Reducible oxides as ultrathin epitaxial films
Paola Luches, Sergio D’Addato
Abstract
This chapter reviews and discusses recent work on two-dimensional films of reducible oxides
supported on metal substrates. In general, peculiar chemical and structural phases, different from
the bulk ones, can be stabilized depending on the oxygen chemical potential, on kinetic processes
and on the specific substrate used. A peculiarity of reducible oxides is that the observed phases can
often be reversibly transformed one into the other by applying reducing and oxidizing treatments.
1. Introduction
An oxide is defined reducible if it can be easily and reversibly reduced depending on the ambient
conditions. Reducibility is linked to the existence of two or more oxidation states with comparable
stability for the cations. Prototypical reducible oxides are represented by cerium and titanium
oxides, although also other transition metal and rare earth oxides can be considered reducible;
among these for example other rare earth oxides, like PrOx, SmOx, TbOx, and other 3d metal oxides,
like VOx, MnOx, FeOx, CoOx, but also HfOx, TaOx, NbOx, WOx and many more.
Indeed, reducibility is very relevant for catalysis, since materials based on reducible oxides can act
as oxygen buffers, which can store and release oxygen and/or charge, promoting redox reactions
with a unique regeneration ability [1]. Furthermore, reducibility is an important property also in
view of the application of oxides in other fields, like for example energy conversion and storage,1
biomedicine [2] and memories [3]. Reducibility is not only linked to the material capacity of easily
forming oxygen vacancies but also to the reversibility of the process and to oxygen transport
properties within the material.
The influence of reduced dimensionality on the reducibility of an oxide is certainly relevant in view
of the optimization of the properties of the material through the understanding of the modifications
which arise with confinement. Pronounced changes of the electronic structure of oxides at reduced
dimensionality have been widely investigated, and interesting properties have been observed on
reducible oxides in particular [4,5]. In cerium oxide particles of nanometric size the oxygen vacancy
formation energy has been found to be greatly reduced and to reach a minimum at a specific size
[4], thereby causing a surprisingly enhanced reactivity of catalysts made of metal nanoparticles
supported on nanocrystalline cerium oxide [6], possibly also facilitated by easy oxygen transfer to
the supported metal nanoparticles [7].
If only one of the dimensions of the considered material is confined to the nanoscale, i.e. if the
system under study is a two-dimensional film, several interesting modifications are introduced and
intriguing phenomena can take place. The stabilization of metastable structural phases by epitaxy or
by the mere spatial confinement, the presence of lattice strain or rumpling, the interaction with the
substrate, in terms of charge transfer but possibly also of interfacial atom exchange, the different
1
stability of structural defects, are all expected to determine relevant changes in the electronic
structure, which give origin to systems with different properties compared to the corresponding bulk
phases.
The studies of two-dimensional reducible oxides are typically carried out on model systems, i.e. on
ultrathin films supported on single crystalline substrates, prepared and analyzed in high or ultrahigh
vacuum. The investigations aim at identifying interesting aspects on simple systems, which can be
analyzed by surface science techniques and understood with the help of theoretical descriptions. A
proper modeling of the systems under investigation often represents a crucial step for a complete
atomic level understanding of the material properties and potential performances. Only through the
fundamental insight of the new properties which arise on simple systems, the much more complex
real systems, based on the same materials, can be understood and optimized in their functionality.
For example, catalysts made of metallic nanoparticles supported on reducible oxides typically
involve some degree of encapsulation of the metal within the oxide and the formation of very thin
oxide layers on the metal nanoparticle surface, through the so called strong metal-support
interaction (SMSI) [8]. Ultrathin oxide layers exposing different facets, with different structures and
morphologies can be prepared in the form of model systems to identify the most active species
towards the considered reaction. The results of these studies can be used to design the catalysts with
an optimized activity and selectivity by finding suitable synthetic routes to maximize the density of
required active sites for the chosen specific functionality.
In this chapter recent studies on the properties of ultrathin films of reducible oxides are reviewed,
focusing on the aspects which influence the reducibility of the materials at the two dimensional
limit. New properties of cerium oxide two-dimensional films are discussed in section 2. Section 3
reports studies of titanium oxide two-dimensional films, which are interesting for comparison and
for complementing the concepts outlined on cerium oxide based systems. Selected studies of
ultrathin films of different reducible transition metal and rare earth oxides are finally reported in
section 4.
2. Cerium oxide two-dimensional films
The most stable cerium oxide phase is CeO2, also known as ceria, which has a fluorite structure
with cerium ions in the 4+ oxidation state. The Ce2O3 phase, with an orthorhombic structure (Atype phase), is relatively less stable at ambient conditions. Several additional metastable phases
with different stoichiometry can be stabilized in specific conditions [9], among them the Ce2O3
bixbyite (c-type) phase deserves a particular relevance, since it can be obtained from the most stable
fluorite-type CeO2 phase by removing 25% of the oxygen lattice atoms in an ordered way and
allowing for a slight structural rearrangement [10].
The (111) surface of the CeO2 phase is the thermodynamically most stable one, followed by the
(100) and (110) surfaces, the latter two being polar [11]. The repeating unit along the (111)
direction is an O-Ce-O triple layer, which in the following will be referred to as a monolayer (ML).
Cerium oxide ultrathin films exposing the (111) surface can be obtained by epitaxial growth on
metallic single crystal substrates with six-fold surface symmetry such as (111) surfaces of cubic
structures [12-18], or hcp (0001) surfaces [14,19,20]. The lattice mismatch between cerium oxide
2
and most of the metal substrates is quite large and it amounts to approximately 30-40%.
Nevertheless, films with a (111) surface orientation and a very good epitaxial quality have been
obtained [15,17,21,22]. The (100) and (110) surface orientation can be stabilized through epitaxy
using substrates with a very small lattice mismatch with CeO2. These are typically non-metallic and
include Si [22,23], YSZ [24] and SrTiO3 [25,26]. However, most of the studies performed using non
metallic substrates focus on thick films with several tens of nm thickness, possibly because of the
difficulties in applying surface science techniques, necessary to obtain information at the ultrathin
limit, on substrates with low conductivity.
The first and main part of this section will be dedicated to studies of cerium oxide (111) ultrathin
films, and the final part to the investigations concerning ultrathin films exposing less stable
surfaces. Important aspects connected to reducibility at the two- dimensional limit will be reviewed
and discussed, with focus on: i) structure, morphology and defectivity, ii) charge transfer and
intermixing at the interface, iii) modifications in structure and morphology induced by reduction,
iv) stabilization of metastable structures induced by dimensionality and/or epitaxial constraints.
Pioneering works studying ultrathin cerium oxide films on the Pt(111) surface date back to the midnineties and they were motivated by the idea that a deep understanding of ceria-based materials
obtained from studies of single crystalline surfaces and films of different thickness, structure and
morphology could help improving the activity of cerium oxide supports in three way catalytic
converters [12,13,27,28]. Cerium oxide films were grown either by deposition of metallic cerium
followed by post-oxidation [12,27] or by oxidation of Ce-Pt surface alloys [13,28]. Already these
early works pointed out that ultrathin films in the ML and sub-ML range contain a non negligible
Ce3+ concentration, while thicker films can be fully oxidized [12,13]. As will be clearer in the
following, this aspect is rather crucial in trying to identify a charge transfer from the metal substrate
and/or a dimensionality-induced decrease of oxygen vacancy formation energy induced by the
lower average O coordination of Ce ions at the early stages of the growth. The structure of the films
was found to be disordered unless thermal treatments at temperatures higher than 700 K, which
stabilize the fluorite (111) phase, are performed [12,13]. Thermally-induced instabilities were
observed above 1000 K [12,13]. Interestingly, films which fully cover the substrate showed a
significant activity towards CO oxidation, higher than the one of the bare Pt(111) surface [12]. An
important role of the metal-oxide interfacial sites and of non-(111) oriented surfaces was also
hypothesized, based on the high activity of films which do not fully cover the substrate [12].
The results of the early studies provided a strong motivation for subsequent works, using more
controlled growth procedures, combined with a variety of techniques for surface characterization
down to the atomic level. Cerium oxide ultrathin films with the (111) fluorite structure were grown
on Pt(111) [17,21,28,29], Rh(111) [15,16], Ru(0001) [14,20], Ni(111) [14], Cu(111) [30,31],
Au(111) [32], Pd(111) [33] and Re(0001) [34] substrates. On one side these studies confirmed the
enhanced reactivity of bidimensional cerium oxide films compared to the clean substrates and the
importance of metal–oxide interface sites [31,35], on the other they clarified important fundamental
aspects linked to the reducibility of low-dimensional supported cerium oxide systems [15,29,30].
In general, the studies of ultrathin supported cerium oxide films and islands were motivated by the
need to identify the active sites in real catalysts made of metallic nanoparticles supported on cerium
oxide. Several works specifically addressed the reactivity of metal-supported cerium oxide systems
3
and towards selected reactions [31,35,36,37]. A catalytically active role of the interface between
cerium oxide and Rh(111) in CO oxidation was suggested for example by Eck et al., who identified
preferential CO adsorption sites near the phase boundaries of the oxide islands [36]. Suchorski et al.
showed that the catalytic activity towards CO oxidation of cerium oxide nanoislands on Pt(111) is
remarkably increased compared to the bare Pt(111) surface [35]. By comparing systems with
different density of step edges they suggested the observed increase to be due to the different
electronic properties at the islands perimeter [35]. A similar role for the oxide metal interface,
combined with the low dimensionality of cerium oxide, was identified for the CO oxidation reaction
and for the water gas shift reaction on the cerium oxide/Cu(111) inverse catalyst model system
[31,37]. Studies of this kind stimulated more accurate work on cerium oxide bidimensional
structures and systematic investigation of the influence of the preparation conditions and of the
substrate used on the structure, morphology, defectivity and reducibility of the investigated systems.
