Bowker M., Davies P.R. (Eds.) Scanning Tunneling Microscopy in Surface Science, Nanoscience and Catalysis

Подождите немного. Документ загружается.

Pd(1 0 0) [146–150] and Pt(1 1 1) [20]. The Neddermeyer group was ﬁrst to

demonstrate that owing to the small lattice mismatch (2%), well-ordered ultrathin

NiO ﬁlms can be prepared on a Au(1 1 1) surface by reactive evaporation of Ni metal

in an oxygen atmosphere [135, 136]. Their STM images revealed tripoid-like ridge

structures with a (2 2) periodicity, corresponding to a p(2 2) reconstructed

NiO(1 1 1) surface, which has been interpreted in terms of Wolfs octopolar

model [151]. The latter consists of {1 0 0} nanofacets, obtained by removing

three-fourth of the surface and one-fourth of the subsurface ions, which results in

a compensation of the polarity of the NiO(1 1 1) surface. More recent glancing

incidence X-ray diffraction (GIXD) study has shown that the (2 2)-reconstructed

NiO(1 1 1) ﬁlms can be grown on the Au(1 1 1) surface up to 8 ML in a layer-by-layer

fashion, and such ﬁlms exhibit both possible terminations (O and Ni), separated by

single-atom steps [152].

The Ag(1 0 0) surface was considered as an ideal candidate for the epitaxial growth

of NiO ﬁlms oriented along the (1 0 0) direction due to the low overlayer/substrate

lattice mismatch ( þ2.2%) and the low reactivity of silver toward oxygen. Interest-

ingly, despite this good structural matching no pseudomorphic NiO(1 0 0) layer

forms for submonolayer coverages. Instead, a 2D oxidic phase with a (2 1)

periodicity grows in the form of two orthogonal domains on the Ag(1 0 0) surface

up to a coverage close to 1 ML, as shown by STM images obtained by Bertrams and

Neddermeyer [138]. On the basis of quantitative LEED analysis, Cafﬁo et al. [142] have

Figure 6.14 Room temperature STM image (400 Å 270 Å,

þ1.0 V, 0.5 nA) after oxidation at 400K. An enlarged image detail

shows the structure of the (2 2), (23), and (3 3) cells in the

NiO(0 0 1) regions. (Reproduced with permission from Ref. [134].)

174

6 STM Imaging of Oxide Nanolayer Model Systems

proposed a structure model of the (2 1) phase similar to the (1 1 1) surface of NiO,

which consists of a buckled NiO bilayer (O-terminated) with a distorted hexagonal

structure. It has been argued that this polar NiO(1 1 1)-like nanolayer is stabilized

by the electrostatic interaction with the metal substrate and is kinetically preferred

over the NiO(1 0 0) structure in the early stages of oxide formation [142]. As the

oxygen dose and/or the ﬁlm thickness increases the (2 1) phase transforms into a

pseudomorphic (1 1)-NiO(1 0 0) phase, which grows in a layer-by-layer mode up to

5 ML [139]. For higher coverages, as demonstrated by SPA-LEED analysis [139], the

strain is relieved by the introduction of misﬁt dislocations with {1 1 0} glide planes,

which leads to the formation of mosaics on the oxide ﬁlm surface.

Recently, we investigated in a series of joint studies [146–150] with the group of

Granozzi in Padua the structure of NiO nanolayers supported on a Pd(1 0 0) surface.

The Pd(1 0 0) surface differs from Ag(1 0 0) by a signiﬁcantly larger lattice mismatch

( þ7.3%) to NiO(1 0 0) and by the predominantly 4d character of the valence band

near the Fermi edge, in contrast to the 5sp character of the silver top valence band.

