Modified surface and bulk properties of Fe-substituted lanthanum titanates enhances catalytic activity for CO + N2O reaction

Article abstract of DOI:10.1016/j.molcata.2010.11.028

Iron substituted lanthanum titanates with nominal composition La ₂Ti_2(1-x)Fe_2xO_7-δ (0.0 ≤ x ≤ 1.0) samples, were synthesized by gel combustion and conventional solid state reaction. The samples have been characterized by XRD, N₂-BET, SEM, FTIR, Mo?ssbauer spectroscopy and TPR. Catalytic activity of these samples was evaluated for a decomposition reaction of N₂O using CO. Transition from La₂Ti₂O₇ (layered perovskite, A₂B₂O₇) to single phase, LaFeO₃ (perovskite, ABO₃) was observed for the samples having ≥40-60% Fe content. The microstructural analysis by SEM revealed homogeneous elongated shaped rods on single phased LTOGC surface, attributed to the layered perovskite whereas spherical grains are due to single perovskite (ABO₃) phase present in LF(0.4)GC samples. Mixed phased samples showed both elongated and spherical shape particles as observed in the SEM images. Among all samples, LF(0.6)GC have shown maximum activity of 100% conversion at 395 °C for decomposition of N₂O using CO and at 325 °C for CO oxidation. The substitution-induced anionic vacancies and asymmetric environment around Fe as suggested by Mo?ssbauer results in partially substituted, LF(0.4)GC/SS and LF(0.6)GC/SS are responsible for considerable lowering in T_max observed in their TPR profiles and enhanced activity for CO + N₂O reaction as compared to pristine samples.

Full text of DOI:10.1016/j.molcata.2010.11.028

Journal of Molecular Catalysis A: Chemical 335 (2011) 158–168
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Modified surface and bulk properties of Fe-substituted lanthanum titanates
enhances catalytic activity for CO + N2O reaction
K.K. Karthaa, M.R. Paia, A.M. Banerjeea, R.V. Paib, S.S. Meenac, S.R. Bharadwaja,∗
a Chemistry Division, Bhabha Atomic Research Centre, Mumbai 400085, India
b Fuel Chemistry Division, Bhabha Atomic Research Centre, Mumbai 400085, India
c Solid State Physics Division, Bhabha Atomic Research Centre, Mumbai 400085, India
a r t i c l e i n f o
a b s t r a c t
Article history:
Iron substituted lanthanum titanates with nominal composition La2Ti2(1−x)Fe2xO7−ı (0.0 ≤ x ≤ 1.0) sam-
ples, were synthesized by gel combustion and conventional solid state reaction. The samples have been
characterized by XRD, N2-BET, SEM, FTIR, Mössbauer spectroscopy and TPR. Catalytic activity of these
samples was evaluated for a decomposition reaction of N2O using CO. Transition from La2Ti2O7 (layered
perovskite, A2B2O7) to single phase, LaFeO3 (perovskite, ABO3) was observed for the samples having
≥40–60% Fe content. The microstructural analysis by SEM revealed homogeneous elongated shaped rods
on single phased LTOGC surface, attributed to the layered perovskite whereas spherical grains are due to
single perovskite (ABO3) phase present in LF(0.4)GC samples. Mixed phased samples showed both elon-
gated and spherical shape particles as observed in the SEM images. Among all samples, LF(0.6)GC have
shown maximum activity of 100% conversion at 395 ◦C for decomposition of N2O using CO and at 325 ◦C
for CO oxidation. The substitution-induced anionic vacancies and asymmetric environment around Fe
as suggested by Mössbauer results in partially substituted, LF(0.4)GC/SS and LF(0.6)GC/SS are responsi-
ble for considerable lowering in Tmax observed in their TPR profiles and enhanced activity for CO + N2O
reaction as compared to pristine samples.
Received 19 February 2010
Received in revised form 17 October 2010
Accepted 25 November 2010
Available online 9 December 2010
Keywords:
Iron substituted lanthanum titanates
Gel combustion
TPR
Mössbauer spectroscopy
CO + N2O
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
graphic shears randomly distributed in the lattice which seems to
affect their properties greatly [1]. La2Ti2O7 (A2B2O7), perovskite
Titanates with nominal oxygen overstoichiometry are espe-
cially interesting due to their very high electronic conductivity
and stability under reducing conditions. Oxides belonging to the
La1−xSrxTiO3+ı system have been studied previously as anode mate-
rial in SOFCs showing good performance comparable to those of the
state of art materials [1,2]. The Ln2Ti2O7 [3] is known to be effec-
tive ferromagnetic material where, Ln = La or Nd and is an ionic
conductor when Ln = Y. Preliminary studies by Happel et al. [4]
reported that LaTiO3 is an active catalyst for the reduction of sul-
fur dioxide to elemental sulfur by means of carbon monoxide with
minimum formation of COS. This makes it a strong candidate for a
variety of applications in electrical, optical devices and as catalysts
for different reactions.
of layered structures built from (1 1 0) perovskite slabs differing in
thickness and bounded by crystallographic shears in the perovskite
[1 0 0] direction. Adjacent slabs are offset from one another by half
of one TiO6 octahedron height and the octahedron connectivity is
broken at the shear interface. In reacting mixtures of La2O3 and
TiO2 a range of compounds is possible. In an unreduced system
La2O3·TiO2, La2O3·2TiO2 and 2La2O3·9TiO2 have been identified by
MacChesney and Sauer [5]. A mixture containing a ratio of La:Ti
slightly lower than unity, if undergoes complete reduction then
the La2O3·2TiO2 crystallizes in the perovskite LaTiO3 phase. Thus,
lanthanum titanate comes close to meeting the requirements for
the ideal cubic perovskite structure either by reducing titanium to
+3 states or by creating A-site vacancies. Much of previous works
lacks detailed structural characterization, and treated these as sim-
ple cubic perovskite in spite of presence of extra oxygen in these
compounds. Both A and B sites in A2B2O7 can be substituted by
diverse component ions to form stable solid solutions. The forma-
tions of stable solid solutions via partial substitutions of Ti sites by
Cr3+, Fe3+ (up to 10%) and Pr3+, Nd3+ at A sites in La2Ti2O7 have
already been reported [1,6,7] for photocatalytic reactions. How-
ever, reports where thermal reactions are studied using lanthanum
titanate as catalyst are very scanty. For proper understanding of the
La2Ti2O7 type compounds are able to accommodate the extra
oxygen beyond the parental perovskite in short-range crystallo-
∗
Corresponding author at: Fuel Cell Materials and Catalysis Section, 3-193
H, Chemistry Division, Modular Labs, Bhabha Atomic Research Centre, Mumbai
400085, Maharashtra, India. Tel.: +91 22 25595100 (off)/25555079 (res);
fax: +91 22 25505331.
E-mail address: shyamala@barc.gov.in (S.R. Bharadwaj).
1381-1169/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2010.11.028

K.K. Kartha et al. / Journal of Molecular Catalysis A: Chemical 335 (2011) 158–168
159
temperature dependent catalytic properties of an oxide, it is impor-
tant to have insight about its bulk properties that enables an oxide
to qualify as a catalyst.
La(NO3)3.6H2O
Fe2(NO3)3
TiOCl2
Citric Acid
To develop catalytic materials that could convert N2O, a
green house gas [8] into environmentally friendly nitrogen using
another exhaust gas CO, metal cation substituted samples based
on lanthanum titanate, La2Ti2O7 were prepared and their surface
properties, bulk characterizations and catalytic activities as a func-
tion of x and synthetic route were studied. In situ FTIR studies on
these oxides for CO and N2O adsorptions to enunciate the role of
carbonate formation in the mechanism of N2O + CO reactions are
discussed in subsequent communications [9]. Kapteijn et al. [10]
reviewed extensively a variety of catalysts, such as metal oxides,
mixed metal oxides for N2O decomposition. Transition metals and
their oxides, in forms of either simple or complex oxides, have also
been investigated thoroughly [11,12]. Many investigations [13–17]
have been made on modification of physico-chemical properties,
so as to augment the activity, selectivity and the thermal stabil-
ity of a particular oxide/mixed oxide catalyst. Recently we have
studied the effect of B-site substitution on thermal–physical bulk
(Fuel)
Stoichiometric amount of
reactants were mixed along with
citric acid (Oxidant to fuel
ratio=1:1.66)
Constant stirring at ~ 200°C upto
gel formation. Heating continued
till black powder is formed.
