Ternary VZrALON oxynitrides - Efficient catalysts for the ammoxidation of 3-picoline

Article abstract of DOI:10.1021/cs500458w

Starting from previous binary VZrON (VAlON) oxynitrides with high (low) activity and low (high) selectivity, a new class of ternary VZrAlON catalysts has been developed for the ammoxidation of 3-picoline to 3-cyanopyridine (3-CP), which combine the benefi

Full text of DOI:10.1021/cs500458w

Research Article
pubs.acs.org/acscatalysis
Ternary VZrAlON Oxynitrides - Efficient Catalysts for the
Ammoxidation of 3‑Picoline
Christiane Janke,† Jorg Radnik,‡ Ursula Bentrup,‡ Andreas Martin,‡ and Angelika Bruckner*,‡
̈
̈
†BASF SE, GCC/PG-M301, 67056 Ludwigshafen, Germany
‡Leibniz Institut fur Katalyse e.V. an der Universitat Rostock, Albert-Einstein-Straße 29a, Rostock, 18059 Germany
̈
̈
S
* Supporting Information
ABSTRACT: Starting from previous binary VZrON (VAlON) oxynitrides with
high (low) activity and low (high) selectivity, a new class of ternary VZrAlON
catalysts has been developed for the ammoxidation of 3-picoline to 3-
cyanopyridine (3-CP), which combine the beneficial properties of the binary
oxynitrides, leading to improved selectivity at retained high activity and to the
highest space-time yield of 3-CP ever measured (488 g L−1 h−1). This is attributed
to the formation of a special −⊡−V5+(O)−N−Al(Zr)− surface moiety consisting
of a V5+O species in the vicinity of a surface nitrogen and an anion vacancy
occupied by an electron, which is supposed to provide optimum conditions for a
double Mars−van Krevelen mechanism comprising activation of gas-phase oxygen
and ammonia via reversible incorporation into the catalyst surface as well as an
efficient electron transport.
KEYWORDS: ammoxidation, 3-picoline, vanadium zirconium aluminum oxynitrides, in situ EPR spectroscopy, XPS,
UV−vis spectroscopy, structure−reactivity relationships
incorporation of N into the catalyst lattice enables a so-called
INTRODUCTION
■
double Mars−van-Krevelen mechanism for the activation of
both oxygen and ammonia.11 On the basis of a comprehensive
analysis of structure−reactivity relationships, it has been
concluded that highly selective and active ammoxidation
catalysts must be able to reversibly incorporate N into their
surface and must expose highly dispersed V species with a mean
surface valence state close to +4 as well as V and N surface sites
in close vicinity (preferentially as V−N−M moieties).9
Moreover, they should be free of pronounced surface acidity.9
Attempts to adjust such properties in VZrPON catalysts were
not successful since the incorporation of P, though enhancing
the V site dispersion, led to a detrimentally high surface
acidity.12
In this work, we follow a new approach aimed at combining
the beneficial properties of both binary catalysts systems. The
objective was to develop a novel class of ternary VZrAlON
catalysts which maintain the high activity of VZrON and the
high selectivity of VAlON. Here we show that this approach
was indeed successful, leading to ammoxidation catalysts with
significantly improved catalytic properties. Besides, relations
between structural peculiarities and catalytic performance have
been established using a variety of analytical techniques (XRD,
XPS, ATR, UV−vis−DRS) as well as by monitoring the
nitridation and ammoxidation process with in situ EPR
spectroscopy.
The gas-phase ammoxidation of methylaromatics and hetero-
aromatics is an ecologically and economically efficient route to
synthesize a variety of different nitriles (eq 1), as long as high
nitrile selectivities and space-time yields (STY) are reached.1,2
Thus, 3-cyanopyridine (3-CP, also known as nicotinonitrile),
being the source for the production of vitamin B3, is obtained
by ammoxidation of 3-picoline over a variety of vanadium-
containing oxide catalysts such as VPO,3−5 VTiO,6 VSnO7 or
VSbTiO.8
R − CH3 + NH3 + 3/2O2 → R − CN + 3H2O
(1)
With binary VAlON and VZrON oxynitrides9 we could
achieve almost 3 times higher STY of 3-CP as obtained with an
industrial SbVTiSiKO benchmark catalysts.10 A maximum STY
of 399 g/L h was reached with VZrON catalysts, yet the 3-CP
selectivity did not exceed 63% at a conversion of 91%. On the
other hand, VAlON catalysts were found to be more selective
but less active than VZrON materials (X3‑PIC = 80%, S3‑CP
=
79%).9 Structural differences have been identified as major
reasons for this. Thus, it was found that VZrON catalysts
contain mainly polymerized VOx species with a high mean
valence state (close to +5) and almost no nitrogen on the
surface under working conditions. In contrast, for VAlON
catalysts not only a lower V site agglomeration and a lower
mean surface V oxidation state but also its ability to incorporate
nitrogen species in the surface has been evidenced. These
properties were regarded as crucial for high 3-CP selectivity,
since it had been shown previously that single VO2+ sites
suppress total oxidation of hydrocarbons to COx and
Received: April 7, 2014
Revised: June 13, 2014
Published: July 5, 2014
© 2014 American Chemical Society
2687
dx.doi.org/10.1021/cs500458w | ACS Catal. 2014, 4, 2687−2695

ACS Catalysis
EXPERIMENTAL SECTION
Research Article
power of 6.3 mW, a modulation frequency of 100 kHz, and a
modulation amplitude of 0.5 mT. The magnetic field was measured
with respect to the standard, 2,2-diphenyl-1-picrylhydrazyl. For in situ
experiments, a homemade quartz plug-flow reactor connected to a
gas/liquid-dosing system (equipped with mass flow controllers for
gases and saturators for liquids) was implemented in the rectangular
cavity of the spectrometer.13 The reactor was loaded with 100 mg of
catalyst particles (315−710 μm). After heating in air/NH3 (4.5/1) to
350 °C, H2O was introduced at the same temperature. Subsequently,
3-PIC was added (3-PIC/air/NH3/H2O = 1/30/6/8) and switched off
again after a certain time. At the end of the experiments, the gas feed
was changed to pure N2, and the reactor was cooled to room
temperature (RT). EPR spectra at 350 °C were recorded until
reaching a steady state. Besides interaction of the oxynitrides with feed
components, also the reaction of the VZrAlO-0.5 oxide precursor with
NH3 was monitored by EPR. In this case, 112 mg particles were first
heated from RT to 120 °C in N2 (10 mL/min) and cooled again to
RT. Then, the gas flow was switched to 20% NH3/N2 (38 mL/min) at
RT, and the reactor was heated stepwisely to a final temperature of
410 °C. Spectra at elevated temperatures were recorded as a function
of time for at least 30 min. Computer simulation of EPR spectra was
performed with the program SIM14S of Lozos et al.14 using the spin
Hamiltonian (eq 2)
■
Catalyst Preparation. The VZrAlO-x oxide precursors, in which x
denotes the theoretical V/(Al+Zr) ratio established during synthesis
from 0.1 to 0.6 at a constant Al/Zr ratio of 1.5, were synthesized by
coprecipitation. The desired amount of NH4VO3 (Aldrich, 99%) was
suspended in distilled water (0.1 M) at 70 °C and subsequently
acidified with 65% HNO3 (Roth, pure) at 70 °C under stirring to
reach a final pH of 3. In a second solution, Al(NO3)3·9 H2O (Sigma-
Aldrich, 98%) and ZrO(NO3)3 (Aldrich, tech.) were dissolved in
distilled water at 70 °C under stirring, adjusting a concentration of
0.15 M for each component. The hot NH4VO3 solution was slowly
added to the hot Al−Zr solution. Subsequently, the pH of the final
solution was raised to 7 at 70 °C by slowly adding NH4OH (Merck,
25%). This led to the formation of a yellow precipitate, which was
filtered, washed twice with distilled water, dried at 120 °C for 16 h and
finally ball-milled to obtain a fine yellow powder of the VZrAlO-x
oxide precursor. About 6 g of each VZrAlO precursor was heated with
5 K/min in NH3 flow (3.8, Air Liquide, 30 L/h) to 600 °C and
nitrided for 6 h, resulting in black VZrAlON-x oxynitride powders.