On most substrates post-growth thermal treatments in O2 have been found to improve the surface
morphology, stoichiometry and structure of the cerium oxide terraces obtained [17,20]. The
ultrathin films have relatively large coincidence cells with the underlying metal surface giving
origin to moiré patterns in scanning tunneling microscopy (STM) images [20,38] or in low energy
electron diffraction (LEED) patterns [15]. The coincidence cells with the lowest strain typically
involve 5×5 ceria surface cells and 7×7 substrate surface cells (hereafter termed 5:7 coincidence),
however at the ultrathin limit smaller coincidence cells, implying a larger strain, can be stabilized.
This phenomenon has been observed for example on the Pt(111) substrate, where ultrathin films
have a compressed surface structure and an interatomic distance comparable with the 3:4
coincidence cell, while thicker films have the relaxed bulk structure [17,21,39]. At the ultrathin
limit a contraction of the in plane lattice parameter has been observed also using Rh(111) as a
substrate [15], and even on Cu(111) substrates [38]. The evidence for a lattice contraction also on
Cu(111), on which cerium oxide could adopt a 3:2 coincidence with the substrate with negligible
strain, demonstrates that the tendency for lattice contraction is probably an intrinsic property linked
to reduced dimensionality [38].
The deposition of cerium oxide amounts in the ML range typically results in the formation of large
flat islands of ML or multi-layer height [15,17,20]. Procedures to obtain continuous films of
monolayer thickness with almost complete coverage have been identified on the Cu(111) substrate,
using low temperature (110 K) growth in O2 followed by post-annealing in O2 at 770 K [18].
Ultrathin cerium oxide films grown in strongly oxidizing conditions contain a non-negligible
concentration of Ce3+ ions on the Pd(111) [33], Pt(111) [17] and on the Rh(111) substrate [15]. In
the latter case a preferential localization of the Ce3+ sites at the interface was deduced by
comparison of x-ray photoemission spectroscopy (XPS) and resonant valence band photoemission
spectra. Valence band resonant photoemission, having a smaller probing depth than XPS, does not
show any significant Ce3+ concentration on the surface. The non-negligible Ce3+ concentration
detected by XPS is therefore ascribed to interface sites [15]. This represents a strong evidence of
possible charge transfer from the metallic substrate towards the cerium oxide films. Effects such as
charge transfer can indeed significantly alter the properties of two-dimensional cerium oxide films,
and a definite assignment of the origin of reduced interfacial states requires a theoretical description
of the system. On the Cu(111) substrate density functional theory (DFT) calculations showed that a
very relevant charge transfer from the metal to the oxide takes place, yielding to the reduction of a
4
full layer of interfacial Ce ions [40]. DFT calculations on a 1 ML cerium oxide film supported on
Pt(111), instead, evidenced electrostatic interactions between the topmost Pt layer and the
interfacial oxygen atoms in the oxide, which cause a significant corrugation of both the film surface
and of the interfacial Pt layer [41]. The charge transfer between Pt and Ce ions is significant only in
the interfacial sites where the distance between the two atoms is short [41].
Not only the size, but also the shape and defects of the terraces obtained on the surface of ultrathin
cerium oxide films seem to depend on the preparation conditions. On Ru(0001), using high
temperature (T>800 K) growth and post-growth annealing (T~1000 K) in oxygen atmosphere
(P>10-7 mbar), ultrathin films with point defects, linear defects ascribed to grain boundaries and
round terrace step edges have been obtained [20]. The presence of grain boundaries, possibly
induced by three-dimensional growth mode at the early stages of film formation [20], is not
suppressed by room temperature growth followed by annealing of ceria films using a Pt(111)
substrate [17]. On this substrate ultrathin films present terraces with straight edges, oriented along
the substrate main symmetry directions [17]. Thicker films instead show more rounded terrace
edges [17]. This is possibly caused by the stabilization of less stable step edges by the proximity of
the metallic substrate in ultrathin films. A detailed characterization of the film step edges for
various preparation conditions showed that depending on the heating temperature different step
types can be obtained. The different steps are rationalized in terms of stability of the exposed facets
and analyzed in terms of electronic properties [42]. One–dimensional electronic states have been
found to develop on specific step types, and at domain boundaries as shown in figure 1, which
reports STM images acquired at different sample biases.
Fig.1: 65 ×65 nm2 STM topographic images of a 6 ML cerium oxide film grown on Ru(0001) with islands and holes
exposing different step edge orientation. Steps indicated as I and domain boundaries show an increased apparent height
at 4.4 V sample bias, while steps indicated as II do not change apparent contrast with bias. Reprinted with permission
from Nilius et al. [42] Copyright 2012 American Chemical Society.
The step edges exposed by cerium oxide two-dimensional films show similarities to those exposed
by bulk ceria surfaces, although in this case depressions rather protrusions are most frequently
observed on the surface [43]. A non-negligible density of point defects has been often observed on
ultrathin film terraces [21]. The defects are very similar to the ones formed on bulk surfaces, the
most common being surface oxygen vacancies, also imaged as trimers and as linear arrays, and
subsurface oxygen vacancies [21]. Defects can also be intentionally induced by electron
bombardment of the film surface [44]. A work by Jerratsch et al. identified some degree of
delocalization of the charge left after O vacancy formation [44]. The delocalization of charge after
O release has been highly debated also on bulk surfaces [45,46], and at reduced dimensionality
some non negligible degree of delocalization has been clearly identified [5,47].
5
Cerium oxide films can be reduced by different procedures, including thermal annealing in vacuum
[15,17,29,35], growth of metallic Ce in low oxygen background pressure [14,16,48], ion
bombardment [49] or exposure to reducing gases [49-51]. Vacuum thermal treatments often induce
also a structural and morphological modification, pointed out to be strongly dependent on the initial
film thickness [15,29]. In particular, the temperature at which reduction starts is lower for thinner
films, which also show an apparently higher final degree of reduction [15,29]. Indeed it is difficult
to determine the absolute amount of Ce3+ ions formed in the different samples by the different
treatments, given the unknown shape of the depth profile of O vacancy concentration, and the
different depth sensitivity of the techniques used to measure the Ce3+/Ce4+ ratio. However, the idea
of having a higher density of Ce3+ sites in the surface layer after reduction is generally accepted
[29]. A dependence of the onset of cerium oxide surface reduction on the growth temperature of the
film, and hence on the density of reduced coordination sites, has been pointed for films grown on
the Cu(111) substrate [38]. A very interesting effect, shown to take place on the surface of a thin
ceria film on Rh(111) after reduction, is the formation of an ordered array of surface defects, clearly
visible in the STM images (figure 2), and ascribed to triple oxygen vacancies [52]. The smaller
energy for the formation of vacancies at specific sites of the 5:7 coincidence lattice is ascribed to
the local surface stress [52].
Fig.2: STM images of a 0.5 ML CeO2-x film grown on the Rh(111) surface after annealing at ~ 900 K. (a) 200 × 200 Å2;
0.93 V; 0.86 nA. (b) 100 × 100 Å2; 0.80 V; 1.05 nA. The grid of black lines evidences the superlattice of defects,
6
ascribed to triple oxygen vacancies. Reprinted with permission from Casterllarin-Cudia et al. [52] Copyright 2004
Elsevier.
The surface of reduced ceria films on Pt(111) instead shows a corrugated morphology with biasdependent features tentatively ascribed to electronic modifications induced by reduction [29].
Interestingly, ultrathin reduced cerium oxide phases show peculiar surface reconstructions (figure
3), whose origin is still under investigation and appears to be linked to the presence of the
underlying substrate [29]. The effect of the heating time and heating rate on the final degree of
reduction of the films has been also pointed out to be very important [29]. Some extra-periodicities
have been observed also on the Cu(111) substrate after deposition of variable amounts of metallic
cerium on a CeO2 buffer layer, followed by annealing in UHV [53]. Some of them are commonly
observed also on the surface of thick films after reduction [54] and correspond to metastable bulk
phases such as the Ce7O12 phase and the c-type bixbyite phase.
Fig.3 LEED patterns (Ep = 80 eV) of a 2 ML cerium oxide film (a) as prepared, (b) after intermediate reduction by
heating in UHV at 770 K for 30 min (cCe3+ ~ 40%, as measured by XPS), showing the (3 × 3) and the 9/4(√3 × √3)R30°
phase (c) after strong reduction by heating in UHV at 1040 K for 15 min (cCe3+~60–80%) showing the 9/4(√3 ×
√3)R30° phase (d) after re-oxidation by heating in O2 at 1040 K. Luches et al. [29] - Reproduced by permission of the
PCCP Owner Societies.
The full reversibility of the reduction process was demonstrated for ceria films of different
thickness on Pt(111) [29]. We note here that this aspect is not trivial, and it is linked to the
reducibility of the films. Cerium oxide films on a Si(111) substrate reduced by vacuum thermal
treatments under conditions similar to those used in ref. 29, showed a non reversible reduction of Ce
ions from the 4+ state to the 3+ state [23]. This process was ascribed to the formation of interface
cerium silicate phases which are not reducible [23]. Interfacial atom exchange during growth was
also observed using Ni(111) and Cu(111) substrates, and they were shown to have a non-negligible
7
influence on the stoichiometry of the films at the ultrathin limit [14,31]. On the Au(111) substrate
the formation of Au/Ce alloys hinders the good ordering of extended two-dimensional films [32].
Substoichiometric films were also obtained by evaporation of Ce on Ru(0001) and on Ni(111) in
low oxygen pressure (PO2~10-8 mbar) and they were found to have the fluorite structure with a high
concentration of oxygen vacancies in at least the topmost oxygen atom layers compatible with a
bixbyite structure rather than to a hexagonal Ce2O3 structure [14].