Calcination at higher temperature
(800°C, 36 hrs).
properties and reduction behavior of In2TiO5 viz.; In2Ti1−xFexO5−ı
,
In2Ti1−xCrxO5−ı (0.0 ≤ x ≤ 0.02) mixed oxides catalysts [18]. Ear-
lier we have reported the role of A/B site substitution on oxide
reducibility, oxygen diffusivity, etc., and correlated with the cat-
alytic activities of Th(VO3)4 and LaMnO3 oxides as a function of
multiple cationic substitutions, for both CO oxidation and methanol
degradation [19–23].
La2Ti2(1-x)Fe2xO7-δ (0.0≤x≤1.0,
typically
x=0.0,0.1,0.2,0.4,0.6,0.8,1.0)
To rule out surface area affect and investigate the role of
modified bulk properties alone on activity of lanthanum titanate
La2Ti2O7, as a result of Fe3+substitution, La2Ti2(1−x)Fe2xO7−ı sam-
ples were synthesized by conventional ceramic route. Secondly
surface area was increased by preparing nanosized crystalline
powders using gel combustion route. In the present report,
La2Ti2(1−x)Fe2xO7−ı samples were synthesized by solid state and
gel combustion routes. Variations in structure and powder prop-
erties as a function of synthetic route and metal ion substitution
were monitored using XRD, FTIR N2-BET and SEM. Mössbauer stud-
ies helped in determining the coordinative atmosphere around Fe
in substituted samples. The effect of Fe-doping on redox behavior
Scheme 1. Flow chart of the gel combustion synthesis.
for 48 h in order to improve the crystallinity of the products
obtained.
2.2. Characterization
The powder XRD patterns were recorded on
diffractometer (model PW 1710), equipped with
a
Philips
graphite
a
monochromator and Ni-filtered Cu K␣ radiation. The crystallite
size, D, was calculated from XRD line width according to the Scher-
rer equation,
of La2Ti2(1−x)Fe2xO7−ı was monitored by temperature programmed
reduction (TPR) profiles. The catalytic properties of nominal com-
positions La2Ti2(1−x)Fe2xO7−ı were evaluated for CO + N2O reaction.
The catalytic activity is evaluated as a function of temperature,
time and correlated with the extent of substitution. Finally, cor-
relations between modified bulk/surface properties – enhanced
catalytic activity of La2Ti2(1−x)Fe2xO7−ı samples were established
considering the role of Fe substitution in lanthanum titanates.
Kꢂ
ꢀ(2ꢁ) =
ˇ(cos ꢁ)
where ꢀ(2ꢁ) is the width at half-maximum intensity (in radians)
and ꢁ is the Bragg angle of the (2 0 3) plane in the diffraction pattern
of the peak, K is a constant, depending on the line shape profile
(currently K = 0.9), ꢂ is the wavelength of the X-ray source (in the
case of Cu radiation, ꢂ = 0.154059 nm) and ˇ is the crystallite size
in nm [19].
2. Experimental
A Quantachrome Autosorb-1 analyzer was employed for mea-
surement of N2-BET surface area by recording the nitrogen
adsorption isotherms. The pore size distribution and pore volume
were calculated from the desorption branch using the BJH method.
The FTIR spectra of the solid samples were recorded in KBr using
2.1. Preparation
Nominal
compositions
La2Ti2(1−x)Fe2xO7−ı
samples
(0.0 ≤ x ≤ 1.0), with typical values of 0.1, 0.2, 0.4, 0.6, 0.8, and 1.0
were synthesized by two routes viz. gel combustion and solid state
routes. Flow chart of step by step synthesis of La2Ti2(1−x)Fe2xO7−ı
samples by gel combustion method using citric acid is shown in
Scheme 1. The most widely used method for the preparations
of polycrystalline mixed metal oxides is the solid state reaction,
where stoichiometric quantities of respective oxides, La2O3, TiO2
and Fe2O3 were mixed and ground thoroughly for at least half an
hour each. The pellets were heated first at a lower temperature
of 700 ◦C for a time period of ∼90 h, followed by calcinations at
900 ◦C for 40 h and 1200 ◦C for 63 h with intermittent grindings.
Final heating was carried out at a higher temperature of 1400 ◦C
a JASCO FTIR (model 610) in the mid IR region (4000–400 cm−1
)
equipped with a DTGS detector having a resolution of 4 cm−1. For
this purpose about 200 mg of dry KBr was mixed with 2–4% of
the sample ground and pressed into a transparent, thin pellet at
5 tons/cm2. These pellets were used for IR spectral measurements.
For microstructural examination under a scanning electron
˚
microscope (SEM) a thin layer of gold (100 A) was coated on the
sample pellets by thermal evaporation in a vacuum coating unit.
A measured quantity of gold wire (99.99% pure) was wrapped on
a tungsten wire which in turn was heated to evaporate the gold
wire. The images of gold coated samples were recorded on scan-

160
K.K. Kartha et al. / Journal of Molecular Catalysis A: Chemical 335 (2011) 158–168
Table 1
Identification of phase in nominal compositions La2Ti2(1−x)Fe2xO7−ı and their abbreviations.
S. no.
Nominal composition
Fe content (x)
Abbreviation LM(x)R (M = Ti/Fe,
x = 0.1–1.0, R = SS/GC)
Synthesis route
Phase identification by XRD
1
2
3
4
5
6
7
8
La2Ti2O7
0.0
0.1
0.2
0.4
0.6
0.8
1.0
0.0
0.1
0.2
0.4
0.6
0.8
1.0
LTOSS
SS
SS
SS
SS
SS
SS
SS
GC
GC
GC
GC
GC
GC
GC
Single phase, La2Ti2O7
Single phase, La2Ti2O7
La2Ti2O7 and LaFeO3
La2Ti2O7 and LaFeO3
LaFeO3
Single phase LaFeO3
Single phase LaFeO3
Single phase, La2Ti2O7
La2Ti2O7 and LaFeO3
La2Ti2O7 and LaFeO3
Single phase LaFeO3
Single phase LaFeO3
Single phase LaFeO3
Single phase LaFeO3
La2Ti1.8Fe0.2O7−ı
La2Ti1.6Fe0.4O7−ı
La2Ti1.2Fe0.8O7−ı
La2Ti0.8Fe1.2O7−ı
La2Ti0.4Fe1.6O7−ı
LaFeO3
LF(0.1)SS
LF(0.2)SS
LF(0.4)SS
LF(0.6)SS
LF(0.8)SS
LFOSS
La2Ti2O7
LTOGC
9
La2Ti1.8Fe0.2O7−ı
La2Ti1.6Fe0.4O7−ı
La2Ti1.2Fe0.8O7−ı
La2Ti0.8Fe1.2O7−ı
La2Ti0.4Fe1.6O7−ı
LaFeO3
LF(0.1)GC
LF(0.2)GC
LF(0.4)GC
LF(0.6)GC
LF(0.8)GC
LFOGC
10
11
12
13
14
ning electron microscope Model Tescan Vega MV 2300T/40 using
an accelerating voltage of 25 kV at the working distance ∼10 mm.
Mössbauer spectra have been obtained using a spectrometer
operated in constant acceleration mode in transmission geometry.
The source employed is Co-57 in Rh matrix of strength 50 mCi. The
calibration of the velocity scale is done using a ␣-Fe metal foil, and
Mössbauer parameters are presented relative to ␣-Fe metal foil.
The outer line width of calibration spectra was 0.29 mm s−1. Möss-
bauer spectra were fitted by a least square fit (MOSFIT) programme
assuming Lorentzian line shapes.
