gold atoms, the reverse reaction can be facilitated as well. This new evidence of the presence of gold ox-ide under conditions relevant for catalysis is important for understanding the mech-anism by which catalysts containing tran-sition metal oxides in combination with Au(111) or gold nanoparticles function, particularly how molecular oxygen is dis-sociated, as well as for developing new, more efficient oxidation catalysts.

Methods

The experimental set-up is described in detail in reference [41]. In short, the UHV system consists of three separable chambers. In a preparation chamber the Au(111) crystal was cleaned by consecu-tive cycles of Ar+ sputtering and anneal-ing to 800 K. Tungsten and rhenium ox-ide respectively were deposited using an Oxford Applied Research e-beam evapo-rator, in which the metal rods were heated to a temperature just underneath the

on-set of glowing in a background of 10−5 mbar of oxygen. The sample can be trans-ferred to the scanning tunnelling micro-scope chamber without exposing it to air. The microscope can be used in UHV or as a reactor STM in a volume of about 0.5 ml of gas closed off by the sample, the STM body and a Kalrez seal. For the high pressure studies the reactor was held at a pressure of 0.5 bar using an oxygen flow of 1 ml/min. The STM tip is a cut platinum/iridium wire, the custom made control system is described in detail in ref-erence [42]. Lastly, the system is equipped with a Specs x-ray photoelectron spec-troscopy system. Peaks were fitted with the CasaXPS software and the energy axis was calibrated using the position of the 4f signal of gold measured on the clean sam-ple. The rhenium 4f XPS signal was cal-ibrated by depositing rhenium metal and comparing the corresponding XPS signal to the metal coverage. For the metal de-position the rhenium rod was heated up to a white glow in UHV. Coverages were determined from STM images using a wa-tershed method in Gwyddion.

Acknowledgements

This project was made possible through the large teaching effort by Rik Mom and the supervision by Irene Groot. I thank all members of the heterogeneous catalysis group for scientific discussions as well as technical support, this includes colleagues from the fine mechanics department and the electronics department of Leiden Uni-versity.

References

[1] G. Bond, C. Louis, and D. Thompson. Catalysis by Gold (Catalytic Science

Series). London: World Scientific Pub-lishing Company, 2006.

[2] Graham J. Hutchings. Gold catalysis in chemical processing. Catal. Today, 72(1-2):11–17, 2002.

[3] Graham J. Hutchings. Catalysis by gold. Catal. Today, 100(1-2):55–61, 2005.

[4] Christopher W. Corti, Richard J. Holliday, and David T. Thompson. Progress towards the commercial ap-plication of gold catalysts. Top. Catal., 44(1-2):331–343, 2007.

[5] Christopher W. Corti, Richard J. Hol-liday, and David T. Thompson. Com-mercial aspects of gold catalysis. Appl. Catal. A Gen., 291(1-2):253–261, 2005.

[6] David T. Thompson. Using gold nanoparticles for catalysis. Nano To-day, 2(4):40–43, 2007.

[7] Don Cameron, Richard Holliday, and David Thompson. Gold’s future role in fuel cell systems. J. Power Sources, 118(1-2):298–303, 2003.

[8] Masatake Haruta, Tetsuhiko Kobayashi, Hiroshi Sano, and Nobumasa Yamada. Novel gold catalysts for the oxidation of carbon monoxide at a temperature far below 0.DEG.C. Chem. Lett., (2):405–408, 1987.

[9] M. Haruta. Gold catalysts pre-pared by coprecipitation for low-temperature oxidation of hydrogen and of carbon monoxide. J. Catal., 115(2):301–309, 1989.

[10] R Meyer, C Lemire, Sh. K Shaikhutdi-nov, and H J. Freund. Surface chem-istry of catalysis by gold. Gold Bull., 37(1-2):72–124, 2004.

[11] Zhen Ma and Sheng Dai. Develop-ment of novel supported gold cata-lysts: A materials perspective. Nano Res., 4(1):3–32, 2011.

[12] J. A. Rodriguez, S. Ma, P. Liu, J. Hrbek, J. Evans, and M. P´erez.