As reported above, several works have shown some degree of reducibility of cerium oxide ultrathin
films, which is typically higher than the bulk, but they also evidenced that fully reduced states are
very difficult to be achieved [15,29]. This may be linked to the fact that it is easy to form oxygen
vacancies in the fluorite structure up to a certain density, or to the fast (partial) reversibility of the
reduction process when high Ce3+ concentrations are obtained. The cubic c-type bixbyite phase,
with 25% oxygen vacancies is in fact unstable in the bulk form and a full cerium oxide reduction
may require a transition to the most stable hexagonal A-type phase. A metastable epitaxial cubic ctype Ce2O3 phase has been stabilized on the Cu(111) surface by deposition of metallic Ce on a
CeO2 buffer layer followed by 900 K annealing [55]. The film shows a well ordered surface with a
(4x4) reconstruction with respect to the CeO2(111) surface (figure 4). The reconstruction is ascribed
to ordered arrays of quadruple oxygen vacancy clusters and corresponds to the bulk termination of
c-Ce2O3 [55]. Although the film thickness is as high as a few nm, the influence of the substrate is
considered to be determinant for the stabilization of the observed phase, ascribed to the effect of the
tetragonal strain induced by the Cu substrate [55]. A phase with a similar c-type bixbyite structure
in the form of an ultrathin film was also stabilized on a Cl-passivated Si(111) surface [22].
8
Fig.4: STM images showing the different stages of formation of c-Ce2O3 films on Cu(111). (a) CeO2 buffer, (b) CeO2
buffer with subsequent metallic Ce deposition, (c) ordered c-Ce2O3 layer obtained by annealing (b) in vacuum at 900 K.
Inset: high-resolution image and surface unit cell (red rhombus) of the c-Ce2O3 layer. Images a−c are to scale. Image
width (a,b) 60 nm, (c) 120 nm, (inset) 6 × 6 nm2. Reprinted with permission from Stetsovych et al. [55] Copyright 2013,
American Chemical Society.
The exposed works show that ultrathin cerium oxide films offer new opportunities in view of
obtaining modifications of the material properties, and in particular for the stabilization of
metastable structural phases. This last issue has been investigated also in a rather recent theoretical
investigation aimed at identifying the most stable Ce2O3 structures at the ultrathin limit [56]. The
work, using simulated mechanical annealing searches and DFT calculations, shows that different
ultrathin film structures are stable at different values of the in plane lattice parameter and it
identifies in particular a new structure, which does not correspond to any known bulk crystalline
polymorph, which is more stable than the A-type and than the c-type Ce2O3 structures at specific,
relatively large, in-plane lattice parameters (figure 5) [56].
9
Fig.5: Results of DFT calculations for 4 ML Ce2O3 films with different structure: A-type (circles), bixbyite (squares)
NF1 (triangles), NF2–4 (diamonds). Energies (relative to that of the optimized bixbyite film) and lattice parameters are
given per Ce2O3 unit. Vertical dotted lines indicate lattice parameters of possible supports for film growth (multiplied
by 3/2 for transition metals – highlighted by italics). Kozlov et al. [56] - Published by The Royal Society of Chemistry.
As mentioned at the beginning of this section, a few studies have also been focused on the
stabilization of ultrathin cerium oxide films exposing surfaces different from the (111) on metal
substrates. For example, the stabilization of cerium oxide nanoislands exposing (100) facets, with
thickness down to 1 ML, has been shown to be possible using a Cu(111) surface and highly
oxidizing growth conditions [57]. The stabilization mechanism invoked is the formation of a copper
oxide with a rectangular unit cell at the interface, which imposes an epitaxial constraint and shares
an O layer with cerium oxide to compensate polarity (figure 6) [57]. On the same substrate the
presence of an interfacial CeO2(100) layer supporting three dimensional CeO2 nanoislands with
(100) surface orientation has also been observed to coexist with the (111) CeO2 orientation [58].
10
Fig. 6: High resolution STM images of CeO2(100) phases prepared by Ce deposition in 5×10-7 mbar of O2 at 650 K on a
copper oxide surface, formed by oxidizing Cu(111) with NO2 at 600 K. The large-scale images with size 40×40 nm2 is
shown in the derivative mode. Reprinted with permission from Yang et al. [57] Copyright 2011 American Chemical
Society.
A recent study pointed out the formation of ceria nanocrystallites exposing (100) facets after
thermal treatments in vacuum of continuous (111) cerium oxide films grown on Ru(0001) [59]. The
nanostructures, of nanometric height and several tens of nm lateral size, compensate their polarity
through surface reconstructions similar to those hypothesized for the bulk (100) orientation [11].
Open aspects which may contribute to a better understanding of less stable surfaces, also in view of
preparing two-dimensional films with less stable orientation, are the understanding of the
mechanisms which preferentially stabilize this phase under highly reducing conditions compared to
other phases, and the possible reasons for the existence of the (001) orientation only at relatively
large heights of a few nm. Although studies of cerium oxide films exposing less stable surfaces at
the two-dimensional limit are in general rather scarce, they have a great potential interest in view of
the expected smaller surface oxygen vacancy formation energy, due to the lower coordination of
surface O atoms. Furthermore, the possible formation of new structural phases driven by the
compensation of polarity at reduced dimensionality may also open up new perspectives, in analogy
with the case of non-reducible oxides [60].
3. Titanium oxide two-dimensional films
Titanium oxide is another case study for reducible oxides. Unlike cerium ions which are stable only
in the 3+ and 4+ oxidation states, titanium ions can have different oxidation states, the most
common ones being Ti2+, Ti3+, Ti4 , giving rise to various oxides. The most stable oxide is TiO2,
11
also known as titania, which can present different structural phases, the most important ones being
rutile, anatase and brookite.
Ultrathin titanium oxide films grown on metal substrates have been the subject of extensive
research over the last twenty years, as they can give rise to novel nanostructures and to a rich
variety of phases which have been studied using surface science techniques. Stoichiometric and
non-stoichiometric phases with different atomic structures can be obtained through the growth on
single crystal metal surfaces, using simple experimental procedures: deposition of Ti in a vacuum
chamber in controlled pressure of residual oxygen, i.e. reactive deposition, post-oxidation of Ti
metal films, reduction in vacuum by heating the films in UHV or in controlled residual hydrogen
pressure. A selection of results reported in the literature, with focus on peculiar properties arising at
two-dimensions and in particular on reducibility-related issues, are hereby reported and discussed.
Experiments on titanium oxide ultrathin films grown on Cu(100) were carried out by different
groups [61-63]. Maeda et al. [61] deposited metallic Ti on the Cu surface previously exposed to O2
(P= 1 × 10-6 mbar at T=330 K), and oxidized the resulting film by post-growth O2 dosage (P= 1 ×
10-7 mbar at T=623 K). A two-dimensional growth of an O2-/Ti4+/O2- trilayer with TiO2
stoichiometry was deduced from XPS data. LEED patterns showed a hexagonal geometry with two
domains rotated by 90º with respect to each other, and an in-plane lattice constant of 0.29 nm. The
study also indicates that this phase is unstable for coverages beyond a single O2-/Ti4+/O2- trilayer.
The model proposed by Maeda et al. [61] for the atomic geometry of the trilayer is analogous to the
one deduced for the quasi-hexagonal structure formed by titanium oxide on the oxidized (110)
surface of the NiTi alloy [64]. A quasi-hexagonal phase was also observed for titanium oxide films
obtained by reactive deposition of Ti (P= 1 × 10-6 mbar, T=573 K) on Cu(100) previously saturated
with chemisorbed oxygen [63,65]. The chemisorption resulted in a (√2 × 2√2) R45º LEED pattern
with Cu missing-row reconstruction. At low Ti coverage (θTi < 0.5 ML) the same group observed a
LEED pattern with a centred rectangular unit cell indicated as c-(√2 × √2) R45º. STM data showed
that the rectangular phase is associated with flat islands of uniform thickness, embedded within the
outermost layer of the substrate. The stabilization of this phase is probably favoured by the missingrow reconstruction of the O-Cu(100) surface used as substrate [63,65]. The quasi-hexagonal phase
is observed at θTi > 0.5 ML (figure 7), however when the whole substrate surface is covered by the
film (θTi=2.0 ML) a regular hexagonal pattern is visible in the LEED, very similar to the one
observed by Maeda et al. [61] An in-deep X-ray photoelectron diffraction (XPD) and LEED
intensity analysis of the quasi-hexagonal phase, compared with DFT calculations, confirmed the OTi-O trilayer model, allowing to identify also the specific registry of the titanium oxide film with
respect to the Cu(100) surface (figure 7) [65]. Interestingly, DFT provided also information about
the electronic properties of the film, showing that the O 2p and Cu 4sp states overlap and that the
film does not show an insulating behaviour because of an upshift of the O 2p bands and a downshift
of the Ti 4s states, compared to the case of the ideal unsupported films.
12
Fig. 7. a) LEED pattern observed a TiO2 film with the quasi-hexagonal structure on the Cu(001) surface. b) Atomistic
model of the quasi-hexagonal phase of the O-Ti-O trilayer, with a rectangular p(2×7) unit cell, showing the registry of
the film with the substrate. Reprinted with permission from Atrei et al. [65] Copyright 2009 American Chemical
Society.
A TiO2 lepidocrocite-like structure was instead obtained at submonolayer coverage when the
substrate used for the growth was Ag(100) [66]. Again, LEED, XPD and STM experiments results
were compared with the results of DFT calculations. At coverages higher than 1 ML, islands with
the rutile (110) surface begin to form [67]. Single domain titania nano-sheets with lepidocrocite
structure were also observed when TiO2 films were grown on the (1×2) Pt(110) surface [68].
