Applied Catalysis A: General 431–432 (2012) 69–78

Contents lists available at SciVerse ScienceDirect

Applied Catalysis A: General

journal homepage: www.elsevier.com/locate/apcata

High-throughput study of the iron promotional effect over Pt/WOx –ZrO2

catalysts on the skeletal isomerization of n-hexane
M.L. Hernandez-Pichardo a , J.A. Montoya de la Fuente b,∗ , P. Del Angel b , A. Vargas b , J. Navarrete b ,
I. Hernandez c , L. Lartundo d , M. González-Brambila c
a
Instituto Politécnico Nacional, ESIQIE, Av. IPN S/N Zacatenco, México D.F., Mexico
b
Instituto Mexicano del Petróleo, DIyP, Eje Central Norte Lázaro Cárdenas 152, 07730 México D.F., Mexico
c
Universidad Autónoma Metropolitana, Azcapotzalco, División de CBI, Av. San Pablo 180 Col. Reynosa, 02200 México D.F., Mexico
d
Instituto Politécnico Nacional, CNMN, Av. IPN S/N Zacatenco, México D.F., Mexico

a r t i c l e i n f o a b s t r a c t

Article history: Pt supported on tungstated zirconia catalysts (Pt/WZ) doped with different Fe contents were evaluated
Received 23 September 2011 in the n-hexane hydroisomerization reaction. High-throughput experimentation (HTE) approaches were
Received in revised form 11 April 2012 used to investigate the effect of the iron incorporation over the catalytic activity. It was found that the
Accepted 13 April 2012
presence of iron at low concentrations (0.5–1 wt%) increases the conversion and the yield to the high
Available online 21 April 2012
octane 2,2-dimethyl-butane (2,2-DMB) isomer. The characterization results suggest that the presence of
iron modifies the nanostructure of the WOx species on the zirconia surface which in turn produces the
Keywords:
generation of active Brønsted acid sites thus optimizing the tungsten content toward the highest catalytic
Iron promoter
High-throughput experimentation
activity. It was also found that the surfactant incorporation further improves the performance of Pt/WZ
Tungstated zirconia catalysts, although, the activity was improved with the Fe addition at lower concentration.
Surfactant © 2012 Elsevier B.V. All rights reserved.
n-Hexane hydroisomerization

1. Introduction activity than the Fe impregnated catalysts. Some other works

The hydrocarbons hydroisomerization process is widely
employed in the production of both fuels and petrochemicals. Par-
ticularly, the n-hexane hydroisomerization allows an increase in
total gasoline pool octane free of aromatics and benzene precursors
due to the production of isomers with high octane number. Solid
acids such as sulfated zirconia (SZ) and tungstated zirconia (WZ)
catalysts have attracted wide interest due to their high activity and
selectivity for n-alkanes isomerization processes [1–3]. In addition,
the use of some promoters has shown to improve the activity of
these catalytic systems; iron- and manganese-promoted sulfated
zirconia for example, catalyzes isomerization reactions two or three
orders of magnitude faster than sulfated zirconia [4]. On the other
hand, it has been found that although Fe and Mn promote sul-
fated zirconia for n-pentane isomerization, they do not promote
tungstated zirconia at the same time [5]. However the individual
role of iron or manganese is still not clearly understood.
In a study of individual promoters, Jatia et al. [6] found that
the incorporation of either Fe or Mn to sulfated zirconia increases
the catalytic activity but the Mn impregnated sample shows lower
∗ Corresponding author. Tel.: +52 55 9175 8375.
E-mail address: amontoya@imp.mx (J.A. Montoya de la Fuente).
have been dedicated to the study of the different modifiers for
nanocrystal zirconia on W/ZrO2 catalysts [7], and recently our
group published some works about the influence of the addition of
Mn in Pt/WZ catalysts; we found through high-throughput experi-
mentation approaches that Mn modifies the interaction of the WOx
species with the support but the catalytic behavior is not a linear
function of the Mn composition [8].
In the particular case of iron as a promoter of WZ catalysts, sev-
eral works have been reported [9–13]. Santiesteban et al. [9] found
that the Fe incorporation by coprecipitation into WZ enhances
the paraffin isomerization activity of these catalysts showing that
the acid sites of these solids are strong enough to perform the
isomerization reaction even by the monofunctional mechanism.
However, none information about the specific role of iron as a
promoter was provided. Additionally, Wong et al. [12] verified an
additional cooperative role of iron promoter in the doubly pro-
moted Pt/FWZ catalyst prepared by impregnation. They suggest
that highly dispersed Fe3+ species promote n-butane isomerization
reaction through a redox effect for the generation of Brønsted acid
sites. It has been also proposed that the promoting effect of Fe is a
combination of various effects such as the decoration of platinum
by iron oxide or allowance of the migration of activated species;
but in general, several studies have suggested a role as redox ini-
tiator for WZ catalysts containing Pt and Fe as promoters [10,11].