This difference in the electronic structure of Pd and Ag is considered to play an

important role in the hybridization of states at the metal–oxide interface and in the

substrate reactivity toward oxygen. The latter was found to strongly affect the

morphology and degree of structural order of the Ni-oxide overlayer in the sub-

monolayer coverage range via the oxide preparation procedure: reactive deposition

(RD) versus postoxidation [148]. Irrespective of the oxide preparation route, an

interface-stabilized c(4 2) phase forms in the early stages of Ni-oxide growth, as

detected by LEED and STM [147, 148]. The PO procedure was shown to be bene ﬁcial

for obtaining a well-ordered wetting 2D layer, which has been imaged with atomic

resolution in STM (Figure 6.15a and inset). A structure model of this c(4 2)

monolayer has been derived by quantitative LEED I–Vanalysis [147] and is presented

in Figure 6.15b. It is based on a NiO(1 0 0) surface with O atoms sitting in on-top Pd

positions and with one-fourth of the Ni atoms missing in a rhombic c(4 2) unit cell,

which corresponds to a formal Ni

stoichiometry. Hybrid-exchange density

functional theory calculations of Pisani and coworkers [149] have conﬁrmed the

LEED I–V model, and it has been argued that the rhombic geometry of Ni vacancies

is electrostatically preferred over a squared one due to the higher average distance

between Ni vacancies in the former structure.

We close this section with highlighting our recent results on the fabrication of

quasi-one-dimensional (1D) Ni-oxide structures on vicinal Rh(1 1 1) surfaces via

decorating their regular step arrays [153, 154]. Figure 6.16 shows STM images of the

0.3 ML Ni/Rh(15 15 13) surface after exposure to 15 l O

at 300



C. Figure 6.16a and b

show straight steps decorated by bright lines of protrusions, which display a linear

1D-(1) and 1D-(2) periodicity in terms of the Rh lattice parameter parallel to the

step edges (Figure 6.16c and d). The oxide-free Rh terraces are covered by (2 1)

domains of chemisorbed oxygen. The structure of Ni-oxide wires at the step edges has

been rationalized in the DFTmodel of Mittendorfer and coworkers [153] for a 0.6 ML

Ni-oxide decorated Rh(5 5 3) surface (Figure 6.16e). Accordingly, the ﬁrst row of

Ni atoms at the step edges is coordinated to four O atoms each, yielding the 1D-(1)

periodicity at the upper and lower step edges. Additional O adatoms are adsorbed in

6.3 Case Studies: Selected Oxide–Metal Systems

175

a(21) lattice on the other Ni rows, resulting in the 1D-(2) lines of protrusions in

the STM images and reproduced well in the simulated STM image in Figure 6.16f.

Although the formal stoichiometry of the NiO stripes on the vicinal Rh(1 1 1)

surfaces is NiO, it has no relation to the bulk NiO, since a fraction of the O atoms is

shared between the Ni and the Rh surface atoms.

Figure 6.15 (a) STM image of 1.25 ML Ni-oxide on Pd(1 0 0)

(500 Å 500 Å, þ2.0 V, 0.1 nA). The inset shows a high-

resolution STM image of the c(4 2) structure (35 Å 35 Å,

þ1.0 V, 1.0 nA). (b) Top and side views of the NiO(1 0 0)-based

model for the c(4 2) overlayer. (Reproduced with permission

from Refs [147, 148].)

176

6 STM Imaging of Oxide Nanolayer Model Systems

6.3.6

Ceria Nanolayers on Metal Surfaces

Cerium oxides are outstanding oxide materials for catalytic purposes, and they are

used in many catalytic applications, for example, for the oxidation of CO, the removal

of SO

from ﬂuid catalytic cracking ﬂue gases, the water gas shift reaction, or in the

oxidative coupling reaction of methane [155, 156]. Ceria is also widely used as an

active component in the three-way catalyst for automotive exhaust pollution control,

Figure 6.16 (a–d) STM images of 0.3 ML Ni-oxide layer on the

Rh(15 15 13) surface: (a) 1000 Å 1000 Å, þ1 V, 0.1 nA;

(b) 150 Å 150 Å, þ0.13 V, 1.0 nA; (c) 18 Å 18 Å, þ0.13 V,

1.0 nA; (d) 45 Å 37 Å, þ0.14 V, 1.0 nA; (e) DFT model of a

0.6 ML Ni-oxide on Rh(5 5 3); and (f) corresponding simulated

STM image. (Reproduced with permission from Ref. [154].)