Rutile TiO2 (110) films were also observed to grow on W(100) [69,70] and O(2×1)-W(100) [71] in
two orthogonal domains along the W[010] and [001] directions. In the first case, the TiO2 films at
coverage values between 5 and 30 ML reverted to the bulk structure, and a systematic splitting of
the spots in the LEED pattern was attributed to the formation of a stepped surface, probably caused
by a strain relaxation mechanism due to the compression of the long axis of the TiO2 unit cell. In
the case of oxygen-reconstructed W(100) surface, STM images showed the formation of titanium
oxide islands at low coverage following the orientation of the original missing rows induced by the
O(2×1) reconstruction (figure 8). At increasing coverage, the islands extend and become higher, but
TiO2 does not form a continuous film. The presence of point defects on the rutile islands,
characteristic of the native rutile TiO2(110)-(1×1) surface was observed in high resolution STM
13
images. TiO2 was also grown by reactive deposition and annealing on Mo(100) [72,73]. At
coverage values below 10 nm, LEED showed a (2√2 × √2) R45º pattern, while STM images showed
ordered atomic rows along the [010] and [001] substrate direction. Annealing procedures allowed
reduction of the film, with the Ti ions oxidation state changing from Ti4+ only, to a mixture of Ti4+,
Ti3+ and Ti2+ valence states, as evidenced by XPS. Similar studies were carried out also on
Mo(110) [74,75], on which films of TiO2 with (100) orientation or Ti2O3(0001) were obtained,
depending on the preparation method. In particular the Ti2O3(0001) phase was stabilized by initial
deposition of a metallic Ti layer on the Mo(110) surface, followed by reactive Ti growth in O2 in
the same conditions which lead to the formation of the TiO2(100) phase [74]. STM images
evidenced smooth surfaces with distinct flat terraces and well-defined step edges after annealing in
O2 at T = 900-1100 K, while XPS data showed that the oxide films remained partially reduced after
the treatment. In contrast, thin films annealed in oxygen at T=1200 K were fully oxidized but they
exhibited a three-dimensional rough surface morphology [75].
Fig. 8. STM image of a 0.2 ML film of TiOx(110) on W(100)-O(2×1). An area of TiOx(110) and an area of the W(100)O(2×1) substrate are indicated. White lines are drawn over some of the W(100)-O(2×1) rows then duplicated, in black,
over the TiOx rows to highlight the similar periodicity. The crystal directions correspond to those of the W(100)
substrate. Reprinted with permission from Pang et al. [71] Copyright 2013 American Chemical Society.
The same group also reported formation of a well-ordered (8×2) TiOx film on Mo(112), obtained by
depositing Ti on SiO2(ML)/Mo(112) followed by oxidation/annealing and a final anneal at 1400 K
to completely remove residual Si [76]. This phase shows peculiar properties, including a strong
interaction with Au, which allowed a complete wetting of its surface, at variance with other oxide
supports [71,76].
14
Fig. 9. Summary of the TiO2 phases grown on Ni(110). Reprinted with permission from Papageorgiou et al. [78]
Copyright 2007 American Chemical Society.
Using a Ni(110) substrate for titanium oxide ultrathin films growth either a quasi-hexagonal phase
or (110) rutile rods on a TiO2 wetting layer could be obtained, depending on the initial titanium
coverage [78,79]. In both cases, XPS and x-ray absorption near edge spectroscopy (XANES) results
demonstrated that Ti was fully oxidized. Figure 9 gives a summary of the TiO2 phases reported in
ref. 76. The rutile and wetting layer phases showed a similar behaviour with respect to exposure to
water at P=10-8 mbar, which resulted in coadsorption of both molecular water and hydroxyl groups.
By dosing water at P=10-6 mbar hydroxylation of the thin films was observed. The process was
found to be reversible by annealing at T=490 K. STM images on rutile rods revealed a (1×2)
reconstruction in some areas, ascribed to a reduced phase, very close to the one observed on the
surface of bulk rutile TiO2(110) single crystals [79]. The (1×2) islands disappear completely after
annealing at T=773 K in O2 [79]. Further reduction of the surface could be induced by annealing in
UHV at T=1110 K, with observation of {132} and {121} families of crystallographic shear planes.
In this respect the behaviour of ultrathin films is analogous to the one of the TiO2 (110) single
crystal surface [80].
The study of titanium oxide films on Pt surfaces is particular relevant, in view of a deeper
understanding of the phases formed in oxide-supported metal catalysts as a consequence of SMSI
[8]. Experiments of Pt nanoparticles grown on TiO2(110) surface revealed that the (111) oriented Pt
islands were encapsulated within TiO2 regular films after annealing at T=1100 K with a subsequent
change of the catalytic activity [81]. The encapsulation was observed and studied in detail also in
15
the case of Pd on (1x2) TiO2(110) [82]. Some of the structures obtained in these systems were also
observed on the reverse catalyst model TiOx/Pt systems.
Matsumoto et al. used a number of experimental techniques for a complete study of titanium oxide
grown on Pt(100) [77,83]. In particular, two different routes for film synthesis were used: in the
first one a Pt3Ti surface alloy was oxidized with O3 and annealed at T=1000K. The result was a
smooth film composed of one layer of Ti2O3 with (3×5) superstructure, with flat terraces without
islands, and with a structure similar to the (1×2) strands formed on the reduced TiO2(110) surface.
The second route was reactive deposition of Ti in O2 (p=6.7 × 10-7 mbar) followed by annealing
above 750 K in vacuum. In this case, a (3×5) structure could be observed for coverages lower than
1 ML, while a (4 × 3√5)R60º structure was obtained after deposition of 2 ML. The proposed model
for the observed reconstruction consists of TiO2 tetragonal nets with some O atoms in the second
layer. The (3×5) structure was re-obtained after annealing above 950 K in vacuum. At increasing
coverage and after annealing a T=1000 K, TiO2 clusters were observed, coexisting with Ti2O3 (3×5)
and clean Pt(100) domains. Finally, after further annealing at T=1300 K the TiO2 clusters
decomposed to form a (2√2 × 2√2)R45º structure, proposed to be Ti5O8, and (3 × 5) domains.
Boffa et al. investigated titanium oxide films grown on Pt(111) for coverages ranging from 1 to 5
ML [84]. Two different structures were obtained depending on the sample pretreatment. In
particular, annealing at temperatures between 770 and 970 K in O2 gave a (√(3 × √43)R7.6º three
fold symmetric structure, with TiO2 stoichiometry and primitive cell 18.2 Å × 18.2 Å size.
Annealing in vacuum (820 -1120 K) gave a second ordered overlayer with primitive cell of 18.2 Å
× 13.9 Å size and Ti4O7 stoichiometry. Partial dissolution of Ti in Pt was observed after annealing
at high temperatures, reducing the thickness of the films to approximately 1 ML [84].
A systematic work on titanium oxide films grown on Pt(111) was carried out during the last years
[85-90]. The study first concentrated in ultrathin layers, with an amount of Ti lower than 1.2 MLE,
where 1 MLE corresponds to the surface density of Pt atoms per unit area on the Pt(111) surface.
By varying the Ti dose and the annealing conditions (temperature and oxygen pressure) six
metastable phases with different long-range order were originally found [85]. The conditions
leading to the different phases are summarized in figure 10.
16
Fig.10: a) Summary of the different ultrathin titanium oxide phases obtained on Pt(111) in different conditions. k
stands for kagomè, z for zig-zag, rect for rectangular and w for wagon-wheels. b) Summary of the structural and
chemical transitions observed to occur under different thermal treatment conditions. The prime symbol indicates a
reduced phase. Reprinted with permission from Sedona et al. [85] Copyright 2005 American Chemical Society.
The different phases are identified by the geometry of the surface structures observed in atomically
resolved STM images. Some of the observed structures include reduced titanium ions. Upon
thermal treatments the films could undergo some structural transformations while in O2-rich and
O2-poor conditions reversible film oxidation and reduction were observed (figure 10). All the
identified phases, with the exception of the stoichiometric TiO2 phase with rectangular surface
symmetry, correspond to a film composed of a Ti-O bilayer, with interfacial Ti. The TiOx phases
wet the Pt substrate, and present long-range order with complex superstructures, observed in LEED
patterns. The combination of LEED, XPD and STM data with DFT calculations allowed to identify
the atomic geometry of all the studied phases [88]. For instance, the z-phase has an
incommensurate unit cell with 6.8 Å × 8.6 Å size, and a Ti6O8 geometry, while the z’-phase has a
commensurate unit cell of 16.6 Å × 14.4 Å size (with Ti24O30), geometry (figure 11). Ti 2p and O1s
core level XPS and valence band photoemission studies [86] allowed also to separate the observed
phases in two main groups: a group of three stoichiometric films (k’, rect and rect’) and a group of
sub-stoichiometric films (z, z’ and w). The valence band photoemission data also revealed some
peculiar aspects of the electronic properties, like for instance the mixing of the Ti-Pt states near the
Fermi level in the sub-stoichiometric films, a clear indication of the presence of the Pt-Ti interface
in this group.
17
Fig. 11 a) atomically resolved image of the z- and z’-TiOx phase on Pt(111). b) simulated STM image and structure of
the z’ phase, as obtained by DFT calculations. Reprinted with permission from Barcaro et al. [88] Copyright 2009
American Chemical Society.
Another interesting phase is the stoichiometric TiO2 rect phase, showing an incommensurate
rectangular unit cell of size 3.8 Å × 3.0 Å. STM revealed also that the apparent height of the rectTiO2 islands can be only explained by a multilayer sequence, with oxygen atoms at the interface
with Pt occupying only top and bridge sites, an arrangement similar to the one proposed for rect VO2 [89]. Recently, other reduced hexagonal phases h(6x6) and h(9x9) were discovered by
annealing z’ phases at high temperature [90].
