Bowker M., Davies P.R. (Eds.) Scanning Tunneling Microscopy in Surface Science, Nanoscience and Catalysis
Подождите немного. Документ загружается.
such as Fe or Ta, becomes feasible. Even deposition of semiconductors such as
Ge from ILs has been successfully demonstrated [99]. Tip-induced nanostructuring
has also been performed in ILs [94, 95]. We end this chapter by showing a ring of
Fe clusters generated on Au(1 1 1) in 1-butyl-3-methyl-imidazolium tetrafluorobo-
rate, a typical and widely used IL, by jump-to-contact (Figure 5.23). Here again, the
unusually high stability of the Fe clusters, which even exceeds that of upd Fe is
noteworthy [95].
Acknowledgments
Parts of this work were supported by the Deutsche Forschungsgemeinschaft through
SFB 569 and the Fonds der Chemischen Industrie. One of us (FS) gratefully
acknowledges a stipend of SFB 569.
References
1 Trasatti, S. (ed.) (2003) Electrocatalysis:
from theory to industrial applications.
Electrochim. Acta, 48, 3727–3974.
2 Ertl, G. (2008) Angew. Chem. Int. Ed., 47,
3524.
3 Markovic, N.M. and Ross, P.N. (2002) Surf.
Sci. Rep., 45, 117.
4 Wintterlin, J., V
€
olkening, S., Janssens,
T.V.W., Zambelli, T., and Ertl, G. (1997)
Science, 278, 1931.
5 Eigler, D.M. and Schweizer, E.K. (1990)
Nature, 344, 524.
6 Pust, S.E., Maier, W., and Wittstock, G.
(2008) Z. Phys. Chem., 222, 1463.
7 Bonnell, D.A. (1993) Scanning Tunneling
Microscopy and Spectroscopy, VCH,
Weinheim.
8 Wiesendanger, R. (1994) Scanning Probe
Microscopy and Spectroscopy, Cambridge
University Press, Cambridge.
9 G
€
untherodt, H-.J. and Wiesendanger, R.
(eds) (1994) Scanning Tunneling Microscopy
I–III, Springer, Berlin.
10 Wintterlin, J., Wiechers, J., Brune, H.,
Gritsch, T., H
€
ofer, H., and Behm, R.J.
(1989) Phys. Rev. Lett., 62, 59.
11 Simmons, J.G. (1963) J. Appl. Phys., 34,
1793.
12 Schmid, M., Stadler, H., and Varga, P.
(1993) Phys. Rev. Lett., 70, 1441.
13 Wouda, P.T., Nieuwenhuys, B.E.,
Schmid, M., and Varga, P. (1996) Surf. Sci.,
359, 17.
14 Ruff, M., Takehiro, N., Liu, P.,
Nørskov, J.K., and Behm, R.J. (2007)
ChemPhysChem., 8, 2068.
15 Engelmann, G.E., Ziegler, J.C., and
Kolb, D.M. (1998) Surf. Sci., 401, L420.
16 Kuk, Y. and Silverman, P.J. (1990)
J. Vac. Sci. Technol., A8, 289.
17 (a) Hugelmann, M. and Schindler, W.
(2003) Surf. Sci., 541, L643; (b) (2004)
J. Electrochem. Soc., 151, E97.
18 Simeone, F.C., Kolb, D.M.,
Venkatachalam, S., and Jacob, T. (2007)
Angew. Chem. Int. Ed., 46, 8903.
19 Simeone, F.C., Kolb, D.M.,
Venkatachalam, S., and Jacob, T. (2008)
Surf. Sci., 602, 1401.
20 Kolb, D.M., Nichols, R.J., and Behm, R.J.
(1992) Electrified Interfaces in Physics,
Chemistry and Biology, NATO ASI, Vol. C
355 (ed. R. Guidelli), Kluwer, Dordrecht,
p. 275.
21 Wintterlin, J. (1988) PhD Thesis,
Free University Berlin.
144
j
5 Characterization and Modification of Electrode Surfaces by In Situ STM
22 Wiechers, J., Twomey, T., Kolb, D.M., and
Behm, R.J. (1988) J. Electroanal. Chem.,
248, 451.
23 Nagahara, L.A., Thundat, T., and Lindsay,
S.M. (1989) Rev. Sci. Instrum., 60, 3128.