Catalytic Tests. Catalytic tests were performed in a fixed bed glass
reactor with an inner diameter of 1.5 cm. Oxynitride powders were
pressed to pellets, crushed, and sieved. Catalyst particles of 1.0−1.25
mm size were diluted with glass beads (1.25−1.55 mm) in a
volumetric ratio of 1:1.7 for each test. The reaction temperature TB
was measured inside the catalyst bed by a thermocouple. Air (5.0, Air
Liquide) and ammonia (3.8, Air Liquide) were dosed by mass flow
controllers (Bronkhorst). Water and 3-picoline (Aldrich, 99%) were
premixed and fed continuously into the gas stream by a syringe pump.
The molar ratio of the feed components was set to 3-PIC/air/NH3/
H2O = 1/28/5/8. Prior to each catalytic test, the catalyst inside the
reactor was heated with 10 K/min in air/NH3 flow to 190 °C, and
subsequently the 3-PIC and H2O were added. The gas stream leaving
the reactor outlet was trapped in ethanol. Concentrations of 3-PIC and
3-CP were analyzed off-line using a GC-17A (Shimadzu) equipped
with an autosampler, a WCOT fused silica CP-SIL 8 CB column
(Varian), and an FID. Concentrations of the byproducts CO2 and CO
were analyzed online using a nondispersive infrared analyzer (BINOS
100 2M, Rosemount). Catalytic tests of VZrAlON-x were performed
at two different gas hourly space velocities (GHSV), either at 2713 h−1
or at 5728 h−1. The reaction temperature TB was varied between 320
and 390 °C. Carbon balances were calculated for each test to judge the
quality of the catalytic test results.
H = μBSgBo + SAI
(2)
in which the symbols have the following meaning: μB - Bohr
magneton, g - g tensor, A - hyperfine tensor, S - electron spin operator,
I - nuclear spin operator and Bo - magnetic field vector. The parameter
Δg∥/Δg⊥, being a measure of the axial distortion of the VO2+ site was
derived by eq 3
Δg /Δg⊥ = (g − ge)/(g⊥ − ge)
(3)
in which g∥ and g⊥ are the parallel and perpendicular components of
the axial g tensor and ge is the g value of the free electron (ge =
2.0023). The in-plane delocalization coefficient β2*2 was calculated
according to eq 4,15 in which A∥ and A⊥ are the principal components
of the hyperfine structure (hfs) tensor and P = 184.5 G is the term for
the dipole−dipole interaction of the magnetic moment of the electron
and the nucleus for the free V(IV) ion.16
β *2 = 7/6Δg − 5/12Δg − 7/6[(A − A⊥)/P]
2
(4)
⊥
Catalyst Characterization. Elemental compositions were deter-
mined by ICP-OES and CHN analysis. ICP-OES measurements were
performed with a Varian 715-ES spectrometer. Ten mg of the catalyst
was dissolved in 4 mL HF and 4 mL aqua regia. These solutions were
treated in a microwave (Anton Paar/PerkinElmer) at ca. 60 bar and
120 °C, diluted to 100 mL, and measured. CHN analyses of 10−30 mg
powder samples were performed with an EA 1110 CHN analyzer (CE
Instrumenta) using a tin crucible. All values were verified by double
determination. BET surface areas (SBET) were measured by krypton
adsorption at 77 K using an ASAP2010 setup.
Powder XRD patterns were recorded with Cu Kα1 radiation on a
Stoe STADI P diffractometer. Processing and assignment of the
powder patterns was done using the software Win X Pow (Stoe) and
the powder diffraction file (PDF) database of the International Center
of Diffraction Data (ICDD).
X-ray photoelectron spectra (XPS) were recorded on a VG
ESCALAB 220iXL instrument with Mg Kα radiation. The peaks
were fitted by Gaussian−Lorentzian curves after a Shirley background
subtraction. The electron binding energies EB were referenced to the
carbon 1s peak at 284.8 eV. For quantitative analysis, the peak areas
were determined and divided by the element-specific Scofield factors
and the analysator-depending transmission function.
FTIR spectra at ambient conditions were recorded in the attenuated
total reflection (ATR) mode using an Alpha-P FTIR spectrometer
(Bruker).
RESULTS AND DISCUSSION
■
Catalytic Performance. Conversion of 3-PIC (X3‑PIC) and
selectivity of 3-CP (S3‑CP) have been measured at two different
space velocities (GHSV = 2713 h−1 and 5728 h−1) as a function
of the experimental bulk V/(Al+Zr) ratio determined by ICP
(Figure 1 A and B). In general, catalysts have been tested on
stream for about 6 h. No deactivation was observed during this
time. The experimental error was 2% for selectivity and even
lower for conversion values. The carbon balance was always in
the range of 96−101%. The bed temperature TB was adjusted
to 360 °C for both space velocities. As expected, the higher the
space velocity the lower X3‑PIC and the higher S3‑CP. In general,
almost equal trends are observed. The X3‑PIC values pass a
maximum at a V/(Al+Zr)ICP ratio of around 0.5, while the S3‑CP
increases almost linearly with rising V/(Al+Zr)ICP ratio for both
high and low GHSV. The best catalytic performance, judged by
the yield Y3‑CP, was obtained with catalyst VZrAlON-0.5 while
sample VZrAlON-0.1 showed the lowest performance. The
marked drop of conversion for the catalyst with V/(Al+Zr) =
0.6 is most probably due to its much lower surface area (SBET
=
55.4 m2 g−1) in comparison to that of the other catalysts (Table
1). Therefore, detailed characterization studies described below
have been focused on these two catalysts.