4. Two-dimensional films of other reducible oxides
4.1 Transition metal oxide two-dimensional films
Indeed, most transition metals, due to their specific electronic structure, can have different oxidation
states and can form oxides which are reducible. Within the wide number of studies done in the field
of two-dimensional reducible transition metal oxide films, we will here focus on those reporting
aspects strictly related to reducibility and we will limit to the strictly two-dimensional case, i.e. to
films with thickness of the order or below 1 nm.
A very interesting example of a study of this kind is represented by the work by Li et al. on
manganese oxide two-dimensional films [91]. By STM, LEED and XPS analyses different low
dimensionality phases were identified in films of approximately 1 ML thickness as a function of the
18
oxygen chemical potential during preparation on a Pd(100) substrate. Figure 12 summarizes the
obtained results. Among the observed structures, some represent the two-dimensional limit of bulk
truncated structures, like the hexagonal structure observed under mildly oxidizing condition, which
recalls the MnO(111) structure. Most of them on the contrary do not resemble any known
manganese oxide phase and they are often found in coexistence on the Pd surface [92].
Fig. 12: Schematic phase diagram of the two-dimensional Mn oxides, presented as a function of the
oxygen pressure p(O2) and of the oxygen chemical potential µO. The nominal coverage of Mn on
Pd(100) is 0.75 ML. © IOP Publishing. Reproduced with permission from Li et al. [91]. All rights
reserved.
Also vanadium cations in vanadium oxide can have different oxidation states (from 2+ to 5+),
giving a wide variety of phases. The most important ones are V2O5, VO2, V2O3 and VO, each one
having a different crystal structure and showing a variety of physical and chemical properties [92].
Furthermore, mixed valence oxides can also be formed, with the cations in two possible oxide
states, like for example V5+ and V4+ in V6O13. Of peculiar interest is the crystal structure of V2O5
(layered orthorhombic), which is essentially composed of zigzag double chains of square VO5
pyramidal units sharing edges and running along the b direction. The VO5 pyramid has the
vanadium atom in the centre, four oxygen atoms in the basal plane and a vanadyl-type oxygen atom
at the apex (see also inset of figure 14d). The layers of chains are stacked along the [0 0 1] (c)
direction. The resulting solid is composed of distorted tetrahedrally coordinated VO6 unit. The VO6
unit is also present in the tetrahedral rutile and monoclinic phases of VO2 which are stable below
and above T=340 K respectively, while the V2O3 phase has a corundum structure above 160 K. All
these structures can be found in ultrathin films grown either on oxides or on metals, but other
phases can also be stabilised by the interaction with the substrate [89,92].
A systematic work was carried out on vanadium oxide nanostructures and films deposited on
Rh(111) by Netzer and co-workers. The experimental results were supported by ab initio DFT
calculations in order to obtain complete information the structure and on the thermodynamics of the
different phases obtained [89,93-96]. A phase diagram of the vanadium oxide nanostructures on
Rh(111) is reported in figure 13. Concentrating on the two-dimensional phases, it was found that (√7
× √7)R 19.1º or (√13 × √13)R 13.8º structures can be formed under highly oxidative conditions,
with PO2= 2 × 10-7 mbar and by keeping the substrate temperature at 670 K.
19
Fig. 13: Phase diagram of vanadium oxide nanostructures on Rh(111) as a function of vanadium coverage
and of substrate temperature. © IOP Publishing. Reproduced with permission from J. Schoiswohl et al.
[89]. All rights reserved.
As obtained by careful analysis of STM images and by DFT calculations, the (√7 × √7)R 19.1º nanolayer
consists on a V3O9 oxide phase with VO5 square pyramids as building blocks. The same holds for the (√13 ×
√13)R 13.8º phase, which has a V6O18 stoichiometry.
By exposing the film to reducing conditions (i. e. to annealing in UHV or in hydrogen atmosphere) reduced
two-dimensional phases can be obtained with V11O23, V13O21, V2O3 and VO stoichiometry. The reduction
process occurs essentially by removing the vanadyl groups.
Zero-dimensional structures could also be obtained by evaporation of small quantities (less than 0.2 ML) of
vanadium on O(2×1)-Rh(111) and by flashing at 250ºC in UHV. The obtained structures are identical planar
star-shaped V6O12 molecules [96]. It was shown also that under high substrate temperature the clusters can
diffuse and assemble in 2-D overlayers with well determined phases, like (5×5) or the (5×3√3)-rect. The
different phases are obtained either in oxidizing or in reducing conditions [89].
20
Fig. 14: a) Large scale (100 x 100 nm2) and b) high resolution STM images of the (√7 X √7)R 19.1º
vanadium oxide nanolayer on Rh(111). Inset simulated STM image obtained by DFT calculations. c) Top
view and d) side view of the structural model obtained by DFT. Inset: the VO5 square pyramid constituting
the building block of the film. © IOP Publishing. Reproduced with permission from J. Schoiswohl et al.
[89]. All rights reserved.
Another interesting example is the growth of V2Oy (y≈5) and V2O5 on Au(111) reported by the Freund group
[97,98]. The films were obtained by physical vapour deposition of vanadium, subsequent oxidation under 50
mbar of oxygen at 670 K and annealing in UHV at 470 K. The high oxygen pressure used allowed to obtain
a high oxidation state (V5+) even for films thicker than 1 MLE (1 MLE corresponding to the same number of
V atoms as one layer of Au(111)), at variance with previous studies, where the standard reactive deposition
and post-oxidation procedures gave V2+, V3+ or V4+ states (see for example ref. 89,93,99-101). At low
coverage, from 0.26 to 1.04 MLE, two different coincidence monolayer structures could be observed by
STM and LEED, which are determined by the interaction with the Au(111) substrate. The first structure has
a rectangular unit cell, with size 3.6 Å × 15 Å, while the second one (observed at 0.52 MLE of V/Au(111)
which correspond to one full layer of oxide film), has a 3.6 Å × 10.8 Å (α = 60º) oblique unit cell. In
analogy with the case of vanadium oxide monolayers on Rh(111) [89,93-96], it was supposed that the full
monolayer film consisted of VO5 pyramid building blocks sharing corners and edges at their bases, and on
the basis of the XPS and XANES measurements it was also shown that the oxidation state in the V cations is
close to V5+, characteristics of V2O5. Increasing the coverage again to 1.05 MLE, a different structure was
observed, which was found to be similar to V6O13 with (001) orientation, while for higher coverage V2O5
islands were formed, extending in size and giving rise eventually to V2O5 (001) films containing a low
number of point defects. These films were composed of large (20 nm size) single crystal domains with some
azimuthal disorder (figure 15) [97].
21
Fig. 15: STM images of a) 100 X 100 nm2, b) 44 X 20 nm2 of a film formed by the oxidation of 1.56 MLE V/Au(111).
The images show the presence of V2O5(100) islands growing on V6O13(001) film. Reprinted with permission from S.
Guimond et al. [97] Copyright 2008 American Chemical Society.
4.2 Rare earth oxide two-dimensional films
In analogy with cerium oxide also in other rare earth oxides (REOs), like praseodymium, terbium
and samarium oxide the cations can have different oxidation states, and the oxides can be defined as
reducible. For samarium oxide the most stable oxidation state is Sm2O3, which can form different
structural phases. Samarium oxide is the only REO which can also form a monoxide, SmO, with a
rock-salt structure, stable under reducing conditions. The most stable valence state for terbium is
3+, leading to the Tb2O3 phase, however also the dioxide TbO2, as well as intermediate phases like
Tb4O7, can form. Praseodymium can have the 2+, 3+ and 4+ oxidation states and the most stable
oxides are Pr2O3, PrO2 and Pr6O11. The sesquioxides are typically stable in the c-type bixbyite
structure, at variance with cerium oxide, for which the hexagonal A-type phase is favoured in the
Ce2O3 stoichiometry.
Some studies report the growth of REO in the form of ultrathin films on metallic substrates and
point out interesting aspects related to those observed in cerium oxide ultrathin films. Temperature
programmed desorption (TPD) allows to monitor oxygen release in these REO films, while in the
case of cerium oxide the release is considered to be too fast to be monitored [102,103].
Samarium oxide in the form of an ultrathin film has been shown to form by controlled oxidation of
surface SmRh and SmRu surface alloys, in turn obtained by heating metallic Sm films deposited on
Rh(100) and Ru(0001) surfaces respectively, leading to a SmOx phase [102-104]. On both
substrates specific sites for CO absorption have been shown to form at the perimeter of the
samarium oxide ultrathin islands, due to the interaction with the underlying metal [102,103]. A
22
more recent study by Jhang et al. investigated samarium oxide films grown on a Pt(111) substrate
by reactive deposition at 600 K followed by annealing in O2 (P~10-7 mbar) at 1000 K [105]. In
close analogy with cerium oxide films grown in similar conditions, the LEED pattern for
submonolayer coverage shows a (1.37 × 1.37) structure. Interestingly, for coverages between 1 and
3 ML the LEED pattern shows additional faint spots in the so called quasi-(3×3) superstructure,
very similar to the one observed in reduced ultrathin cerium oxide phases on Pt(111) (figure 16)
[17,29]. The authors ascribe this features to a Sm2O3(111) phase with a defective fluorite structure,
in which the oxygen vacancies are randomly distributed within the crystal, and to the formation of a
8:11 coincidence with the underlying Pt. A similar origin may be invoked also for the LEED pattern
of reduced cerium oxide films, although in the latter case a similar coincidence may be expected
also when the films are oxidized, and the reasons for its appearance only when the films are reduced
are not clear. Reduction by thermal treatments in UHV at 1000 K for 30 min of the Sm2O3 films
lead to the coexistence of Sm2O3(111) phase and of a SmO(100) phase, giving a superstructure
rotated by 30° with respect to the (111) spots in the LEED pattern [105]. The reversibility of the
structural and morphological modifications is observed also in this case [105].