24 Bach, C.E., Nichols, R.J., Beckmann, W.,
Meyer, H., Schulte, A., Besenhard, J.O.,
and Jannakoudakis, P.D. (1993)
J. Electrochem. Soc., 140, 1281.
25 See, for example, Dakkouri, A.S. (1996)
PhD Thesis, University Ulm.
26 Kolb, D.M. (2002) Advances in Electro-
chemical Science and Engineering., Vol. 7
(eds R.C. Alkire and D.M. Kolb), Wiley-
VCH Verlag, Weinheim, p. 107.
27 Pohl, D.W. (1986) IBM J. Res. Dev.,30, 417.
28 Bockris, J.OM., Devanathan, M.A.V., and
M
€
uller, K. (1963) Proc. Roy. Soc. Lond.,
A274, 55.
29 Gerischer, H. (1997) The CRC Handbook
of Solid State Electrochemistry, Chapter 2
(eds P.J. Gellings and H.J.M. Bouw-
meester), CRC Press, Boca Raton.
30 Xie, Z.-X. and Kolb, D.M. (2000)
J. Electroanal. Chem., 481, 177.
31 Stimming, U., Vogel, R., Kolb, D.M., and
Will, T. (1993) J. Power Sources, 43/44, 169.
32 Adzic, R.R., Tripkovic, A.V., and
OGrady, W.E. (1982) Nature, 296, 137.
33 Schultze, J.W. and Dickertmann, D. (1976)
Surf. Sci., 54, 489.
34 J
€
uttner, K. and Lorenz, W.J. (1980) Z. Phys.
Chem. N.F., 122, 163.
35 Hubbard, A.T. (1973) Crit. Rev. Anal.
Chem., 3, 201.
36 OGrady, W.E., Woo, M.Y.C., Hagans, P.L.,
and Yeager, E. (1977) J. Vac. Sci. Technol.,
14, 365.
37 Kolb, D.M. (1987)
Z. Phys. Chem. N.F., 154,
179.
38 Clavilier, J., Faure, R., Guinet, G., and
Durand, R. (1980) J. Electroanal. Chem.,
107, 205, 211.
39 Hamelin, A., Doubova, L., Wagner, D., and
Schirmer, H. (1987) J. Electroanal. Chem.,
220, 155.
40 Wasberg, M., Palaikis, L., Wallen, S.,
Kamrath, M., and Wieckowski, A. (1988) J.
Electroanal. Chem., 256, 51.
41 Kibler, L.A., Cuesta, A., Kleinert, M., and
Kolb, D.M. (2000) J. Electroanal. Chem.,
484, 73.
42 Kibler, L.A. Preparation and Character-
ization of Noble Metal Single-Crystal
Electrodes, http://www.uni-ulm.de/
echem/index.html?id¼1023001.
43 Hoyer, R. (2004) PhD Thesis,
University Ulm.
44 Magnussen, O.M. (2002) Chem. Rev., 102,
679.
45 Kibler, L.A., Kleinert, M., and Kolb, D.M.
(2000) Surf. Sci., 461, 155.
46 Waibel, H.-F., Kleinert, M., Kibler, L.A.,
and Kolb, D.M. (2002) Electrochim. Acta,
47, 1461.
47 Cuesta, A., Kleinert, M., and Kolb, D.M.
(2000) Phys. Chem. Chem. Phys., 2, 5684.
48 Schweizer, M. and Kolb, D.M. (2003) Surf.
Sci., 544, 93.
49 Kibler, L.A., Kleinert, M., Randler, R., and
Kolb, D.M. (1999) Surf. Sci., 443, 19.
50 Maroun, F., Ozanam, F., Magnussen, O.,
and Behm, R.J. (2001) Science, 293, 1811.
51 Varga, P., Schmid, M., and Redinger, J.
(2000) Physik in Unserer Zeit, 31, 215.
52 Diemant, T., Hager, T., Hoster, H.E.,
Rauscher, H., and Behm, R.J. (2003)
Surf.
Sci., 541, 137.
53 Kibler, L.A. and Kolb, D.M. (2003) Z. Phys.
Chem., 217, 1265.
54 El-Aziz, A.M. and Kibler, L.A., in pre-
paration.