6.3 Case Studies: Selected Oxide–Metal Systems

177

where it is believed to play a crucial role as an oxygen storage-release agent to

control the oxygen concentration on the catalyst surface during the different cycles of

the operating engine. The latter involves oxygen transport from and to the surface,

which is determined to a large extent by the formation, concentration, and mobility

of oxygen vacancy defects at the surface and in the bulk. The ease of formation and

annihilation of oxygen vacancy defects in cerium oxides is of course intimately

connected to the propensity of Ce ions to undergo valence changes from þ4to þ3

and vice versa.

Therefore, it is not surprising that there has been and still is a strong interest in the

atomic structure of cerium oxide surfaces and of the defects encountered thereon.

There has been considerable effort to elucidate the surface structures of the low-index

single-crystal surfaces of CeO

, which crystallizes in the cubic ﬂuorite structure.

According to theory, the most stable low-index surface is the (1 1 1) surface with an

oxygen-terminating outer layer [157]. Note that in the (1 1 1) orientation, the CeO

crystal may be regarded as a stack of hexagonal OCeO trilayer planes. N

€

orenberg

and Briggs [158] were the ﬁrst to apply STM to the study of CeO

(1 1 1) surfaces, and

they indeed observed images with atomically resolved features in a hexagonal

arrangement. Because CeO

is a wide-band insulator with a bandgap of approxi-

mately 6 eV, the experiments had to be performed on partially reduced CeO

2x

surfaces to enhance the electron conductivity for the STM experiments; the reduced

CeO

2x

surfaces were obtained by annealing the samples in vacuum at high

temperatures (1000



C). On the basis of their tunneling conditions (negative sample

bias of several V, thus tunneling out of the ﬁlled states of the surface), N

€

orenberg and

Briggs concluded that they are imaging the oxygen atoms of the CeO

(1 1 1)-1 1

surface. The typical defect structures observed at the partially reduced CeO

2x

surfaces displayed triangular and linear dark contrast features, which were inter-

preted by N

€

orenberg and Briggs in terms of multiple oxygen vacancies [158]. The

defect formation process on CeO

(1 1 1) single-crystal surfaces has recently been

addressed in more detail using STM by Esch et al. [159], again on partially reduced

samples. In combination with DFTcalculations, the local structure of surface defects

has been ascribed to multiple surface and subsurface oxygen vacancies in various

combinations.

A better approach than using STM to study defects at CeO

single-crystal surfaces,

independent of the overall reduction state, is to use atomically resolved atomic force

microscopy. Iwasawa and coworkers [156, 160, 161] have imaged the oxygen atoms

on CeO

(1 1 1) surfaces with atomic resolution with noncontact AFM and have

observed several kinds of defects, both point defects and multiple defects such as

the line and triangular structures, which were also observed in STM before [158]. The

creation of oxygen vacancies leads to the formation of reduced Ce 3 þ ions in the

vicinity of the defect, which take up the two negative charges left behind per oxygen

vacancy. This leads to local reconstructions of the atoms around the defect that can be

visualized in the AFM images and that contribute to the stability of the defect.

By means of successive AFM measurements of the same area on slightly reduced

CeO

2x

surfaces, the dynamics of surface O atoms and the healing of defects by

exposure to O

from the gas phase has been observed directly by Namai et al. [161].

178

6 STM Imaging of Oxide Nanolayer Model Systems

High-quality resolved images and atomic details of surface features on CeO

(1 1 1) in

various oxidation states have also been obtained recently by dynamic AFM measure-

ments by Gritschneder et al. [162].