0926-860X/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apcata.2012.04.023
70 M.L. Hernandez-Pichardo et al. / Applied Catalysis A: General 431–432 (2012) 69–78
However, just as in SZ case, the nature of the promoting effect of
these metals is still very controversial. On the other hand, the struc-
tural and electronic properties of iron in WZ catalysts has been
studied on samples with similar compositions (about 15 wt% W
and 1 wt% Fe); such studies have shown that iron is present on
the surface as a mixture of ␣-Fe2 O3 , isolated Fe3+ anchored on WOx
surface, and Fe3+ oxide clusters located in the surface layer forming
“rafts” [12,13]. Some other studies propose that iron predominantly
forms a surface solid solution and a minor amount of iron is in the
form of small iron oxide particles [11,14]. Apparently all these dif-
ferences are produced by the synthesis method employed, as well
as by the tungsten-iron ratio used.
Therefore, there is a need for the study of the role of iron as
promoter on tungstated zirconia catalysts at different Fe and W
contents. This variation of the different synthesis parameters is pos-
sible by using high-throughput approaches. These techniques have
been used to study and develop multicomponent catalysts [15,16];
thus, this work is focused on the study of the influence of Fe on
the Pt/WZ catalysts performance prepared with different tungsten
and iron compositions over the n-hexane hydroisomerization reac-
tion. By using HTE techniques it is possible to compare the catalytic
properties of these materials under exactly the same conditions. In
addition, it has been found that WZ catalysts present high activity
despite their low surface area, therefore, we also evaluate the influ-
ence of the incorporation of surfactant in different compositions
over the textural properties of these materials.
2. Experimental
2.1. Catalysts preparation
High-throughput experimentation techniques were used for
the synthesis of 36 catalysts by surfactant-assisted coprecipitation
similar to the procedure described previously [8]. The synthesis
variables include: (a) tungsten content (10–20 wt% W), (b) iron
content (0–2 wt% Fe) and (c) surfactant content in terms of sur-
factant/zirconia molar ratio (Sz : 0–2) to give Fe/W ratios among
0.025–0.2. Iron promoted tungstated zirconia samples (FWZ)
were prepared from zirconyl chloride ZrOCl2 ·xH2 O, iron nitrate
Fe(NO3 )3 ·9H2 O, ammonium metatungstate (NH4 )6 W12 O39 ·xH2 O,
and cetyl-trimethylammonium bromide (CTAB) as surfactant, all
chemicals from Aldrich. Briefly, zirconia, tungsten and iron precur-
sors in aqueous solutions were coprecipitated with the surfactant
using NH4 OH and the precipitates were then aged in their mother
liquor at 80 ◦ C for 16 h washed and dried at 110 ◦ C for 24 h. Cal-
cination treatment was performed in static air at 800 ◦ C for 4 h.
Samples were labeled as Fx ZWy -Sz , where the subscripts (x) and
(y) represent the Fe and W contents in wt% respectively, and
Sz denotes the surfactant content as surfactant/zirconia molar
ratio.
Platinum impregnated catalysts (0.3 wt%) were prepared by
conventional impregnation of the Fx ZWy -Sz samples calcined at
800 ◦ C with a H2 PtCl6 ·6H2 O aqueous solution, followed by dried at
110 ◦ C and finally calcined at 400 ◦ C in static air for 3 h. The impreg-
nation procedure was also performed using the robot and only the
sample needed for the evaluation (100 mg) was impregnated with
Pt. In analogy with the previous nomenclature, the resulting mate-
rials were denoted as P/Fx ZWy -Sz , indicating that these materials
were impregnated with 0.3 wt% of platinum.
2.2. Catalytic testing
Catalytic activity was measured in the n-hexane hydroisomer-
ization reaction in a Combinatorial Multi Channel Fixed Bed Reactor
(MCFBR) (Symyx). This apparatus is appropriate to evaluate the
activity and selectivity of 48 samples in parallel by applying high-
throughput testing techniques with a matrix arrangement of 6 × 8.
