Physico-chemical properties of unusual Ga2O3 polymorphs

Article abstract of DOI:10.1007/s00706-015-1628-z

A comparative structural and spectroscopic study, combined with reactivity tests in (inverse) water-gas shift and methanol steam reforming reaction, has been performed on various Ga₂O₃ polymorphs with special focus on δ-Ga₂O₃ and ε-Ga₂O₃ with the aim of highlighting the eventual intrinsic physico-chemical properties of the latter two. Using this comparative approach, the question whether δ-Ga₂O₃ in fact is a nanocrystalline modification of ε-Ga₂O₃, a mixture of ε-Ga₂O₃ and β-Ga₂O₃, or a single polymorphic form could be answered: especially Raman spectroscopy measurements, alongside reactivity tests, indicate that δ-Ga₂O₃ exhibits lots of properties of ε-Ga₂O₃. In fact, particularly in Raman measurements it appears as a mixture of ε-Ga₂O₃ and β-Ga₂O₃.

Full text of DOI:10.1007/s00706-015-1628-z

Monatsh Chem
DOI 10.1007/s00706-015-1628-z
ORIGINAL PAPER
Physico-chemical properties of unusual Ga2O3 polymorphs
Simon Penner1 Chen Zhuo1 Ramona Thalinger1 Matthias Gru¨nbacher1
•
•
•
•
Clivia Hejny2 Stefan Vanicek3 Michael Noisternig4
•
•
Received: 15 October 2015 / Accepted: 3 December 2015
Ó Springer-Verlag Wien 2015
Abstract A comparative structural and spectroscopic
study, combined with reactivity tests in (inverse) water–gas
shift and methanol steam reforming reaction, has been
performed on various Ga2O3 polymorphs with special
focus on d-Ga2O3 and e-Ga2O3 with the aim of highlighting
the eventual intrinsic physico-chemical properties of the
latter two. Using this comparative approach, the question
whether d-Ga2O3 in fact is a nanocrystalline modification
of e-Ga2O3, a mixture of e-Ga2O3 and b-Ga2O3, or a single
polymorphic form could be answered: especially Raman
spectroscopy measurements, alongside reactivity tests,
indicate that d-Ga2O3 exhibits lots of properties of e-
Ga2O3. In fact, particularly in Raman measurements it
appears as a mixture of e-Ga2O3 and b-Ga2O3.
Graphical abstract
Keywords Catalysis Á Material science Á Oxides Á
Nanostructures
Introduction
Polymorphism is one of the most important features in
oxide chemistry. On the one hand, it significantly adds to
the complexity of understanding of the fundamental issues
of operation; on the other hand—due to the different crystal
structures and physico-chemical properties—it opens
exciting new possibilities in their use also as technologi-
cally relevant materials [1, and references therein].
Furthermore, many of these usually metastable polymor-
phic modifications can be stabilized even under ambient
conditions in their nanostructured forms [1, and references
therein]. In that respect, synthesis of these materials typi-
cally is a major challenge and often external modifiers or
stabilizers are used. One of the best known examples is the
ZrO2 structure, whose metastable cubic or tetragonal
polymorphs are usually stabilized by the deliberate
Electronic supplementary material The online version of this
article (doi:10.1007/s00706-015-1628-z) contains supplementary
material, which is available to authorized users.
& Simon Penner
simon.penner@uibk.ac.at
1
Institute of Physical Chemistry, University of Innsbruck,
Innrain 80-82, 6020 Innsbruck, Austria
2
Institute of Mineralogy and Petrography, University of
Innsbruck, Innrain 52d, 6020 Innsbruck, Austria
3
Institute of General and Inorganic Chemistry, University of
Innsbruck, Innrain 80-82, 6020 Innsbruck, Austria
4
Institute of Pharmaceutical Technology, University of
Innsbruck, Innrain 52c/II, 6020 Innsbruck, Austria
123

S. Penner et al.
incorporation of anions or cations in its crystal lattice [2].
From both a fundamental scientific and more application-
oriented viewpoint, the study of oxides with a broad variety
of polymorphic modifications is most rewarding. Examples
include WO3, MoO3, Al2O3, or HfO2, all exhibiting four or
more polymorphic structures with different physico-
chemical properties [1, and references therein]. In general,
synthesis procedures for almost all structures exist, as in
most cases are the stability limits and the transformation
temperatures and conditions of the individual modifica-
tions. Although in some of the best characterized systems
also physico-chemical and even catalytic properties have
been examined [1, and references therein], there appears to
be a profound general lack of knowledge of these proper-
ties of most of the structures. This is sometimes owing to
the difficulty in applying a synthesis routine giving rise to
quantitative amounts of materials, especially when no
straightforward wet chemical preparation procedures exist
and, more importantly, the polymorphic forms cannot be
isolated as pure materials [3, 4]. One system, which shows
most of these difficulties, but appears also very rewarding,
regarding its potential use in technology, is the Ga2O3
system [4–19]. Its thermodynamically most stable mono-
clinic b-modification acts as an efficient sensor material
[20] or exhibits activity in many catalytic reactions (e.g.,
methanol reforming and synthesis, benzene hydrogenation
[21–23], or (inverse) water–gas shift reaction [24]).
Although the polymorphic system of Ga2O3 has been
scrutinized both experimentally and theoretically [4, 5, 7–
9, 11–14], there is still some debate about the existence of
several structures [4]. It is widely accepted that cubic a-
and c-Ga2O3 modifications exist [4]. Reliable synthesis
routines to the latter two phases exist [14] and the surface
chemistry [5, 9, 12, 19, 25] as well as their stability limits
and catalytic properties, e.g., in methanol steam reforming
[6], propane dehydrogenation [18], or photocatalysis [16]
are well known. They even served as supporting material
for small, catalytically CO2-selective intermetallic Pd2Ga
particles in methanol steam reforming, giving rise to a
catalytic bifunctional oxide-intermetallic synergism [6,
26]. In contrast, the physico-chemical properties of most of
the remaining polymorphic forms are less clear and even
their existence, or at least their synthesis routes, under
question [4]. The cubic d-Ga2O3 polymorph is repeatedly
reported to be synthesized by prolonged heating of gallium
nitrate hydrate at 493 K for several hours [4, 14, 16, 18].
There is some ongoing discussion, whether d-Ga2O3
exhibits a bixbyite structure and hence bears some resem-
blance to other X2O3 oxides like Mn2O3 or In2O3
crystallizing in the same structure, or if it indeed is only a
nanostructured form of another Ga2O3 polymorph, namely
e-Ga2O3 [4]. Nevertheless, propane dehydrogenation
activity of d-Ga2O3 has been reported, as are hydrogen
uptake measurements [18]. Even more under question is
the e-polymorph, which has been reported by Roy et al.
almost 50 years ago to be synthesized by simple high-
temperature annealing of d-Ga2O3 at temperatures
T [ 773 K [14]. Structural similarities with hexagonal e-
Fe2O3 have thereby been stressed [14]. However, repeated
attempts to re-synthesize this phase failed [16], but some
authors were finally successful in synthesizing e-Ga2O3
[13], also in its nanorod morphology [15]. Recently, the
initial X-ray diffractogram of Roy et al. was shown to be in
fact a mixture of e-Ga2O3 and b-Ga2O3, to which e-Ga2O3
at high temperatures inevitably transforms [4]. Playford
et al. were able to calculate the unit cell of e-Ga2O3 from
neutron and X-ray diffraction data and, hence, the presence
of this phase with a hexagonal structure in the Ga–O phase
diagram seems to be verified [4]. The latter authors were
also able to detect a transient new polymorph, j-Ga2O3,
upon oxidation of metallic Ga but, to date, isolation of this
phase repeatedly failed [4]. In due course, hardly anything
is known about the physico-chemical properties of these
materials, which is a particular pity, since the correlation of
their properties with other Ga2O3 polymorphs or isostruc-
tural compounds could help in resolving some of the
pending questions in this system.
As a consequence, the aim of the present study is to
provide data on the structure, morphology, surface chem-
istry, reducibility and electronic and catalytic properties of
d-Ga2O3 and e-Ga2O3 to eventually answer the specific
questions:
1. Given the contradicting results in literature, can the e-
Ga2O3 phase be reproducibly prepared?
2. Are both stable enough to ensure the properties can be
studied without unwanted phase transformation?
3. What are the intrinsic physico-chemical properties of
the d-Ga2O3 and e-Ga2O3 phase?
4. Does the d-Ga2O3 phase exhibit properties of e-
Ga2O3?
A multitude of structural and spectroscopic characteri-
zation tools (Raman spectroscopy, X-ray diffraction,
scanning electron microscopy), adsorption techniques
(volumetric adsorption, surface area and porosity mea-
surements according to BET and BJH), reactivity tests
(influence of water and hydrogen, catalytic testing in
methanol steam reforming and water–gas shift reaction)
and contact angle measurements will be applied to fulfill
these tasks. Particular emphasis will be given to a detailed
comparison to other yet well-established Ga2O3 polymor-
phic forms and to bixbyite-type In2O3.
It is be noted at this point that especially the catalytic
profiles could be of particular help in pointing towards the
intrinsic properties, since the potentially suspected
isostructure of d-Ga2O3 and cubic In2O3 due to the
123