EPR spectra in X-band (ν ≈ 9.5 GHz) were recorded on an
ELEXSYS 500-10/12 cw spectrometer (Bruker) using a microwave
2688
dx.doi.org/10.1021/cs500458w | ACS Catal. 2014, 4, 2687−2695

ACS Catalysis
Research Article
values measured for the best catalyst of the ternary VZrAlON
series studied in this work. As already described in ref 9, the
VZrON-0.25 catalyst shows a much higher 3-PIC conversion
but lower 3-CP selectivity than catalyst VAlON-0.5. For this
reason the space velocity had to be raised to 5728 h−1 for
VZrON catalysts, to control the exothermicity of the reaction
and to minimize undesired total oxidation to COx.
Interestingly, addition of Zr to VAlON-0.5 improves the
activity but lowers also the selectivity, as can be seen from a
comparison of VAlON-0.5 and VZrAlON-0.5 under the same
reaction conditions at TB = 360 °C and GHSV = 2713 h−1
(Figure 2). However, at almost equal conversion (82% for
VZrAlON-0.5 and 80% VAlON-0.5) which was reached at the
same bed temperature of 360 °C but with different space
velocities, VZrAlON-0.5 is only marginally less selective than
VAlON-0.5. Interestingly, for VZrAlON-0.5 a 2-fold higher
GHSV could be applied, which enabled a markedly higher STY
without significant loss of selectivity compared to that with
VAlON-0.5. Catalysts VZrAlON-0.5 and VZrON-0.25 reach
almost the same conversion (96 vs 94%) at a GHSV of 5728
h−1. However, a higher bed temperature of 382 °C was needed
for sample VZrAlON-0.5 compared to that for catalyst VZrON-
0.25 (360 °C), since the former catalyst is less active (but more
selective).
Nevertheless, a remarkable improvement of the space-time
yield (STY) was achieved with the ternary VZrAlON-0.5
catalyst. Its maximum value of 488 g L−1 h−1 is now the highest
STY ever measured in gas-phase ammoxidation of 3-PIC to 3-
CP and markedly exceeds our previous result of STY = 399 g
L−1 h−1 obtained for the binary VZrON-0.25 catalysts9. To
elucidate possible structural reasons for these differences, the
best and the worst ternary VZrAlON catalysts have been
comprehensively characterized ex situ before and after use in
the catalytic reaction as well as by in situ EPR during
ammoxidation. Additionally, the latter technique was also
applied to follow nitridation of the respective oxide precursors
in NH3 flow. The obtained results are discussed in comparison
to those of the binary VZrON and VAlON materials already
published in ref 9.
Figure 1. Catalytic performance of VZrAlON catalysts measured at a
bed temperature of TB = 360 °C and space velocities of GHSV = 5728
h−1 (A) and 2713 h−1 (B) as a function of the V/(Al+Zr)ICP ratio.
Table 1. Bulk and Surface V/(Al+Zr), Al/Zr, and N/V
Elemental Ratios, N Contents and Specific Surface Areas
a
(SBET) of the Fresh and Used Catalysts
SBET
N [wt [m2/
sample
V/(Al+Zr)
0.15 (0.10)
Al/Zr
N/V
%]
g]
VZrAlON-
0.1 fresh
1.55 (1.99)
0.68 (0.00)
1.51
199
VZrAlON-
0.1 used
0.16 (0.09)
0.54 (0.20)
0.52 (0.28)
1.54 (1.85)
1.51 (1.96)
1.52 (2.17)
0.44 (0.70)
0.54 (0.35)
0.34 (0.32)
0.84
3.07
1.84
167
153
110
VZrAlON-
0.5 fresh
VZrAlON-
0.5 used
a
Surface ratios derived by XPS are given in brackets. Bulk ratios and N
contents were obtained from ICP-OES and CHN analysis,
respectively.
In Figure 2, conversion and selectivity for the best catalysts
from the binary VAlON and VZrON series (values for VZrON-
0.25 and VAlON-0.5 taken from ref 9) are compared with the
Bulk and Surface Composition and Structure of
VZrAlON-0.1 and VZrAlON-0.5 Catalysts. Both catalysts
do not show any reflections in the XRD powder patterns
neither before nor after use in the ammoxidation reaction. This
indicates that they are X-ray-amorphous and the incorporation
of Zr does not lead to crystallization of ZrO2 as observed
previously in binary VZrON samples,9 but rather to a well
intermixed Zr−O−Al network. The bulk properties and surface
elemental ratios for fresh and used VZrAlON samples are listed
in Table 1.
It is obvious that the V/(Al+Zr) ratios on the surface are
lower than those in the bulk, indicating that vanadium is
partially enriched in the volume of the catalysts. The opposite
trend is observed for the Al/Zr ratios, suggesting that Al sites
are more abundant on the surface. However, both the V/(Al
+Zr) as well as the Al/Zr ratios do not change much for both
catalysts after use in the catalytic reaction, neither in the bulk
nor on the surface. More significant are differences related to
the nitrogen content. It can be seen that the amount of
incorporated nitrogen increases with the total V content
suggesting that V and N sites might be related to each other,
yet about 40% of the nitrogen incorporated during the
nitridation process is lost after ammoxidation catalysis for
both fresh catalysts. Interestingly, the surface N/V ratio stays
Figure 2. Conversion (X3‑PIC) and selectivity (S3‑CP) at different bed
temperatures (TB) and space velocities (GHSV) for the VZrON-0.25,
VZrAlON-0.5, and VAlON-0.5 catalysts. Values of VZrON-0.25 and
VAlON-0.5 are taken from ref 9.
2689
dx.doi.org/10.1021/cs500458w | ACS Catal. 2014, 4, 2687−2695

ACS Catalysis
Research Article
Table 2. Binding energies EB [eV] and surface percentages (given in brackets) for fresh and used VZrAlON catalysts derived
from XPS
a
a
b
c
sample
Al 2p
Zr 3d3/2
N 1s
V 2p3/2
ΔEB
mean V valence
VZrAlON-0.1 fresh
VZrAlON-0.1 used
VZrAlON-0.5 fresh
72.8 (19.5)
73.8 (15.5)
73.8 (12.7)
181.1 (12.3)
181.8 (10.8)
182.0 (10.9)
−
517.1 (3.6)
516.6 (2.8)
516.8 (3.8)
518.2 (2.7)
13.6
13.3
14.3
12.9
4.3
4.5
3.8
4.8
399.5 (2.0)
397.6 (1.2)
401.0 (0.6)
403.4 (0.6)
399.2 (1.2)
400.8 (1.0)
VZrAlON-0.5 used
74.7 (8.0)
182.5 (7.5)
514.8 (2.6)
517.4 (4.1)
14.8
12.2
3.4
5.2
a
b
c
Values in brackets denote the ratio between different species of the same element. ΔEB = EB(O 1s) − EB(V 2p3/2). Derived from ΔEB (see ref
31).
almost the same before and after use for VZrAlON-0.5. In
contrast, no surface N sites are detected in fresh VZrAlON-0.1.