Temperature programmed reduction/oxidation (TPR/O) run was
recorded on a TPDRO-1100 analyzer (Thermo Quest, Italy) under
the flow of (H2 (5%) + Ar) alternatively O2 (5%) + He at a gas flow
rate of 20 ml min−1, in temperature range 25–1100 ◦C, at a heating
rate of 6 ◦C min−1. H2O produced was removed from the evolved
gases through a trap attached prior to TCD detector.
ovskite, La2Ti2O7 phase as a function of calcination temperature
was monitored by recording XRD diffraction pattern after calci-
nations at different temperatures. As the calcination temperature
increased to 1200 ◦C the intensity of the unreacted phases due
to TiO2 and La(OH)3 decreased, and the counts due to La2Ti2O7
phase increased. The extra lines which were seen in XRD pat-
tern of La2Ti2O7 calcined at 900 ◦C appeared at 2ꢁ = 25.2◦, 28.2◦,
29.2◦, and 39.6◦. These extra lines are attributed to the pres-
ence of free reactants in the sample corresponding to La(OH)3
(JC-PDS card no. 36-1481) and TiO2 (Brookite, JC-PDS card no. 29-
1360). Also shift in d value by 0.055 in most intense reflection
from h k l of 2 1 2 planes at d value of 2.995 of parent phase is
observed. These observations suggested that the solid state reac-
tion is incomplete even at 1200 ◦C. For completion of diffusion
controlled solid state reaction, all samples were further calcined
at higher temperatures. The solid state reaction of La2O3 + TiO2,
and henceforth single phase synthesis of lanthanum titanate was
observed at 1400 ◦C which crystallized in A1112BIV2O7 lattice type
having layered perovskite structure. The XRD patterns of samples
(0.0 ≤ x ≤ 1.0) synthesized by solid state reaction calcined at 1400 ◦C
are shown in Fig. 1(a–g). For modifying the powder/morphological
properties of crystalline La2Ti2(1−x)Fe2xO7−ı samples, an alterna-
tive preparation route, gel combustion using citric acid as fuel was
adopted. For comparison, the XRD patterns of corresponding sam-
ples synthesized by gel combustion are shown in Fig. 1(h–n). The
XRD pattern of sample with x = 0.0 as shown in curve a of Fig. 1
matches with monoclinic La2Ti2O7 (JC-PDS card no. 28-0517) hav-
2.3. Catalytic activity
A fixed-bed continuous-flow tubular reactor made of quartz was
employed for the measurement of catalytic activity of different
samples for CO oxidation and N2O + CO reaction. For activity eval-
uation, 0.1 g of each sample was placed between two quartz wool
plugs (bed length ∼1 mm), the reactant (CO:O2:He = 2:1:17 for CO
oxidation and CO:N2O:He = 2:2:16 for CO + N2O reaction) flow rate
being at 40 ml min−1. Volume reduction glass rods were inserted
on both sides of the quartz wool plugs. Effluents were sampled and
analyzed on a gas chromatograph (Porapak-Q column and ther-
mal conductivity detector). The temperature dependent catalytic
˚
˚
˚
ing cell parameters as a = 13.015 A, b = 5.5456 A and c = 7.817 A, z = 4,
3
˚
cell volume = 557.8 A . However by gel combustion route the sin-
gle La2Ti2O7 phase was obtained when calcined at 800 ◦C as shown
in curve h in contrast to 1400 ◦C (Fig. 1a) by solid state route. The
crystal structure was also highly dependent on the doping content
of Fe. As expected the substitution of lower valent Fe3+ in place of
Ti4+ will decrease the extra oxygen in the La2Ti2O7 phase. This is
reflected in the change of layered perovskite to perovskite phase
for the solid state samples having Fe more than 60% substitution.
The samples having x ≥ 0.6 (Fig. 1e–g) crystallizes in rhombohedral
activity of all La2Ti2(1−x)Fe2xO7−ı samples was evaluated under con-
tinuous flow of reactant mixtures in range of 50–600 ◦C, by holding
1 h at each temperature. Each sample was initially pretreated at
350 ◦C for 2 h under N2 flow prior to activity measurements. The
catalytic conversions were calculated on the basis of product yield
CO2 for 0.1 g of each sample. The void volume fraction estimated for
LTOGC and LF(0.4)GC were found to be 0.80, and 0.89, respectively.
˚
˚
LaFeO3 phase having unit cell parameters a = 5.566 A, b = 7.854 A,
3. Results and discussion
and c = 5.56 A, Vol. = 242.81 A3 (JC-PDS card no. 37-1493). The crys-
tallization of LaFeO3 phase was induced even by meagre amounts
of Fe doping as suggested by the presence of weak XRD line at
2ꢁ = 32.19◦ marked as * in LF(0.1)SS/GC sample as shown by curve
b of Fig. 1. The intensity of this line increased with increase in Fe
content as seen in curves b–g and i–n in Fig. 1. At the same time
the intensity of reflection due to 2 1 2 plane (100% line marked
as #) in pyrochlore phase was found to diminish with increase
in Fe content. As seen in curve d of Fig. 1 the intensity of XRD
˚
˚
3.1. Phase characterization by XRD
Table 1 lists the abbreviations and phases identified from
XRD patterns of all the nominal compositions La2Ti2(1−x)Fe2xO7−ı
(0.0 ≤ x ≤ 1.0) prepared by two different routes viz. solid state and
gel combustion. Henceforth in the present communication, all the
samples will be referred by their abbreviation as mentioned in
Table 1. The crystal growth of the lanthanum titanate, layered per-

K.K. Kartha et al. / Journal of Molecular Catalysis A: Chemical 335 (2011) 158–168
161
Fig. 1. XRD patterns of nominal compositions La2Ti2(1−x)Fe2xO7−ı synthesized by solid state (SS) reaction (a–g) and gel combustion (GC) route (h–n) for different samples,
LTO (a, h), LF(0.1) (b, i), LF(0.2) (c, j), LF(0.4) (d, k), LF(0.6) (e, l), LF(0.8) (f, m), LFO (g, n). *XRD peak at 2ꢁ = 32.17 attributed to perovskite LaFeO3; #XRD peak at 2ꢁ = 29.8
corresponds to La2Ti2O7 phase.
line due to parent phase was drastically reduced to ∼30.4% from
100% in LF(0.4)SS was absent in LF(0.8)SS and LFOSS (Fig. 1f and
g). Thus Fe doping by solid state reaction stabilized the lanthanum
titanate in perovskite lattice above 60% of substitution by lowering
down the excess oxygen of the system. The single phased compo-
sitions in La2Ti2(1−x)Fe2xO7−ı (A2B2O7) where x ≥ 0.6 may be better
represented as 2LaTi1−xFexO3+ı (ABO3) perovskites. For instance,
La2Ti0.8Fe1.2O7−ı, can be represented by 2La(Ti0.4Fe0.6)O3+ı, oxide.
The XRD patterns of samples synthesized by gel combustion route
also revealed nearly same crystal structures as by ceramic route
with some differences. Firstly, the XRD lines in curves i–k are very
broad suggesting decrease in crystallite size of samples synthesized
by gel combustion route (Table 1). Homogeneous mixture of reac-
tants in gel combustion helped in achieving single phase, LF(0.4)GC
sample whereas that synthesized by solid state route is biphasic
with a mix of both the extreme phases. Also the unsubstituted LTO
single phase was obtained at much lower temperature (800 ◦C) as
compared to solid state reaction (1400 ◦C).
3.2. Crystallite size, surface area and pore size distribution
The increasing order of crystallite size as was calcu-
lated using Scherrer’s equation observed from Table
2 is
LF(0.4) ≈ LF(0.2) < LF(0.6) ≈ LF(0.8) < LTO < LFO. The cationic substi-
tution of Fe3+ in La2Ti2O7 lattice was accompanied with drastic
decrease in crystallite size and formation of pores. Nanosized pow-
ders have been obtained by gel combustion route as compared to
bulk oxides synthesized by solid state reaction (Table 2). For exam-
ple LF(0.4)GC has a crystallite size of 14–17 nm as compared to
38–40 nm size of its solid state analogue LF(0.4)SS. It is interest-

162
K.K. Kartha et al. / Journal of Molecular Catalysis A: Chemical 335 (2011) 158–168
Table 2
Crystallite size and N2-BET surface area of nominal compositions La2Ti2(1−x)Fe2xO7−ı as calculated from XRD line width using Scherrer’s equation.