This system consists of six reactor heads each one containing
eight micro-reactors of approximately 4 mm of inner diameter and
47 mm of length. The six reactor heads are connected indepen-
dently to six chromatographs (Agilent, 6850 Series) equipped with
a SPB-1 capillary column (Supelco) with a length of 100 m and a
flame ionization detector (FID) for the analysis of products. Cata-
lysts sample of 100 mg diluted in 200 mg of inert silicon carbide
were loaded in each well and fixed into the reactor heads. The
injection of reactant and products to the GC’s was carried out at
the reaction pressure. The reliability of the quantitative method
was evaluated with an uncertainty of the conversion ±2% mea-
surable through the 48 wells. Pretreatment of the catalysts was
carried out in situ prior to the activity test and it consisted in a
drying-reduction program, drying the samples at 260 ◦ C for 2 h in
helium (200 cm3 min−1 ) followed by reduction in a hydrogen flow
(200 cm3 min−1 ) at 450 ◦ C for 3 h. Hydrogen and n-hexane flows
were adjusted to give a H2 /n-C6 = 1.47 molar ratio. The reaction was
conducted at 260 ◦ C, 0.689 MPa, 3.7 h−1 WHSV and using a mixture
of 100 cm3 min−1 of H2 and 0.4 cm3 min−1 of n-hexane fed with an
HPLC pump.
2.3. Catalyst characterization
Fe–WOx –ZrO2 samples were characterized by X-ray powder
diffraction (XRD), UV–vis, Raman spectroscopy and scanning and
transmission electron microscopy (STEM). X-ray diffraction pat-
terns were obtained in a Bruker-Axs D8 Discover with GADDS
(General Area Detector Diffraction Systems, two-dimensional
detector) diffractometer fitted with a Cu tube (40 kV, 40 mA) using
HT approaches for both measurements and patterns evaluation.
Ultraviolet–visible (UV–vis) spectra were obtained using a Varian
(Cary 1G) spectrophotometer for the data handling with an inte-
gration sphere accessory and Raman spectra were recorded in the
100–1200 cm−1 wave number range using a ThermoNicolet Raman
apparatus (Almega model) equipped with a Nd:YVO4DPSS laser
source. The excitation line of the laser was 532 nm and the laser
power was of 25 mW.
The nitrogen adsorption isotherms and BET surface areas were
measured at 77 K with a Micromeritics, ASAP 2405. Prior to the
measurements, the samples were out-gassed at 350 ◦ C for 15 h in
vacuum. The pore size distribution was calculated following the
BJH method using the desorption branch of the isotherm.
FTIR of adsorbed pyridine experiments were performed in order
to determine the type and amount of surface acid sites of WOx –ZrO2
and some Fe–WOx –ZrO2 solid acids. The analyses were carried out
by using a Fourier transform infrared (FTIR) NICOLET spectrome-
ter Model MAGNA 560. Pretreatment of the samples prior to the
adsorption of pyridine consisted in outgassing followed by heating
to 500 ◦ C at 20 ◦ C min−1 and cooling to room temperature. After
the pretreatment, the samples were exposed to saturated pyridine
vapor for 10 min. IR spectra were recorded after desorption at room
temperature (RT), 100, 200, 300, and 400 ◦ C respectively. XPS spec-
tra of the catalysts were recorded on a K-Alpha Thermo Scientific
apparatus after excitation with a monochromatic Al K˛ radiation.
Calibration of the energy position of an XPS peak was performed by
using the binding energy of adventitious carbon 1s peak at 284.8 eV.
FTIR analyses for acidity measurements were carried out using the
“In-Situ” mode and Transmission-Absorption methodology with a
very thin self-supported wafer, in order to allow the IR beam trans-
mittance. The calculation of Brønsted and Lewis acid sites were
obtained by the method described by Emeis [17] where the molar
extinction coefficients were used to measure the acidity of the cat-
alysts using the intensities of the bands at 1545 and 1450 cm−1
in the mid-IR spectra. The Beer–Lambert law is the basis for this
 M.L. Hernandez-Pichardo et al. / Applied Catalysis A: General 431–432 (2012) 69–78 71