Physico-chemical properties of unusual Ga2O3 polymorphs
potentially similar high CO2 selectivity in methanol steam
reforming (MSR) will be eventually directly revealed. In
addition, all Ga2O3 profiles in MSR are very characteristic
due to the different extent of water–gas shift reactivity and
intermediate formic acid formation [6].
that is reportedly assigned to the unequivocal presence of
e-Ga2O3. Specifically, only after heating d-Ga2O3 for 60 h
at 600 K after 14 days at 663 K, characteristic features of
e-Ga2O3 all appear in the diffractogram. Note that treat-
ments under conditions, e.g., reported in Ref. [4] or [13]
(cf. diffractogram b), show the splitting of the group of
peaks between 60° and 65°, but not lead to full crystal-
lization of the e-Ga2O3 phase. As reported earlier (and will
be discussed later on), the e polymorph is always con-
taminated with b-Ga2O3 and, therefore, the measured XRD
patterns are always a mixture of both. In that respect (as
marked in Fig. 1 in diffractogram e), the peaks at
2h = 35.5°, 36.9°, 38.4°, 40.8°, 62.2°, and 64.3° corre-
spond to the (100), (101)/(10-1), (004)/(00-4), (102)/(10-2),
(105)/(10-5), and (2-10) reflections of the e-Ga2O3 phase
([4] and a theoretically calculated pattern using the VestaÓ
program; Table S1 gives a full overview of the peaks),
while the ones at 2h = 30.8°, 31.7°, 33.5°, 48.7°, and 57.4°
correspond to the most intense reflections (-401), (002), (-
111), (510), and (-313) of the b-Ga2O3 structure [27].
Figure 2a (d-Ga2O3) and 2B (e-Ga2O3) show clear dif-
ferences in the morphologies of both Ga2O3 modifications.
Whereas the latter shows a distinct and pronounced
microporous structure, d-Ga2O3, after the chosen prepara-
tion pathway, consists of a network of characteristically
randomly oriented elongated grains of 50–100 nm length.
No pronounced porous morphology as for e-Ga2O3 is
observed. It should be noted that effective control of the
preparation and synthesis parameters upon dehydration of
gallium nitrate induces the formation of a number of dif-
ferent morphologies, including nanorods, for several
different polymorphic forms of Ga2O3, including, e.g., a-
Ga2O3 and e-Ga2O3 [15]. In fact, the morphology differ-
ences in the as-prepared states are significant. We note at
this stage, that it seems at first contradicting, that the sur-
face area of e-Ga2O3, which is prepared by prolonged
heating at higher temperatures than used for d-Ga2O3, is
actually higher than that of d-Ga2O3. Substantial sintering,
therefore, is not an issue. Rather, the transformation from
d-Ga2O3 into e-Ga2O3 must be associated with significant
recrystallization, leading to the observed morphology in the
SEM images, being fully consistent with the BET results.
Results
Formation of the phases and structural
characterization
Figure 1 reveals the successful preparation of the modifi-
cations under question. d-Ga2O3 is represented by its
characteristic fingerprint consisting of its broad (222) and
(400) peaks at 2h = 31.3° and 2h = 35.8° and (622) at
2h = 62.5° [14, 18]. Note that this assignment is essen-
tially based on that reported by Roy et al. [14], which is in
fact problematic due to the reasons discussed below. To
ensure the validity of the presence of the phase known as d-
Ga2O3 and to directly connect to the experiments reported
by other authors, e.g., [18], this assignment is kept for the
sake of clarity. A typical treatment of prolonged heating in
air at 493 K for 12 h was necessary for single phase for-
mation. Preparation of e-Ga2O3 did not prove to be
straightforward. More attempts—under, however, different
experimental conditions than previously reported [4, 13,
15]—were necessary to reproduce the XRD diffractogram
Spectroscopic characterization
Fig. 1 X-ray diffraction patterns collected after various annealing
treatments of the Ga(NO3)3ÁxH2O precursor in air and hydrogen.
a 12 h at 493 K (‘‘d-Ga2O3’’), b subsequent heating (after performing
treatment a) for 14 days at 673 K, c after performing the temperature-
programmed hydrogen treatment at 480 K with the sample prepared
via a, d 30 min at 800 K after 14 days at 673 K and e 60 h at 800 K
after 14 days at 663 K (‘‘e-Ga2O3’’). Peaks of ‘‘d-Ga2O3’’ have been
directly assigned [14, 18], those of ‘‘e-Ga2O3’’ (and b-Ga2O3)
indicated as ticks with further assignment in the text (according to
[4, 27])
For further detailed structural analysis, Raman spectra have
been collected for d-Ga2O3 and e-Ga2O3, as well as for
comparison, b-Ga2O3 and a-Ga2O3. The Raman spectra of
the latter are well known both from the experimental point
of view and theoretical group symmetry analysis and will,
therefore, serve as reference [16, 17, 28].
In short, the Raman spectra of b-Ga2O3 and a-Ga2O3
perfectly agree with those reported in the literature [16, 17,
123