Activity of CeO. Science,

318(December):1757–1760, 2007. [13] M. Li and E. I. Altman. Shape,

mor-phology, and phase transitions dur-ing Co oxide growth on Au(111). J. Phys. Chem. C, 118(24):12706–12716, 2014.

[14] J ¨urgen Biener, Enrique Farfan-Arribas, Monika Biener, Cynthia M Friend, and Robert J Madix. Syn-thesis of TiO2 nanoparticles on the Au(111) surface. J. Chem. Phys., 123(9):94705, 2005.

[15] Su Ying Quek, Monika M. Biener, Juergen Biener, Cynthia M. Friend, and Efthimios Kaxiras. Tuning elec-tronic properties of novel metal oxide nanocrystals using interface interac-tions: MoO3 monolayers on Au(111). Surf. Sci., 577(2-3):L71–L77, 2005. [16] Byoung Koun Min and Cynthia M.

Friend. Heterogeneous gold-based catalysis for green chemistry: Low-temperature CO oxidation and propene oxidation. Chem. Rev., 107(6):2709–2724, 2007.

[17] Hendrik Str¨ater, Hanna Fedder-witz, Boris Groß, and Niklas Nilius. Growth and Surface Properties of Cuprous Oxide Films on Au(111). J. Phys. Chem. C, 119(111):5975–5981, 2015.

10 [18] Xingyi Deng and Christopher

Matranga. Selective Growth of Fe2O3 Nanoparticles and Islands on

Au(111). Signals, (111):11104–11109, 2009.

[19] Chen Wu and Martin R. Castell. Ba and BaOx surface structures on Au(111). Surf. Sci., 606(3-4):181–185, 2012.

[20] Daniel Duprez. Handbook of Ad-vanced Methods and Processes in Oxi-dation Catalysis: From Laboratory to In-dustry. Number 2. 2014.

[21] Juan C. Fierro-Gonzalez and Bruce C. Gates. Evidence of active species in CO oxidation catalyzed by highly dispersed supported gold. Catal. To-day, 122(3-4):201–210, 2007.

[22] Juan C. Fierro-Gonzalez, Javier Guz-man, and Bruce C. Gates. Role of cationic gold in supported CO oxida-tion catalysts. Top. Catal., 44(1-2):103– 114, 2007.

[23] Juan C. Fierro-Gonzalez and Bruce C. Gates. Catalysis by gold dis-persed on supports: the importance of cationic gold. Chem. Soc. Rev., 37(9):2127–34, 2008.

[24] Qi Fu, Howard Saltsburg, and Maria Flytzani-Stephanopoulos. Active nonmetallic Au and Pt species on ceria-based water-gas shift catalysts. Science, 301(5635):935–8, 2003.

[25] S Minico, S Scire, C Crisafulli, A M Visco, and S Galvagno. FT-IR study of Au/Fe2O3 catalysts for CO

oxida-tion at low temperature. Catal. Let-ters, 47(1-2):273–276, 1997.

[26] Javier Guzman and Bruce C. Gates. Catalysis by Supported Gold: Cor-relation between Catalytic Activity

for CO Oxidation and Oxidation States of Gold. J. Am. Chem. Soc., 126(9):2672–2673, 2004.

[27] Alexander Yu. Klyushin, Mark T. Greiner, Xing Huang, Thomas Lunkenbein, Xuan Li, Olaf Timpe, Matthias Friedrich, Michael H¨avecker, Axel Knop-Gericke, and Robert Schl ¨ogl. Is Nanostructuring Sufficient To Get Catalytically Active Au? ACS Catal., pages 3372–3380, 2016.

[28] V Schwartz, D R Mullins, W Yan, B Chen, S Dai, and S H Overbury. XAS Study of Au Supported on TiO 2 : Influence of Oxidation State and Particle Size on the Catalytic Activ-ity. J. Phys. Chem. B, 108:15782–15790, 2004.

[29] N. Weiher, E. Bus, L. Delannoy, C. Louis, D. E. Ramaker, J. T. Miller, and J. A. van Bokhoven. Struc-ture and oxidation state of gold on different supports under various CO oxidation conditions. J. Catal., 240(2):100–107, 2006.