Table 1
Parameters used in the screening of the Pt/Fe–WOx –ZrO2 catalysts composition.

Parameter Component Content (wt%)

Acidity enhancer W 10, 15, 20

Promoter Fe 0, 0.5, 1, 2
Surfactant CTAB 0, 1, 2 a
Metallic promoter Pt 0.3 b
a
Surfactant/zirconia molar ratio.
b
Constant content.

calculation but it needs to be modified for heterogeneous systems,

then the Eq. (1) was used to calculate the integrated molar extinc-
tion coefficients (IMEC):
 S 1
IMEC = (IA) ; (1)
W n
where, IA = integrated absorbance in the peak region, S = geometric
area of the support wafer, W = weight of adsorbing sample, and
n = normalized amount of pyridine per unit weight; integrated
Fig. 1. High-throughput results of the catalytic activity of the 36 Pt/FWZ samples in
adsorption intensity. The values of n are constants that were cal-
the n-hexane hydroisomerization reaction (lead catalyst without Fe: red circle and
culated for the Brønsted or Lewis acid sites. The final values of the lead Fe promoted catalysts: rectangle). (For interpretation of the references to color
integrated adsorption intensity coefficients are: n = 1.3159 ␮mol/g in this figure legend, the reader is referred to the web version of the article.)
for Brønsted and n = 1.4158 ␮mol/g for Lewis. The wafer’s diame-
ter was 1.3 cm and the thin wall was 80 ␮m (measured by MEB) in
order to allow the transmission of IR beam. since the other bi-ramified isomer, the 2,3-dimethyl-butane (2,3-
Finally, scanning and transmission electron microscopy (STEM) DMB), reaches the equilibrium faster. Fig. 1 showed the complete
was used to image the catalyst with a JEOL JEM-2200FS with Schot- data of the catalytic activity for the 36 catalysts complete library.
tky type field emission gun operating at 200 kV. HAADF-STEM The catalytic activity was expressed as 2,2-DMB yield in order to
detector was used to acquire the images. The elemental com- follow the catalytic evolution according to the synthesis parame-
position and mapping were obtained by energy dispersive X-ray ters considered in the library (Table 2). After that, the lead catalysts
spectroscopy (EDS) with an NORAN spectrometer fitted to the were distinguished with conversion values higher than 52% and
STEM. 5.2% mol of selectivity towards 2,2-DMB (that corresponds to the
lead catalyst without Fe). The lead catalysts, which corresponded to
3. Results and discussion the iron promoted, 0.5 and 1% Fe, were indicated inside the rectan-
gle. Nevertheless, catalysts with similar selectivity level and higher
3.1. Library description conversions were also considered because are in the down limit of
the upright corner.
A library with 36 catalysts was prepared by coprecipitation with Fig. 2 showed the results obtained with catalysts prepared with
three variables: tungsten, iron and surfactant contents, to give Fe/W three different tungsten contents, Figs. 2(a) 10% W, (b) 15% W,
ratios from 0.025 to 0.2. Table 1 depicts the variables of the library and (c) 20% W, promoted with three surfactant contents and Fe/W
for the catalytic activity screening, and Table 2 showed the synthe- ratios. It was found that there was an optimum Fe/W ratio at about