S. no.
Value of x
Sample
Crystallite size (nm)
Gel combustion
Surface area (m2 g−1
Gel combustion
)
Pore size (Å)
Solid state calcined at 1400 ◦
C
1
2
3
4
5
6
0.0
0.2
0.4
0.6
0.8
1.0
LTO
60–62
14–16
15–17
40–42
44–46
77–80
82–84
51–53
38–40
50–52
151–153
100–102
12
∼400
LF(0.2)
LF(0.4)
LF(0.6)
LF(0.8)
LFO
14
68
47
9
∼300
∼20
∼20
7
ing to note that both the solid state reaction and citrate complex
decomposition route has led to same trend in crystallite size and
surface area. The surface area of LF(0.4)GC and LF(0.6)GC calcined at
800 ◦C was found to be 68 and 47 m2 g−1, respectively, considerably
higher than unsubstituted LTOGC as well as LFOGC samples, with
12 and 7 m2 g−1, respectively. Enhanced surface area of LF(0.6)GC
(47 m2 g−1) sample as compared to LF(0.2)GC of 14 m2 g−1 is not in
agreement with the trend followed by their crystallite size (Table 2).
This anomaly in trends can be explained from their pore size dis-
tribution curves (Fig. 2) where LF(0.6)GC sample have maximum
[22] and in In2TiO5 [18] has played a significant role to decrease
the crystallite size of La0.8Sr0.2Mn1−xFexO3−ı (LSMF) and indium
titanate samples, respectively. Similar results have been obtained
by other researchers, for instance, Taguchi et al. [26] have demon-
strated that the crystallite size decreased with increasing x in the
perovskite-type (La1−xCax)FeO3 samples and such samples with
smaller particle size exhibited a higher catalytic activity.
3.3. Scanning electron microscopy
˚
pores of radii ∼20–30 A exists, whereas larger pores with radii
Fig. 3a–c shows the microstructures of the LTOGC, LF(0.2)GC,
LF(0.4)GC pellets synthesized by gel combustion method. The
micrograph of LTOGC (Fig. 3a) reveals the presence of elongated
rod like structures. While, 40–60% Fe substitution resulted in more
or less spherical shaped geometry as observed in the SEM micro-
graph of LF(0.4)GC (Fig. 3c). The inhibition of grain size observed
as a result of Fe substitution may be due to the creation of defects
in the doped samples. The LTOGC and LF(0.4)GC crystallizes in lay-
ered perovskite structure and latter sample is ideal perovskite as
confirmed by their XRD patterns. The rod like structures forming
layers can be attributed to the layered perovskite, La2Ti2O7 phase
present on the surface of LTOGC sample (Fig. 3a) whereas spheri-
cal grains are due to perovskite, ABO3 arrangement of cations and
anions in LF(0.4)GC (Fig. 3c). The mixed phases are observed in case
of LF(0.2)GC as observed in its SEM image (Fig. 3b) where both rod
like and spherical grains are present. The crystallite sizes reported
in Table 2 are calculated from the FWHM of the 100% peak in the
XRD pattern of the sample which gives us the average crystallite
size using Scherrer’s equation. While in SEM, we have reported the
microstructure of the sample pellets revealing their grain size. So
it is the grain size which is reflected in the SEM.
˚
>200 A are observed in LF(0.2)GC sample. Citrate nitrate gel com-
bustion method has previously yielded porous powders with high
surface areas [24,25]. However, further addition of Fe3+ in Ti4+
site has resulted in increase in the particle size as can be seen in
Table 2 for LF(0.8)GC and LFOGC samples which are having crys-
tallite size of the order of 40–42 and 77–80 nm, respectively. This,
suggest that the substitution-induced microstructural defects and
nonstoichiometry created on lattice reorganization during synthe-
sis also play a role in deciding the morphology of the sample.
This result is well in agreement with our earlier studies, where
Fe substitution at B-site cation in La0.8Sr0.2MnO3−ı lattice (LSM)
0.0006
0.0005
0.14
0.12
LF(0.2)GC
0.0004
0.0003
0.0002
0.0001
0.10
0.08
0.06
0.04
0.02
303A°
3.4. FTIR
Infrared spectroscopy can be used to monitor chemical and
structural changes in the mixed metal oxides. The FTIR spectra of all
the prepared solid state oxides in KBr are shown in Fig. 4. Table 3
shows the characteristic IR bands (cm−1) and their assignments
observed in FTIR spectra recorded for various samples. A prominent
absorption band at 552 cm−1, a broad band at 618 cm−1, several
weak bands between 400 and 500 cm−1 in addition to three bands
above 700 cm−1 are observed in the FTIR spectrum (Fig. 4a) of pris-
tine lanthanum titanate, LTOSS. The vibration bands for absorption
bands of TiO6 octahedra are generally lower than 700 cm−1. The
bands below 405 cm−1 are assigned to Ti–O bending vibration and
those above 405 cm−1 are attributed to Ti–O stretching vibration
[27–29]. On the other hand, rare earth oxygen stretching modes
generally lie in the region of 300–500 cm−1. Thus, the bands in
the 400–500 cm−1 region in LTOSS can be attributed to complex
motions involving the participation of both La and Ti cations [30].
The strong peak at 550 cm−1 can be assigned to Ti–O vibration in
TiO6 octahedra present in the structure. There are two additional
bands at 750 cm−1 and 805 cm−1 which are of higher frequencies
generally observed in M2TiO5 compounds at about 775 cm−1 due
0.0000
0.0016
0.00
0.20
0
100
200
300
400
500
600
700
0.0014
0.0012
0.0010
0.0008
0.0006
0.0004
0.0002
0.0000
22A°
0.15
0.10
0.05
0.00
LF(0.6)GC
0
100
200
300
400
500
Pore radius (A°)
Fig. 2. BJH pore size distribution and cumulative volume curves of LF(0.2)GC and
LF(0.6)GC samples.

K.K. Kartha et al. / Journal of Molecular Catalysis A: Chemical 335 (2011) 158–168
163
Fig. 4. FTIR spectra of nominal compositions La2Ti2(1−x)Fe2xO7−ı as a function of Fe
content. (a) LTO, (b) LF(0.1)SS, (c) LF(0.2)SS, (d) LF(0.4)SS, (e) LF(0.6)SS, (f) LF(0.8)SS,
and (g) LFOSS.
to one short Ti–O bond in TiO5 groups present in such compounds
[31] and also in tetrahedrally coordinated TiO4 units, as smaller the
coordination number, the shorter the bond length and higher the
vibrational frequencies [32]. No such possibilities arise in LTOSS
as the lattice consists of slabs of corner sharing TiO6 octahedra.
Therefore, vibrational bands observed at higher wave numbers in
LTOSS can be assigned to Ti–O stretching where the oxygen ion
is shared between two octahedrally coordinated titanium ion and
two nine-fold coordinated lanthanum ion.
As we substitute Fe at Ti site in LTOSS the higher vibrational
bands (>700 cm−1) gradually weakens (Fig. 4b–g) and ultimately
disappears in LF(0.8)SS which further confirms the XRD observa-
tions corresponding to the transformation from layered perovskite
to perovskite lattice. Besides, due to the combined effect of Ti–O
stretching and Fe–O stretching the band at ∼550 cm−1 broadens
with increase in Fe content (Fig. 4b–g). The peak at ∼ 416 cm−1
sharpens with increase in Fe substitution because of the pres-
ence of O–Fe–O deformation vibration as observed in curves b–g
of Fig. 4. The other extreme LFOSS (Fig. 4g) shows two bands, one
at 415 cm−1 which can be attributed to Fe–O–Fe deformation vibra-
tion and another very prominent and broad band at 567 cm−1 which
can be assigned to Fe–O stretching vibration [33].