55 Nichols, R.J., Bunge, E., Meyer, H., and
Baumg
€
artel, H. (1995) Surf. Sci., 335, 110.
56 Hoyer, R., Kibler, L.A., and Kolb, D.M.
(2003) Electrochim. Acta, 49, 63.
57 Somorjai, G.A. (1977) Adv. Catal., 26,1.
58 Lehmpfuhl, G., Uchida, Y., Zei, M.S., and
Kolb, D.M. (1999) Imaging of Surfaces and
Interfaces (eds J. Lipkowski and P.N.
Ross), Wiley-VCH Verlag, Weinheim.
59 Cerviño, A.M., Triaca, W.A., and Arvia, A.J.
(1985) J. Electrochem. Soc., 132, 266.
60 Canullo, J.C., Triaca, W.E., and Arvia, A.J.
(1984) J. Electroanal. Chem., 175, 337.
61 Canullo, J., Uchida, Y., Lehmpfuhl, G.,
Twomey, T., and Kolb, D.M. (1987) Surf.
Sci., 188, 350.
References
j
145
62 Twomey, T.A. (1989) J. Electroanal. Chem.,
270, 465.
63 Schneeweiss, M.A. and Kolb, D.M. (1997)
Solid State Ionics, 94, 171.
64 Stickney, J.L., Villegas, I., and Ehlers, C.B.
(1989) J. Am. Chem. Soc., 111, 6473.
65 Will, T., Dietterle, M., and Kolb, D.M.
(1995) Nanoscale Probes of the Solid/Liquid
Interface, NATO ASI, Vol. E288
(eds A.A. Gewirth and H. Siegenthaler),
Kluwer, Dordrecht, p. 137.
66 Crommie, M.F., Lutz, C.P., and
Eigler, D.M. (1992) Science, 262, 218.
67 Meyer, G., Bartels, L., Z
€
ophel, S., and
Rieder, K.H. (1999) Appl. Phys. A, 68, 125.
68 Cuberes, M.T., Schlitter, R.R., and
Gimzewski, J.K. (1997) Surf. Sci., 371,
L231.
69 Li, W., Virtanen, J.A., and Penner, R.M.
(1992) Appl. Phys. Lett., 60, 1181.
70 Nyffenegger, R.M. and Penner, R.M.
(1997) Chem. Rev., 97, 1195.
71 Homma, T., Kubo, N., and Osaka, T. (2003)
Electrochim. Acta, 48, 3115.
72 Kubo, N., Homma, T., Hondo, Y., and
Osaka, T. (2005) Electrochim. Acta, 51, 834.
73 LaGraff, J.R. and Gewirth, A.A. (1994)
J. Phys. Chem., 98, 11246.
74 Petri, M. and Kolb, D.M. (2002) Phys.
Chem. Chem. Phys., 4, 1211.
75 Schindler, W., Hofmann, D., and
Kirschner, J. (2001) J. Electrochem. Soc.,
148, C124.
76 Schindler, W., Hofmann, D., and
Kirschner, J. (2000)
J. Appl. Phys., 87, 7007.
77 Schindler, W., Hugelmann, P.,
Hugelmann, M., and K
€
artner, F.X. (2002)
J. Electroanal. Chem., 522, 49.
78 Garcia, S.G., Salinas, D.R., Mayer, C.E.,
Lorenz, W.J., and Staikov, G. (2003)
Electrochim. Acta, 48, 1279.
79 Lang, G., Bakos, I., and Horanyi, G. (2000)
J. Electroanal. Chem., 493, 141.
80 Kirchner, V., Cagnon, L., Schuster, R., and
Ertl, G. (2001) Appl. Phys. Lett., 79, 1721.
81 Kirchner, V., Xia, X.H., and Schuster, R.
(2001) Acc. Chem. Res., 34, 371.
82 Schuster, R. and Ertl, G. (2003) Catalysis
and Electrocatalysis at Nanoparticle Surfaces
(eds E. Savinova C. and Vayenas), Marcel
Dekker, New York.
83 Avouris, P., Hertel, T., and Martel, R.
(1997) Appl. Phys. Lett., 71, 285.
84 Meier, J., Friedrich, K.A., and
Stimming, U. (2002) Faraday Discuss., 121,
365.
85 Bard, A.J. and Mirkin, M.V. (eds) (2001)
Scanning Electrochemical Microscopy,
Marcel Dekker, New York.