sis parameters of the resultant matrix. 0.05–0.07 regardless of the tungsten or surfactant content. More-
over, the results in Fig. 2 also indicates that the catalysts prepared
3.2. Catalytic activity with surfactant presented higher activity than the catalysts with-
out this texture promoter, being the intermediate ratio Sz = 1 the
The catalytic activity of Pt/ZW and Pt/ZW-Fe samples at different optimum surfactant content, for the three W contents. It is impor-
Fe compositions was evaluated in the n-hexane hydroisomerization tant to notice that the optimal W content was from 10 to 15%,
at 260 ◦ C. Among the reaction products, the bi-ramified isomers as it was found previously [8]. However, the catalysts promoted
present the higher octane number. In this regard, the catalytic with Fe between 0.05 and 0.07 Fe/W ratios showed the highest 2,2-
results were compared in terms of the influence of the Fe/W ratio DMB yield. In contrast, if Fe/W ratio is higher than 0.1 produces the
as a function of the yield to 2,2-dimethyl-butane (2,2-DMB) isomer decrease of the catalytic activity.
The influence of Fe content over the catalysts with 10, 15 and
20 W, without the surfactant effect is shown in Fig. 3. The indi-
Table 2
vidual effect of Fe over the 2,2-DMB yield can be observed. It was
Matrix of the Fe/W ratios used in the screening of the activity of the Pt/Fe–WOx –ZrO2
catalysts.
found that the yield of the catalysts promoted with 0.5% showed
the highest yield and decreases linearly when the tungsten con-
0% Fe 0.5% Fe 1% Fe 2% Fe tent increases. However, when Fe content was 1.0, the behavior
10%W 0 0.05 0.10 0.20 changed, and a maximum was observed at 15% W. A drastic 2,2-
15%W 0 0.03 0.07 0.13 DMB yield decrease was observed for the catalysts containing 2.0%
20%W 0 0.025 0.05 0.10
of Fe, even below the unpromoted catalysts.
10%W 0 0.05 0.10 0.20
15%W 0 0.03 0.07 0.13 The analysis of the lead catalysts (Table 3) on the n-hexane
20%W 0 0.025 0.05 0.10 isomerization reaction also indicates that the most active formula-
10%W 0 0.05 0.10 0.20 tions, in terms of the conversion and selectivity to the 2,2-DMB
15%W 0 0.03 0.07 0.13 isomer, are mainly catalysts with an iron–tungsten Fe/W ratio
20%W 0 0.025 0.05 0.10
between 0.05 and 0.07 that is to say catalysts between 10 and
Sz = 0, Sz = 1 and Sz = 2. 15% W promoted with 0.5–1% Fe. These samples are then closer
72 M.L. Hernandez-Pichardo et al. / Applied Catalysis A: General 431–432 (2012) 69–78

Fig. 2. Catalytic performance of the Pt/FWZ materials in the n-hexane hydroisomerization reaction at different tungsten contents (3.7 h−1 WHSV, 260 ◦ C and 0.69 MPa): (a)
10% W, (b) 15% W, (c) 20% W and (d) Global results: (—) catalyst with 10 and 15% W and (- - -) catalysts with 20% W.

5.0 (3 MP), 2,2-dimethyl butane (2,2-DMB) and 2,3-dimethyl butane

(2,3-DMB), as well as the light products formed by cracking and
4.5
0.5% Fe Sz=0 also probably by n-hexane hydrogenolysis. Thus, the catalysts syn-
4.0 thesized with low iron contents Fe/W (0.5–1% Fe) present the
Yield to 2,2-DMB (%)