S. Penner et al.
Fig. 3 Raman spectra of different Ga2O3 polymorphs. Black b-Ga2O3,
red d-Ga2O3, blue e-Ga2O3, magenta a-Ga2O3. The inset shows a direct
comparison between d-Ga2O3 and e-Ga2O3 (color figure online)
for b-Ga2O3 and the peaks observed for d-Ga2O3. The
relationship between d-Ga2O3 and e-Ga2O3 is best seen in
the inset of Fig. 3, highlighting the direct comparison of
Raman spectra. This strongly implies—in perfect
agreement with previous findings by XRD and neutron
diffraction [4]—that d-Ga2O3 is in fact a mixture of e-
Ga2O3 and b-Ga2O3, and the peaks of d-Ga2O3 are those of
the e-Ga2O3 structure.
Thermogravimetric analysis
To ensure the phase stability of the materials under ques-
tion, their behavior during an annealing cycle from room
temperature (298 K) to 873 K was tested by thermogravi-
metric analysis, coupled with calorimetry and mass
spectrometry measurements. Especially the latter will
directly reveal, if and how d-Ga2O3 and e-Ga2O3 transform
to the thermodynamically more stable b-modification.
Coupled with mass spectrometry, also chemical informa-
tion on eventual phase transformation can be deduced.
Figure 4 shows that within the temperature regime under
scrutiny both phases do not show exothermic signals due to
phase transformations to b-Ga2O3.
Fig. 2 SEM images of d-Ga2O3 (a) and e-Ga2O3 (b) in the as-
prepared state
28]. The Raman spectrum of d-Ga2O3 mainly consists of
broad features (found at 245, 530, 635, 718, and 808 cm-1
)
(cf. inset of Fig. 3). Note that in early publications, d-
Ga2O3 was reported to crystallize in the bixbyite structure
ꢀ
(space group Ia3) [4, 14]. Implying that it would occur as a
single phase, according to symmetry analysis, this should
give rise to the following number of Raman-active phonon
modes:
This holds for the annealing, the isothermal and the re-
cooling periods. While the mass loss of e-Ga2O3 is negli-
gible during the experiment, d-Ga2O3 loses about 0.5 mg
total mass. Following the mass loss trace of the latter,
several steps can be noticed: the first one, leading to
*0.21 mg mass loss, is associated with loss of water,
whereas the second (*0.22 mg) and the third (*0.05 mg)
are connected for the removal of CO2 and CO. This follows
from, in parallel, recorded ATR infrared experiments
(Fig. S1) and mass spectrometry data (Fig. S2), clearly
showing peaks in the carbonate region between 1500 and
1000 cm-1. In addition, the ATR spectra also directly
C ¼ 4Ag þ 41Eg þ 42Eg þ 14Tg
By comparison to the Raman spectrum of e-Ga2O3,
however, it is obvious that some common Raman modes
for both structures exist. In fact, all modes observed for d-
Ga2O3 can be also found for e-Ga2O3. The peaks for which
this can be best seen are marked by asterisks in the main
panel of Fig. 3. Most importantly, the Raman spectrum of
e-Ga2O3 is essentially a superposition of the one observed
123

Physico-chemical properties of unusual Ga2O3 polymorphs
Fig. 4 Thermogravimetric analysis of a d-Ga2O3 and b e-Ga2O3
during an annealing cycle in He from 298 K and 873 K. Black trace
mass loss, red trace temperature, blue trace heat flow. Heating rate
2 K min-1 (color figure online)
reveal the presence of a more or less hydroxylated surface
after preparation.
Volumetric adsorption
To further test the reactivity of both Ga2O3 modifications,
hydrogen consumption measurements and thermodesorp-
tion spectroscopy were performed. This is of particular
help in judging the behavior of d-Ga2O3 and e-Ga2O3,
since especially for b-Ga2O3, but also for the other modi-
fications, a more or less large data set is already available
[16, 19].
Fig. 5 Volumetric H2 adsorption uptake (a) and subsequent H2
thermodesorption (b) on d-Ga2O3 (blue) and e-Ga2O3 (green).
Heating rate in both cases 10 K min-1. Starting pressure in the
volumetric experiments 100 mbar (color figure online)
the 10-min isothermal period, the uptake still increases,
indicating considerable kinetic limitations of hydrogen
uptake. Upon cooling, no hydrogen desorption is observed,
indicating irreversible hydrogen adsorption. Pronounced
hydrogen consumption on e-Ga2O3 only starts at 573 K,
with a small decreased rate compared to d-Ga2O3 at around
623 K. Hence, in the temperature region between 523 and
743 K, most differences between the two modifications are
noticeable. Also on e-Ga2O3 hydrogen consumption at the
highest temperatures has not yet reached equilibrium and
most hydrogen is again irreversibly adsorbed.
Figure 5a shows the hydrogen uptake curves from room
temperature to 733 K. The temperature was restricted
generally to below 773 K, to avoid potential partial phase
transformations.
Both samples were routinely checked by XRD after use
(for d-Ga2O3, see Fig. 1) and still show the as-prepared
structures. Initially, both traces follow almost the same trend
and a slight increase in hydrogen uptake is only noticeable
starting at 423 K. d-Ga2O3 exhibits a second small step in
uptake at around 540 K, after which yet another plateau
region is reached. Only after 623 K, the hydrogen con-
sumption increases considerably up to 773 K. Also during
The corresponding thermodesorption spectra, taken
directly after cooling down the samples to room
123

