An oxygen-deficient perovskite as selective catalyst in the oxidation of alkyl benzenes

Article abstract of DOI:10.1002/anie.201007941

Extraordinary reaction rates and selectivity for the conversion of alkyl aromatic compounds, such as p-xylene, in terephthalic acid were observed for a heterogeneous perovskite catalyst of composition (La,Sr)_0.5(Mn,Co) _0.5O_2.38 (see picture). The oxygen-deficient catalyst was prepared by reduction of the oxygen-stoichiometric parent compound. Copyright

Full text of DOI:10.1002/anie.201007941

DOI: 10.1002/anie.201007941
Perovskite Catalysts
An Oxygen-Deficient Perovskite as Selective Catalyst in the Oxidation
of Alkyl Benzenes**
A. Aguadero, H. Falcon, J. M. Campos-Martin, S. M. Al-Zahrani, J. L. G. Fierro, and
J. A. Alonso*
The aerobic oxidation of hydrocarbons is a very important
commercial process for the production of oxygen-containing
compounds such as alcohols, aldehydes, ketones, carboxylic
acids, and epoxides.[1–4] However, the reaction conditions are
often harsh, the reagent mixtures are corrosive (ClÀ or BrÀ),
and the reactions are frequently unselective. Therefore, the
selective transformation of hydrocarbons, especially saturated
hydrocarbons such as alkanes, to valuable oxygenated com-
pounds constitutes an extremely important area of the
contemporary industrial chemistry. Today, over 90% of
organic chemicals are derived from petroleum, whose main
components are saturated hydrocarbons.
An innovation in the aerobic oxidation of hydrocarbons
through catalytic carbon radical generation under mild
conditions has been achieved through the use of N-hydroxy-
phthalimide (NHPI) as a key compound, albeit requiring the
presence of a transition-metal compound.[5–11] In this oxida-
tion, the phthalimide N-oxyl radical (PINO) generated in situ
abstracts the hydrogen atom from the hydrocarbon to form an
alkyl radical, which is readily trapped by dioxygen to form
oxygenated compounds. The NHPI method is applicable in a
broad variety of organic syntheses via carbon radical inter-
mediates.[5–11] A key challenge is to develop an inexpensive
heterogeneous catalyst for liquid-phase oxidation of alkyl
aromatics with an oxygen containing gas with high conversion
rate and selectivity.
These phases have been usually described with oxygen
stoichiometries equal to or higher than 2.5 in compositions
ranging from ABO3 to ABO2.5. It is generally observed that
these deviations of stoichiometry can be supported almost
without changing the building blocks of the perovskite
structure even though the anionic defects may form disor-
dered or vacancy-ordered structures. Only a few members of
the wide family of perovskite oxides have been obtained with
values of d > 0.5, in which the three dimensional array of the
perovskite structure is maintained; most of them are Co-
based perovskites of the Sr1ÀyBayCoO3Àd family.[14–17]
In this work we have developed a new oxygen-defective
(La,Sr)0.5(Mn,Co)0.5O3Àd heterogeneous catalyst with high
conversion and selectivity in the heterogeneous oxidation of
alkyl aromatics. The obtained perovskite presents a wide
range of oxygen stoichiometries and is ferromagnetic at room
temperature.
The oxygen-stoichiometric (La,Sr)0.5(Mn,Co)0.5O3 perov-
skite oxide was obtained as a well-crystallized black poly-
crystalline powder. Figure 1 displays the thermogravimetric
curve corresponding to the reduction process to obtain the
hypostoichiometric phase; a weight loss of d = 0.62 oxygen
atoms per formula unit was observed. The XRD patterns of
the stoichiometric and hypostoichiometric samples are
included in Figure S1a,b in the Supporting Information.
