Oxygen non-stoichiometry and reducibility of B-site substituted lanthanum manganites

Article abstract of DOI:10.1016/S0040-6031(00)00532-3

LaMn_0.8B'_0.2O₃₊(δ) (B'=Ni, Zn, Cu) and LaMn_0.5Cu_0.5O₃₊(δ) perovskites were studied by temperature programmed reduction (TPR). Oxygen non-stoichiometry of the perovskite samples calcined in

Full text of DOI:10.1016/S0040-6031(00)00532-3

Thermochimica Acta 360 (2000) 71±76
Oxygen non-stoichiometry and reducibility of B-site
substituted lanthanum manganites
F. Patcas*, F.C. Buciuman, J. Zsako
Department of Chemistry and Chemical Engineering, `Babes-Bolyai' University of Cluj-Napoca, 3400 Cluj-Napoca, Romania
Received 31 January 2000; received in revised form 18 April 2000; accepted 26 April 2000
Abstract
LaMn0.8B00.2
O
3‡d (B0ˆNi, Zn, Cu) and LaMn0.5Cu0.5
O
3‡d perovskites were studied by temperature programmed reduction
(TPR). Oxygen non-stoichiometry of the perovskite samples calcined in air (PO ˆ0.21 bar) at 7508C decreased upon Zn2‡
2
(dˆ0.03), Cu2‡ (dˆ0.04) and Ni3‡ (dˆ0.08) substitution as compared to LaMnO3.16. LaMn0.5Cu0.5O3‡d had reductive non-
stoichiometry and contained low amounts of La2CuO4 and CuO. The reducibility at low and mean temperatures (t<5008C)
increased in the sequence (Mn, Ni)ꢀ(Mn, Zn)<Mn<(Mn, Cu). # 2000 Elsevier Science B.V. All rights reserved.
Keywords: Catalyst; Lanthanum manganite; Perovskite; Non-stoichiometry; TPR
1. Introduction
proposed by Chan et al. [4] as reasons for the enhanced
activity.
Copper-substituted
lanthanum
manganites
In spite of these promising results with copper-
substituted lanthanum manganites, less is known
about manganese substitution with other transition
metals [3]. The aim of this paper is to compare the
effects of Ni, Cu and Zn substitution on the oxidative
non-stoichiometry, the Mn4‡/(Mn4‡‡Mn3) ratio and
reducibility of lanthanum manganite, by means of the
temperature programmed reduction (TPR) technique.
Nickel, though divalent in most of its oxides, was
found to adopt the trivalent state in the perovskite
LaNiO3 [5]. Copper was found by X-ray photoelectron
spectroscopy (XPS) to adopt the divalent state in
substituted lanthanum manganite [3]. As for zinc, it
is known as a divalent cation. It is therefore expected
that substitution of Cu2‡ and Zn2‡ for Mn3‡ should
lead to a decrease in the oxidative non-stoichiometry
and/or an increase of the Mn4‡ content. In order to
facilitate the assignment of the TPR peaks, reduction
runs with NiO, CuO and ZnO were also carried out.
LaMn1 xCuxO3‡d exhibit high catalytic activity for
CO oxidation [1±4] and NO±CO reaction [2]. The
rates of these reactions have been found to correlate
well with the reducibility and CO adsorption capacity
of the catalyst [2,3] but also with the oxygen avail-
ability measured by temperature programmed deso-
rption (TPD) [4]. The pronounced synergetic effect
that arises upon copper substitution was ascribed in
[2,3] to the combined high CO adsorption capacity of
copper sites and the high reactivity of oxygen bonded
to manganese. The increased mobility of lattice
oxygen in copper-substituted lanthanum manganites
and the change to a high activity state by surface
reduction in the reactive environment have been
* Corresponding author.
E-mail address: fpatcas@chem.ubbcluj.ro (F. Patcas)
0040-6031/00/$ ± see front matter # 2000 Elsevier Science B.V. All rights reserved.
PII: S 0 0 4 0 - 6 0 3 1 ( 0 0 ) 0 0 5 3 2 - 3