S. Penner et al.
temperature, in contrast, show qualitatively a very similar
behavior (Fig. 5b). For both samples, two distinct desorp-
tion peaks have been observed, one at 540 K being present
as a broad shoulder, and a second one, more pronounced, at
*673 K. For d-Ga2O3, this peak is shifted to slightly
smaller temperatures (653 K). In essence, this indicates
that hydrogen is bound to two different surface sites, which
seem to be very similar on both modifications.
disentangling the relationship between d-Ga2O3, e-Ga2O3,
and b-Ga2O3.
Figure 6 shows the reactivity of both polymorphs in
methanol steam reforming. On d-Ga2O3 (Fig. 6a) methanol
conversion starts at about 560 K, with a parallel increase of
the CO2 signal.
Up to 670 K, the CO2 selectivity is almost constant at
about 40 %. Methanol conversions of *80 % at the
highest reaction temperatures are obtained. Two features
are worth noting: at first, a small signal of intermediate
formic acid is observed, with a maximum at around 560 K.
Secondly, hydrogen desorption is observed prior to actual
methanol conversion (at around 520 K), but the hydrogen
signal strongly increases at around 560 K, i.e., when the
Putting these results into perspective of previous work,
the following picture arises: On b-Ga2O3, hydrogen con-
sumption starts at much lower temperature (*373 K) with
a rather steep increase [19]. Qualitatively, the overall fea-
ture of hydrogen consumption is pretty similar. Kinetic
limitations of irreversible hydrogen uptake still prevail at
the highest temperatures. The corresponding TPD spectra
taken on b-Ga2O3 after performing hydrogen treatments
under comparable conditions are also similar [19], both
with respect to the number and temperature of the observed
desorption peaks and their relative intensity. A pronounced
peak at *700 K and a small shoulder at 500 K were also
observed on b-Ga2O3, indicating similar hydrogen binding
sites. The thermodesorption spectra of d-Ga2O3 are also in
line with those reported by Zheng et al. reporting on the
dehydrogenation of propane over different Ga2O3 poly-
morphs. In that work, for all modifications small hydrogen
adsorption peaks at *473 K were reported, decreasing
from a- over b- and c- to d-Ga2O3. A high-temperature
peak at 633 K was interestingly only observed (without
interpretation) on the latter [18]. As for interpretation of
these peaks, one may safely assume that the peak at lower
temperatures is connected with surface reduction (being
obviously qualitatively similar on all modifications) and
the high-temperature peak is associated either with reduc-
tion of surface-near or bulk regions.
Note, however, that the picture of hydrogen adsorption
especially on b-Ga2O3 is very complicated due to its
dependence on surface area and reduction degree of the
surface and bulk regions, as well as the propensity of dis-
sociative hydrogen adsorption, subsequently giving rise to
directed Ga-H bonds [19]. The required temperatures are
nevertheless hardly accessible on the two Ga2O3 modifi-
cations under question, due to their inherent structural
instability at higher temperatures.
Reactivity tests
One of the most promising techniques to unravel the
properties of the Ga2O3 polymorphs under scrutiny is to
test their reactivity in C1 chemistry. Particularly useful in
this respect are methanol steam reforming and (inverse)
water–gas shift reaction experiments, because each poly-
morph exhibits a characteristic fingerprint in activity and
selectivity. This, in turn, provides a reliable basis for
Fig. 6 Temperature-programmed methanol steam reforming reaction
on d-Ga2O3 (a) and e-Ga2O3 (b). Sample mass 50 mg. Temperature
ramp 5 K min-1
.
Reaction conditions: methanol:water = 1:2,
50 mbar total reaction mixture with addition of 10 mbar Ar and He
to 1 bar total pressure
123

Physico-chemical properties of unusual Ga2O3 polymorphs
steam reforming reaction starts. After that the reaction
obeys the stoichiometry of the methanol steam reforming
reaction. e-Ga2O3 (Fig. 6b) shows a similar reaction pat-
tern in methanol steam reforming. The onset of reaction is
more or less equal (at around 540–560 K), as is the CO2
selectivity (increasing slightly from about 48 to 55 %
during reaction). Intermediate formic acid formation is also
observed.
In comparison to the other Ga2O3 polymorphs, a-Ga2O3 is
clearly different, with a much higher CO2 selectivity. Both d-
Ga2O3 and e-Ga2O3 bear some similarities to c-Ga2O3 in
terms of onset of reaction, intermediate formic acid forma-
tion and CO2 selectivity. Of paramount importance,
however, are the correlating experiments to b-Ga2O3. This
polymorph exhibits a very similar reaction profile as com-
pared to d-Ga2O3 and e-Ga2O3, also with a very comparable
onset of reaction and intermediate formic acid formation.
Additional hints towards the intrinsic catalytic properties
will be given by a direct comparison of reactivities, the
extrapolated CO2-TOF values and apparent activation
energies. On the basis of active Ga–O entities [6], the TOF
values at 500 K (for better comparison) were calculated as
1.1 9 10-6 s-1 and 4.1 9 10-6 s-1, for d-Ga2O3 and e-
Ga2O3, respectively. These are in the range of b-Ga2O3
(6.0 9 10-6 s-1), but higher as determined for a-Ga2O3 and
c-Ga2O3 (\1.0 9 10-6 s-1). The corresponding activation
energies, on the basis of an Arrhenius rate analysis, were
determined as 142 and 134 kJ mol-1, which are also well in
the range of those measured for other polymorphs [6].
Further corroborating the reactivity similarities of d-
Ga2O3 and e-Ga2O3, Fig. 7 reveals that the reactivity
profile in the reaction CO ? H2O ? CO2 ? H2 is almost
identical. Activity starts by CO consumption and in parallel
formation of CO2 at *570 K. As no pronounced inverse
water–gas shift activity (CO2 ? H2 ? CO ? H2O; insets)
is observed, but only a more or less drastic H2 consumption
(at *640 K for both samples) without reaction of CO2, the
reaction of CO in the former case could also be explained
by a simple non-catalytic removal of oxygen from the
lattice and the subsequent formation of CO2, i.e., a Mars–
van-Krevelen—type reaction mechanism.
Fig. 7 Temperature-programmed water–gas shift reaction on d-
Ga2O3 (a) and e-Ga2O3 (b). Sample mass 50 mg. Temperature ramp
5 K min-1. Reaction conditions 25 mbar CO and 25 mbar H2O
(water–gas shift reaction), 50 mbar CO2 and 50 mbar H2 (inverse
water–gas shift reaction, inset) with addition of 10 mbar Ar and He to
1 bar total pressure
valuable catalysts, knowledge on the surface chemical
properties is imperative. This is even more important in
reforming reactions or any reaction which includes water
as a reagent. Under normal conditions, the surfaces of most
metal oxide materials are hydroxylated. These hydroxyl
groups may come in differently bound species, either
bridged or isolated. In due course, their thermal stability is
different, crucially influencing, e.g., the material’s reduc-
tion or catalytic behavior. Discriminating between the
different contributions of the components of the surface
free energy will, therefore, give not only information about,
e.g., the dispersive or polar nature of the surface, but also
about specific bonding affinities to different reactants.
Finally, we note that both polymorphic forms remain
structurally stable after reaction. All the experiments have
been tested on one sample (with re-oxidation steps at
673 K in between each step). Routinely collected XRD
diffractograms after all reactions still show the initial
structures and especially no signs of re-crystallization or
phase transformations (Fig. 8).
Surface chemical characterization
As the chemical nature of especially oxide systems is of
paramount importance for their technological use, e.g., as
123