The crystal structure of the oxidized sample was success-
ꢀ
Oxygen-deficient ABO3Àd perovskites have been reported
fully Rietveld-refined in the rhombohedral R3c space group
to present high catalytic activity in oxidation reactions.[12,13]
with oxygen content of 2.98(2) per formula unit (Figure 2a)
from neutron powder diffraction (NPD) data, revealing that
ꢀ
the material is not ordered in a cubic Fm3m space group as
previously stated from X-ray diffraction results.[18,19] As for
[*] Dr. A. Aguadero, Prof. Dr. J. A. Alonso
Instituto de Ciencia de Materiales de Madrid, CSIC
Cantoblanco, 28049 Madrid (Spain)
Fax: (+34)91-372-0623
E-mail: ja.alonso@icmm.csic.es
Homepage: http://www.icmm.csic.es/jaalonso
Dr. H. Falcon, Dr. J. M. Campos-Martin, Prof. Dr. J. L. G. Fierro
Sustainable Energy and Chemistry Group (EQS)
Instituto de Catꢀlisis y Petroeloquꢁmica, CSIC
Marie Curie, 2 Cantoblanco, 28049 Madrid (Spain)
Prof. Dr. S. M. Al-Zahrani
Chemical Engineering Department, College of Engineering
King Saud University, Riyadh (Kingdom of Saudi Arabia)
[**] We thank our research sponsor The King Saud University, Riyadh
(Saudi Arabia). We acknowledge the financial support of the
Spanish Ministry of Innovation and Science for the projects
(MAT2010-16404, CT02010-19157-C03-01). We are grateful to
Institut Laue-Langevin for making facilities available. A.A. thanks
MICINN for a “Juan de la Cierva” contract.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201007941.
Figure 1. Thermogravimetric analysis under 5%H2/95%N2 flow of the
stoichiometric (La,Sr)0.5(Mn,Co)0.5O3.
Angew. Chem. Int. Ed. 2011, 50, 6557 –6561
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6557

Communications
The re-oxygenation of the sample happens in at least two
subsequent processes (Figure S3 in the Supporting Informa-
tion) with concomitant oxidation of the transition-metal
cations from an average oxidation state of + 2.54 for the
partially oxidized sample to + 3.42 in the intermediate stage
and + 3.56 in the totally oxidized sample. These oxidation
states were confirmed by XPS analysis (Figure S4 and
Table S1 in the Supporting Information).
The reduction of the sample is mirrored by the shift of the
binding energy of Mn2p3/2 and Co2p3/2 core levels from 642.0
to 641.2 eVand 781.1 to 779.2 eV, respectively. Figure 3 shows
Figure 3. Magnetization versus field isotherms at T=4 K and 300 K for
the reduced phase.
Figure 2. Observed (circles), calculated (full line), and difference
(bottom) NPD Rietveld profiles for a) (La,Sr)0.5(Mn,Co)0.5O2.98(2) (air-
ꢀ
prepared) at room temperature, after the refinement in the R3c space
the magnetization isotherms of the reduced sample at 4 and
300 K; both curves are characteristic of a ferromagnet with a
Curie temperature above room temperature. The oxidized
sample is ferromagnetic below 225 K as previously
reported;[18,19] it seems that the reduction of the sample
promotes the increment of the magnetic interactions and the
Curie temperature, despite the presence of a large number of
oxygen vacancies that would perturb the double exchange
À
group; (Mn,Co) O distance 1.9280(5) ꢂ; b) (La,Sr)0.5(Mn,Co)0.5O2.52(2)
(reduced) at 423 K, refined in the Pbnm space group; average distance
À
(Mn,Co) O=1.948(1) ꢂ. Vertical lines correspond to the Bragg posi-
tions for pure compound (top) and LaSrMnO4 secondary phase
ꢀ
(bottom). The insets illustrate the rhombohedral R3c and orthorhom-
bic Pbnm crystal structures for the oxidized and reduced perovskite
oxides.