86 Zhou, J., Zu, Y., and Bard, A.J. (2000)
J. Electroanal. Chem., 491, 22.
87 Baltes, N., Thouin, L., Amatore, C., and
Heinze, J. (2004) Angew. Chem. Int. Ed., 43,
1431.
88 Ufheil, J., Hess, C., Borgwarth, K., and
Heinze, J. (2005) Phys. Chem. Chem. Phys.,
7, 3185.
89 Landman, U., Luedtke, W.D., Burnham,
N.A., and Colton, R.J. (1990) Science,
248, 454.
90 Kolb, D.M., Ullmann, R., and Will, T.
(1997) Science, 275, 1097.
91
Kolb, D.M. (2001) Angew. Chem. Int. Ed.,
40, 1162.
92 Kolb, D.M. and Simeone, F.C. (2005)
Electrochim. Acta, 50, 2989.
93 Kolb, D.M., Engelmann, G.E., and
Ziegler, J.C. (2000) Solid State Ionics, 131,
69.
94 Wang, J.G., Tang, J., Fu, Y.C., Wei, Y.M.,
Chen, Z.B., and Mao, B.W. (2007)
Electrochem. Commun., 9, 633.
95 Wei, Y.M., Zhou, X.S., Wang, J.G., Tang, J.,
Mao, B.W., and Kolb, D.M. (2008) Small, 4,
1355.
96 Kolb, D.M., Engelmann, G.E., and Ziegler,
J.C. (2000) Angew. Chem. Int. Ed., 39, 1123.
97 Engelmann, G.E. (1997) PhD Thesis,
University Ulm.
98 Endres, F., Abbott, A.P., and MacFarlane,
D.R. (eds) (2008) Electrodeposition from
Ionic Liquids, Wiley-VCH Verlag,
Weinheim.
99 Endres, F. (2004) Z. Phys. Chem., 218, 255.
146
j
5 Characterization and Modification of Electrode Surfaces by In Situ STM
6
STM Imaging of Oxide Nanolayer Model Systems
Falko P. Netzer and Svetlozar Surnev
6.1
Introduction
Scanning tunneling microscopy (STM) has become an invaluable tool for the study
of atomic processes at surfaces. Amongst other advantages, such as resolution at
the atomic level, high sensitivity, and providing information on a local scale, seeing
the atomic structures directly in real space at surfaces is a significant benefit for
fueling the inspiration of scientific intuition. The application of STM to problems in
heterogeneous catalysis has thus been widely recognized since it allows one to gain
insight into the nanoscience aspects behind catalysis, with the ultimate goal of
imaging surface reactions with atomic or molecular resolution. The application of
STM to catalysis requires suitable model systems, preferably of planar geometry,
which are often based on oxide model surfaces that are covered in this chapter.
Metal oxides are ubiquitous in heterogeneous catalysis as support surfaces in
oxide-supported metal catalysts or as active surface phases in oxide catalysis.
Epitaxially grown thin films of oxides on metal substrates constitute the most
promising route to generate clean, well-ordered oxide surfaces, which exhibit a
number of technical advantages over their corresponding bulk-terminated oxide
single-crystal surfaces. Oxide single crystals of suitable size are often difficult to
obtain, sample handling including heating and cooling is difficult, and the charging
problem of insulating oxide samples upon exposure to charged particle probes
(such as electrons) is experimentally difficult to overcome. These problems can be
avoided by using thin oxide films grown on metallic substrates for the investigation of
surface properties [1, 2], and thin films have therefore become the preferred oxide
phases employed in model studies of fundamental catalytic research. Originally, it
was thought that thin films of oxides mimic closely the respective properties of the
corresponding bulk-terminated surfaces [1]; however, more recently it has become
clear that this is true only if the films are beyond a certain critical thickness [3]. Oxide
films on metal supports thinner than this critical thickness – which is typically of
the order of 2–3nm – display novel properties due to confinement and interfacial
j
147
proximity effects, which make them very different from their respective bulk
counterparts [3, 4]. These ultrathin films – for convenience we will call them oxide
nanolayers – may be regarded as new kinds of material, which do not occur in nature
but are characterized by novel emergent properties of structure, electronic and
magnetic behavior, and chemical reactivity. The nanometer scale in the thickness
dimension means that the oxide nanolayers are only a few unit cell thick and that
the interfacial bonding to the metal substrate and charge transfer effects across the
metal–oxide interface may be important ingredients in determining their structure
and stability. Moreover, low-dimensionality aspects may play a role in the advent of
emergent phenomena.