72
F. Patcas et al. / Thermochimica Acta 360 (2000) 71±76
2. Experimental
The perovskites have been prepared by decomposi-
tion of metal nitrates Mn(NO3)2Á4H2O (Fluka),
La(NO3)3Á6H2O (Fluka), Ni(NO3)2Á6H2O (Fluka),
Zn(NO3)Á4H2O (Fluka), Cu(NO3)2Á3H2O (Riedel de
Haen), after the method given in [6,7]. Calculated
amounts of dried nitrates were dissolved in distilled
water and glucose was added to enhance the surface
area of the catalysts. The solution was evaporated to
dryness and the solid material was calcined at 7508C
in air for 6 h. The simple oxides have been prepared by
decomposition of the corresponding metal nitrates in
air at 6008C (NiO), 5008C (CuO), except for ZnO
which was of Merck quality.
Cu Ka X-ray analysis were performed on a HZ G4
diffractometer for all samples. TPR experiments were
carried out by using a ¯ow device equipped with a
thermal conductivity detector [8]. The gas ¯ow
(30 ml/min NTP with 10% H2 in N2) passed over
about 50 mg of sample in a quartz reactor, then
through a cold trap prior to the detector. The tem-
perature was raised at a linear rate of 10 K/min
from 20 to 9508C, then it was kept at 9508C for
20 min. To calibrate the TPR curves, hydrogen pulses
of known volumes were injected into a nitrogen ¯ow.
The hydrogen consumption was calculated from the
peak area and expressed as an equivalent amount of
oxygen removed from the sample. The calibration
method was checked up by reducing a sample of
copper oxide (CuO) in the same device. A hydrogen
consumption equivalent to 0.98 mol O/mol CuO was
found.
Fig. 1. TPR pro®les of the catalyst samples: (a) ZnO; (b) Mn2O3;
(c) NiO and (d) CuO.
nese oxides displayed a two-stage reduction pattern
according to the mechanism proposed by Kapteijn
et al. [10]. The ®rst reduction temperature of Mn4‡ in
MnO2 was 302±3318C [9] while for Mn3‡ in Mn2O3 it
was 4458C (Fig. 1). The reduction of nickel oxide
occurs approximately in the same temperature range
as that of the trivalent manganese oxide, i.e. at higher
temperatures than the tetravalent manganese. The
shoulder at 3678C arises from the reduction of
Ni3‡, which is always present in NiO samples, cal-
cined in air at 6008C [8]. Zinc oxide is essentially not
reduced in the temperature range of 20±9508C under
the conditions of our TPR runs. Copper oxide is
reduced in a single step at 3308C to metallic copper,
in the same temperature range as Mn4‡
.
3. Results and discussion
The reduction pro®les of the perovskites are dis-
played in Fig. 2 and the amounts of lattice oxygen
removed corresponding to the reduction peaks are
given in Table 1.
The XRD-patterns of LaMn0.8B00.2O3‡d (B0ˆNi,
Zn, Cu) samples revealed only features of the
perovskite phase. For LaMn0.5Cu0.5
O3‡d, additional
lines of La2CuO4 and CuO were found, though
the total amount of these phases (estimated from
the ratios of the most intense XRD lines) did not
exceed 10%.
The TPR pro®les of the simple NiO, CuO, Mn2O3
and ZnO oxides are given in Fig. 1. Those of La2O3
and MnO2 were discussed in a previous work [9].
Essentially, La2O3 was not reducible and the manga-
3.1. LaMnO3‡d
The discussion of the TPR spectra, leading to the
formula LaMnO3.16 and the ratio Mn4‡
/
(Mn4‡‡Mn3‡) of 0.32 (32%), is given in a previous
paper [9]. The main reduction stages are Mn4‡ to
Mn3‡ (3618C) and Mn3‡ to Mn2‡ (7908C).