3.5. Mössbauer spectroscopy
Fig. 5 presents room temperature Mössbauer spectra of nominal
compositions La2Ti2(1−x)Fe2xO7−ı typically for values of x = 0.1, 0.4
and 0.8 prepared by solid state (Fig. 5a–c) and gel combustion route
(Fig. 5d–g). The corresponding values of Mössbauer parameters:
Fig. 3. Scanning electron micrographs of (A) LTOGC, (B) LF(0.2)GC and (C) LF(0.4)GC
samples.
Table 3
Position and assignments of the observed vibrational bands for nominal compositions La2Ti2(1−x)Fe2xO7−ı with different values of x.
S. no.
x
Sample
Vibrations (cm−1) and their assignments
Complex vibrations involving both
ꢃTiO/ꢃFeO
ꢃTiO [30]
ꢃTiO [31]
La and Ti/ıFeOFe [29]
1
2
3
4
5
6
7
0.0
0.1
0.2
0.4
0.6
0.8
1.0
LTOSS
417
415
416
416
416
417
415 [32]
464
464
458
456
450w
490w
490
475sh
552vs
568
570vs
571vs
575vs
590vs,b
567vs,b [32]
618sh,b
616sh
620sh
615sh
750s
762
760w
770w
805s
807
804w
800w
LF(0.1)SS
LF(0.2)SS
LF(0.4)SS
LF(0.6)SS
LF(0.8)SS
LFOSS
450w
474w

164
K.K. Kartha et al. / Journal of Molecular Catalysis A: Chemical 335 (2011) 158–168
Fig. 5. Mössbauer spectra of nominal compositions La2Ti2(1−x)Fe2xO7−ı {LF(x)} synthesized by (a–c) solid state route and (d–g) gel combustion route. Component spectra are
shown by dotted lines. (a) LF(0.1)SS, (b) LF(0.4)SS, (c) LF(0.8)SS, (d) LF(0.1)GC, (e) LF(0.4)GC, (f) LF(0.6)GC, and (g) LF(0.8)GC.
magnetic hyperfine values (Hhf), isomer shift (ı), quadrupole split-
ting (ꢀEQ), and area in percentage of tetrahedral and octahedral
sites of Fe3+ ions at room temperature for La2Ti2−xFexO7−ı sam-
ples synthesized by both these routes gel combustion and solid
state are listed in Tables 4 and 5. Mössbauer spectra reveal that Fe
is present in trivalent state in all the substituted samples but its
coordinative environment depended upon Fe content. Mössbauer
spectrum of LF(0.1)SS could be fitted into one symmetric paramag-
netic doublet, with a large quadrupole splitting (ꢀEQ):0.643 mm/s
and isomer shift (ı) of ∼0.358 mm/s, line width (ꢄ ):0.474 mm/s
(Fig. 5a) as mentioned in Table 4. This paramagnetic doublet is
attributed to the presence of Fe3+ ions in an octahedral oxygen lat-
tice distributed homogeneously in the sample [34]. With increasing
iron content in samples LF(0.4)SS and LF(0.8)SS a gradual transi-
tion from paramagnetic to magnetic (ferro- or antiferro-) can be
noted. This is reflected clearly in the presence of two hyperfine-
split sextets in LF(0.4)SS and LF(0.8)SS as shown in curves b and
c, respectively, of Fig. 5. The isomer shift for these sextets varies
Table 4
Mössbauer parameters: magnetic hyperfine values (Hhf), isomer shift (ı), quadrupole splitting (ꢀEQ), and area in percentage of tetrahedral and octahedral sites of Fe3+ ions
at room temperature for nominal compositions La2Ti2(1−x)Fe2xO7−ı synthesized by solid state reaction.
La2Ti2(1−x)Fe2xO7−ı (SS)
Sample
Iron sites
Isomer shift (ı)
Quadrupole splitting
(ꢀEQ) (mm/s)
Hyperfine values
(Hhf) (kG)
Width (ꢄ ) (mm/s)
Area (%)
100
x = 0.1
x = 0.4
LF(0.1)SS
LF(0.4)SS
Doublet
0.358
0.643
–
0.474
Doublet
Sextet 1
Sextet 2
0.380
0.403
0.330
0.575
0.029
0.002
–
0.495
1.031
1.938
44.1
24.9
31.0
432.3
371.1
x = 0.8
LF(0.8)SS
Sextet 1
Sextet 2
0.378
0.356
0.043
0.041
492.6
464.9
0.478
1.020
31.1
68.9
Table 5
The magnetic hyperfine values (Hhf), isomer shift (ı), quadrupole splitting (ꢀEQ), and area in percentage of tetrahedral and octahedral sites of Fe3+ ions at room temperature
for nominal compositions La2Ti2(1−x)Fe2xO7−ı synthesized by gel combustion route.
La2Ti2(1−x)Fe2xO7−ı GC
Sample
Iron sites
Isomer shift (ı)
Quadrupole splitting
(ꢀEQ) (mm/s)
Hyperfine field
(Hhf) (kG)
Width (ꢄ ) (mm/s)
Area (%)
100
x = 0.1
x = 0.4
LF(0.1)GC
LF(0.4)GC
Doublet
0.313
0.677
–
0.474
Doublet
Sextet 1
0.364
0.468
0.579
0.414
–
0.479
0.512
79.65
20.35
516.6
x = 0.6
x = 0.8
LF(0.6)GC
LF(0.8)GC
Doublet
Sextet 1
Sextet 2
0.360
0.395
0.416
0.571
0.032
0.090
–
0.497
0.512
1.211
34.19
22.24
43.57
515.3
429.7
Doublet
Sextet 1
Sextet 2
0.193
0.399
0.383
0.001
0.033
0.029
–
1.237
0.532
1.211
03.31
30.61
66.08
510.7
474.2

K.K. Kartha et al. / Journal of Molecular Catalysis A: Chemical 335 (2011) 158–168
165
Fig. 6. The area under the peak corresponds to quantitative esti-
mation of H2 consumed for the reduction of Fe3+ → Fe0 in the
substituted samples. The amount of H2 consumption increases with
increase in Fe content in substituted samples. Partial Fe substitution
helped in facilitating completion of reduction cycle below 1000 ◦C,
which was extended beyond 1100 ◦C in parent samples, LTOSS and
LFOSS (Fig. 6a and f). Poor reducibility of LTOSS is in agreement
with the observation commonly seen in scheelites and pyrochlore
structures, oxygen rich compositions in which diffusivity of oxy-
gen is restricted. The changes in the TPR profiles as a result of Fe
substitution can be ascribed to the nonstoichiometry and imperfec-
tions generated in the single phased compositions. The decrease in
Tmax in LF(0.4)SS sample is attributed to greater non-stoichiometry
caused by substitution of an aliovalent cation at B site facilitat-
ing the lattice oxygen diffusion via anionic vacancies generated
in the process. The reduction behavior of the samples having sin-
gle LaFeO3 phase thus show reduction behavior of metal–oxygen
bond typically as that in perovskites. In these structures, oxy-
gen overstoichiometry is accommodated within the perovskite
framework in randomly distributed short range linear defects as
recently reported [1,7]. The two peak pattern observed at ∼820 ◦C
and ∼960 ◦C in TPR profile of LF(0.8)SS (Fig. 6e) is in accordance
with its Mössbauer spectra (Fig. 5c) where existence of Fe in two
distinct sites, i.e., octahedral and tetrahedral is revealed. Reduc-
tion of Fe in octahedral sites occurs at lower temperature than
tetrahedral sites as bulk oxygen from Fe–O bond in FeO6 coordi-
nation is labile as compared to desorptivity of oxygen from Fe–O
bond in FeO4 coordination. The Tmax of the individual peaks in
LF(0.4)SS (Fig. 6c), was found to be at ∼785 ◦C as compared to
∼822 ◦C in sample having 80% Fe substitution, i.e. LF(0.8)SS (Fig. 6e).