À À
through (Mn,Co) O (Mn,Co) paths.
the reduced compound, the NPD study reveals a slight
deviation from the cubic symmetry that can be structurally
defined in the Pbnm space group, with a random distribution
of the metals over the A and B positions of the perovskite and
oxygen vacancies (Figure 2b). A minor LaSrMnO4 impurity
with K2NiF4-type structure was observed. It is worth noting
that the oxygen-hypostoichiometric perovskite is easily
oxidized even at ambient conditions, where it superficially
uptakes some oxygen; for this reason the NPD data collected
at room temperature showed an admixture with the oxidized
phase (Figure S2 in the Supporting Information). However,
the NPD pattern collected at 423 K under vacuum conditions
corresponded to a single reduced perovskite phase with slight
amount of LaSrMnO4 as secondary phase. The quantity of
oxygen refined in the structure is O2.52(2) with high values of
isotropic thermal factors (2.7 (2) ꢀ2 for O1 and 5.3 (1) ꢀ2 for
O2), which might indicate a huge mobility of oxide ion in this
structure; standard values of thermal displacements of oxygen
in perovskite oxides are around 1 ꢀ2 at room temperature.
These materials have been used as catalysts in the
oxidation of several alkyl aromatic compounds. Very inter-
esting results were obtained in the oxidation of p-xylene. This
reaction was carried out using various catalytic systems—
(La,Sr)0.5(Mn,Co)0.5O3Àd/NHPI,
(La,Sr)0.5(Mn,Co)0.5O3/
NHPI—and, for comparative purposes to benchmark the
present state of the art, we employed Co(OAc)2/Mn(OAc)2/
NHPI, which is the best catalytic system described in
literature for soft reaction conditions.[6–8] The p-xylene con-
version was 100% for all the systems (Table 1). However, the
product distribution is highly affected by the catalyst used. In
all cases, three main products were obtained: p-toluic acid
(p-TOA), 4-carboxybenzaldehyde, and the desired product
terephthalic acid (TPA). TPA is the industrial product of
interest because it is one of the reactants used in the
production of PET (polyethylene terephthalate). The most
interesting results were obtained with the hypostoichiometric
sample, which promotes a great yield to terephthalic acid
(99%) produced with only 0.9% of p-toluic acid (Table 1).
6558
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 6557 –6561

Table 1: Products of the oxidation of p-xylene for 5 h.
The large difference in reactivity towards TPA formation
of both perovskite catalysts seems to be related to their ability
to favor the oxidation of p-TOA and thus it should be related
to their oxygen vacancy concentration and the oxidation
states of Co and Mn. The hypostoichiometric perovskite
((La,Sr)0.5(Mn,Co)0.5O3Àd) is able to accommodate a wide
variety of oxygen stoichiometries and therefore different
oxidation states for Co and Mn, favoring the charge transfer
in the oxidation reaction. In contrast, the presence of a high
number of oxygen vacancies could favor the activation of the
NHPI to form the phthalimide N-oxyl radical (PINO),
promoting the absorption of the hydrogen atom from the
hydrocarbon to form an alkyl radical, which reacts with O2
molecules to afford oxygenated compounds.[6–11]
The hypostoichio-
Catalyst
p-Xylene
Product
distribution [%]
p-TOA
conversion [%]
TPA
NHPI/Co(OAc)2/Mn(OAc)2
NHPI/(La,Sr)0.5(Mn,Co)0.5O3
NHPI/(La,Sr)0.5(Mn,Co)0.5O3Àd
100.0
100.0
100.0
19.0
33.0
0.9
81.0
67.0
99.1
Figure 4 illustrates the temporal product distribution
obtained by examining the concentration time course profiles
of each product. The p-xylene consumption rate was high and
similar for all catalysts, and no p-xylene was detected at
reaction times over 1 h. The concentration of the intermediate
metric perovskite was
also very efficient in
the aerobic oxidation
of other alkyl aromatics
(toluene or ethylben-
zene; Table 2). In both
cases, the conversion of
alkyl aromatics was very
high (> 99.5%). The
selectivity in the oxida-
tion of toluene was 98%
and 2% to benzoic acid
and
respectively.