F. Patcas et al. / Thermochimica Acta 360 (2000) 71±76
73
oxygen non-stoichiometry, and consequently the
amount of tetravalent manganese, are diminished. A
possible explanation of this phenomenon is the
absence of the Jahn±Teller distortion of the Ni3‡ ions,
which renders unnecessary the transformation of a
part of the distorted Mn3‡ cations into undistorted
Mn4‡ in order to achieve an energetically more stable
con®guration. The appearance of the TPR-pro®le
suggests that only about a half of the Ni2‡ cations
are reduced below 5608C, as in LaNiO3, while the rest
undergo reduction at higher temperatures (>6008C)
together with Mn3‡ cations.
3.3. LaMn0.8Zn0.2O3‡d
The TPR pro®le of this perovskite is identical to that
of LaMnO3‡d, the only difference being the relative
amount of hydrogen consumption corresponding to
the peak at 3618C (reduction of Mn4‡). This points out
to the same reduction sequence as in LaMnO3.16
4‡
3‡
2‡
La3‡Mn0:2‡2d Mn0:6
Zn0:2 O3‡d
2d
Fig. 2. TPR pro®les of oxide catalysts with perovskite structure:
(a) LaMnO3; (b) LaMn0.8Ni0.2O3; (c) LaMn0.8Zn0.2O3; (d)
LaMn0.8Cu0.2O3 and (e) LaMn0.5Cu0.5O3.
! La3‡Mn0:8 Zn0:2 O2:9
3‡
2‡
! La3‡Mn0:8 Zn0:2 O2:5
2‡
2‡
 ꢁꢂ 0:5La2O3 ‡ 0:8MnO ‡ 0:2ZnO†
The value of oxygen non-stoichiometry is thus
(2)
3.2. LaMn0.8Ni0.2O3‡d
0.5‡dˆ0.53, leading to the formula LaMn0.8Zn0.2
-
)
The total amount of removed oxygen points to a
reduction process that brings nickel to the metallic
state. LaNiO3 is also known to be reduced to La2O3
and Ni, in two stages, the ®rst one between 150 and
4008C (Ni3‡!Ni2‡), and the second one between 400
and 5608C (Ni2‡!Ni0) [5]. The reduction of the
substituted perovskite can be thus written as
O3.03 and the corresponding Mn4‡/(Mn4‡‡Mn3‡
ratio of (0.2‡2d)/0.8ˆ0.325 (32.5%). Substitution
of manganese for zinc reduces the oxidative non-
stoichiometry while keeping the Mn4‡ content con-
stant. This ®nding is similar to the effect of lanthanum
substitution for divalent alkaline-earth cations [9]. The
presence of Zn2‡ ions also causes an increased reduc-
tion of Mn3‡ at mean temperatures (the peak at
4608C), which indicates an increased anionic mobility.
A possible explanation for this peculiarity could be the
strong distortion around the Zn2‡ cations (usually in
the tetrahedral co-ordination) situated in the octahe-
dral sites of the perovskite lattice.
4‡
3‡
3‡
La3‡Mn2d Mn0:8
Ni0:2 O3‡d
2d
3‡
3‡
! La3‡Mn0:8 Ni0:2 O3
3‡
2‡
! La3‡Mn0:8 Ni0:2 O2:9
3‡
0
! La3‡Mn0:8 Ni0:2 O2:7
2‡
0
! La3‡Mn0:8 Ni0:2 O2:3
 ꢁꢂ 0:5La2O3 ‡ 0:8MnO ‡ 0:2Ni†
(1)
3.4. LaMn0.8Cu0.2O3‡d and LaMn0.5Cu0.5O3‡d
The amount of removed oxygen of 0.7‡dˆ0.78
leads to the formula LaMn0.8Ni0.2O3.08 and the corre-
sponding Mn4‡/(Mn4‡‡Mn3‡) ratio of 2d/0.8ˆ0.2
(20%). By introducing nickel, a cation of the same
valence as manganese, in the perovskite lattice the
The TPR spectra and the corresponding H2-con-
sumption indicate a complete reduction of the copper
ions to the metallic state. Hence the reduction
process for the 20% substituted perovskite can