Oxide nanolayers may be classified into two categories: (i) surface oxides, which
are generated by the oxidation of the corresponding bulk metal samples and (ii)
nanolayers of particular oxides supported on a different metal. Surface oxides have
attracted considerable interest recently as intermediate phases between the chemi-
sorbed oxygen layers on metal surfaces and the corresponding bulk oxides [5–7], and
their particular significance in some catalytic reactions has been pointed out [8, 9].
Oxide nanolayers supported on different metals constitute a different kind of
systems, which allow one to design nanostructured materials with tunable properties
via appropriate choices of metal substrate and oxide overlayer [10, 11]. They thus form
versatile and novel catalytic model systems, both as supports for the deposition of
metal nanoparticles and as active catalytic phases themselves. Of particular interest
are the epitaxially ordered oxide nanolayers that grow on noble metal single-crystal
surfaces. It is this class of oxide nanolayers and their emergent physical and chemical
properties as probed by STM that are addressed in this chapter.
A special case of nanostructured oxide-metal systems can be created by submo-
nolayer coverages of oxides on metal single crystals, where the surfaces expose areas
of both bare metal substrate and oxide islands. This forms the basis of the so-called
inverse model catalyst concept [12]. An inverse model catalyst surface is a planar
system consisting of a metal single-crystal surface that is decorated with oxide island
structures, that is, it is the reverse of a real catalyst where the metal component is
oxide supported. This inverse catalyst system contains both metal and oxide surface
sites and sites at the metal-oxide phase boundary; it is therefore an attractive model
system to study the role of the metal-oxide interface in catalytic reactions, as involved,
for example, in the promotor effect of oxide minority components that leads to the
enhancement of the reactivity of the underlying metal [13]. It has been shown that
the length of the metal-oxide phase boundary can be controlled by careful tuning of
deposition parameters of the oxide, thus allowing the assessment of the kinetics of
the promotion effect [14].
Since the main topic of this review is STM imaging, growth properties, surface
morphology, and atomic structures of oxide nanosystems are the central themes.
Oxide nanolayers on noble metal surfaces often display very complex structural
arrangements, as illustrated in the following sections. The determination of the
surface structure of a complex oxide nanophase by STM methods is, however, by no
means trivial; resolution at the atomic scale in STM is a necessary but not sufficient
condition for elucidating the atomic structure of an oxide nanophase. The problem
148
j
6 STM Imaging of Oxide Nanolayer Model Systems
with interpreting the STM images of oxide surfaces is that the physical and chemical
nature of the (bright or dark) contrast in the images is not always clear: bright
protrusions may be due to metal or oxygen atoms, they may represent single atoms
or groups of atoms. Since novel metal – oxygen coordination spheres and building
blocks may be present in oxide nanostructures that do not occur in the known bulk
oxide compounds, it is necessary to combine the STM data with as much structural
and spectroscopic information as possible; results from other experimental techni-
ques are necessary to obtain complementary information on long-range order and
lattice parameter (LEED), oxidation state, oxygen–metal coordination and stoichi-
ometry (XPS, NEXAFS) of the oxide phases. Moreover, the experimental data have to
be combined with high-level theoretical modeling, including both oxide overlayer
and support surfaces, to resolve the complicated oxide phases. Since very large unit
cells are frequently involved containing several tens of atoms, advanced DFT
methodology is necessary. Indeed, the progress of the last decade in resolving
complex oxide structures has been made possible by this close combination of
high-level experimental and theoretical methods [15].
The structure of the review is organized as follows. In Section 6.2, we will address
experimental aspects concerning apparatus developments and oxide nanolayer
preparation methods, and briefly comment on the interplay between experimental
and theoretical results. Section 6.3 constitutes the main body of this chapter, where
we present case studies of selected oxide-metal systems. They have been chosen
according to their prototypical oxide nanosystem behavior and because of their
importance in catalysis. We conclude with a synopsis and a brief outlook speculating
on future developments.