This may be due to the fact that the surface V content of
sample VZrAlON-0.1 is much smaller than that of VZrAlON-
0.5. Thus, the number of V sites on the surface of VZrAlON-0.1
might be too small for stabilizing sufficient N surface sites that
exceed the detection limit of XPS. Interestingly, surface N sites
in sample VZrAlON-0.1 become abundant after usein
contrast to the binary VZrON catalysts, in which no surface
N species could be detected neither before nor after their use in
the catalytic reaction.9 This suggests that the incorporation of
Al in the surface of VZrAlON catalysts favors the incorporation
of surface N species, probably in the form of V−N−Al bridges
as found accordingly in binary VAlON catalysts.9 Moreover,
different reactions involving N sites of the oxynitride may
compete during ammoxidation: (1) N sites move from the bulk
to the surface and are partially released as N2 or transformed to
surface NHx (x = 1−4) sites and (2) N sites are incorporated
into the catalyst surface from gaseous NH3 in the feed.
Moreover, a certain loss of the SBET surface area is observed for
both samples after use (see Table 1).
binding energies of ZrN (397.3−397.6 eV),22,23 AlN (396.5
eV)18,19 and VN (397.4 eV),24 site N1 is assigned to a nitride-
like species N3−. Species N2 with EB values of 400 to 401 eV
represent most likely NHx (x = 1 or 2) sites25−27 while species
28
+
N3 might be assigned to NH4 . In addition, the presence of
NH4+ species has been confirmed by FTIR-ATR measurements
showing a characteristic vibration at 1400 cm−1 (Figure S4 in
SI). After ammoxidation the nitride-like species (N1) are no
longer seen, and the surface of both used catalysts contains
merely NHx species reflected by EB values 399.2 and 400.8.
Simultaneously, the bulk N content decreases for both samples
after use (Table 1). This suggests that N3− species from the
bulk move to the surface under reaction conditions where a
part of them is transformed into surface NHx sites while
another part might be converted to molecular N2 as observed
accordingly also for several other metal oxynitrides.29,30 While
only one N site is seen in the used sample VZrAlON-0.1, two N
sites are detected in the used sample VZrAlON-0.5. This might
be due to their coordination to different V sites.
From Table 2 it is evident that the fresh VZrAlON-0.1
catalyst contains only one surface V site, while two of such
species are exposed in the fresh sample VZrAlON-0.5. It has
been shown previously for a variety of different vanadium oxide
samples that reliable information on the mean V valence state
can be obtained from the difference of the binding energies of
the O 1s and the V 2p3/2 electrons, ΔEB = EB(O 1s) − EB(V
2p3/2), rather than from EB(V 2p3/2) alone.31 Those ΔEB values
are shown in Table 2 together with the mean V valence state
derived from the correlation between ΔEB and V valence found
for VxOy samples in ref 31 (Figure S3 in SI).
It can be seen that the V surface species in the fresh sample
VZrAlON-0.1 has a mean valence of 4.3 and is slightly oxidized
during ammoxidation to V4.5. On the fresh catalyst VZrAlON-
0.5 a more reduced site V1 (V3.8) and an almost completely
oxidized site V2 (V4.8) coexist on the surface which get even
more reduced and oxidized, respectively, during ammoxidation.
This could be an indication that V4+ on the surface of this
catalyst is depleted under reaction conditions and is trans-
formed into V3+ and V5+. A further indication for this may also
be derived from in situ EPR measurements discussed below.
Based on this result, it is probable that the N species with the
lower EB value (399.2 eV) are connected with V3+ sites, while
the ones with EB = 400.8 eV are rather attached to V5+. On the
other hand, the single N species in catalyst VZrAlON-0.1 has
an intermediate binding energy of 399.5 eV suggesting its
location in the vicinity of V4+, which would agree with the
presence of the single V surface species detected in this catalyst
(Table 2).
At this point it is useful to consider the binding energies of
the different surface species in more detail (Table 2,
corresponding XP spectra shown in Figure S1 and S2 of
Supporting Information [SI]). The EB values of the Al sites in
the fresh catalysts are significantly lower than those reported for
Al2O3.17 In VZrAlON-0.5, this might be due to the presence of
nitride-like sites in the vicinity of Al (see Discussion below),
since the value of 73.8 eV is close to the one observed in AlN
(EB = 73.5−73.6).18,19 This is also probable for the used
VZrAlON-0.1 catalysts, since its surface is significantly
populated by N (Table 1). However, for the low Al 2p value
in the fresh VZrAlON-0.1 another reason must be responsible
since its surface is completely free of N sites. The Zr 3d3/2 value
of 181.1 eV in this catalyst is markedly lower than the one
found in ZrO2 (182.4 eV20). This suggests a partial reduction of
surface Zr4+ to lower-valent Zr species, e.g. Zr3+.21 A Zr3+
species is supposed to attract the electrons in a Zr−O−Al
junction weaker than a Zr4+ species, which might lead in turn to
the lower binding energy for the Al site in this sample. The Zr
3d3/2 binding energy for the used VZrAlON-0.1 catalyst (181.8
eV) is close to that of the fresh VZrAlON-0.5 sample (182.0
eV). It is higher than expected for Zr3+ but lower than for ZrO2.
Considering that the surface of the VZrAlON-0.5 catalyst
contains nitrogen, it may be explained by the partial formation
of Zr4+−N moieties.
Three different N surface sites with binding energies at 397.6
eV (N1), 401.0 eV (N2), and 403.4 eV (N3) are detected for
the fresh VZrAlON-0.5 sample. In agreement with N 1s
2690
dx.doi.org/10.1021/cs500458w | ACS Catal. 2014, 4, 2687−2695

ACS Catalysis
Research Article
In addition to XPS, EPR and UV−vis diffuse reflectance
spectra were recorded to obtain further information on
structure and valence states of bulk V sites in fresh VZrAlON
catalysts (Figure 3). In the EPR spectra, a multiline signal
Among the different possible valence states of vanadium,
only V4+ is detected by EPR under the given conditions.
Despite the higher V content, the total signal intensity of the
fresh VZrAlON-0.5 catalyst is much lower compared to
VZrAlON-0.1. This can either be due to a higher percentage
of V5+ or of V3+ in the former sample, both species being EPR-
silent.