benzaldehyde,
Surpris-
ingly, the oxidation of
ethylbenzene produced
mainly benzoic acid
(63%) and acetophe-
none (37%); this prod-
uct distribution is differ-
ent from that usually
published, whereby the
main product is aceto-
Figure 4. Concentration profiles of the main components identified in the oxidation of p-xylene at 20 bar O2 and
363 K using: a) NHPI/Co(OAc)2/Mn(OAc)2, b) NHPI/(La,Sr)0.5(Mn,Co)0.5O3, and c) NHPI/(La,Sr)0.5(Mn,Co)0.5O3Àd
catalytic systems.
p-TOA product increases quickly at short reaction times; in
parallel, the p-xylene consumption reaches a maximum in a
time interval of 0.5–1 h and then decreases for reaction times
longer than 1 h. This behavior is typical of consecutive
reactions in which one or more intermediate products are
formed before they are transformed into the final one, namely
TPA in this case. The peak concentration and the shape of
p-TOA and TPA profiles depend on the catalyst used. The
p-TOA concentration decreases quickly on the oxidized
(La,Sr)0.5(Mn,Co)0.5O3 catalyst but more slowly than observed
with the homogenous Co(OAc)2/Mn(OAc)2 reference cata-
lyst.
phenone, a stable product against soft oxidation condi-
tions.[20,21] These results indicate again that the hypostoichio-
metric perovskite is more effective than standard catalytic
systems for the oxidation of organic substrates, because with
NHPI/(La,Sr)0.5(Mn,Co)0.5O3Àd we were able to continue the
oxidation of ethylbenzene beyond acetophenone to yield
benzoic acid.
The hypostoichiometric perovskite was reused twice in
the oxidation of ethylbenzene; the conversion rate was
virtually unchanged, and the selectivity continued to be
reasonably good (Figure S5). XRD analysis of the used
sample (Figure S1c) showed a perovskite structure with
unit-cell parameters corresponding to the reduced specimen.
In conclusion, we have prepared an extremely oxygen-
defective perovskite oxide of composition (La,Sr)0.5-
(Mn,Co)0.5O2.38 by topotactic reduction of an oxygen-stoi-
chiometric specimen (La,Sr)0.5(Mn,Co)0.5O3. The crystal struc-
tures of both materials were studied by NPD analysis; the
structure of the reduced oxide presents a subtle orthorhombic
distortion, and the metals as well as the oxygen vacancies are
distributed randomly over the corresponding sublattices. The
Interestingly,
the
hypostoichiometric
(La,Sr)0.5-
(Mn,Co)0.5O3Àd catalyst counterpart displays the highest
p-TOA disappearance rate and subsequently the highest
TPA formation rate. According to these profiles, the forma-
tion rate of TPA follows the trend: (La,Sr)0.5(Mn,Co)0.5O3 <
Co(OAc)2/Mn(OAc)2 < (La,Sr)0.5(Mn,Co)0.5O3Àd
.
These
results clearly indicate that the hypostoichiometric perovskite
facilitates the global oxidation process; summing up, it is a
more efficient oxidizing catalyst.
Angew. Chem. Int. Ed. 2011, 50, 6557 –6561
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6559

Communications
Magnetic measurements were carried out in a Quantum-Design
SQUID magnetometer; magnetization isotherms were collected at 4
and 300 K between À50 kOe and + 50 kOe. Photoelectron spectra
were recorded with a VG Escalab 200R electron spectrometer
provided with MgKa X-ray source and a hemispherical electron
analyzer. The binding energies were referenced to the C1s peak at
284.9 eV. Data processing was performed with the XPS peak
program, and the spectra were decomposed with the least-squares
fitting routine provided with the software with Gaussian/Lorentzian
(90/10) lines and after subtracting a Shirley background.
Table 2: Products of the oxidation of alkyl aromatics at 20 bar O2 after
3 h of reaction over NHPI/(La,Sr)0.5(Mn,Co)0.5O3À&#948;
.