6.2
Experimental Aspects and Technical Developments
Since the invention of the STM by Binnig and Rohrer about 25 years ago [16], this
technique has become one of the most powerful tools for surface structure obser-
vation, in real space and at atomic length scales. The ongoing technical development
and improvement of STM instruments has largely extended the scope of their
application in a number of different disciplines – physics, chemistry, biology, and
engineering. One of the major apparatus developments was the building of variable-
temperature STMs, which allow imaging at temperatures well above or below room
temperature. In many surface phenomena, for example, surface phase transitions,
surface diffusion, or crystal growth, the temperature plays a key role. However, STM
instruments are susceptible to temperature variations of the sample–tip unit, causing
large thermal drifts (both parallel and perpendicular to the sample surface), which
prevent from keeping the same surface area within the microscopes view. Recently,
several variable-temperature STMs have been designed by various research groups
and commercial companies (e.g., Omicron, JEOL, RHK Technology, WA Technol-
ogy), where the thermal drift has been minimized in one or the other way (the
number of the various technical modifications has grown enormously over the past
6.2 Experimental Aspects and Technical Developments
j
149
decade, and their review is beyond the scope of this paper). The strengths of
employing a variable-temperature STM instrument in the field of the oxide nanolayer
growth are obvious and have been demonstrated recently in the study of oxidation–
reduction processes [17], structural phase transitions [18], and surface diffusion [19]
of vanadium oxide nanostructures on Pd(1 1 1) and Rh(1 1 1) surfaces, as well as for
monitoring in situ the growth of nickel oxide films on a Pt(1 1 1) substrate at elevated
temperature [20]. At the other extreme, the use of STM at liquid He temperatures
allows one to immobilize the surface species and thus investigate their local
electronic density of states (LDOS) by scanning tunneling spectroscopy (STS). This
technique has recently been applied to study the homogeneity of the surface
potential [21] or the appearance of Kondo-like resonances [22] on FeO bilayers on
Pt(1 1 1), as well as for identifying the electronic signatures of different building units
of vanadium oxide phases on Rh(1 1 1) [23].
Because the tunneling effect is not restricted to a vacuum environment, the STM
has the potential ofovercoming the so-called pressure gap in heterogeneous catalysis.
This has been realized in diverse high-pressure STM designs [9, 24–27],which enable
one to achieve atomic resolution during adsorption/desorption and reaction experi-
ments on single-crystal surfaces up to atmospheric pressures. In close relation to the
topic of the present review, Frenken and coworkers [9, 28] have performed experi-
ments in a high-pressure flow reactor and shown that surface oxides, forming on the
Pt(1 1 0) and Pt(1 0 0) surfaces under oxygen-rich conditions, have a dramatic effect
on the CO oxidation rate; these surface oxides have been suggested to be the
catalytically active phase. High-pressure STMs, when combined with the recent
development of the control electronics allowing high-speed scanning [29], will make
the STM an invaluable tool for monitoring surface dynamic processes or catalytic
reactions in real time.
Typically, ultrathin oxide films have been prepared by physical vapor deposition of
metal atomsonto a dissimilarsingle-crystal metal substrate in an oxygen atmosphere:
this is the so-called reactive evaporation (RE) method. Alternatively, the metal atoms
can be deposited first in UHV onto the metal substrate and subsequently oxidized;
this is the so-called postoxidation (PO) method. Which of these two preparation
techniques is applied with success in a specific case depends on the interfacial
chemistry, where factors such as the rate of oxygen dissociation on the metal
substrate, oxidation reaction, and bulk diffusion need to be carefully considered.
In most of the cases, molecular oxygen has been used as an oxidant in both the RE and
the PO methods, and the oxygen content in the oxide nanolayer has been controlled
by the oxygen pressure and the substrate temperature. Although the reactivity of
many metals (e.g., Mg, Al, or Ni) is sufficient to form fully oxidized films using
molecular oxygen in a high-vacuum environment (10
6
mbar), this is not the case
for other metals (e.g., Fe). The alternative is to use atomic oxygen sources, such as
an electron cyclotron resonance (ECR) oxygen plasma generator or gaseous NO
2
[30].