S. Penner et al.
necessarily include an in-depth study of especially the acid–
base properties as well as the electron-donor and -acceptor
parameters. The latter are of paramount importance in oxide
chemistry, as they also strongly depend on, e.g., the
hydroxylation degree of the surface, among others.
Discussion
Following the introductory four specific research questions,
the discussion is now grouped into three parts that will
provide detailed answers to these questions. For a better
correlation of results, the discussion of the physico-chem-
ical properties is merged in one comparative part. As our
results in part strongly deviate from what has been outlined
in previous literature (and this in fact seems to be an ‘‘in-
trinsic’’ property of the Ga–O system), especially regarding
phase formation and stability, we in the first two sections
provide a thorough discussion of the similarities and dif-
ferences in phase formation and stability and try to
establish a reliable basis of understanding.
Fig. 8 X-ray diffraction patterns of d-Ga2O3 and e-Ga2O3 in their
initial states (a, c) and after performing the reactivity tests to ensure
structural integrity of the samples (b, d)
Together with reaction profiles discussed above, this will
lead to a complete picture of the reactivity of the different
Ga2O3 surfaces and to eventually reveal and explain their
intrinsic catalytic properties.
The surface energies and, more specifically, the contri-
butions of the dispersive and polar components for all the
studied Ga2O3 polymorphs are jointly summarized in
Table 1. These data were extracted from direct contact
angle measurements as outlined in the ‘‘Experimental’’
section. Although it is known that the surface chemistry of
all the Ga2O3 polymorphs is very variable and ranges from
strongly basic surface sites on a-Ga2O3 [5, 6] to strongly
acidic sites, e.g., on b-Ga2O3 [5], the data do not indicate
substantial variations in the resulting surface free energy.
Respective energies between 67 and 75 mN m-1 are
obtained. In due course, neither the resulting dispersive nor
the polar components do vary substantially between the
different modifications. However, all the modifications
exhibit a significantly higher dispersive fraction to the free
surface energy, implying that the polar contribution is of
less importance to the surface chemistry.
Formation of the individual phases
The behavior at lower annealing temperatures and gener-
ally the transformation of the hydrated gallium nitrate
precursor over the d-Ga2O3 phase into e-Ga2O3 is proven to
be most interesting. However, in this respect, most dis-
crepancies arise. While the formation of e-Ga2O3 is known
for about 50 years now, starting with the initial experi-
ments of Roy et al., the reproduction of this phase proved
difficult. Several groups reported the formation of e-Ga2O3
under comparable conditions: Playford et al. after heating
gallium nonahydrate nitrate at 493 K for 12 h, followed by
heating at 673 K for 14 days. The same result was obtained
by Nishi et al. [13] after heating at 823 K for 4 h (in dry
air) or by Roy et al. at 798 K for 30 min [14] (in air). In
contrast, Zheng et al. [18] and Hou et al. [16] failed to
prepare e-Ga2O3 under comparable experimental condi-
tions. The reported preparation pathways are in any case
preparation procedures in principle following the initial
routine of Roy et al. [14]. e-Ga2O3 was also prepared in its
nanorod morphology by a dedicated microwave-assisted
synthesis routine including urea, L-cysteine, and EDTA as
additives [15]. The poorly crystalline material of e-Ga2O3
in our case was only obtained after a final treatment of d-
Ga2O3 in air at about 798 K for 60 h. Note that d-Ga2O3 is
always the intermediate phase on the way to e-Ga2O3 [4,
13, 14]. The discussion indicates that the formation of the
latter phase obviously very much depends on the effective
control of the experimental parameters, in particular on the
precursor state and heating rate or chemical gas phase
environment. We also might infer a considerable influence
Knowing that the catalytic behavior of the polymorphs is
indeed different [6], as are the acid–base properties of the
individual modifications [6], more detailed investigations of
this particular phenomenon are indicated. This would
Table 1 Dispersive and polar components of the total free surface
energy, as deduced from the contact angle measurements, given in
mN m-1
Phase
Dispersive
Polar
Free surface energy
a-Ga2O3
b-Ga2O3
c-Ga2O3
d-Ga2O3
e-Ga2O3
44.80
44.10
45.80
43.39
45.48
27.86
26.90
29.18
24.01
28.97
72.67
71.00
74.98
67.41
74.46
123