To clarify this issue, UV−vis spectra were recorded (Figure
3B). It has been shown previously that V5+ species in oxide
catalysts give rise to strong charge-transfer bands in the low
wavelength range whereby the band position depends on the
coordination number as well as on the agglomeration degree of
the V site: CT bands of tetrahedral V5+O4 species appear at
lower wavelength than those of octahedral V5+O6 species and a
red shift is observed with increasing number of V−O−V
bridges.34 Thus, CT bands of single and polymerized V5+O4
species have been observed below 300 and between 300 and
350 nm, respectively, while those of single and weakly
polymerized octahedral V5+O6 species fall around 400 nm
respectively between 400 and 430 nm.35,36 For V2O5
nanocrystallites with the highest polymerization degree, light
absorption extends into the visible range up to 485 nm, due to
delocalized acceptor levels.37 Bands arising from d-d transitions
of reduced V4+ and V3+ species are usually broad and less
intense compared to CT bands of V5+ since such d-d transitions
are frequently symmetry-forbidden. Thus, d-d transitions of the
VO(H2O)52+ complex ion fall at 365, 609, and 840 nm38 while
they have been observed at 625 and 769 nm in V-silicalite.34a d-
d transitions of V3+ in vanadium phosphates occurred at 980,
694, and 444 nm in the case of V(PO3)339 as well as at 500 nm
in case of reduced (VO)2P2O7.40 From this consideration it is
readily evident that in V-containing solid catalysts a distinct
assignment of absorbance in the high wavelength range to
specific d-d transitions of V4+ and V3+ is very difficult when
both species superimpose in the UV−vis spectrum. This is the
reason why frequently only a more or less featureless increase
of absorbance is observed above 400 nm in partially reduced V-
containing solids.41 This is also evident from the UV−vis
spectra of VZrAlON catalysts in Figure 3B. The spectrum of
VZrAlON-0.1 is dominated by a strong charge-transfer band of
V5+ below 300 nm, pointing to the dispersed character of the V
sites, and a slight increase of absorbance above 400 nm which
confirms the presence of some reduced V species. In sample
VZrAlON-0.5 with the higher V content, yet the smaller EPR
signal (Figure 3A) light absorption is most pronounced in the
higher wavelength range above 400 nm which indicates the
presence of reduced V4+ and V3+ species. Although, as explained
above, a clear distinction of these both valence states just based
on UV−vis spectra is not possible, comparison of UV−vis and
EPR spectra suggests that the lower total EPR intensity of
sample VZrAlON-0.5 might not be related to V5+ (which
should contribute to the CT band intensity below 400 nm) but
to V3+, which contributes to the absorbance above 500 nm in
the UV−vis spectrum of Figure 3B but not to the
corresponding EPR spectrum in Figure 3A. This is confirmed,
too, by XPS for the catalyst surface (Table 2).
Figure 3. EPR (A) and UV−vis−DR spectra (B) of the fresh catalysts
VZrAlON-0.1 and VZrAlON-0.5.
resulting from the hyperfine structure (hfs) coupling of the
unpaired electron with the nuclear spin of a V4+ single site is
superimposed on a broad isotropic background signal which
arises from magnetically interacting V4+ species. The general
shape of the EPR spectra is similar to those of the binary
VAlON catalysts in which the V sites were found to be highly
dispersed, but it differs significantly from the spectra of the
binary VZrON catalysts in which hfs was hardly seen due to
dominating polymerized V species.9 This shows clearly that
incorporation of Al into the VZrON network increases the
percentage of highly dispersed single V sites. Another
interesting feature, which occurs only in the VZrAlON-0.5
but not in the VZrAlON-0.1 catalyst with the lower N content,
is a very narrow isotropic line at g = 2.00, seen particularly well
in the second derivative of the EPR spectrum (Figure 3a, inset).
This signal is characteristic for an anionic vacancy occupied by a
single electron (also known as F-center).32,33 Probably, such F-
centers are formed during nitridation since this reaction can be
formally understood as the replacement of three O2− anions by
two N3− anions leaving behind an anionic vacancy which can
then accommodate a single electron, possibly from a nearby V4+
species as proposed below (eq 5). In catalyst VZrAlON-0.1
with the lower N content, the amount of such defects is
obviously not yet high enough to create an EPR signal of
sufficient intensity.
The spin Hamiltonian parameters g∥ g⊥, A∥ and A⊥ derived
by spectra simulation for the single VO2+ sites (Table 3) can
provide information on differences of their local environment
in the two catalysts. The quotient Δg∥/Δg⊥ (eq 3) is a measure
for the axial distortion of the VO2+ site; the higher this value,
the shorter the axial VO bond and/or the longer the
equatorial V−O bonds of the VO2+ site. The in-plane
2691
dx.doi.org/10.1021/cs500458w | ACS Catal. 2014, 4, 2687−2695

ACS Catalysis
Research Article
Table 3. Spin Hamilton parameters g∥, g⊥, A∥, A⊥, Δg∥/Δg⊥ (eq 3) and In-Plane Delocalization Coefficients β2*2 (eq 4) for
Single VO2+ Derived from EPR Spectra Simulation (for a comparison of selected experimental and simulated spectra see SI,
Figure S5); Parameters for the Broad Isotropic Background and the F Center Signal Are Not Included
sample
A∥ [G]
A⊥ [G]
g∥
g⊥
Δg∥/Δg⊥
β2*2
VZrAlON-0.1
1.962
fresh, 20 °C
−176
−54
1.936
1.935
1.923
1.935
1.933
1.939
1.65
0.71
0.70
0.69
0.70
0.68
0.72
air/NH3, 350 °C
air/NH3/H2O/3-PIC, 350 °C
air/NH3, 350 °C
N2, 350 °C
−176
−187
−177
−177
−177
−55
−66
−55
−58
−53
1.968
1.978
1.968
1.970
1.962
1.94
3.20
1.94
2.14
1.58
N2, 20 °C
VZrAlON-0.5
fresh, 20 °C
site 1
−193
−180
−181 (−183)
−61
1.920
1.930
1.968
1.966
2.41
0.75
0.61
site 2
−72
2.01
a
a
a
a
a
a
air/NH3, 350 °C
air/NH3/H2O/3-PIC, 350 °C
air/NH3, 350 °C
N2, 350 °C
−64 (−65)
1.921 (1.932)
1.920
1.976 (1.981)
1.976
3.10 (3.30)
3.10
0.66 (0.67)
0.67
−183
−184
−176
−182
−64
−64
−63
−61
1.920
1.973
3.10
0.67
1.914
1.972
3.28
0.65
N2, 20 °C
1.921
1.973
2.73
0.69
a
Values in brackets taken from ref 9 for the binary VAlON-0.5 catalyst under the same conditions.
delocalization coefficient β2*2 (eq 4) reflects the extent to
which the single electron is delocalized away from the V center
toward the ligands in the basal plane of the VO2+ site. For a
pure VO2+ ion, β2*2 is equal to unity and decreases with
electron delocalization, i. e., with rising covalent character of the
V-ligand bonds in the basal plane.
While the experimental EPR spectrum of the fresh
VZrAlON-0.1 catalyst reflects only one type of single VO2+
site, two of such species S1 and S2 are present in sample
VZrAlON-0.5 (Table 3). The ratio of their relative intensities
S2/S1 = 1.6 is almost equal to the Al/Zr ratio. This could
suggest that site 1 may be connected with Zr, while site 2 may
be rather localized in the vicinity of Al. This is also supported
by the similarity of the spin Hamiltonian parameters of site 2
with those observed for single VO2+ species in the binary
VAlON-0.5 given in brackets in Table 3. Apart from the X-ray-
amorphous nature of the catalysts, this is another hint for the
presence of a well intermixed V−Al−N(O)−Zr network.