Fig.16: LEED patterns obtained after growing Sm2O3(111) films on Pt(111) to the coverages indicated followed by
annealing in 7 × 10−7 mbar of O2 at 1000 K for 10 min: (a) 0.4 ML of Sm2O3 shows a hexagonal (1.37 × 1.37)
superstructure in registry with the Pt(111) (1 × 1) spots, E = 58 eV, (b) 1.1 ML, E = 48 eV, (c) 1.7 ML, E = 56 eV, (d)
3.0 ML, E = 52 eV exhibit a quasi-(3 × 3) superstructure, (e) 4.2 ML, E = 56 eV; the quasi-(3 × 3) becoming blurry, and
(f) shows a schematic representation of the quasi-(3 × 3) pattern that is determined from FFT analysis of a structural
model of superposed, hexagonal Sm and Pt lattices that form a hexagonal (8 × 8) coincidence lattice with respect to the
Sm lattice, i.e. (11 × 11) with respect to the Pt lattice. Reprinted with permission from Jhang et al. [105] Copyright 2013
American Chemical Society.
Terbium oxide films were prepared with similar methods, i. e. using reactive deposition and post
growth annealing in O2 [106]. The LEED pattern up to 4 ML thickness shows the (1.32 × 1.32)
structure, compatible with the terbium oxide lattice parameter of the Tb2O3 phase in a fluorite
structure with disordered oxygen vacancies in close analogy with the case of samarium oxide [107].
In this case however diffraction from possible coincidence superlattices could not be observed. The
23
films could not be oxidized by using thermal treatments in O2 at P ~ 7×10-7 mbar and temperatures
from 95 to 1000 K, while plasma-generated atomic oxygen was shown to be effective in oxidizing a
film at 300 K to TbO2 and in generating weakly bound surface O species. The reduced Tb2O3 phase
could be re-obtained by annealing TbO2 in vacuum at 1000 K [106].
Praseodymium oxide has been studied only in the form of relatively thick films on Si(111)
substrates [107-109]. As for terbium oxide thin films, also in this case, an oxidizing treatment in
oxygen plasma was shown to be necessary to oxidize the Pr2O3 phase to the PrO2 fluorite phase
[107]. UHV thermal treatments to reduce the PrO2 phase have been shown to give origin to a phase
including a mixture of Pr6O11 and Pr5O9, the latter being unstable in the bulk phase [109].
5. Conclusions
The studies discussed in this chapter show that reducible oxides at two dimensionality may show
important modifications of their properties, and indeed also of the reversible reduction and
oxidation processes. Ultrathin films, typically supported on metal substrates, at different degree of
reduction show new structural and chemical phases, often unstable in the bulk, which have been
deeply characterized by surface science techniques, combined with the extremely important support
of theoretical modeling. In some cases the different phases can be reversibly transformed one into
the other by reducing and oxidizing treatments. The studies point out interesting challenging
aspects, on which future studies might be focused, such as for example the stabilization of phases
with less stable surface orientation, or more complex systems such as ternary two-dimensional
compounds or mixed oxide phases. We believe that investigations following these lines may open
up unexpectedly interesting aspects within the field of reducible oxide based materials.
Acknowledgements
The authors gratefully acknowledge the support by the Italian MIUR through the FIRB Project
RBAP115AYN ‘‘Oxides at the nanoscale: multifunctionality and applications’’ and by the COST
Action CM1104 “Reducible oxide chemistry, structure and functions”.
References
1. Trovarelli A, Fornasiero P (2013) Catalysis by Ceria and Related Materials. Catalytic Science, 2nd edn.
Imperial College Pr., London
2. Celardo I, Pedersen JZ, Traversa E, Ghibelli L (2011) Pharmacological potential of cerium oxide
nanoparticles. Nanoscale 3 (4):1411-1420
3. Waser R, Dittmann R, Staikov G, Szot K (2009) Redox-Based Resistive Switching Memories - Nanoionic
Mechanisms, Prospects, and Challenges. Adv Mater 21 (25-26):2632
24
4. Migani A, Vayssilov GN, Bromley ST, Illas F, Neyman KM (2010) Dramatic reduction of the oxygen
vacancy formation energy in ceria particles: a possible key to their remarkable reactivity at the nanoscale. J
Mater Chem 20 (46):10535-10546
5. Cafun JD, Kvashnina KO, Casals E, Puntes VF, Glatzel P (2013) Absence of Ce3+ Sites in Chemically
Active Colloidal Ceria Nanoparticles. Acs Nano 7 (12):10726-10732
6. Carrettin S, Concepcion P, Corma A, Nieto JML, Puntes VF (2004) Nanocrystalline CeO2 increases the
activity of an for CO oxidation by two orders of magnitude. Angew Chem Int Edit 43 (19):2538-2540
7. Vayssilov GN, Lykhach Y, Migani A, Staudt T, Petrova GP, Tsud N, Skala T, Bruix A, Illas F, Prince KC,
Matolin V, Neyman KM, Libuda J (2011) Support nanostructure boosts oxygen transfer to catalytically
active platinum nanoparticles. Nat Mater 10 (4):310-315
8. Tauster SJ, Fung SC, Garten RL (1978) Strong Metal-Support Interactions - Group-8 Noble-Metals
Supported on Tio2. J Am Chem Soc 100 (1):170-175
9. Bevan DJM (1955) Ordered intermediate phases in the system CeO2-Ce2O3. J Inorg Nucl Chem 1 :49
10. Skorodumova NV, Simak SI, Lundqvist BI, Abrikosov IA, Johansson B (2002) Quantum origin of the
oxygen storage capability of ceria. Phys Rev Lett 89 (16):166601
11. Conesa JC (1995) Computer Modeling of Surfaces and Defects on Cerium Dioxide. Surf Sci 339
(3):337-352
12. Hardacre C, Ormerod RM, Lambert RM (1994) Platinum-Promoted Catalysis by Ceria - a Study of
Carbon-Monoxide Oxidation over Pt(111)/CeO2. J Phys Chem-Us 98 (42):10901-10905
13. Schierbaum KD (1998) Ordered ultra-thin cerium oxide overlayers on Pt(111) single crystal surfaces
studied by LEED and XPS. Surf Sci 399 (1):29-38-X
14. Mullins DR, Radulovic PV, Overbury SH (1999) Ordered cerium oxide thin films grown on Ru(0001)
and Ni(111). Surf Sci 429 (1-3):186-198
15. Eck S, Castellarin-Cudia C, Surnev S, Ramsey MG, Netzer FP (2002) Growth and thermal properties of
ultrathin cerium oxide layers on Rh(111). Surf Sci 520 (3):173-185
16. Wilson EL, Chen Q, Brown WA, Thornton G (2007) CO adsorption on the model catalyst Pd/CeO2x(111)/Rh(111). J Phys Chem C 111 (38):14215-14222
17. Luches P, Pagliuca F, Valeri S (2011) Morphology, Stoichiometry, and Interface Structure of CeO2
Ultrathin Films on Pt(111). J Phys Chem C 115 (21):10718-10726
18. Staudt T, Lykhach Y, Hammer L, Schneider MA, Matolin V, Libuda J (2009) A route to continuous
ultra-thin cerium oxide films on Cu(111). Surf Sci 603 (23):3382-3388
19. Xiao WD, Guo QL, Wang EG (2003) Transformation of CeO2(111) to Ce2O3(0001) films. Chem Phys
Lett 368 (5-6):527-531
20. Lu JL, Gao HJ, Shaikhutdinov S, Freund HJ (2006) Morphology and defect structure of the CeO2(111)
films grown on Ru(0001) as studied by scanning tunneling microscopy. Surf Sci 600 (22):5004-5010
21. Grinter DC, Ithnin R, Pang CL, Thornton G (2010) Defect Structure of Ultrathin Ceria Films on Pt(111):
Atomic Views from Scanning Tunnelling Microscopy. J Phys Chem C 114 (40):17036-17041
25
22. Flege JI, Kaemena B, Gevers S, Bertram F, Wilkens T, Bruns D, Batjer J, Schmidt T, Wollschlager J,
Falta J (2011) Silicate-free growth of high-quality ultrathin cerium oxide films on Si(111). Phys Rev B 84