Physico-chemical properties of unusual Ga2O3 polymorphs
of the humidity on the phase formation, since as shown in
the corresponding ATR spectra, the samples are hydroxy-
lated to a certain extent. However, as the humidity was not
controlled and the synthesis routine basically focussed on
those in the literature [4, 14, 18], further studies are nec-
essary to fully clarify this specific issue. In any way, as the
starting material in almost all cases is a hydrate gallium
nitrate precursor, the initial phase occurring after annealing
must be a de-nitrated gallia hydroxo-species, which further
transforms into the respective Ga2O3 polymorphs. So far,
no experimental proof of this intermediate phase has been
provided. The first species that arises in all experiments is
the d-Ga2O3 phase as poorly crystalline material. Note that
for other polymorphic forms of Ga2O3, such precursors
have been already detected: GaOOH for a-Ga2O3 and
Ga5O7(OH) for j-Ga2O3. In the latter case, the Ga5O7(OH)
intermediate was shown to bear significant structural
resemblance also to e-Ga2O3, which suggests inter-trans-
formation between different polymorphs [14]. It should be
noted that this behavior is by no means restricted to the
Ga–O system: similar observations have also been made in
the Zr–O system, where structural similarities between
hydroxy-gels and the tetragonal ZrO2 polymorph have been
held accountable for the favored crystallization of the latter
phase upon annealing [29]. In line with these experiments,
also for formation of the rhombohedral In2O3 phase,
structural similarities to InOOH precursors are of para-
mount importance [30].
hydroxylation degree is obviously of paramount impor-
tance also for the observed phase transformations and not
only for phase formation. Secondly, it appears that the
phase transformation itself seems to be dominated by
kinetic effects, as a crucial temperature has to be overcome
to induce the structural changes. Unfortunately—although
some structural key features between the associated
hydrated species and the resulting structure of the Ga2O3
polymorphs are known [4]—the detailed mechanism of the
phase transformation remains unclear up to now. This
would also explain why seemingly similar experimental
procedures lead to varying results in extent of phase
transformation. Careful and reproducible control of the
experimental parameters would be necessary for reliable
phase preparation. Again, cross-correlation to results on the
phase transformation in the Zr–O system helps in getting a
clearer picture: In this specific system, the transformation
between the metastable tetragonal and thermodynamically
more stable monoclinic ZrO2 phase is extremely dependent
on grain size, defectivity of the structure, heating rate, and
chemical potential of the gas phase during annealing. The
amount of monoclinic ZrO2 could, e.g., be steered between
0 and 80 % by controlling the annealing temperature
between 673 K and 1273 K and the gas phase environ-
ment. Oxygen-rich conditions are reported to induce the
formation of a higher amount of monoclinic ZrO2, whereby
annealing in vacuum (i.e., more reducing conditions) leads
to stabilization of the associated tetragonal phase. Hence,
the extent of phase transformation is strongly connected
with the defectivity of the material [29]. Note that due to
the high purity of the materials (99.99 %), a substantial
influence of external stabilizers can be excluded. In due
course, similar behavior is expected in the Ga–O system
and this is strongly backed by the outlined results and
previous studies. In the Ga–O system, the picture is further
complicated by the still existing discrepancies whether
some polymorphs (especially d-Ga2O3) are true modifica-
tions or just nano-crystalline variants of the same structure.
Stability of the individual phases
Most notably, both structures proved to be very stable, both
in hydrogen up to about 773 K and in helium up to 873 K.
No substantial transformation into b-Ga2O3 has been
observed. This is insofar remarkable, as—according to
previous works—these temperatures are in principle high
enough to induce the corresponding phase transformation.
Playford et al., for example, reported transformation of e-
Ga2O3 into b-Ga2O3 at temperatures T [ 773 K [14]. Nishi
et al. observed full transformation into b-Ga2O3 at 923 K
(4 h) [13]. Most importantly—in line with the argumen-
tation outlined in the preceding section—Roy et al.
reportedly observed significant differences in the phase
stability upon treatment under dry and moist conditions. In
fact, the stability of d-Ga2O3 under wet conditions directly
produced b-Ga2O3 at T [ 573 K. Under dry conditions, e-
Ga2O3 is formed above T [ 773 K, whereby final b-Ga2O3
formation is observed at temperatures higher than 1143 K.
Note that a very sharp transformation temperature of 773 K
for e-Ga2O3 has been observed, but no transformation was
observed at slightly lower temperatures, even when held
for a week [14]. Two conclusions can be immediately
drawn from these two experiments: Firstly, the
Physico-chemical properties of the d-Ga2O3 and e-
Ga2O3 phase
Revisiting the initially introduced questions about the
physico-chemical properties of the d-Ga2O3 and e-Ga2O3
as to whether especially the former is an individual poly-
morph in the Ga–O system, the following picture arises:
Especially Raman measurements and volumetric adsorp-
tion give clear indications that d-Ga2O3 in fact exhibits
many features of e-Ga2O3. This is in line with previous
XRD and neutron diffraction experiments, which strongly
indicate that d-Ga2O3 is in fact a nanocrystalline variant of
e-Ga2O3 [4]. Furthermore, the Raman spectra of e-Ga2O3
support the assumption that e-Ga2O3 is always
123

S. Penner et al.
3. Surface properties of e-Ga2O3 phase are similar to
other polymorphic forms, which are also true for
reaction profiles in methanol steam reforming and
(inverse) water–gas shift reaction. In this respect, our
physico-chemical characterization results corroborate
previous suggestions that e-Ga2O3 is always contam-
inated with b-Ga2O3. This is especially derived from
Raman measurements.
contaminated with the thermodynamically more stable b-
Ga2O3 modification (as it also follows from XRD). As has
been discussed in the catalytic section, the reactivity in the
water–gas shift equilibrium and CO2 selectivity in metha-
nol steam reforming are clearly different for the different
modifications [6]. a-Ga2O3, for example, is a strongly basic
oxide yielding high amounts of CO2 in the reactant feed
[6]. In contrast, b-Ga2O3, due to the presence of acidic and
basic surface sites, is also highly reactive in the inverse
water–gas shift reaction, therefore, yielding only decreased
CO2 selectivity [24]. The presence of formic acid during
methanol steam reforming indicates also the presence of
strong acidic surface sites. This correlation of surface
chemistry and product selectivity in methanol synthesis
and steam reforming is highlighted in a concept recently
introduced by Tatibouet [31] and is, therefore, also valid
for d-Ga2O3 and e-Ga2O3: The acidic character of the
surface is much more pronounced than its associated basic
character. The CO2 selectivity in MSR is, therefore, not
very pronounced.
4. Our results strongly indicate that d-Ga2O3 phase
exhibits properties of e-Ga2O3. Apart from Raman
measurements, especially the steam reforming exper-
iments do not show the characteristic high CO2
selectivity being typical for bixbyite-type basic oxides
like In2O3. On the contrary, the catalytic profile is very
similar to e-Ga2O3.
Experimental
Preparation of materials
Preparation of the materials was done using standard rou-
tines outlined in recent literature [4, 13, 14]. Starting
material in all cases was Ga(NO3)3ÁxH2O (99.99 % Alfa
Aesar). A muffle furnace operated under ambient condi-
tions was used to decompose Ga(NO3)3ÁxH2O and to
subsequently form the desired Ga2O3 modifications. d-
Ga2O3 was prepared by annealing Ga(NO3)3ÁxH2O for 12 h
at 493 K. The XRD diffractogram of the resulting white
powder (cf. Fig. 1) coincides well with those reported for
‘‘d-Ga2O3’’ in the literature [4, 13, 14, 18]. Preparation of
the corresponding ‘‘e-Ga2O3’’ phase proved more
demanding, since a simple annealing treatment of d-Ga2O3
for 14 days at 673 K, as reported earlier [4, 14], did not
lead to reproducible crystallization of e-Ga2O3. Only after
a prolonged treatment (60 h) at 800 K, the XRD fingerprint
of e-Ga2O3 was finally obtained (cf. Fig. 1) [13, 15] (vide
supra). BET areas were measured using a Quantachrome
2000 Pore Size Analyzer and yielded 17 and 54 m2 g-1 for
d-Ga2O3 and e-Ga2O3, respectively. Further details of N2
sorption analysis (pore size diameter and volume) are given
jointly in Table 2.
Conclusion
Exemplarily shown for the Ga2O3 polymorphic system, the
advantages of a concerted methodological approach in
resolving pending questions of phase stability, structure
and physico-chemical properties are highlighted. In the
present case, some of the most crucial structural informa-
tion is derived indirectly from characterization of the
physico-chemical properties, including catalytic reactivity
profiles. In line with the four questions raised in the
introductory chapter, we are now able to give detailed
answers:
1. e-Ga2O3 can be reproducibly prepared, but the details
of preparation (e.g., reaction temperature and time, gas
atmosphere, etc.) are different from those in literature
and very much depend on the specific reaction
conditions. In that respect, e-Ga2O3 in the present
case needs experimental conditions, where other
authors already observed beginning phase transforma-
tion. The mechanism of stabilization is unclear up to
now, but given the high surface area, particle size
effects, a common phenomenon in oxide phase stabi-
lization, cannot be excluded.
In summary, the experimental conditions to obtain the
Ga2O3 modifications discussed in here are as follows:
2. Both polymorphs are stable enough to ensure that the
properties can be studied without unwanted phase
further transformations, although, as reported earlier,
the final phase of both samples is always b-Ga2O3.
This in particular refers to stability in methanol steam
reforming and (inverse) water–gas shift reaction
mixtures.
Table 2 N2-physisorption data for d-Ga2O3 and e-Ga2O3
Sample
Surface area-
BET/m2 g-1
Pore volume/
cm3 g-1
Pore radius/nm
d-Ga2O3
e-Ga2O3
17
54
0.022
0.121
2
2
123