In Situ EPR Studies during Nitridation. To follow the
evolution of the V sites during nitridation, the oxide precursor
VZrAlO-0.5 was first heated in N2 flow from 20 to 120 °C
before switching to 20% NH3/N2. EPR spectra have been
recorded until reaching a steady state before the temperature
was ramped to the next level. From Figure 4 it can be seen that
the total EPR signal intensity increases strongly upon raising
the temperature from 120 to 300 °C and decreases again upon
further heating to 410 °C. This points to a gradual reduction of
EPR-silent V5+ in the fresh oxide precursor, first to EPR-active
V4+ at moderate temperature and further to EPR-silent V3+ at
410 °C. At this temperature, also the narrow F-center signal
appears, confirming that the formation of these vacancies is
indeed a consequence of increasing replacement of O by N.
Most likely, the F-centers are located in the vicinity of V5+
centers, since the magnetic interaction with the unpaired d-
electrons of V4+ and/or V3+ sites would broaden the line
beyond detection. Considering the fact that these F-centers are
only formed with increasing incorporation of N species and the
latter are themselves stabilized by V, it may be assumed that a
process depicted in eq 5 occurs during nitridation, in which ⊡
Figure 4. EPR spectra measured during nitridation of the VZrAlO-0.5
oxide precursor: (A) at 120 °C after heating the sample in N2 followed
by 1 h treatment in 20% NH3/N2 flow, (B) as a function of time after
reaching T = 300 °C in 20% NH3/N2 flow, and (C) as a function of
time after reaching T = 410 °C in 20% NH3/N2 flow.
denotes an oxygen vacancy occupied by a single electron (F-
center) and M is Al or Zr.
In Situ EPR Studies during Ammoxidation. To probe
the impact and role of paramagnetic sites in the oxynitrides
during the catalytic reaction, in situ EPR measurements of
samples VZrAlON-0.1 and VZrAlON-0.5 have been performed
2692
dx.doi.org/10.1021/cs500458w | ACS Catal. 2014, 4, 2687−2695

ACS Catalysis
Research Article
Figure 5. EPR spectra of the as-prepared catalysts VZrAlON-0.1 (A) and VZrAlON-0.5 (B) Recorded at RT and after heating in air/NH3 (5.6/1)
flow to the reaction temperature of 350 °C.
Figure 6. EPR spectra of VZrAlON-0.1 (A) and VZrAlON-0.5 (B) recorded at 350 °C under air/NH3 (black, identical with blue spectra in Figure 5)
and after subsequent treatment for 45 min in the total feed (3-PIC/air/NH3/H2O = 1/30/6/8) at 350 °C.
during subsequent treatment with feed components. The
catalysts have been first heated in air/NH3 (4.5/1) flow to the
reaction temperature of 350 °C (Figure 5). Then subsequently
H2O and 3-PIC have been added to the feed stream at 350 °C
(3-PIC/air/NH3/H2O = 1/28/5/8) (Figure 6). After reaching
a steady state, H2O and 3-PIC were removed from the feed
again.
A significant difference between the two catalysts is already
evident upon heating in air/NH3 flow to the reaction
temperature of 350 °C (Figure 5), since the total EPR signal
intensity decreases for VZrAlON-0.1, while it increases for
VZrAlON-0.5. In the former case, the loss of intensity might be
just a temperature effect, since Curie’s law predicts an intensity
ratio I(350 °C)/I(20 °C) of 0.5, which is indeed the case for
VZrAlON-0.1, suggesting that the concentration of V4+ in the
sample remains constant during heating in air/NH3. In
contrast, the EPR intensity of sample VZrAlON-0.5 increases
strongly (Figure 4b). This indicates that V3+ sites formed upon
nitridation are at least partially reoxidized in air/NH3 flow at
350 °C. Moreover, the spin Hamiltonian parameters in Table 3
indicate that the local structure of the VO2+ sites in VZrAlON-
0.1 remains almost unchanged during this treatment, while
significant changes of VO2+ sites are detected in VZrAlON-0.5.
This catalyst contains two different VO2+ single sites in its
initial state, but site 1 (tentatively assigned as attached to Zr) is
no longer seen after treatment in air/NH3, probably due to
2693
dx.doi.org/10.1021/cs500458w | ACS Catal. 2014, 4, 2687−2695

ACS Catalysis
Research Article
oxidation. In accordance with the strong intensity increase
(Figure 4b) and the similarity of spin Hamiltonian parameters
of site 2 with those of the binary VAlON-0.5 catalyst (values in
brackets in Table 3), in can be concluded that V3+ sites in the
vicinity of Al are reoxidized and stabilized as VO2+ while those
in the vicinity of Zr might be converted to VO3+. This agrees
well with previous observations on binary VZrON catalysts, in
which V sites were found to prefer a valence state of +5 under
reaction conditions.9 From Table 3 it is evident that the
quotient Δg∥/Δg⊥ of sites 2 increases during treatment in air/
NH3, indicating that their local geometry becomes more
distorted. This could suggest that already under these
conditions nitride-like sites are transformed to NHx sites as
also indicated by XPS results. In addition, also the
concentration of F-centers decreases. This is evident from the
ratio of the intensities of the subsignals needed to reproduce
the experimental EPR spectrum of catalyst VZrAlON-0.5 in air/
NH3 flow at 350 °C. The relative contribution of the F-center
signal to the total EPR intensity decreases by a factor of about
6. A possible reason may be that gas-phase oxygen takes up the
electrons in these vacancies upon adsorption and is converted
into O2− ions (eq 6) which fill the vacancies.42 The reverse
process would restore the F-centers and liberate oxygen which
then reacts with 3-PIC. This is the well-known Mars−van
Krevelen mechanism.
Lewis surface site followed by H abstraction from the CH3
group at the neighboring Vn+O surface site, which gets
reduced to V(n−1)+−OH. It is clear that such a step is not
probable on a V3+ surface site since reduction to V2+ might be
energetically unfavorable. Therefore, V3+ surface species can be
ruled out as active sites, as it was also shown experimentally.44
The conversion of −CH3 to −CHO and further to −CN
implies the participation of lattice O and N and, consequently,
requires several subsequent reduction/reoxidation cycles of the
involved vanadyl site (see Figure 9 in ref 37). It has been
concluded that the periodic electronic distortion of this vanadyl
site is diminished by delocalization of its electron density within
the respective vanadyl chain, which was considered as a major
reason for the beneficial catalytic effect of interconnected VO2+
structural units.