(23):235418
23. Pagliuca F, Luches P, Valeri S (2013) Interfacial interaction between cerium oxide and silicon surfaces.
Surf Sci 607:164-169
24. Putna ES, Bunluesin T, Fan XL, Gorte RJ, Vohs JM, Lakis RE, Egami T (1999) Ceria films on zirconia
substrates: models for understanding oxygen-storage properties. Catal Today 50 (2):343-352
25. vanWijck MAAM, Verhoeven MAJ, Reuvekamp EMCM, Gerritsma GJ, Blank DHA, Rogalla H (1996)
CeO2 as insulation layer in high T-c superconducting multilayer and crossover structures. Appl Phys Lett 68
(4):553-555
26. Overbury SH, Huntley DR, Mullins DR, Ailey KS, Radulovic PV (1997) Surface studies of model
supported catalysts: NO adsorption on Rh/CeO2(001). J Vac Sci Technol A 15 (3):1647-1652
27. Hardacre C, Roe GM, Lambert RM (1995) Structure, Composition and Thermal-Properties of Cerium
Oxide-Films on Platinum(111). Surf Sci 326 (1-2):1-10
28. Berner U, Schierbaum KD (2002) Cerium oxides and cerium-platinum surface alloys on Pt(111) singlecrystal surfaces studied by scanning tunneling microscopy. Phys Rev B 65 (23):235404
29. Luches P, Pagliuca F, Valeri S (2014) Structural and morphological modifications of thermally reduced
cerium oxide ultrathin epitaxial films on Pt(111). Phys Chem Chem Phys 16 (35):18848-18857
30. Matolin V, Libra J, Matolinova I, Nehasil V, Sedlacek L, Sutara F (2007) Growth of ultra-thin cerium
oxide layers on Cu(111). Appl Surf Sci 254 (1):153-155
31. Yang F, Graciani J, Evans J, Liu P, Hrbek J, Sanz JF, Rodriguez JA (2011) CO Oxidation on Inverse
CeOx/Cu(111) Catalysts: High Catalytic Activity and Ceria-Promoted Dissociation of O-2. J Am Chem Soc
133 (10):3444-3451
32. Ma SG, Rodriguez J, Hrbek J (2008) STM study of the growth of cerium oxide nanoparticles on
Au(111). Surf Sci 602 (21):3272-3278
33. Škoda M, Libra J, Šutara F, Tsud N, Skála T, Sedláček L, Cháb V, Prince KC, Matolín V (2007) A
resonant photoemission study of the Ce and Ce-oxide/Pd(1 1 1) interfaces. Surf Sci 601 (21):4958-4965
34. Grinter DC, Yim CM, Pang CL, Santos B, Mentes TO, Locatelli A, Thornton G (2013) Oxidation State
Imaging of Ceria Island Growth on Re(0001). J Phys Chem C 117 (32):16509-16514
35. Suchorski Y, Wrobel R, Becker S, Weiss H (2008) CO Oxidation on a CeOx/Pt(111) Inverse Model
Catalyst Surface: Catalytic Promotion and Tuning of Kinetic Phase Diagrams. J Phys Chem C 112
(50):20012-20017
36. Eck S, Castellarin-Cudia C, Surnev S, Prince KC, Ramsey MG, Netzer FP (2003) Adsorption and
reaction of CO on a ceria–Rh(1 1 1) “inverse model catalyst” surface. Surf Sci 536 (1–3):166-176
37. Rodriguez JA, Graciani J, Evans J, Park JB, Yang F, Stacchiola D, Senanayake SD, Ma S, Pérez M, Liu
P, Sanz JF, Hrbek J (2009) Water-Gas Shift Reaction on a Highly Active Inverse CeOx/Cu(111) Catalyst:
Unique Role of Ceria Nanoparticles. Angewandte Chemie International Edition 48 (43):8047-8050
26
38. Dvorak F, Stetsovych O, Steger M, Cherradi E, Matolinova I, Tsud N, Skoda M, Skala T, Myslivecek J,
Matolin V (2011) Adjusting Morphology and Surface Reduction of CeO2(111) Thin Films on Cu(111). J
Phys Chem C 115 (15):7496-7503
39. Luches P, Pagliuca F, Valeri S, Boscherini F (2013) Structure of Ultrathin CeO2 Films on Pt(111) by
Polarization-Dependent X-ray Absorption Fine Structure. J Phys Chem C 117 (2):1030-1036
40. Szabova L, Stetsovych O, Dvorak F, Camellone MF, Fabris S, Myslivecek J, Matolin V (2012) Distinct
Physicochemical Properties of the First Ceria Monolayer on Cu(111). J Phys Chem C 116 (11):6677-6684
41. Spiel C, Blaha P, Suchorski Y, Schwarz K, Rupprechter G (2011) CeO2/Pt(111) interface studied using
first-principles density functional theory calculations. Phys Rev B 84 (4):045412
42. Nilius N, Kozlov SM, Jerratsch JF, Baron M, Shao X, Vines F, Shaikhutdinov S, Neyman KM, Freund
HJ (2012) Formation of One-Dimensional Electronic States along the Step Edges of CeO2(111). Acs Nano 6
(2):1126-1133
43. Torbrugge S, Cranney M, Reichling M (2008) Morphology of step structures on CeO(2)(111). Appl Phys
Lett 93 (7):073112
44. Jerratsch JF, Shao X, Nilius N, Freund HJ, Popa C, Ganduglia-Pirovano MV, Burow AM, Sauer J (2011)
Electron Localization in Defective Ceria Films: A Study with Scanning-Tunneling Microscopy and DensityFunctional Theory. Phys Rev Lett 106 (24):246801
45. Esch F, Fabris S, Zhou L, Montini T, Africh C, Fornasiero P, Comelli G, Rosei R (2005) Electron
localization determines defect formation on ceria substrates. Science 309 (5735):752-755
46. Ganduglia-Pirovano MV, Da Silva JLF, Sauer J (2009) Density-Functional Calculations of the Structure
of Near-Surface Oxygen Vacancies and Electron Localization on CeO2(111). Phys Rev Lett 102 (2):026101
47. Loschen C, Bromley ST, Neyman KM, Illas F (2007) Understanding ceria nanoparticles from firstprinciples calculations. J Phys Chem C 111 (28):10142-10145
48. Kaemena B, Senanayake SD, Meyer A, Sadowski JT, Falta J, Flege JI (2013) Growth and Morphology of
Ceria on Ruthenium (0001). J Phys Chem C 117 (1):221-232
49. Pfau A, Schierbaum KD (1994) The Electronic-Structure of Stoichiometric and Reduced Ceo2 Surfaces an Xps, Ups and Hreels Study. Surf Sci 321 (1-2):71-80
50. Matolin V, Libra J, Skoda M, Tsud N, Prince KC, Skala T (2009) Methanol adsorption on a
CeO2(111)/Cu(111) thin film model catalyst. Surf Sci 603 (8):1087-1092
51. Senanayake SD, Stacchiola D, Evans J, Estrella M, Barrio L, Perez M, Hrbek J, Rodriguez JA (2010)
Probing the reaction intermediates for the water-gas shift over inverse CeOx/Au(111) catalysts. J Catal 271
(2):392-400
52. Castellarin-Cudia C, Surnev S, Schneider G, Podlucky R, Ramsey MG, Netzer FP (2004) Strain-induced
formation of arrays of catalytically active sites at the metal-oxide interface. Surf Sci 554 (2-3):L120-L126
53. Duchon T, Dvorak F, Aulicka M, Stetsovych V, Vorokhta M, Mazur D, Veltruska K, Skala T,
Myslivecek J, Matolinova I, Matolin V (2014) Ordered Phases of Reduced Ceria As Epitaxial Films on
Cu(111). J Phys Chem C 118 (1):357-365
27
54. Wilkens H, Schuckmann O, Oelke R, Gevers S, Reichling M, Schaefer A, Baumer M, Zoellner MH, Niu
G, Schroeder T, Wollschlager J (2013) Structural transitions of epitaxial ceria films on Si(111). Phys Chem
Chem Phys 15 (42):18589-18599
55. Stetsovych V, Pagliuca F, Dvorak F, Duchon T, Vorokhta M, Aulicka M, Lachnitt J, Schernich S,
Matolinova I, Veltruska K, Skala T, Mazur D, Myslivecek J, Libuda J, Matolin V (2013) Epitaxial Cubic
Ce2O3 Films via Ce-CeO2 Interfacial Reaction. J Phys Chem Lett 4 (6):866-871
56. Kozlov SM, Demiroglu I, Neyman KM, Bromley ST (2015) Reduced ceria nanofilms from structure
prediction. Nanoscale 7 (10):4361-4366
57. Yang F, Choi Y, Agnoli S, Liu P, Stacchiola D, Hrbek J, Rodriguez JA (2011) CeO2 <-> CuOx
Interactions and the Controlled Assembly of CeO2(111) and CeO2(100) Nanoparticles on an Oxidized
Cu(111) Substrate. J Phys Chem C 115 (46):23062-23066
58. Stetsovych O, Beran J, Dvorak F, Masek K, Myslivecek J, Matolin V (2013) Polarity driven morphology
of CeO2(100) islands on Cu(111). Appl Surf Sci 285:766-771
59. Pan Y, Nilius N, Stiehler C, Freund HJ, Goniakowski J, Noguera C (2014) Ceria Nanocrystals Exposing
Wide (100) Facets: Structure and Polarity Compensation. Adv Mater Interfaces 1 (9):1400404
60. Noguera C, Goniakowski J (2013) Polarity in Oxide Nano-objects. Chem Rev 113 (6):4073-4105
61. Maeda T, Kobayashi Y, Kishi K (1999) Growth of ultra-thin titanium oxide on Cu(100), Fe/Cu(100) and
ordered ultra-thin iron oxide studied by low-energy electron diffraction and X-ray photoelectron
spectroscopy. Surf Sci 436 (1-3):249-258
62. Passeggi MCG, Vergara LI, Mendoza SM, Ferron J (2002) Passivation and temperature effects on the
oxidation process of titanium thin films. Surf Sci 507:825-831
63. Finetti P, Caffio M, Cortigiani B, Atrei A, Rovida G (2008) Mechanism of growth and structure of
titanium oxide ultrathin films deposited on Cu(001). Surf Sci 602 (5):1101-1113
64. Atrei A, Bardi U, Rovida G (1997) Structure and composition of the titanium oxide layers formed by