Table 1
TPR results and characterisation of Mn-substituted perovskite oxides regarding their reducibility and oxygen non-stoichiometry
Oxide
H2-consumption expressed as O-removed per mole oxide
Total O-removed O-removed below ꢁMn4‡†=
d
(mol/mol)
5008C (mol/mol) ꢁMn4‡ ‡ Mn3‡† (%)
Peak 1
Peak 2
Peak 3
TR1 (8C)
mol/mol
TR2 (8C)
mol/mol
TR3 (8C)
mol/mol
LaMnO3‡d
361
348
361
248
273
0.185
0.07
0.09
0.07
0.52
460
414
460
348
361
0.07
0.16
0.14
0.28
0.08
790
790
790
790
740
0.405
0.55
0.30
0.39
0.24
0.66
0.78
0.53
0.74
0.84
0.255
0.23
0.23
0.35
0.60
32
20
0.16
0.08
0.03
0.04
0.16
LaMn0.8Ni0.2O3‡d
LaMn0.8Zn0.2O3‡d
LaMn0.8Cu0.2O3‡d
LaMn0.5Cu0.5O3‡d
32.5
35
a
36
a All quantities are referred to the formula LaMn0.5Cu0.5O3‡d though the real oxide contains some La2CuO4 and CuO.

F. Patcas et al. / Thermochimica Acta 360 (2000) 71±76
75
be written as
amount of oxygen during copper reduction. Generally
speaking, the stability of this perovskite against reduc-
tive environment seems to be poor, though its ability to
release a great part of the lattice oxygen could make it
very suitable for intrafacial catalysis.
4‡
3‡
2‡
La3‡Mn0:2‡2d Mn0:6
Cu0:2 O3‡d
2d
4‡
3‡
1‡
! La3‡Mn0:2‡2d Mn0:6‡2d Cu0:2 O2:9‡d
4‡
3‡
0
! La3‡Mn0:2‡2d Mn0:6‡2d Cu0:2 O2:8‡d
3‡
0
! La3‡Mn0 Cu0:2 O2:7
2‡
0
! La3‡Mn0 Cu0:2 O2:3
4. Conclusions
 ꢁꢂ 0:5La2O3 ‡ 0:8MnO ‡ 0:2Cu†
(3)
In conclusion, the following remarks regarding the
features of B-site substituted manganites in relation to
their potential use as oxidation catalysts may be stated
on the basis of TPR studies.
The oxygen excess is dˆ0.04, that means LaMn0.8-
Cu0.2O3.04 and the ratio Mn4‡/(Mn4‡‡Mn3‡) of 0.35
(35%). From this point of view, copper substitution
produces almost the same effect as zinc substitution.
Unlike copper oxide, where Cu2‡ is reduced to Cu0
in a single step, in the 20%-substituted lanthanum
manganite the copper cations appear to follow a two-
step reduction with Cu1‡ as intermediate, as indicated
by the size of the ®rst reduction peak. The second step
1. Manganese substitution for divalent (Zn2‡, Cu2‡
)
cations in lanthanum manganite leads to the
lowering of oxidative non-stoichiometry by keep-
ing the Mn4‡/(Mn4‡‡Mn3‡) ratio constant, while
trivalent cation (Ni3‡) substitution produces the
diminution of both oxidative non-stoichiometry
and Mn4‡ content. Therefore the latter, containing
a higher amount of Mn3‡ sites is likely to adsorb
gas-phase oxygen to a higher extent.
occurs at the same time with Mn4‡ reduction to Mn3‡
.
The reduction of Mn3‡ seems to take place entirely at
high temperatures. The relative good stability of cop-
per in the perovskite lattice against low temperature
(<3008C) reduction to metallic copper can be regarded
as a convenient feature of the low substituted manga-
nites, by avoiding sintering of copper during oscilla-
tions of working environment.