Thin ﬁlms of CeO

2x

with (1 1 1) surface orientation have b een grown epitaxially

on a variet y of metal substrates. Mullins et al. [163] have prepared highly ordered

CeO

2x

(111)-typeﬁlms, up to 10 ML thick, by reactive evaporation of Ce in oxygen

atmosphereonRu(0001)andNi(111)surfaces.Theoxideoverlayerswerealigned

parallel to the princip al azimut hs of the substrates as judged from ion s cattering

spectroscopy and LEED measurements, and the degree of oxidation (i.e., the x in

CeO

2x

) could be adjusted to some exte nt by the oxygen pressure during evapo-

ration. No direct imaging information on the morphology and structure of these

ﬁlms was available in the early work of Mullins et al. [163]. However, Lu et al. [164]

recently took up and modiﬁed the recipe of Mullins for the preparation of CeO

(1 1 1)

ﬁlms on Ru(0 0 0 1) and have shown that higher oxidation temperatures lead to ﬂat

ceria ﬁlms with a low defect density. Schierbaum and Berner [165–167] have used a

different app roach to generate epitaxial ceria nanolayers on Pt(1 1 1), namely, the

oxidation of Pt–Ce alloy surfaces; the latter were formed by depositing Ce metal

onto P t(1 1 1) surfaces followed by a high-temperature (1000 K) annealing

treatment. The surface structures of the oxide nanolayers formed were consistent

with CeO

(1 1 1) as revealed by STM, but heating of the oxide phase in vacuum at

900 K lead to a different structure, which was attributed to a Ce

surface

phase [167].

Epitaxial cerium oxide nanolayers have been deposited by reactive evaporation on

Rh(1 1 1) by Eck et al. [168]. STM and LEED indicated that the ceria grows in the form

of ordered CeO

double-layer islands, with (1 1 1) faces parallel to the surface and

orientationally aligned to the main azimuthal directions of the substrate. While the

nanolayer ﬁlms contained signiﬁcant amounts of reduced Ce 3 þ species, for the

thicker ﬁlms (>6 ML) stoichiometric CeO

was detected in XPS. Vacuum annealing of

the CeO

/Rh(1 1 1) submonolayer nanostructures produced morphologically well-

deﬁned hexagonal islands (see Figure 6.17a), which displayed a characteristic Moir



pattern in STM at their surfaces (Figure 6.17b and c) [169]; the latter is the result of

the interfacial lattice mismatch and the interference effects between the CeO

and the

Rh lattices, which give rise to a (5 5) coincidence lattice of a compressed cerium

oxide overlayer (5 a

CeO

7 a

). When subjected to more reducing conditions,

for example, by additional vacuum annealing at >600



C, an ordered array of surface

defects has been observed on the ceria nanoislands as displayed in the STM images of

Figure 6.18. In contrast to the STM images obtained from bulk-terminated CeO

surfaces [158, 159], the STM images from the CeO

nanolayers on Rh (Figures 6.17

and 6.18) have been recorded with positive bias voltages [169], where the empty states

are imaged. Ab initio DFTcalculations have revealed that the density of states within

1 eVabove the Fermi level in the ﬁrst CeO

trilayer of the ceria–Rh system is made up

by Ce 4f states, which are hybridized to a certain extent with O 2p states; according to

the calculations, the bright maxima in the corresponding STM images correspond,

therefore, to the cerium sites [169]. Using STM images with atomic resolution

in combination with DFT calculations, the defects seen in Figure 6.18 have been

6.3 Case Studies: Selected Oxide–Metal Systems

179

Figure 6.17 Constant current topographic STM images of 0.5

monolayer (ML) of CeO

on Rh(1 1 1). (a) 1850 Å 1850 Å,

þ1.99 V, 1 nA; (b) 200 Å 200 Å, þ0.89 V, 0.83 nA;

(1 1 1)1 1 unit

cell (u) is indicated on the image (c). The inset to (a) shows

a corresponding LEED pattern (electron energy ¼77 eV).

(Reproduced with permission from Ref. [169].)