3.5
higher formation of bi-ramified isomers, being an indication of
the higher activity of the Fe-promoted catalysts. These results
3.0 also show that the tungsten content can be optimized by the
1.0% Fe samples promotion with iron, since the samples with low tung-
2.5
sten content (10% W) promoted with 0.5% Fe show high catalytic
2.0 activity even superior to the catalysts with 15% W without Fe
0 % Fe promoter.
1.5
Moreover, some authors have proposed that the hydroisomer-
1.0 ization of n-alkanes on Pt/ZW catalysts proceeds through the
0.5 2.0 % Fe nonclassical bifunctional mechanism [18], since the most impor-
tant role of platinum on the isomerization activity of Pt/WZ
0.0
catalysts is to stabilize the catalytic activity, or to supply disso-
8 10 12 14 16 18 20 22
ciated hydrogen necessary for the generation of Brønsted acid sites
W wt. % on the partially reduced WOx domains [9,19], and other function
is to supply enough protons to olefins avoiding the cracking and
Fig. 3. Correlation between Fe content and the yield to the 2,2-DMB as function of
the tungsten content. the formation of carbon, which in turn is going to block the acid
sites. So, even though the catalytic activity requires the presence
of Pt, the reaction is mainly controlled by the acidic properties
to the tungsten loadings for the theoretical polytungstate mono- of the WZ. Then, it was observed from the results of the catalytic
layer needed for the generation of catalytically active sites in evaluation that for improving the catalytic performance the main
coprecipitated catalysts. The products distribution results, Table 3, variables are the iron incorporation along with the surfactant-
indicate that all catalysts show sufficient activity for the formation assisted synthesis, since the Pt content was maintain constant at
of the four isomers: 2-methyl-pentane (2 MP), 3-methyl-pentane 0.3%.

Table 3
Non-promoted and lead catalysts in the n-hexane isomerization reaction at 260 ◦ C: composition, iron/tungsten ratio, conversion (% mol) and products distribution.