Given that the same reaction mechanism holds also for the
ammoxidation of 3-PIC on VZrAlON-0.5, the active sites
should provide functions similar to those in vanadyl
pyrophosphate, namely a Lewis site for adsorption of the
aromatic ring with a nearby redox-active vanadyl site that can
delocalize its fluctuating electron density. This may be realized
in a proper way by the −⊡−V5+(O)−N−Al(Zr)− moiety
formed upon nitridation in VZrAlON-0.5, but not in the poorly
active VZrAlON-0.1catalyst (eq 5). In the mechanism proposed
in ref 37 it is assumed that the methyl group of the substrate is
first oxidized to adsorbed aldehyde which subsequently reacts
e−
e−
2e−
+
O2 ⇆ O2 ⇆ O2 ⇆ 2O− [ooZ 2O2−
−
2−
with surface NH4 species. It is probable that N within the
(6)
surface −⊡−V5+(O)−N−Al(Zr)− moiety represents an NHx
species, which may react with an aldehyde intermediate
according to the mechanism in ref 37. The XPS results point
to a dynamic behavior of the N species during reaction which
may be promoted by H2O in the feed. It was found that, on the
one hand, N sites move from the bulk to the surface where they
are partially converted to surface NHx (x = 1−4) sites, and on
the other hand, such sites are formed by interaction of the
surface with gaseous NH3 in the feed.
A similar behavior was observed for F-centers in an amorphous
iron vanadate phase which catalyzed the selective gas-phase
oxidation of fluorene to 9-fluorenone very selectively. The high
selectivity was attributed to the fast conversion of nonselective
electrophilic oxygen intermediates to nucleophilic oxide ions
able to insert into the aromatic ring without destroying it.43
When H2O is added to the air/NH3 flow at 350 °C, the EPR
spectra did not change and are therefore not shown. However,
upon admixing 3-PIC, which results in the complete feed
mixture (3-PIC/air/NH3/H2O = 1/30/6/8), the EPR signal
rises strongly for sample VZrAlON-0.1 but only marginally for
sample VZrAlON-0.5 (Figure 6). This intensity increase is
most probably due to the reduction of V5+ to V4+ by 3-PIC
which itself is oxidized in turn. The spin Hamiltonian
parameters of the VO2+ site in sample VZrAlON-0.1 point to
an increasing distortion of the local environment, reflected by a
rise of the Δg∥/Δg⊥ ratio from 1.94 to 3.20, which is completely
reversible when H2O and 3-PIC are removed from the feed
again (Table 3). Given the fact that H2O caused no changes of
the VO2+ sites, this suggests that the (amm)oxidation product
of 3-PIC is adsorbed on the latter.
In contrast to VZrAlON-0.1, the total EPR intensity
increases only very slightly and the spin Hamiltonian
parameters of the VO2+ sites do not change at all (Table 3).
This indicates that the local environment of these sites remains
unchanged upon contact of the catalyst surface with 3-PIC,
probably, since they are not exposed on the surface. This would
agree with the XPS results, which show that the two kinds of
surface V species are mainly trivalent and pentavalent,
respectively, but not tetravalent (Table 2).
The F-center (⊡) can activate gas-phase oxygen (eq 6), Al
and/or Zr can act as Lewis adsorption sites, while the V5+O
can abstract H from CH3 in the initial step, forming V4+−OH.
V4+ can delocalize its electron to the neighboring vacancy
(restoring the F-center for another redox cycle) after O and N
leave their surface sites for reaction with the adsorbed
intermediate.
CONCLUSIONS
■
We have developed a new class of ternary VZrAlON
oxynitrides for the ammoxidation of 3-picoline to 3-
cyanopyridine which combine the beneficial properties of the
highly active but poorly selective VZrON and the highly
selective but poorly active VAlON binary counterparts, namely
an activity which is only slightly lower than that of VZrON and
a selectivity almost as high as that obtained with VAlON. With
the best VZrAlON-0.5 catalyst, the highest space-time yield
ever measured in gas-phase ammoxidation of 3-PIC to 3-CP
(STY = 488 g L−1 h−1) has been obtained, which is markedly
higher than our best previous result for VZrON-0.25 (STY =
399 g L−1 h−1)9 and all the more that of the industrial
SbVTiSiXO benchmark catalyst (STY = 150 g L−1 h−1).10 This
might be due to the following reasons:
Mechanistic Considerations. Previously it was found that
ammoxidation of methylaromatics proceeds very effectively on
vanadyl pyrophosphate that contains O-bridged ladder-like
double chains of VO2+ sites.37 The proposed reaction
mechanism implies the adsorption of the aromatic ring at a
In comparison to VZrON materials, which contain no surface
nitrogen and are dominated by crystalline ZrO2 and Vx5+Oy
clusters,9 the addition of Al to the ZrO matrix prevents
crystallization of ZrO2 and leads to an X-ray-amorphous well-
2694
dx.doi.org/10.1021/cs500458w | ACS Catal. 2014, 4, 2687−2695

ACS Catalysis
Research Article
(10) Hippel, L. v.; Neher, A.; Amtz, D. (Degussa-Huls AG). EP
0726092 B1 1999.
(11) Martin, A.; Zhang, Y.; Zanthoff, H. W.; Meisel, M.; Baerns, M.
Appl. Catal., A 1996, 139, L11−16.
(12) Janke, C.; Schneider, M.; Bentrup, U.; Radnik, J.; Martin, A.;
̈
mixed Al−O−Zr network which incorporates preferentially
highly dispersed vanadyl sites and exposes surface N sites. In
addition to a mean surface V valence close to +4, the latter two
features were previously related to the high selectivity of
VAlON catalysts9 and might be responsible, too, for the
improved selectivity of the present VZrAlON-0.5 catalyst. In
contrast to the less active VAlON-0.5 catalyst, which contained
V4+ and V3+ but no V5+ surface sites,9 the surface of VZrAlON-
0.5 exposes mainly V5+ and V3+ but no V4+. V3+ surface sites are
inactive in both cases. Thus, one reason for the improved
activity of the VZrAlON-0.5 catalyst might be, apart from the
somewhat higher surface area (SBET =153 m2 g−1 for VZrALON
vs 98 m2 g−1 for VAlON), the exposure of V5+ which has a
higher redox potential than V4+. However, in contrast to
VZrON, which exposes abundant V5+ besides some V4+ (mean
V valence = 4.7) mainly in the form of VxOy clusters,9 the active
V5+ surface sites on VZrALON-0.5 are separated by inactive V3+
species (mean V valence = 4.2) as evidenced by XPS. This
“dilution” might be a reason why total oxidation is suppressed.
It is anticipated that the junction of a V5+ site with an F-center
and a nearby N species on the surface of VZrAlON-0.5
provides optimum conditions for double Mars−van Krevelen
mechanism which requires both the activation of gas-phase
oxygen and ammonia via reversible incorporation into the
catalyst surface as well as an efficient electron transport, the
latter being realized by shuttling electrons between V5+, the F-
centers, and the reactants.
Scholz, G.; Bruckner, A. J. Catal. 2011, 277, 196−207.
̈
(13) Bruckner, A. Chem. Commun. 2005, 13, 1761−1763.
̈
(14) Lozos, G. P.; Hofman, B. M.; Franz, C. G. Quantum Chemistry
Program Exchange; 1973; 265.
(15) Novinska, K.; Wieckowski, A. Z. Phys. Chem, Neue Folge 1989,
162, 231−244.
(16) McGarvey, B. R. J. Phys. Chem. 1967, 71, 51−67.