180

6 STM Imaging of Oxide Nanolayer Model Systems

associated with oxygen vacancies. It has been suggested that this self-assembly of

oxygen vacancies is facilitated by the lattice mismatch between the oxide overlayer

and the metal substrate and that it may be ascribed to a strain-related reduction in the

vacancy formation energy [169].

Figure 6.18 STM images of the 0.5 ML CeO

2x

–Rh(1 1 1) surface

after annealing to 625



C. (a) 200 Å 200 Å, þ0.93 V, 0.86 nA;

(b) 100 Å 100 Å, þ0.80 V, 1.05 nA. The grid of thin lines

illustrates the superlattice of defects. (Reproduced with

permission from Ref. [169].)

6.3 Case Studies: Selected Oxide–Metal Systems

181

An interesting point is the fact that the epitaxial growth ofCeO

overlayers has been

observed on a number of noble metal surfaces [163, 165, 168], all of which display a

large lattice mismatch of the order of at least 30% at the respective interfaces. It has

been proposed that coincidence lattices with relatively low interfacial strain are

formed in all these cases, in a manner similar to the one described above for the

CeO

–Rh(1 1 1) interface, thus creating ordered low-energy interfaces that mediate

the heteroepitaxial growth further [169].

6.4

Synopsis and Outlook

As in other ﬁelds of nanoscience, the application of STM techniques to the study

of ultrathin oxide layers has opened up a new era of oxide materials research.

New emergent phenomena of structure, stoichiometry, and associated physical and

chemical properties have been observed and new oxide phases, hitherto unknown in

the form of bulk material, have been detected in nanolayer form and have been

elucidated with the help of the STM. Some of these oxide nanolayers are and will be of

paramount interest to the ﬁeld of advanced catalysis, as active and passive layers in

catalytic model studies, on the one hand, and perhaps even as components in real

nanocatalytic applications, on the other hand. We have illustrated with the help of

prototypical examples the growth and the structural variety of oxide nanolayers on

metal surfaces as seen from the perspective of the STM. The selection of the

particular oxide systems presented here reﬂects in part their relevance in catalysis

and is also related to our own scientiﬁc experience.

We believe that the ﬁeld of oxide nanolayer characterization is still in its early

stages and that many novel oxide phases will be detected in the years to come. While

the way to their structural characterization is essentially paved, the elucidation of

other properties such as catalytic chemistry or physical properties including mag-

netic behavior is largely unexplored. The commercial availability of high-quality

low-temperature STM instruments has opened the way to the observation of

chemical reactions on a single-atom basis – keyword: single-atom catalysis – that

is of fundamental scientiﬁc interest. The local spectroscopy capability of STS in

probing the electronic structure at the single-atom level is central to this single-atom

chemistry approach. Bond formations induced by the STM tip [170] or STM-induced

catalysis [171] are directions of chemical nanostructure research, where increasing

activities can be expected in the future. The reduction of the dimensionality of

metal-supported oxide nanostructures from three- and two-dimensional to one-

dimensional (monoatomic oxide line structures [153]) or quasi-zero-dimensional

(oxide quantum dots), thus creating low-dimensional oxide–metal nanoscale hybrid

structures, is another area of scientiﬁc endeavor with promise of ﬁnding novel

chemical and physical behavior. These latter systems will allow us to study, in a

controlled way and at the atomic scale, the chemical reactivity of low-coordinated

surface sites, which are abundantly present in practical catalyst systems but which

escape direct scientiﬁc characterization due to the inherent lack of control at the

182

6 STM Imaging of Oxide Nanolayer Model Systems

atomic level. In all these areas, STM techniques will play a central role, and the future

of STM applications in the ﬁeld of surface chemistry and catalysis, and oxide catalysis

in particular, is bright and wide open.