More recently, a high-pressure cell has been employed by Kuhlenbeck and
coworkers [31, 32], where oxygen pressures up to 50 mbar have been applied to
fully oxidize ultrathin V-oxide films to a V
2
O
5
stoichiometry. A modification of the
traditional PO approach has been reported recently by Matsumoto et al. [33, 34], who
150
j
6 STM Imaging of Oxide Nanolayer Model Systems
grew TiOx nanolayers on a Pt(1 0 0)surface, by formingan ordered Pt
3
Ti surface alloy
prior to the oxidation step with ozone. This procedure is to some extent similar to the
method used to form Al-oxide nanolayers via oxidation of NiAl(1 1 0) and Ni
3
Al(1 1 1)
bulk alloy crystal surfaces, as mentioned in Section 6.3.1, and was found to be
beneficial for the growth of well-ordered Ti-oxide overlayers, with an atomically flat
surface morphology, compared to films prepared by the conventional RE method
with molecular oxygen. An original approach for growing two-dimensional (2D)
oxide overlayers has been developed by Niehus et al. [35–37]. Here, a Cu
3
Au(1 1 0)
alloy surface has been used as a substrate, which when bombarded by O
þ
ions
and subsequently annealed in UHV becomes oxygen-terminated. This smooth
Cu
3
Au(1 1 0)-O surface, on the one hand, prevents the intermixing between the
evaporated metal atoms and the substrate, and, on the other hand, significantly acts
as an oxygen reservoir with subsurface oxygen stored near the surface, which can be
released in a controlled fashion by thermal treatment. This preparation technique
has been successfully applied for growing atomically flat VOx [35, 36], NbOx, and
MoOx [37] overlayers.
As mentioned in the introduction, the determination of oxide overlayer structures
beyond the simplest cases requires the combination of a high-level experimental and
theoretical methodology. As experimentalists, we refrain from a discussion of the
theoretical aspects, but some words on how the models derived by the theoretical
calculations can be checked against the experiments may be appropriate. For a given
structure model, which has been found to be stable on total energy grounds,
simulations of the STM images, calculation of core-level binding energies and X-ray
absorption profiles, and of phonon frequencies give suitable data for comparison
with experimental values. Core-level binding energies of ultrathin oxide overlayers on
metal surfaces as obtained by X-ray photoelectron spectroscopy (XPS) are problem-
atic to compare with corresponding values from bulk oxide samples because of the
proximity of the underlying metal surfaces. The latter gives rise to interfacial bonding
and screening effects of the photoionized final state that modify the measured
binding energy; typically, lower core-level binding energies of metal cations than in
the corresponding bulk oxides are obtained, making the oxidation state difficult to
evaluate. Here, calculations are often necessary to make progress and, conversely, the
agreement between theoretical predictions and experimental findings give credence
to a derived model. The phonon frequencies of oxide nanolayers have been found to
be particularly diagnostic [18] because they can be interpreted to good approximation
in a local building block picture. The agreement between calculated and measured
phonon frequencies, the latter obtained in, for example, high-resolution electron
energy loss spectroscopy experiments, is a stringent test on the validity of a structural
model.
The combination of state-of-the-art first-principles calculations of the electronic
structure with the Tersoff–Hamann method [38] to simulate STM images provides
a successful approach to interpret the STM images from oxide surfaces at the
atomic scale. Typically, the local energy-resolved density of states (DOS) is evaluated
and isosurfaces of constant charge density are determined. The comparison between
simulated and measured high-resolution STM images at different tunneling
6.2 Experimental Aspects and Technical Developments
j
151
voltages, therefore corresponding to the imaging of the spatial distribution of
different electronic states, is a very rigorous test for a given structural model.
6.3
Case Studies: Selected Oxide–Metal Systems
6.3.1
Alumina Nanolayers on NiAl Alloys
Ultrathin alumina layers on metallic substrates play a key role in many technologies,
for example, as dielectric and diffusion barrier layers in electronic devices and as
corrosion resistant coatings; and in model studies of catalytic processes, they are
widely used as stable supports for metal cluster growth [39]. The selective oxidation of
alloys provides a very promising route to synthesize highly ordered thin alumina
films [40], and in 1991 the Freund group has published a comprehensive paper on the
formation of a well-ordered aluminum oxide overlayer by high-temperature oxidation
of NiAl(1 1 0) [41], taking up earlier reports in the literature [42, 43]. Jaeger et al. [41],
applying a variety of electron spectroscopic techniques including LEED, concluded
that the oxide film is about 5 Å thick and most likely consists of two Al layers and
two quasi-hexagonal oxygen layers with oxygen surface termination. A distorted
hexagonal Al
2
O
3
-type surface such as a-Al
2
O
3
(0 0 0 1) or g-Al
2
O
3
(1 1 1) has been
proposed. Further progress has been made by Libuda et al. by the application of spot
profile analysis low-energy electron diffraction (SPA-LEED) and STM [44], where
two reflection domains inclined by 24
with respect to the [1 1 0] direction of the
NiAl(1 1 0) substrate have been identified. The alumina overlayer structure is
commensurate along the NiAl [1 1 0] direction but incommensurate along [0 0 1].