The increase in Tmax from LF(0.4)SS to LF(0.8)SS and LFOSS indi-
cates that lesser number of the defects is created in the sample
as the ideal perovskite is reached in samples having Fe content
higher than 80%. The mobility of lattice oxygen is restricted in the
ideal perovskite as compared to disordered lattice as evident by
Mössbauer results and therefore increase in Tmax is observed for
LF(0.8)SS and LFOSS samples as compared to disordered LF(0.4)SS
perovskite.
Fig. 6. TPR cycles of nominal compositions La2Ti2(1−x)Fe2xO7−ı synthesized by solid
state route. (a) LTOSS, (b) LF(0.2)SS, (c) LF(0.4)SS, (d) LF(0.6)SS, (e) LF(0.8)SS, and (f)
LFOSS.
between 0.3 and 0.4, the quadrupole splitting and values of other
Mössbauer parameters as mentioned in Table 4. These sextets may
be ascribed to Fe3+ cations located at distorted octahedral sites,
thus suggesting the existence of domains/regions with varying local
fields. This distortion in symmetry in LF(0.4)SS sample may arise
due to anisotropic deformation of the environment of the Fe3+ ions
due to oxygen vacancies generated by substitution at Ti4+ sites.
The sample LF(0.8)SS exhibits only two magnetic sextets (Fig. 5c)
with hyperfine magnetic field (Hhf) values are 492.6 kG (octahedral
sites) and 464.9 kG (tetrahedral sites), revealing the relative con-
centration of Fe in octahedral and tetrahedral sites, respectively,
as shown in Table 4. These results thus reveal the presence of iron
in at least two distinct environments, the relative abundance of
which may depend upon Fe-content. The Mössbauer spectroscopy
results also show that the hyperfine field (Hhf) increased progres-
sively, as one goes to LF(0.8)SS (Fig. 5c) from LF(0.4)SS (Fig. 5b),
indicating that Fe is more magnetically ordered in LF(0.8)SS. This
is further supported by the lower ꢀEQ (quadrapole splitting) val-
ues for LF(0.8)SS samples as compared to LF(0.4)SS as seen from
Table 4.
Similarly corresponding spectra recorded for samples prepared
by gel combustion route are shown in Fig. 5d–g, reveals that inten-
sity of the magnetic component increases with the increase in Fe
substitution. 60–80% Fe substituted samples prepared by gel com-
bustion shows the existence of Fe in three distinct environments
which is evident by presence of one paramagnetic doublet and two
magnetic sextets as shown in Fig. 5f and g. The large quadrupole
splitting (ꢀEQ) as observed from Table 5 for gel combustion sam-
ples is characteristic of very small crystallites (nanocrystalline
nature), their large surface to volume ratio causing large lattice
strain and a correspondingly large electric field gradient at the iron
nuclei [35].
The effect of modified surface properties of powder oxides and
higher surface area on the reduction behavior is shown in Fig. 7.
Typical first TPR cycle of LF(0.4)GC and LF(0.4)SS sample is shown
in Fig. 7. LF(0.4)GC sample synthesized by gel combustion route
shows a reduction peak at 510 ◦C in addition to main reduction band
at ∼800 ◦C as seen in Fig. 7. The higher surface area (68 m2 g−1),
and decreased particle size (14–17 nm) of sample as mentioned in
Table 2, synthesized by gel combustion route induces ease in dif-
fusivity and reactivity of hydrogen lowering the activation barrier,
induces ease in reducibility of sample at lower temperature.
3.7. Phase identification of reduced samples
The XRD patterns of TPR residue of LTOSS and Fe substituted
LF(0.8)SS samples after reduction are compared with the fresh sam-
ples in Fig. 8. The reduction in 5% H2/Ar did not affect the XRD of
these samples significantly as evident from Fig. 8. Thus the oxygen
excess titanate LTOSS show considerable structural and phase sta-
bility even after reduction. Since there is no significant change in
lattice after reduction for LF(0.8)SS, it gives an indication that along
with Fe3+, Ti4+ is also undergoing reduction and converts to metal-
lic Fe and Ti3+. This indicates that hydrogen swipes away the extra
oxygen from the system and stabilizes the reduced species Ti3+ in
the ABO3 type of lattice. The considerable phase stability of LTOGC
on reduction makes it a preferred prospective support [36] for dis-
persion of metal ions where it can show excellent oxygen storage
capacity.
3.6. Temperature programmed reduction (TPR)
The TPR profiles of nominal compositions La2Ti2(1−x)Fe2xO7−ı
(0.0 ≤ x ≤ 1.0) are presented in Fig. 6. Pristine LTOSS sample shows
a weak reduction profile with minimal reduction (Fig. 6a). Substitu-
tion by reducible cation, Fe3+ in place of Ti4+ enhanced reducibility
in temperature range of 700–800 ◦C as shown in curves b–f of

166
K.K. Kartha et al. / Journal of Molecular Catalysis A: Chemical 335 (2011) 158–168
Fig. 9. The temperature dependent activity of nominal compositions
La2Ti2(1−x)Fe2xO7−ı synthesized by solid state (a–g) and gel combustion route
(h–m) for CO + N2O reaction.
Fig. 7. Effect of synthetic route on reduction behavior of the sample is shown by
comparing first TPR profile of LF(0.4)GC (a) and LF(0.4)SS (b) prepared by gel com-
bustion and solid state route, respectively.
and gel combustion (h–m) for decomposition of N2O using CO
viz.; CO + N2O reaction is exhibited in Fig. 9. The reactant mix-
ture of CO:N2O:He = 2:2:16 was fed to the catalyst at a rate of
40 ml/min. Partially Fe substituted samples have shown better
activity as compared to unsubstituted samples. Unsubstituted
samples, La2Ti2O7 (LTOSS) and LaFeO3 (LFOSS) have shown poor
activity of ∼80% and 100% at 700 ◦C as shown in Fig. 9a and b,
respectively. LF(0.1)SS and LF(0.2)SS (curves c and d) samples have
shown slightly better activity than unsubstituted LTOSS but similar
to LFOSS phase. Whereas LF(0.4)SS sample resulted in consider-
able rise in activity as shown in curve e with 80% conversion at
450 ◦C in contrast to mere 10% and 35% conversion by unsub-
stituted samples. Maximum conversion of 100% was observed
over the LF(0.4)SS sample at 550 ◦C for CO + N2O → CO2 + N2 reac-
tion (Fig. 9e). The light off temperature for LF(0.4)SS sample
was at ∼275 ◦C much lower as compared to ∼400 ◦C in case
of unsubstituted sample LTOSS. LF(0.6)SS and LF(0.8)SS with
higher Fe content have considerably activity (Fig. 9f and g)
but lower than LF(0.4)SS sample. The decreasing order of activ-
ity LF(0.4)SS > LF(0.6)SS > LF(0.8)SS > LF(0.2)SS > LFOSS > LF(0.1)SS >
LTOSS was observed in solid state samples.
Gel combustion synthesis has imparted enhanced catalytic
activity to these La2Ti2(1−x)Fe2xO3+ı samples as compared to their
solid state analogues (Fig. 9h–m). Catalytic activities of unsubsti-
tuted LTOGC and LFOGC with single phase La2Ti2O7 and LaFeO3
phases, respectively, have shown mere ∼20% and ∼45% conversion
at 400 ◦C, respectively, in curves h and m of Fig. 9. Thus, pure
single phase are not responsible for high activity. LF(0.2)GC and
LF(0.8)GC (curves i and m) samples have shown 100% conversion
at ∼450 ◦C. LF(0.4)GC sample resulted in considerable rise in
activity as shown in curve j with 100% conversion at 425 ◦C in
comparison to other samples. Among all samples, maximum
conversion of 100% occurred over LF(0.6)GC at ∼395 ◦C as shown in
Fig. 9k for CO + N2O → CO2 + N2 reaction. The catalytic activity for
CO + N2O reaction on gel combustion synthesized samples follows,
Fig. 8. XRD patterns of unsubstituted La2Ti2O7 (LTOSS) and Fe substituted
La2Ti0.4Fe1.6O7−ı {LF(0.8)SS} samples before and after First TPR cycle.