[30] Peng Jiang, Soeren Porsgaard, Fer-enc Borondics, Mariana Kober, Al-fonso Caballero, Hendrik Bluhm, Flemming Besenbacher, and Miquel Salmero. Room-temperature reac-tion of oxygen with gold: An in situ ambient-pressure x-ray photo-electron spectroscopy investigation. J. Am. Chem. Soc., 132(9):2858–2859, 2010.

[31] B. M. W. Trapnell. The activities of evaporated metal films in gas chemisorption. Proceedings of the Royal Society of London A: Mathemat-ical, Physical and Engineering Sciences, 218(1135):566–577, 1953.

[32] Xingyi Deng, Byoung Koun Min, Amado Guloy, and Cynthia M Friend. Enhancement of O2

Disso-ciation on Au ( 111 ) by Adsorbed Oxygen : Implications for Oxidation Catalysis. JACS Artic., (111):9267– 9270, 2005.

[33] B. K. Min, A. R. Alemozafar, M. M. Biener, J. Biener, and C. M. Friend. Reaction of Au(111) with Sulfur and Oxygen: Scanning Tunneling Micro-scopic Study. Top. Catal., 36(Au-gust):77, 2005.

[34] Thomas A Baker, Bingjun Xu, Xiaoy-ing Liu, Efthimios Kaxiras, and Cyn-thia M Friend. Nature of Oxidation of the Au ( 111 ) Surface : Experimen-tal and Theoretical Investigation. So-ciety, (111):16561–16564, 2009.

[35] Jeffery M. Hawkins, Jason F. Weaver, and Aravind Asthagiri. Density functional theory study of the initial oxidation of the Pt(111) surface. Phys. Rev. B, 79(12):125434, 2009.

[36] D. J. Miller, H. ¨Oberg, S. Kaya, H. Sanchez Casalongue, D. Friebel,

T. Anniyev, H. Ogasawara,

H. Bluhm, L. G M Pettersson, and A. Nilsson. Oxidation of Pt(111) under near-ambient conditions. Phys. Rev. Lett., 107(19):1–5, 2011. [37] H Shi and C Stampfl. First-principles

investigations of the structure and stability of oxygen adsorption and surface oxide formation at Au(111). Phys. Rev. B, 76(7):–, 2007.

[38] Zhi Pan Liu, P. Hu, and Ali Alavi. Catalytic role of gold in gold-based

catalysts: A density functional theory study on the CO oxida-tion on gold. J. Am. Chem. Soc., 124(49):14770–14779, 2002.

[39] Haidong Zheng, Jian Zhen Ou, Michael S. Strano, Richard B. Kaner, Arnan Mitchell, and Kourosh Kalantar-Zadeh. Nanostructured tungsten oxide - Properties, synthe-sis, and applications. Adv. Funct. Mater., 21(12):2175–2196, 2011.

[40] J. Haber, J. Janas, M. Schiavello, and R. J D Tilley. Tungsten oxides as cat-alysts in selective oxidation. J. Catal., 82(2):395–403, 1983.

[41] C. T. Herbschleb, P. C. Van Der Tu-ijn, S. B. Roobol, V. Navarro, J. W. Bakker, Q. Liu, D. Stoltz, M. E. Ca ˜nas Ventura, G. Verdoes, M. A. Van Spronsen, M. Bergman, L. Crama, I. Taminiau, A. Ofitserov, G. J C Van Baarle, and J. W M Frenken. The Re-actorSTM: Atomically resolved scan-ning tunneling microscopy under high-pressure, high-temperature cat-alytic reaction conditions. Rev. Sci. In-strum., 85(8):0–10, 2014.

[42] M. J. Rost, L. Crama, P. Schakel, E. Van Tol, G. B E M Van Velzen-Williams, C. F. Overgauw, H. Ter Horst, H. Dekker, B. Okhuijsen, M. Seynen, A. Vijftigschild, P. Han, A. J. Katan, K. Schoots, R. Schumm, W. Van Loo, T. H. Oosterkamp, and J. W M Frenken. Scanning probe mi-croscopes go video rate and beyond. Rev. Sci. Instrum., 76(5), 2005.