low-pressure oxidation of the Ni94Ti6(110) surface. Surf Sci 391 (1-3):216-225
65. Atrei A, Ferrari AM, Finetti P, Beni A, Rovida G (2009) LEED and DFT Study of the Quasihexagonal
TiO2 Structure on Cu(001). J Phys Chem C 113 (45):19578-19584
66. Atrei A, Ferrari AM, Szieberth D, Cortigiani B, Rovida G (2010) Lepidocrocite-like structure of the
TiO2 monolayer grown on Ag(100). Phys Chem Chem Phys 12 (37):11587-11595
67. Atrei A, Cortigiani B, Ferrari AM (2012) Epitaxial growth of TiO2 films with the rutile (110) structure
on Ag(100). J Phys-Condens Mat 24 (44):445005
68. Orzali T, Casarin M, Granozzi G, Sambi M, Vittadini A (2006) Bottom-up assembly of single-domain
titania nanosheets on (1x2)-Pt(110). Phys Rev Lett 97 (15):156101
69. McCavish ND, Bennett RA (2003) Ultra-thin film growth of titanium dioxide on W(100). Surf Sci 546
(1):47-56
70. Bennett RA, Mulley JS, Newton MA, Surman M (2007) Spectroscopy of ultrathin epitaxial rutile
TiO2(110) films grown on W(100). J Chem Phys 127 (8):084707
28
71. Pang CL, Grinter DC, Matharu J, Thornton G (2013) A Scanning Tunneling Microscopy Study of
Ultrathin Film Rutile TiO2(110) Supported on W(100)-O(2 x 1). J Phys Chem C 117 (48):25622-25627
72. Oh WS, Xu C, Kim DY, Goodman DW (1997) Preparation and characterization of epitaxial titanium
oxide films on Mo(100). J Vac Sci Technol A 15 (3):1710-1716
73. Chen MS, Goodman DW (2008) Ultrathin, ordered oxide films on metal surfaces. J Phys-Condens Mat
20 (26):264013
74. Guo Q, Oh WS, Goodman DW (1999) Titanium oxide films grown on Mo(110). Surf Sci 437 (1-2):4960
75. Lai X, Guo Q, Min BK, Goodman DW (2001) Synthesis and characterization of titania films on
Mo(110). Surf Sci 487 (1-3):1-8
76. Chen MS, Goodman DW (2004) The structure of catalytically active gold on titania. Science 306
(5694):252-255
77. Matsumoto T, Batzill M, Hsieh S, Koel BE (2004) Fundamental studies of titanium oxide-Pt(100)
interfaces - I. Stable high temperature structures formed by annealing TiOx films on Pt(100). Surf Sci 572
(2-3):127-145
78. Papageorgiou AC, Cabailh G, Chen Q, Resta A, Lundgren E, Andersen JN, Thornton G (2007) Growth
and reactivity of titanium oxide ultrathin films on Ni(110). J Phys Chem C 111 (21):7704-7710
79. Papageorgiou AC, Pang CL, Chen Q, Thornton G (2007) Low-dimensional, reduced phases of ultrathin
TiO(2). Acs Nano 1 (5):409-414
80. Bennett RA, Poulston S, Stone P, Bowker M (1999) STM and LEED observations of the surface
structure of TiO2(110) following crystallographic shear plane formation. Phys Rev B 59 (15):10341-10346
81. Dulub O, Hebenstreit W, Diebold U (2000) Imaging cluster surfaces with atomic resolution: The strong
metal-support interaction state of Pt supported on TiO2(110). Phys Rev Lett 84 (16):3646-3649
82. Bennett RA, Pang CL, Perkins N, Smith RD, Morrall P, Kvon RI, Bowker M (2002) Surface structures in
the SMSI state; Pd on (1 x 2) reconstructed TiO2(110). J Phys Chem B 106 (18):4688-4696
83. Matsumoto T, Batzill M, Hsieh S, Koel BE (2004) Fundamental studies of titanium oxide-Pt(100)
interfaces - II. Influence of oxidation and reduction reactions on the surface structure of TiOx films on
Pt(100). Surf Sci 572 (2-3):146-161
84. Boffa AB, Galloway HC, Jacobs PW, Benitez JJ, Batteas JD, Salmeron M, Bell AT, Somorjai GA (1995)
The Growth and Structure of Titanium-Oxide Films on Pt(111) Investigated by Leed, Xps, Iss, and Stm. Surf
Sci 326 (1-2):80-92
85. Sedona F, Rizzi GA, Agnoli S, Xamena FXLI, Papageorgiou A, Ostermann D, Sambi M, Finetti P,
Schierbaum K, Granozzi G (2005) Ultrathin TiOx films on Pt(111): A LEED, XPS, and STM investigation. J
Phys Chem B 109 (51):24411-24426
86. Sedona F, Agnoli S, Granozzi G (2006) Ultrathin wagon-wheel-like TiOx phases on Pt(111): A
combined low-energy electron diffraction and scanning tunneling microscopy investigation. J Phys Chem B
110 (31):15359-15367
29
87. Finetti P, Sedona F, Rizzi GA, Mick U, Sutara F, Svec M, Matolin V, Schierbaum K, Granozzi G (2007)
Core and valence band photoemission Spectroscopy of well-ordered ultrathin TiOx films on Pt(111). J Phys
Chem C 111 (2):869-876
88. Barcaro G, Agnoli S, Sedona F, Rizzi GA, Fortunelli A, Granozzi G (2009) Structure of Reduced
Ultrathin TiOx Polar Films on Pt(111). J Phys Chem C 113 (14):5721-5729
89. Schoiswohl J, Surnev S, Netzer FP, Kresse G (2006) Vanadium oxide nanostructures: from zero- to
three-dimensional. J Phys-Condens Mat 18 (4):R1-R14
90. Barcaro G, Cavaliere E, Artiglia L, Sementa L, Gavioli L, Granozzi G, Fortunelli A (2012) Building
Principles and Structural Motifs in TiOx Ultrathin Films on a (111) Substrate. J Phys Chem C 116
(24):13302-13306
91. Li F, Parteder G, Allegretti F, Franchini C, Podloucky R, Surnev S, Netzer FP (2009) Two-dimensional
manganese oxide nanolayers on Pd(100): the surface phase diagram. J Phys-Condens Mat 21 (13):134008
92. Surnev S, Ramsey MG, Netzer FP (2003) Vanadium oxide surface studies. Prog Surf Sci 73 (4-8):117165
93. Surnev S, Kresse G, Ramsey MG, Netzer FP (2001) Novel interface-mediated metastable oxide phases:
Vanadium oxides on Pd(111). Phys Rev Lett 87 (8):086102
94. Schoiswohl J, Sock M, Eck S, Surnev S, Ramsey MG, Netzer FP, Kresse G (2004) Atomic-level growth
study of vanadium oxide nanostructures on Rh(111). Phys Rev B 69 (15):155403
95. Schoiswohl J, Surnev S, Sock M, Eck S, Ramsey MG, Netzer FP, Kresse G (2005) Reduction of
vanadium-oxide monolayer structures. Phys Rev B 71 (16):165437
96. Schoiswohl J, Kresse G, Surnev S, Sock M, Ramsey MG, Netzer FP (2004) Planar vanadium oxide
clusters: Two-dimensional evaporation and diffusion on Rh(111). Phys Rev Lett 92 (20):206103
97. Guimond S, Gobke D, Romanyshyn Y, Sturm JM, Naschitzki M, Kuhlenbeck H, Freund HJ (2008)
Growth and characterization of ultrathin V(2)O(y) (y approximate to 5) films on Au(111). J Phys Chem C
112 (32):12363-12373
98. Guimond S, Sturm JM, Gobke D, Romanyshyn Y, Naschitzki M, Kuhlenbeck H, Freund HJ (2008) Wellordered V2O5(001) thin films on Au(111): Growth and thermal stability. J Phys Chem C 112 (31):1183511846
99. Petukhov M, Rizzi GA, Granozzi G (2001) Spectroscopic and structural characterisation of a VOx (x
approximate to 1) ultrathin epitaxial film on Pt (111). Thin Solid Films 400 (1-2):154-159
100. Petukhov M, Rizzi GA, Granozzi G (2001) Ultrathin film growth and spectroscopic characterization of
VOx (0.8 <= x <= 1.3) on Pt(111). Surf Sci 490 (3):376-384
101. Surnev S, Kresse G, Sock M, Ramsey MG, Netzer FP (2001) Surface structures of ultrathin vanadium
oxide films on Pd(111). Surf Sci 495 (1-2):91-106
102. Jiang ZQ, Zhou WP, Tan DL, Zhai RS, Bao XH (2004) Evidence for perimeter sites over SmOxmodified Rh(100) surface by CO chemisorption. Surf Sci 565 (2-3):269-278
30
103. Kuriyama T, Kunimori K, Nozoye H (1998) Interaction of CO with the SmOx/Ru(001) interface. Surf
Sci 402 (1-3):178-181
104. Zhu JF, Ma YS, Zhuang SX (2009) Exploring the role of samarium in the modification of rhodium
catalysts through surface science approach. Surf Sci 603 (10-12):1802-1811
105. Jhang JH, Schaefer A, Cartas W, Epuri S, Baeumer M, Weaver JF (2013) Growth and Partial Reduction
of Sm2O3(111) Thin Films on Pt(111): Evidence for the Formation of SmO(100). J Phys Chem C 117
(41):21396-21406
106. Cartas W, Rai R, Sathe A, Schaefer A, Weaver JF (2014) Oxidation of a Tb2O3(111) Thin Film on
Pt(111) by Gas-Phase Oxygen Atoms. J Phys Chem C 118 (36):20916-20926
107. Schaefer A, Gevers S, Zielasek V, Schroeder T, Falta J, Wollschlager J, Baumer M (2011)
Photoemission study of praseodymia in its highest oxidation state: The necessity of in situ plasma treatment.
J Chem Phys 134 (5):054701
108. Gevers S, Weisemoeller T, Schaefer A, Zielasek V, Baumer M, Wollschlager J (2011) Structure of
oxygen-plasma-treated ultrathin praseodymia films on Si(111). Phys Rev B 83 (19):193408
109. Wilkens H, Gevers S, Rohe S, Schaefer A, Baumer M, Zoellner MH, Schroeder T, Wollschlager J
(2014) Structural Changes of Ultrathin Cub-PrO2(111)/Si(111) Films Due to Thermally Induced Oxygen
Desorption. J Phys Chem C 118 (6):3056-3061
31