Physico-chemical properties of unusual Ga2O3 polymorphs
CCD detector. The spectral resolution, determined by mea-
suring the Rayleigh line, was better than 2 cm-1. The spectra
were recorded in unpolarized mode at ambient conditions. To
determine the accuracy of wavelength measurement, a cali-
bration using an Si-Wafer (522 cm-1) was performed.
Table 3 Overview of the known polymorphic forms of Ga2O3
˚
Lattice parameters/A
Sample
Space group
Remark
[6]
ꢀ
R3c
a-Ga2O3
a = 4.98
c = 13.43
a = 12.23
b = 3.04
b-Ga2O3
C2/m
[27]
Thermogravimetric analysis
c = 5.80
ꢀ
Fd3m
c-Ga2O3
d-Ga2O3
e-Ga2O3
a
a = 8.24
[4]
[14]a
The DTA experiment was performed on a Setaram Set-
sysEvolution 2400 instrument, using a TGA-DTA1600
transducer, equipped with Pt/Pt–Rh S-type thermocouples.
The instrument is coupled with an OmniStar QMS200,
Pfeifer Vacuum quadrupole mass spectrometer. Several
milligrams of the powdered samples were filled in a
100 mm3 Al2O3 crucible. Subsequently, after a routine pre-
run including evacuation and flooding the sample chamber
with He, the DTA experiment was carried out by heating
with a rate 2 K min-1 to a maximum temperature of 873 K
in He atmosphere and simultaneous recording of selected
masses (m/z = 16, 17, 18, 28, and 44 corresponding to
H2O, CO, CO2 and OH fragments).
ꢀ
Ia 3
a = 9.99
P63mc
a = 2.903 c = 9.255
[4]
Note the discussion of the structure in the ‘‘Discussion’’ section
1. d-Ga2O3: annealing Ga(NO3)3ÁxH2O for 12 h at 493 K
in air under ambient conditions,
2. e-Ga2O3: as d-Ga2O3, but subsequently, the sample
was treated at additional 14 days at 663 K followed by
60 h at 800 K,
3. a- and c-Ga2O3 were synthesized according to recipes
further described in [6], and
4. b-Ga2O3 is
99.999 %).
a commercial sample (Alfa Aesar
Reactivity tests
Table 3 gives an overview of the structural parameters
of the polymorphs under question. The chemical homo-
geneity of the samples has been checked by EDXRFA
measurements (Spectro-XEPOS) and revealed a 99.99 %
purity of the synthesized materials (cf. Table S2).
Tests in methanol steam reforming and (inverse) water–gas
shift reaction were performed in an NI Labview-automa-
tized re-circulating batch reactor of about 13 cm3 volume
[6]. A quadrupole mass spectrometer (Balzers QMG 311)
attached to the circulating batch quartz glass reactor via a
capillary leak was used for analysis. Details of the basic
catalytic setup have been outlined elsewhere [6]. For the
specific reaction details, we refer to the respective fig-
ure captions. Note that for all experiments, about 10 mbar
Ar was added to account for the decrease of the mass
spectrometer signal due to the continuous gas withdrawal
through the leak and for correction of thermally induced
expansion. Finally, He was added to 1 bar total pressure.
All mass spectrometer signals were externally calibrated
and corrected for fragmentation in the mass spectrometer.
To account for the partial adsorption of some reactants, all
catalytic measurements include a 30 min equilibration
period in the starting mixture prior to each measurement. In
each case, the catalyst was exposed to the reaction mixture
and the temperature was ramped with 5 K min-1 to the
final value of 673 K. Blind reactivity was corrected by the
activity of the catalyst holder containing only quartz wool
resulting in an almost negligible, at maximum 1 % con-
version based on CO2 formation after 1 h (for methanol
steam reforming). For data evaluation, the relative inten-
sities of the mass spectrometer signals were converted into
partial pressures via external calibration using gas mixtures
of defined partial pressures.
Structural and morphological characterization
X-ray diffraction experiments were performed ex situ
under ambient conditions using a Siemens D5000 Spec-
˚
trometer and Cu-Ka radiation (1.54178 A) at 300 K in the
2h range 10°–70° with a step size of 0.02°.
For SEM experiments, an SM 982 GEMINI ZEISS field
emission scanning electron microscope was used. Prior to
SEM imaging, the samples were coated with 10 nm Au/Pd to
improve its conductance and fixed with conducting carbon
paste.
Spectroscopic characterization
Confocal Raman spectra in the range of 50–3800 cm-1 were
recorded with a Horiba Jobin-Yvon Labram-HR 800 Raman
micro spectrometer. The samples were excited using the
532 nm (2.33 eV) emission line of a frequency-doubled
25 mW Nd:YAG laser under an Olympus 1009 objective lens
with a numerical aperture of 0.9. The size of the laser spot on
the surface was approximately 1 lm in diameter. The scattered
light was dispersed by an optical grating with 1800 li-
nes mm-1 and collected by a 1024 9 256 open-electrode
123