Substrate
Conv. [%]
Product
distribution [%]
99.7
98.1
63.0
1.9
The oxidation of alkyl aromatics was performed in a 100 mL steel
autoclave (autoclave Engineers). In a typical run, 1.46 g of substrate,
0.435 g of N-hydroxyphthalimide, 50 mg of perovskite or 0.0166 g of
cobalt(II) acetate, 0.0115 g of manganese(II) acetate, and 20 mL of
acetic acid were mixed together. The reactor was pressurized with
oxygen at 20 bar and the temperature was set to 363 K. Then, the
stirring was started and set at 1000 rpm. The pressure was maintained
at 20 bar, and fresh oxygen was fed in as it was consumed. The liquid
was analyzed by gas chromatography with a flame ionization detector
(GC-FID) on an Agilent 6850 device equipped with an HP-WAX
column. Crystallographic data for oxidized (La,Sr)0.5(Mn,Co)0.5O3:
99.8
100
37.0
ꢀ
R3c, hexagonal setting, a = 5.4463(5) ꢀ, c = 13.2885(2) ꢀ, V=
341.355(6) ꢀ3, O1 (x,0,1= ) x = 0.4745(1), 1calcd = 8.386 gcmÀ3, Z = 6,
4
R
wp = 3.19%, Rp = 2.42%, Rexptl = 1.95%, c2 = 2.68, RBragg = 2.55%,
99.1
0.9
measurement range 108 ꢀ 2q ꢀ 1608, 3198 data points, 80 observed
reflections, 68 parameters refined. Crystallographic data for reduced
(La,Sr)0.5(Mn,Co)0.5O3Àd at 423 K: Pbnm, a = 5.5071(4) ꢀ, b =
5.4873(5) ꢀ, c = 7.7898(6) ꢀ, V= 235.40(3) ꢀ3, 1calcd = 7.890 gcmÀ3
,
hypostoichiometric material shows an extraordinary reaction
rate and selectivity for the conversion of alkyl aromatics, for
example, p-xylene in TPA, with respect to the standard
homogeneous catalyst. The fact that this (La,Sr)0.5-
(Mn,Co)0.5O2.38 perovskite phase is ferromagnetic above
room temperature constitutes an added advantage, since it
could be separated from the solid product with a magnet; the
mother liquor could be removed by filtration to yield pure
TPA.
Z = 4, Rwp = 3.22%, Rp = 2.48%, Rexptl = 1.85%, c2 = 3.05, RBragg
=
8.38%, measurement range 108 ꢀ 2q ꢀ 1608, 3198 data points, 254
observed reflections, 23 parameters refined. Further details on the
crystal structure investigation may be obtained from the Fachinfor-
mationszentrum Karlsruhe, 76344 Eggenstein-Leopoldshafen, Ger-
many (fax: (+ 49)7247-808-666; e-mail: crysdata@fiz-karlsruhe.de),
on quoting the depository numbers CSD-422410 and -422411 for the
oxidized and reduced phases, respectively.
Received: December 15, 2010
Revised: April 13, 2011
Published online: May 31, 2011
Experimental Section
Keywords: heterogeneous catalysis · magnetic properties ·
neutron diffraction · oxidation · perovskite phases
The oxygen-stoichiometric (La,Sr)0.5(Mn,Co)0.5O3 oxide was obtained
by a nitrate–citrate route followed by a heat treatment at 11508C 12 h
in air. The subsequent reduction of the polycrystalline powder at
5208C for 8 h in H2/N2 (5%/95%) forming gas promotes the
formation of the extremely oxygen-defective (La,Sr)0.5(Mn,Co)0.5O2.38
perovskite. Thermal analysis was carried out in a Mettler TA3000
system equipped with a TC10 processor unit. Thermogravimetric
(TG) curves were obtained in a TG50 unit, working at a heating rate
of 108CminÀ1 in an H2/N2 (5%/95%) or O2 flow of 0.3 LminÀ1. The
initial characterization of the products was carried out by laboratory
X-ray diffraction (XRD) with a Bruker-axs D8 Advanced diffrac-
tometer (40 kV, 30 mA) with Cu Ka radiation (l = 1.5418 ꢀ) and a
PSD (position sensitive detector). The crystallographic structures of
the oxidized and reduced samples were refined from high-resolution
neutron powder diffraction (NPD) patterns collected at the Institut
Laue-Langevin (ILL) in Grenoble (France), acquired at 295 and
423 K at the D2B diffractometer with l = 1.594 ꢀ. The high-intensity
mode was used; the collection time was 2 h. The refinement of the
crystal structure was performed by the Rietveld method, using the
FULLPROF refinement program[22] and its internal tables for the
coherent scattering lengths of the corresponding elements. The peak
profiles were fitted by the Thompson–Cox–Hastings pseudo-Voigt
function corrected for axial divergence asymmetry. The following
parameters were refined in the final run: scale factor, background
coefficients, zero-point error, pseudo-Voigt corrected for asymmetry
parameters, positional coordinates, and anisotropic thermal factors.