(17) Hess, A.; Kemnitz, E.; Lippitz, A.; Unger, W. E. S.; Menz, D. H.
J. Catal. 1994, 148, 270−280.
(18) Liu, H.; Bertolet, D. C.; Rogers, J. W., Jr. Surf. Sci. 1995, 340,
88−100.
(19) Rosenberger, L.; Baird, R.; McCullen, E.; Auner, G.; Shreve, G.
Surf. Interface Anal. 2008, 40, 1254−1261.
(20) Chris, B. V. In Handbooks of Monochromatic XPS Spectra; XPS
International LLC: Mountain View, 2005; Vol. 2, p 828.
(21) Wang, Q. M.; Gong, J.; Sun, C.; Lee, H. W.; Kim, K. H. J.
Ceram. Proc. Res. 2011, 12, 259−264.
(22) Prieto, P.; Galan
395−399.
(23) Milose
Interface Anal. 1996, 24, 448−458.
́
, L.; Sanz, J. M. Surf. Interface Anal. 1994, 21,
̌
̌ ̌
v, I.; Strehblow, H. H.; Gaberscek, M.; Navinsek, B. Surf.
(24) Hendrickson, D. N.; Hollander, J. M.; Jolly, W. L. Inorg. Chem.
1969, 8, 2642−2647.
(25) Wiame, H.; Bois, L.; Lharidon, P.; Laurent, Y.; Grange, P. Solid
State Ionics 1997, 101−103, 755−759.
(26) Truica-Marasescu, F.; Wertheimer, M. R. Plasma Proc. Polym.
2008, 5, 44−57.
ASSOCIATED CONTENT
* Supporting Information
■
(27) Ando, R. A.; do Nascimento, G. M.; Landers, R.; Santos, P. S.
Spectrochim. Acta, Part A 2008, 69, 319−326.
(28) Florea, M.; Prada Silvy, R.; Grange, P. Appl. Catal., A 2005, 286,
1−10.
S
XPS spectra and data, ATR-FTIR spectra, calculated EPR
spectra. This material is available free of charge via the Internet
at http://pubs.acs.org.
(29) Wiame, H.; Centeno, M.-A.; Picard, S.; Bastians, P.; Grange, P. J.
Eur. Ceram. Soc. 1998, 18, 1293−1299.
AUTHOR INFORMATION
Corresponding Author
*E-mail: angelika.brueckner@catalysis.de
■
(30) Le Gendre, L.; Marchand, R.; Laurent, Y. J. Eur. Ceram. Soc.
1997, 17, 1813−1818.
(31) Mendialdua, J.; Casanova, R.; Barbaux, Y. J. Electron Spectrosc.
Relat. Phenom. 1995, 71, 249−261.
Notes
(32) Cooke, D. W.; Blair, M. W.; Smith, J. F.; Bennett, B. L.;
Jacobsohn, L. G.; McKigney, E. A.; Muenchausen, R. E. IEEE Trans.
Nucl. Sci. 2008, 55, 1118−1122.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
(33) Schimpf, S.; Lucas, M.; Mohr, C.; Rodemerck, U.; Bruckner, A.;
̈
The authors thank Deutsche Forschungsgemeinschaft (Grant
No. BR 1380-15-1) and Max-Buchner-Forschungsstiftung for
financial support.
Radnik, J.; Hofmeister, H.; Claus, P. Catal. Today 2002, 72, 63−78.
(34) Burcham, L. J.; Deo, G.; Gao, X.; Wachs, I. E. Top. Catal. 2000,
11−12, 85−100.
(35) Centi, G.; Perathoner, S.; Trifiro, F.; Aboukais, A.; Aissi, C. F.;
Guelton, M. J. Phys. Chem. 1992, 96, 2617−2629.
REFERENCES
(1) Grasselli, R. K.; Trifiro,
Catalysis, 2nd ed.; Ertl, G., Knozinger, H., Schuth, F., Weitkamp, J.,
Eds.; Wiley-VCH, Weinheim, 2008; Vol. 7, p 3517.
■
(36) Bruckner, A.; Rybarczyk, P.; Kosslick, H.; Wolf, G.-U.; Baerns,
̈
̀
F.; Forni, L. In Handbook of Heterogeneous
M. Stud. Surf. Sci. Catal. 2002, 142, 1141−1148.
̈
̈
(37) Morey, M.; Davidson, A.; Eckert, H.; Stucky, G. Chem. Mater.
1996, 8, 486−492.
(2) Lucke, B.; Martin, A. In Fine Chemicals through Heterogeneous
̈
(38) Ballhausen, C. J.; Gray, H. B. Inorg. Chem. 1962, 1, 111−122.
(39) Rojo, J. M.; Mesa, J. L.; Lezama, L.; Rojo, T.; Olazcuaga, R.;
Guillen, F. Ann. Chim. Sci. Mater. 1998, 23, 107−l 11.
Catalysis, 1st ed.; Sheldon, R., van Bekkum, H., Eds.; Wiley-VCH,
Weinheim, 2001, 527.
(3) Centi, G.; Perathoner, S. Catal. Rev. Sci. Eng. 1998, 40, 175−208.
̀
(40) Cavani, F.; Ligi, S.; Monty, T.; Pierelli, F.; Trifiro, F.; Albonetti,
(4) Martin, A.; Hannour, F. K.; Bruckner, A.; Lucke, B. React. Kinet.
̈
̈
S.; Mazzoni, G. Catal. Today 2000, 61, 203−210.
Catal. Lett. 1998, 63, 245−251.
(41) Bruckner, A.; Kondratenko, E. Catal. Today 2006, 113, 16−24.
̈
(5) Martin, A.; Lucke, B. Catal. Today 2000, 57, 61−70.
̈
(42) Bielanski, A.; Haber, J. Oxygen in Catalysis; Marcel Dekker: New
York, 1991; p 150.
(6) Narayana, K. V.; Venugopal, A.; Rama Rao, K. S.; Venkat Rao, V.;
Masthan, S. K.; Kanta Rao, P. Appl. Catal., A 1997, 150, 269−278.
(7) Sai Prasad, P. S.; Lingaiah, N.; Khaja Masthan, S.; Rama Rao, K.
S.; Kanta Rao, P. Catal. Lett. 1996, 36, 195−199.
(43) Bruckner, A.; Wolf, G. U.; Meisel, M.; Stosser, R.; Mehner, H.;
Majunke, F.; Baerns, M. J. Catal. 1995, 154, 11−23.
(44) Bruckner, A. Catal. Rev. Sci. Eng. 2003, 45, 97−150.
̈
(8) Martin, A.; Narayana Kalevaru, V.; Lucke, B.; Bruckner, A. Appl.
̈
̈
Catal., A 2008, 335, 196−203.
(9) Janke, C.; Radnik, J.; Bentrup, U.; Martin, A.; Bruckner, A.
̈
ChemCatChem. 2009, 1, 485−491.
2695
dx.doi.org/10.1021/cs500458w | ACS Catal. 2014, 4, 2687−2695