Acknowledgments

This work has been supported by the Austrian Science Funds within the National

Research Network Nanoscience on Surfaces. FPN acknowledges with gratitude

the excellent hospitalities of Professor Wolf-Dieter Schneider, EPFL Lausanne, and

Professor Charlie Campbell, University of Washington, during his sabbatical stays in

Lausanne and Seattle in 2008.

References

1 Freund, H.-J., Kuhlenbeck, H., and

Staemmler, V. (1996) Rep. Prog. Phys., 59,

283.

2 Street, S.C., Xu, C., and Goodman, D.W.

(1997) Annu. Rev. Phys. Chem., 48, 43.

3 Freund, H.-J. (2007) Surf. Sci., 601, 1438.

4 Schoiswohl, J., Surnev, S., Netzer, F.P.,

and Kresse, G. (2006) J. Phys.: Condens.

Matter, 18, R1.

5 Lundgren, E., Kresse, G., Klein, C.,

Borg, M., Andersen, J.N., De Santis, M.,

Gauthier, Y., Konvicka, C., Schmid, M.,

and Varga, P. (2002) Phys. Rev. Lett., 88,

246103.

6 Gustafson, J., Mikkelsen, A., Borg, M.,

Lundgren, E., K

€

ohler, L., Kresse, G.,

Schmid, M., Varga, P., Yuhara, J.,

Torrelles, X., Quiros, C., and Andersen,

J.N. (2004) Phys. Rev. Lett., 92, 126102.

7 Lundgren, E., Mikkelsen, A., Andersen,

J.N., Kresse, G., Schmid, M., and

Varga, P. (2006) J. Phys.: Condens. Matter,

18, R481.

8 Over, H., Kim, Y.D., Seitsonen, A.P.,

Wendt, S., Lundgren, E., Schmid, M.,

Varga, P., Morgante, A., and Ertl, G.

(2000) Science, 287, 1474.

9 Hendriksen, B.L.M. and Frenken, J.W.M.

(2002) Phys. Rev. Lett., 89, 046101.

10 Giordano, L., Cinquini, F., and Pacchioni,

G. (2005) Phys. Rev., B73, 045414.

11 Giordano, L. and Pacchioni, G. (2006)

Phys. Chem. Chem. Phys., 8, 3335.

12 Sock, M., Surnev, S., Ramsey, M.G., and

Netzer, F.P. (2001) Top. Catal., 14, 15.

13 Boffa, A., Lin, C., Bell, A.T., and Somorjai,

G.A. (1994) J. Catal., 149, 149.

14 Schoiswohl, J., Surnev, S., and Netzer,

F.P. (2007) J. Phys. Chem. C, 111, 10503.

15 Surnev, S., Kresse, G., Sock, M., Ramsey,

M.G., and Netzer, F.P. (2001) Surf. Sci.,

495, 91.

16 Binnig, G., Rohrer, H., Gerber, Ch., and

Weibel, E. (1982) Phys. Rev. Lett., 49, 57.

17 Surnev, S., Schoiswohl, J., Kresse, G.,

Ramsey, M.G., and Netzer, F.P. (2002)

Phys. Rev. Lett., 89, 246101.

18 Schoiswohl, J., Surnev, S., Sock, M.,

Eck, S., Ramsey, M.G., Netzer, F.P., and

Kresse, G. (2005) Phys. Rev., B71, 165437.

19 Schoiswohl, J., Kresse, G., Surnev, S.,

Sock, M., Ramsey, M.G., and Netzer, F.P.

(2004) Phys. Rev. Lett., 92, 206103.

20 Hagendorf, Ch., Shantyr, R., Nedder-

meyer, H., and Widdra, W. (2006) Phys.

Chem. Chem. Phys., 8, 1575.

21 Rienks, E.D.L., Nilius, N.,Rust, H.-P., and

Freund, H.-J. (2005) Phys. Rev. B, 71,

241404(R).

22 Rienks, E.D.L., Nilius, N., Giordano, L.,

Goniakowski, J., Pacchioni, G.,

Felicissimo, M.P., Risse, T., Rust, H.-P.,

References

183