3.8. Catalytic activity
3.8.1. CO + N2O reaction
The temperature dependent catalytic activity of nominal com-
positions La2Ti2(1−x)Fe2xO3+ı synthesized by solid state route (a–g)
LF(0.6)GC > LF(0.4)GC > LF(0.8)GC ≈ LFOGC > LF(0.2)GC > LTOGC
a

K.K. Kartha et al. / Journal of Molecular Catalysis A: Chemical 335 (2011) 158–168
167
100
J
g
d
e
f
90
80
h
c
b
70
60
i
50
40
30
20
10
0
a
a) LTOSS
b) LFOSS
c) LT(0.2)SS
d) LT(0.4)SS
e) LT(0.6)SS
f) LTOGC
g) LF(0.2)GC
h) LF(0.4)GC
i) LF(0.8)GC
J) LF(0.6)GC
200
300
400
500
600
700
Temperature °C
Fig. 11. Temperature dependent catalytic activity of nominal compositions
La2Ti2(1−x)Fe2xO7−ı samples synthesized by solid state and gel combustion routes
for CO + O2 reaction.
Fig. 10. Effect of Fe substitution on T50 (a, c)and T100 (b, d), temperatures at
which 50% and 100% conversions were obtained over nominal compositions
La2Ti2(1−x)Fe2xO7−ı synthesized by gel combustion (a, b) and solid state (c, d) routes
for CO + N2O reaction.
reaction, partially Fe substituted samples have shown better activ-
ity as compared to both the unsubstituted phase LTOSS and
LFOSS. Maximum conversion of 100% at ∼325 ◦C with an onset
at ∼175 ◦C over LF(0.6)GC combustion route samples (Fig. 11j).
Whereas among solid state samples maximum activity of 100%
conversion at 550 ◦C, with the onset at 250 ◦C for CO2 forma-
tion over LF(0.4)SS (curve d) was observed. All other samples,
particularly unsubstituted phases, have shown reduced activity
(Fig. 11). The decreasing order of activity for CO + O2 reaction
is LF(0.6)GC > LF(0.4)GC ≈ LF(0.8)GC > LF(0.2)GC > LTOGC for gel
combustion and LF(0.4)SS > LF(0.6)SS > LF(0.2)SS > LFOSS > LTOSS
for the solid state synthesized samples as revealed from
Fig. 11.
trend similar as above for solid state samples at lower activity level
with an exception for the most active sample (Fig. 9).
Plot of T50 and T100 values vs. x (extent of substitution) in nom-
inal compositions La2Ti2(1−x)Fe2xO3+ı is shown in Fig. 10. T50 and
T100 suggests the temperatures at which 50% and 100% conversion
for CO + N2O reaction achieved in all substituted and unsubstituted
samples synthesized by both the routes. Both T100 and T50 values
are lowest for LF(0.4)SS and LF(0.6)GC suggesting that 40–60% Fe
substitution leads to catalytically active compositions synthesized
by either route for CO + N2O reaction. Gel combustion synthesis has
improved activity in temperature range of 395–525 ◦C lower by
∼150–225 ◦C than corresponding solid state samples (550–750 ◦C)
although there is not much decrease in the onset of reaction as
evident by curves b and d of Fig. 10. The difference in T50 and T100
values (curves a and b) of samples synthesized by GC route is ∼28 ◦C
much lower as compared to corresponding curves (c and d) of solid
state synthesized samples (250 ◦C) in Fig. 10. Thus, after the onset
of reaction, rate increases steeply with rise in temperature over
fine powders of La2Ti2(1−x)Fe2xO3+ı synthesized by gel combus-
tion, in contrast to sluggish rate observed in solid state analogues
(Figs. 9 and 10).
3.8.3. Surface–bulk-activity correlations
The decreasing order of activity as revealed from
Figs. 9–11 for CO + N2O as well as CO + O2 reaction is:
LF(0.4)SS > LF(0.6)SS > LF(0.2)SS > LFOSS > LTOSS in case of solid
state prepared samples and LF(0.6)GC > LF(0.4)GC ≈ LF(0.8)GC >
LF(0.2)GC > LTOGC for combustion samples. The activity order is
similar for the samples synthesized by same route irrespective of
the reaction. In both the reactions, CO oxidation occurs either by
N2O or by O2, the maximum activity was obtained over LF(0.6)GC
and LF(0.4)SS samples. Substituted samples with 40–60% Fe are
most active catalytic formulations irrespective of the preparation
route. However, 100% conversion for CO oxidation occurs at 325 ◦C
as compared to 395 ◦C for CO + N2O reaction over LF(0.6)GC, the
most active sample, suggesting that CO oxidation is the driving
force for the reduction of N2O by CO. The difference in activity
can be attributed to presence of oxygen in CO oxidation which
assists the oxidation reaction over LF(0.6)GC sample, also O2 is
stronger oxidizing agent as compared to N2O. Whereas in CO + N2O
reaction, CO is oxidized by facile lattice oxygen as suggested by
catalytic activity results which is further confirmed by in situ FTIR
studies presented elsewhere [9].
3.8.2. CO oxidation
Nominal compositions La2Ti2(1−x)Fe2xO3+ı were also evalu-
ated for CO oxidation reaction to support and establish the
structure–activity correlations. In this case, the reactant mix-
ture comprising of CO:O2:He = 2:1:17 was fed to the catalyst at
a rate of 40 ml/min. Fig. 11 presents the temperature depen-
dent catalytic activity of La2Ti2(1−x)Fe2xO3+ı solid state (a–e)
and combustion route (f–j) samples for CO + O2 reaction. Sam-
ples prepared by softer route shifted the temperature window
from 400–700 ◦C (Fig. 11a–e) to 200–400 ◦C (curves f–j). Gel
combustion synthesis has induced activity in all Fe substituted
samples even at 325 ◦C, whereas corresponding solid state sam-
ples do not show any conversion at 325 ◦C. Similar to CO + N2O
Catalytic activity of solid state samples calcined at 1400 ◦C
with negligible surface area could be attributed only to crystal
structure and modified bulk properties as a result of Fe substi-
tution. Substitution by more reducible cation, i.e., Fe3+ at Ti4+

168
K.K. Kartha et al. / Journal of Molecular Catalysis A: Chemical 335 (2011) 158–168
sites resulted increased reducibility with decreased Tmax observed
in TPR profile of LF(0.4)SS. Studies investigated by Mössbauer
spectroscopy also corroborate these results wherein asymme-
try around the Fe3+ ions imparted by nonstoichiometry-induced
anionic vacancies is clearly evident (Fig. 4d) and are responsi-
ble for maximum conversion of 100% at 550 ◦C observed over
LF(0.4)SS. In case of samples prepared by gel combustion, surface
surface area. This suggests that along with enhanced adsorption
active sites on the surface, crystal structure also plays an impor-
tant role in deciding their catalytic properties. While modified bulk
properties, like maximum asymmetric distortion around Fe3+ atom,
enhanced reducibility and substitution induced-anionic vacancies
in nonstoichiometric LF(0.4)SS sample has maximized its activity
for both the reactions.