The first atomically resolved images of AlO
x
on NiAl(1 1 0) have been recorded with
a low-temperature STM by Kulawik et al. [45] using a sample bias of 0.5 V, that is,
a voltage that is within the bandgap of the oxide overlayer (see Figure 6.1). The
tunneling current is thus mediated by the substrate, with the oxide overlayer exerting
only a modulating influence. It was suggested that mainly the topmost oxygen layer
atoms are being imaged, whereas the contributions from the Al
3 þ
sublattice are very
small, but a possible influence of the properties of the tip has been emphasized [45].
Nevertheless, the atomic structure of the alumina overlayer on NiAl(1 1 0) remained
obscure and could not be resolved in this study.
The structure of alumina on NiAl(1 1 0) was the subject of a surface X-ray
diffraction study by Stierle et al. [46]. The model derived by Stierle et al. from the
analysis of the X-ray diffraction data was based on a strongly distorted double layer
of hexagonal oxygen ions, where the Al ions are hosted with equal probability on
octahedral- and tetrahedral-coordinated sites; the resulting film structure was closely
related to bulk k-Al
2
O
3
. An attractive feature of Stierles model was that it provided
a natural explanation of the domain structure of the alumina overlayer, which is
induced by a periodic row matching between film and substrate lattices. However,
as pointed out recently by Kresse et al. [47], this structure model has two bonds with
152
j
6 STM Imaging of Oxide Nanolayer Model Systems
unphysically short AlAl and AlO lengths, and was found to be unstable in ab initio
DFT calculations.
The riddle of the structure of alumina on NiAl(1 1 0) was finally solved by Kresse
and collaborators [47] with the help of STM data and extensive theoretical DFT
modeling calculations. The essential breakthrough was that the idea of hexagonal
oxygen planes, taken previously from the structure of practically all known Al
2
O
3
bulk phases, has been given up and has been replaced by a square arrangement of
oxygen atoms. The latter has been suggested after a close inspection of STM images.
Figure 6.2 shows the STM results and the resulting DFT-derived model. As suggested
before by Kulawik et al. [45], the positions of the oxygen atoms in the surface layer
have been taken from the room-temperature STM images (Figure 6.2b and c),
whereas the Al surface atoms, almost coplanar with these O atoms, have been
associated with bright protrusions in low-temperature STM images under certain
tunneling conditions (Figure 6.2e). It was also pointed out by Kresse [47] that the
atomic resolution in the low-temperature STM images at moderately high tunneling
voltages may not be simply because of LDOS effects but rather caused by an adsorbate
at the tip interacting with surface Al atoms. The surface thus consists of oxygen atoms
arranged in squares and triangles and Al atoms in between slightly below but almost
coplanar with the oxygen layer. The oxide film also has another oxygen and Al layer
at the interface with the substrate. The stoichiometry of the film is Al
4
O
6
Al
6
O
7
in
Figure 6.1 Atomically resolved STM image of a
thin Al
2
O
3
film on NiAl(1 1 0), taken at 4 K with
U ¼500 mV, I ¼1 nA (58 Å 40 Å). All atoms
exhibit a quasi-hexagonal coordination, where
the three characteristic directions are marked
with arrows. The unit cell was determined from
the repetition of atoms. Here, gray-filled circles
mark the most prominent atomic features,
thus characterizing six positions of the cell.
Furthermore, a zigzagged pattern is observed,
where the atoms appear alternately larger and
smaller. Differences in the apparent height
are emphasized with black and gray circles,
respectively. (Reproduced with permission from
Ref. [45].)
6.3 Case Studies: Selected Oxide–Metal Systems
j
153