2. The Ni3‡ and Zn2‡ substituted manganites display
a poorer reducibility at low and mean tempera-
tures (<5008C) as compared to non-substituted
lanthanum manganite, while Cu2‡ substituted
perovskites release at relatively low (250±
2708C) temperature an important amount of lattice
oxygen, which increases with the ratio Cu/Mn.
The ease of reduction, which is desirable from the
point of view of intrafacial catalytic reactions,
represents at the same time a disadvantage for the
stability of the catalyst in reducing atmospheres,
especially for the higher substituted perovskites.
The accurate interpretation of the TPR curve
of LaMn0.5Cu0.5O3‡d is rendered dif®cult by the
presence of La2CuO4 and CuO. If the ®nal reduc-
tion point is considered 0.5La2O3‡0.5MnO‡0.5Cu
(BLaMn0.5Cu0.5O2), the total amount of removed
oxygen is 1‡dˆ0.84, that is dˆ 0.16, which leads
to the stoichiometry LaMn0.5Cu0.5O2.84. The real
value of d should be lower in absolute value, since
the La:O ratio in La2CuO4 is 1:2 whereas in LaMnO3
it is 1:3, which means that a part of this apparent
oxygen de®ciency is due to the presence of La2CuO4.
In fact, the stoichiometry found by Gallagher et al. [1]
References
[1] P.K. Gallagher, D.W. Johnson, E.M. Vogel, J. Am. Ceram.
Soc. 60 (1977) 28.
for the pure perovskite phase was LaMn0.5Cu0.5O2.93
,
while Rojas et al. [11] have prepared LaMn0.5-
Cu0.5O3.00. By examining the TPR curve it appears
that copper cations in high-substituted manganites are
reduced to metal in a single step similar to copper
oxide but at lower temperatures. Another striking
feature is the lowering of the reduction temperature
of Mn3‡, which can be caused by the high perme-
ability of the lattice arising from the removal of a great
[2] N. Mizuno, Y. Fujiwara, M. Misono, J. Chem. Soc., Chem.
Commun. (1989) 316.
[3] H. Yasuda, Y. Fujiwara, N. Mizuno, M. Misono, J. Chem.
Soc., Faraday Trans. 90 (1994) 1183.
[4] K.S. Chan, J. Ma, S. Jaenicke, G.K. Chuah, J.Y. Lee, Appl.
Catal. A 107 (1994) 201.
[5] T.-R. Ling, Z.-B. Chen, M.-D. Lee, Catal. Today 26 (1995) 79.
Á Â
[6] S. ZuÈhlke, H.-G. Lintz, Recents progres en genie des
 Â
procedes, No. 54, Vol. 11, Paris, 1997, p. 13.

76
F. Patcas et al. / Thermochimica Acta 360 (2000) 71±76
[9] F.C. Buciuman, F. Patcas, J. Zsako, J. Thermal Anal.,
[7] F.C. Buciuman, T. Hahn, Wiss. Abschlussber. Int. Semin.
Forschung Lehre Chemieingenieurwes, Tech. Phys. Chem.,
30th, University Karlsruhe, Germany, p. 125.
accepted for publication.
[10] F. Kapteijn, L. Singoredjo, A. Andreini, J.A. Moulijn, Appl.
Catal. B 3 (1994) 173.
[8] F. Patcas, Heterogen katalysierte Gasphasen-Oxidehydrierung
È
von Cyclohexan an Nickeloxid-Tragerkatalysatoren, Disser-
[11] M.L. Rojas, J.L.G. Fierro, L.G. Tejuca, A.T. Bell, J. Catal.
124 (1990) 41.
È
tation, Technische Universitat Chemnitz, 1998, p. 54.