Reaction mechanism investigated by in situ FTIR studies by the
authors over these samples for the decomposition of N2O using CO
is recently published elsewhere [9].
area is greatly enhanced, particularly in LF(0.4)GC (∼68 m2 g−1
)
and LF(0.6)GC (∼47 m2 g−1), both are active but LF(0.6)GC shows
maximum activity as against their surface area trend. This sug-
gests that along with enhanced adsorption active sites on the
surface, crystal structure also plays an important role in decid-
ing their catalytic properties. Both LF(0.4)GC and LF(0.6)GC are
single phased LaFeO3 samples, where LF(0.6)GC sample is more
crystalline. An increased crystallinity ensures easy diffusion path of
lattice oxygen via anionic vacancies properly generated in nonsto-
ichiometric LF(0.6)GC sample. The Mössbauer spectroscopy results
also show that the hyperfine field (Hhf) increased progressively,
as one goes to LF(0.8)SS (Fig. 5c) from LF(0.4)SS (Fig. 5b), indi-
cating that Fe is more magnetically ordered in LF(0.8)SS. The
increase in Tmax from LF(0.6)SS to LFOSS sample (Fig. 6d and f,
respectively) may be attributed to increase in symmetric order-
ing of Fe in case of higher substituted LF(0.8)SS sample, as evident
by Mössbauer results (Fig. 3) resulting in decreased catalytic
activity of single phased crystalline LF(0.8)SS and LFOSS sam-
ples as compared to LF(0.4)SS. Thus, enhanced catalytic activity of
LF(0.6)GC is attributed to more active LaFeO3 phase with increased
surface area. While modified bulk properties, like maximum
asymmetric distortion around Fe3+ atom, enhanced reducibility
and substitution-induced anionic vacancies in nonstoichiometric
LF(0.4)SS sample has maximized its activity for both the reac-
tions.
References
[1] J. Canales Vázquez, S.W. Tao, J.T.S. Irvine, Solid State Ionics 159 (2003) 159–
165.
[2] W.F. Libby, Science 171 (1971) 499–500.
[3] M.M. Milanova, M. Kakihana, M. Arima, M. Yashima, M. Yoshimura, J. Alloys
Compd. 242 (1996) 6–10.
[4] J. Happel, M.A. Hnatow, L. Bajars, M. Kundrath, Ind. Eng. Chem. Prod. Res. Dev.
14 (1975) 154–158.
[5] J.G. MacChesney, H.A. Sauer, J. Am. Ceram. Soc. 45 (1962) 416–422.
[6] D.W. Hwang, H.G. Kim, J.S. Lee, J. Kim, W. Li, S.H. Oh, J. Phys. Chem. B 109 (2005)
2093–2102.
[7] J. Canales-Vazquez, J.C. Ruiz-Morales, J.T.S. Irvine, W. Zhou, J. Electrochem. Soc.
152 (7) (2005) A1458–A1465.
[8] C. Brink, C. Kroeze, Z. Klimont, Atmos. Environ. 35 (2001) 6299–6312.
[9] M.R. Pai, A.M. Banerjee, K. Kartha, R.V. Pai, V.S. Kamble, S.R. Bharadwaj, J Phys.
Chem. B 114 (2010) 6943–6953.
[10] F. Kapteijn, J.R. Mirasol, J.A. Moulijn, Appl. Catal. B 9 (1996) 25–64.
[11] P. Pomonis, D. Vatts, A. Lycourghiotis, C. Kordulis, J. Chem. Soc. Faraday Trans.
1 81 (1985) 2043–2051.
[12] A.K. Ladavos, P.J. Pomonis, Appl. Catal. B 2 (1993) 27–47.
[13] R.J.H. Voorhoeve, D.W. Johnson, J.P. Remeika, P.K. Gallagher, Science 195 (1977)
827–833.
[14] L.G. Tejuca, J.L.G. Fierro, J.M.D. Tascon, Adv. Catal. 36 (1989) 237–328.
[15] H. Tanaka, M. Misono, Curr. Opin. Solid State Mater. Sci. 5 (2001) 381–387.
[16] C.S. Swamy, J. Christopher, Catal. Rev. Sci. Eng. 34 (1992) 409–425.
[17] B. Viswanathan, Catal. Rev. Sci. Eng. 34 (1992) 337–352.
[18] M.R. Pai, A.M. Banerjee, S.R. Bharadwaj, S.K. Kulshreshtha, J. Mater. Res. 22
(2007) 1787–1796.
[19] M.R. Pai, B.N. Wani, N.M. Gupta, J. Mater. Sci. Lett. 21 (2002) 1187–1190.
[20] M.R. Pai, B.N. Wani, N.M. Gupta, J. Chem. Phys. Solids 67 (2006) 1502–1509.
[21] M.R. Pai, B.N. Wani, A.D. Belapurkar, N.M. Gupta, J. Mol. Catal. A: Chem. 223
(2004) 275–283.
[22] M.R. Pai, B.N. Wani, B. Shreedhar, S. Singh, N.M. Gupta, J. Mol. Catal. A: Chem.
246 (2006) 128–135.
[23] A.M. Banerjee, M.R. Pai, K. Bhattacharya, A.K. Tripathi, V.S. Kamble, S.R. Bharad-
waj, S.K. Kulshreshtha, Int. J. Hydrogen Energy 33 (2008) 319–326.
[24] X.-H. Huang, J. Chang, Mater. Chem. Phys. 115 (2009) 1–4.
[25] Z. Qiu, Y. Zhou, M. Lu, A. Zhang, Q. Ma, Solid State Sci. 10 (2008) 629–633.
[26] H. Taguchi, Y. Masunaga, K. Hirota, O. Yamaguchi, Mater. Res. Bull. 40 (2005)
773–780.
[27] R.A. Nyquist, R.O. Kajel, Infrared Spectra of Inorganic Compounds, Academic
Press, New York, 1971.
[28] D. Chen, X. Jiao, R. Xu, Mater. Res. Bull. 34 (1999) 685–691.
[29] J.T. Last, Phys. Rev. 105 (1957) 1740–1750.
4. Conclusion
Role of surface/bulk properties on catalyst performance was
emphasized in the present study by applying both citrate com-
bustion synthesis and conventional routes of synthesis. Here,
replacement of Ti4+ by lower valence Fe3+ cation, created anionic
vacancies that might favor lattice oxygen mobility and activity.
Combustion route helped in achieving single phase, nanosized
porous LF(0.4)GC and LF(0.6)GC samples at 800 ◦C in contrast
to 1400 ◦C required for solid state synthesis. Fe substitution led
to considerable rise in surface areas of order of ∼68 m2 g−1 and
∼47 m2 g−1 in LF(0.4)GC and LF(0.6)GC samples accompanied with
˚
formation of pores of 20–30 A radius, as compared to unsubstituted,
[30] U. Balachandran, N.G. Eror, J. Mater. Res. 4 (1989) 1525–1528.
[31] M.T. Paques-Ledent, Spectrochim. Acta Part A 32 (1976) 1339–1344.
[32] P. Tarte, Spectrochim. Acta Part A 23 (1967) 2127–2143.
[33] W. Zheng, R. Liu, D. Peng, G. Meng, Mater. Lett. 43 (2000) 19–22.
[34] M. Abdelmola, M. Petitjean, G. Caboche, J.-M. Genin, L.C. Dufour, Hyperfine
Interact. 156/157 (2004) 299–303.
[35] N.S. Gajbhiye, S. Bhattacharya, G. Balaji, R.S. Ningthoujam, R. Kumar Das, S.
Basak, J. Weissmuller, Hyperfine Interact. 165 (2005) 153–159.
[36] M.R. Pai, K.K. Kartha, A.M. Banerjee, A. Bokare, R. Tewari, G.K. Dey, V.S. Kamble,
S.R. Bharadwaj, 19th National Symposium on Catalysis (CATSYMP-19), held at
National Chemical Laboratory, Pune, India, during Jan 18–21, 2009.
sample, LTOGC with poor surface area of ∼12 m2 g−1. The increase
in Tmax from LF(0.6)SS to LFOSS (Fig. 6d and f, respectively) may be
attributed to increase in symmetric ordering of Fe in case of higher
substituted LF(0.8)SS, as evident by Mössbauer results (Fig. 4) also
manifested in their lower catalytic activity. Among all samples, the
maximum conversion of 100% occurred over LF(0.6)GC at ∼395 ◦C
for CO + N2O and at ∼325 ◦C for CO oxidation reaction. In conclu-
sion we may state that enhanced catalytic activity of LF(0.6)GC is
attributed to nonstoichiometric single LaFeO3 phase with higher