.
[1] I. Hermans, J. Peeters, P. A. Jacobs, Top. Catal. 2008, 50, 124.
[2] P. P. Toribio, J. M. Campos-Martin, J. L. G. Fierro, J. Mol. Catal.
A 2005, 227, 101.
[3] P. P. Toribio, J. M. Campos-Martin, J. L. G. Fierro, Appl. Catal. A
2005, 294, 290.
[4] L. Barrio, P. P. Toribio, J. M. Campos-Martin, J. L. G. Fierro,
Tetrahedron 2004, 60, 11527.
[5] P. P. Toribio, A. Gimeno-Gargallo, M. C. Capel-Sanchez, M. P.
de Frutos, J. M. Campos-Martin, J. L. G. Fierro, Appl. Catal. A
2009, 363, 32.
[6] Y. Ishii, S. Sakaguchi, T. Iwahama, Adv. Synth. Catal. 2001, 343,
393.
[7] R. A. Sheldon, I. W. C. E. Arends, Adv. Synth. Catal. 2004, 346,
1051.
[8] Y. Ishii, S. Sakaguchi, Catal. Today 2006, 117, 105.
[9] H. Falcon, J. M. Campos-Martin, S. M. Al-Zahrani, J. L. G.
Fierro, Catal. Commun. 2010, 12, 5.
[10] R. A. Sheldon, I. W. C. E. Arends, J. Mol. Catal. A 2006, 251,
200.
[11] F. Recupero, C. Punta, Chem. Rev. 2007, 107, 3800.
[12] L. G. Tejuca, J. L. G. Fierro, J. M. D. Tascon, Adv. Catal. 1989, 36,
237.
6560 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 6557 –6561

[13] M. A. Peꢁa, J. L. G. Fierro, Chem. Rev. 2001, 101, 1981.
[14] Y. Takeda, R. Kanno, T. Takada, O. Yamamoto, Z. Anorg. Allg.
Chem. 1986, 540, 259.
[18] A. Androulakis, N. Katsarakis, J. Giapintzakis, N. Vouroutzis, E.
Pavlidou, K. Chrissafis, E. K. Polychroniadis, V. Perdikatsis,
J. Solid State Chem. 2003, 173, 350.
[15] V. V. Vashook, M. V. Zinkevich, Y. G. Zonov, Solid State Ionics
1999, 116, 129.
[16] C. De La Calle, J. A. Alonso, A. Aguadero, M. T. Fernandez-
Diaz, F. Porcher, Z. Kristallogr. 2010, 225, 209.
[17] C. de La Calle, J. A. Alonso, A. Aguadero, M. T. Fernandez-
Diaz, Dalton Trans. 2009, 4104.
[19] R. N. Mahato, K. Sethupathu, V. Sankaranarayannan, K. K.
Bharathi, R. Nirmala, A. K. Nigam, S. K. Malik, IEEE Trans.
Magn. 2009, 45, 4271.
[20] K. R. Seddon, A. Stark, Green Chem. 2002, 4, 119.
[21] T. Mallat, A. Baiker, Chem. Rev. 2004, 104, 3037.
[22] J. Rodrꢂguez-Carvajal, Physica B 1993, 192, 55.
Angew. Chem. Int. Ed. 2011, 50, 6557 –6561
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6561