S. Penner et al.
Volumetric adsorption
References
1. Sood S, Gouma P (2013) Nanomater Energy 2:82
2. Duwez P, Brown FH, Odell F (1951) J Electrochem Soc 98:356
3. Bekheet MF, Schwarz MR, Lauterbach S, Klebe H, Knoll P,
Riedel R, Gurlo A (2013) Angew Chem Int Ed 52:6531
4. Playford HY, Hannon AC, Barney ER, Walton RI (2013) Chem
Eur J 19:2803
5. Lavalley JC, Daturi M, Montouillout V, Clet G, Otero-Arean C,
Rodriguez-Delgado M, Sahibed-Dine A (2003) Phys Chem Chem
Phys 5:1301
The volumetric adsorption measurements were conducted
in an all-quartz apparatus equipped with metal bellow
valves (Witeg), a Baratron pressure transducer (MKS),
mass flow controllers (MKS), and a Balzers QMA125
quadrupole mass analyzer. The furnace was a Linn model
operating up to 1473 K. Before each experiment and to
ensure identical starting conditions, the samples were
fully oxidized by heating for 1 h in 1 bar of dry flowing
O2 to 733 K. The samples were cooled to 298 K in dry
flowing O2 followed by evacuating the cell to about
5 9 10-7 mbar. The temperature-programmed reduction
experiments in hydrogen involved heating from 298 K up
to the final temperature of 733 K at a linear rate of
10 K min-1 in a defined amount of *100 mbar H2,
maintaining the maximum temperature for 10 min, and
then re-cooling to 298 K at a similar rate. For better
comparison also to literature-reported data, all mass
spectrometer data (in mbar)—except H2—were converted
into lmol based on the ideal gas equation, with subse-
quent normalization to sample mass and surface area.
Hydrogen in the TPR experiments was quantified via the
overall system pressure in the cell.
¨
6. Lorenz H, Thalinger R, Kock EM, Kogler M, Mayr L, Schmid-
mair D, Bielz T, Pfaller K, Klotzer B, Penner S (2013) Appl Catal
¨
A 453:34
7. Playford HY, Hannon AC, Tucker MG, Dawson DM, Ashbrook
SE, Kastiban RJ, Sloan J, Walton RI (2014) J Phys Chem C
118:16188
8. Yoshioka S, Hayashi H, Kuwabara A, Oba F, Matsunaga K,
Tanaka IJ (2007) Phys Condens Matter 19:346211
9. Otero-Arean C, Lopez Bellan A, Penarroya Mentruit M, Rodri-
guez-Delgado M, Turnes Palomino
Mesoporous Mater 40:35
G (2000) Microporous
10. Luenangchaichaweng W, Brooks NR, Fiorilli S, Gobechiya E,
Lin K, Li L, Parres-Esclapez S, Javon E, Bals S, Van Tendeloo G,
Martens JA, Kirschhock CEA, Jacobs PA, Pescarmona PP (2014)
Angew Chem Int Ed 53:1585
11. Meriaudeau P, Primet M (1990) J Mol Catal 62:227
12. Vimont A, Lavalley JC, Sahibed-Dine A, Otero-Arean C,
Rodriguez-Delgado M, Daturi M (2005) J Phys Chem B
109:9656
The temperature-programmed experiments were con-
cluded by a subsequent thermodesorption measurement,
where the amount of hydrogen being irreversibly adsorbed
after the experiments was deliberately desorbed. This
included heating the sample at a linear rate of 10 K min-1
in the evacuated cell up to 753 K. After these experiments,
the structural integrity of the samples was ensured by XRD
measurements (cf. Fig. 1).
13. Nishi K, Shimizu K, Takamutsu M, Yoshida H, Satsuma A,
Tanaka T, Yoshida S, Hattori T (1998) J Phys Chem B 102:10190
14. Roy R, Hill VG, Osborn EF (1952) J Am Chem Soc 74:719
15. Ge S, Zhang L, Jia H, Zheng Z (2009) J Mater Res 24:2268
16. Hou Y, Wu L, Wang X, Ding Z, Li Z, Fu X (2007) J Catal 250:12
17. McMahon D, McMillan PF, Xu B, Dong J (2006) Phys Rev B
73:094125
18. Zheng B, Hua W, Yue Y, Ghao Z (2005) J Catal 232:143
¨
19. Jochum W, Penner S, Fottinger K, Kramer R, Rupprechter G,
¨
Klotzer B (2008) J Catal 256:268
20. Weh T, Frank J, Fleischer M, Meixner H (2001) Sens Actuators
78:202
Surface chemical characterization
21. Ono Y (1992) Catal Rev Sci Eng 34:179
22. Meitzner GD, Iglesia E, Baumgartner JE, Huang ES (1993) J
Catal 140:209
Contact angle measurements were performed with a Drop
Shape Analyzer DSA 25E (Kru¨ss) applying the sessile drop
method with double distilled water and diiodomethane at
ambient conditions. To achieve perfect smooth surfaces,
10–70 mg powder was compressed with the hydraulic
press PW 10 (Paul Weber) using flat-faced punches (8 mm
diameter) and a maximum pressing force of 25 kN. To
calculate the surface free energy, the OWRK-Model
(Owens–Wendt–Rabel and Kaelble) provided by the Soft-
ware Advance 1.2 (Kru¨ss) is used.
23. Takahara I, Saito M, Inaba M, Murata K (2004) Catal Lett 96:29
¨
24. Jochum W, Penner S, Fottinger K, Kramer R, Rupprechter G,
¨
Klotzer B (2008) J Catal 256:278
25. Collins SE, Baltanas MA, Bonivardi AL (2005) Langmuir 2:962
¨
¨
26. Li L, Zhang B, Kunkes E, Fottinger K, Armbruster M, Su DS,
¨
Wie W, Schlogl R, Behrens M (2012) ChemCatChem 4:1764
27. Grier D, McCarthy G, North Dakota State University, ICDD
Grant-in-Aid 1991; Powder Diffraction File, International Center
for Diffraction Data 1994, pattern 00-043-1012
28. Dohy D, Lucazeau G, Revolevschi A (1982) J Solid State Chem
45:4524
¨
29. Kogler M, Kock EM, Vanicek S, Schmidmair D, Gotsch T,
Stoger-Pollach M, Hejny C, Klotzer B, Penner S (2014) Inorg
¨
¨
¨
Acknowledgments The work was performed within the platform
‘‘Material- and Nanoscience’’ at the University of Innsbruck. Finan-
cial support of the SFB project F45, funded by the Austrian Science
Fund, is greatly acknowledged. We thank R. Tessadri for assistance
with the EDXRFA measurements.
Chem 53:13247
30. Epifani M, Siciliano P, Gurlo A, Barsan N, Weimar U (2004) J
Am Chem Soc 126:4078
31. Tatibouet JM (2008) Appl Catal A 148:34
123