Catalyst %W % Fe Sz a Fe/W XA b
(% mol) Products Distribution

2,2-DMB 2,3-DMB 2DMP 3DMP C5 - c

P/F0 ZW10 -S0 10 0 0 0 39 3.3 9.3 52.2 34.6 0.6

P/F0 ZW15 -S0 15 0 0 0 53 5.4 10.3 50.4 33.5 0.4
P/F0 ZW20 -S0 20 0 0 0 37 3.6 11.0 50.7 33.5 0.2
P/F1 ZW15 -S1 15 1 1 0.07 71 6.2 12.3 47.7 31.4 2.4
P/F0.5 ZW10 -S1 10 0.5 1 0.05 68 6.5 12.6 45.8 30.0 5.1
P/F0.5 ZW10 -S2 10 0.5 2 0.05 61 6.9 12.1 47.8 31.5 1.7
P/F0.5 ZW10 -S0 10 0.5 0 0.05 69 5.7 12.7 47.7 31.1 2.8
P/F1 ZW15 -S2 15 1 2 0.07 70 5.5 12.6 47.7 31.5 2.7
P/F1 ZW15 -S0 15 1 0 0.07 63 5.2 12.5 47.5 31.2 3.6
P/F1 ZW20 -S1 20 1 1 0.05 62 5.2 12.1 48.0 31.4 3.3
P/F1 ZW10 -S1 10 1 1 0.10 61 5.7 12.0 48.3 31.7 2.3
P/F0.5 ZW15 -S1 15 0.5 1 0.03 58 5.1 11.1 48.3 31.8 3.7
P/F1 ZW10 -S2 10 1 2 0.10 56 4.0 12.9 48.8 31.8 2.5
a
Surfactant content (zirconia/surfactant molar ratio).
b
n-hexane conversion (% mol).
c
Light products (C5 -) from cracking and hydrogenolysis.
 M.L. Hernandez-Pichardo et al. / Applied Catalysis A: General 431–432 (2012) 69–78
WO3 WO3
intensity (a. u.)
15 20 25 30 35
Fig. 4. XRD patterns of the Fe–WOx –ZrO2 samples calcined at 800 ◦ C, at different tungsten and iron contents, synthesized without surfactant (Sz = 0).
These results suggest that an intermediate tungsten coverage
on the zirconia surface is needed, which could be reached by con-
trol the tungsten content and the textural properties of samples
to produce the catalytically active Brønsted acid sites generated
by the slight reduction of the WOx domains. Recently Zhou et al.
[20] has shown that the incorporation of zirconium cations into the
WOx clusters form the most active species in tungstated zirconia
catalysts; thus it is possible that the addition of iron to the cat-
alytic system at low contents promotes the incorporation of some
zirconium cations to the WOx clusters and modifies the tungsten
surface density and as a consequence, the formation of the acid
sites. Hence, FTIR of adsorbed pyridine analysis were performed in
order to determine the type and amount of surface acid sites, as
well as other characterization results of these catalysts which are
discussed in Section 3.3.
3.3. Characterization results
3.3.1. Effect of the iron incorporation
The X-ray diffraction patterns of Fe-doped WZ samples without
surfactant (Sz = 0) at different tungsten and iron contents, calcined
at 800 ◦ C are shown in Fig. 4. It is important to mention that
these diffraction patterns were measured using HTE equipment,
therefore, the reflection positions and intensity can be directly com-
pared. All of these samples present the characteristic reflections of
zirconia in tetragonal phase, with some traces of the monoclinic
phase. It is also possible to observe the reflections due to the tung-
sten oxide in monoclinic phase (2 ∼ = 23–25); most of the patterns
present the formation of WO3 crystallites indicating a tungsten
surface density superior to the monolayer (7–10 ats W nm−2 ), with
WO3 domains larger than 3 nm, which is the detection limit of XRD.
The comparison between these patterns show that the growth of
the WO3 crystallites depends on the tungsten content, however as
the iron content increases, the intensity of the WO3 diffractions
also increases, suggesting that one of the iron effects is to facilitate
the growing of the WOx domains on the zirconia surface at higher
Fe contents (1–2% Fe). On the other hand, none of these patterns
present any evidence of the formation of iron oxides, indicating
the formation of a solid solution or highly dispersed species of iron
oxides.
The XRD evidence is supported by the UV–visible analyses; this
technique has been used to correlate the optical absorption edge
73
Sz= 0
: Tetragonal zirconia
%Fe
20% W - 2% Fe
15% W - 2% Fe
10% W - 2% Fe
20% W - 1% Fe
15% W - 1% Fe
10% W - 1% Fe
20% W - 0.5% Fe
15% W - 0.5% Fe
10% W - 0.5% Fe
20% W - 0% Fe
15% W - 0% Fe
10% W - 0% Fe
40 45 50 55 60 65 70 75 80
2θ
energy with the average degree of aggregation of WOx and MoOx
clusters [21]. This approach was used in this work in order to com-
pare the dispersion degree of the WOx phase. Fig. 5 shows the
UV–vis spectra of samples with 10%W without surfactant (Fig. 5a)
and 15%W with surfactant (Fig. 5b). The values of the electronic
edge energy (Eg) were determined by finding the energy intercept
of a straight line fitted trough the low energy rise in the graphs
of [F(R∞ ) x h]1/2 vs. photon energy. It is observed that the Eg
decreases to lower energy values from 2.7 to about 1.5 eV with
increasing Fe loading, in samples with the same tungsten content.
These Eg values correspond to about 6 covalent bridging W O W
bonds of the central W(VI) cation, according to the empirical corre-
lation between Eg and the number of nearest cations surrounding
the central W cation proposed by Ross-Medgaarden and Wachs
[22]. Therefore, spectra in Fig. 5 suggest that the iron incorpora-
tion increases the aggregation degree of the WOx species on the
zirconia surface; the intensity of the absorption band decreases
with a higher Fe concentration as it was expected due to the growth
and change of coordination symmetry of the polytungstate species.
The decrease in energy absorption edge indicates that the addi-
tion of Fe by coprecipitation facilitates the electronic transfer from
the valence band to the conduction band, probably because of the
incorporation of this cation to the zirconia lattice that hinders the
incorporation of tungsten and generates WOx species with larger
domain size. Martínez et al. [19] found that the WOx clusters in WZ
samples prepared by coprecipitation are characterized by a lower
UV–vis absorption edge energy and enhanced reducibility com-
pared with those generated by impregnation. They proposed that it
was due to enhanced formation of anion vacancies by substitutional
intergrowth at the interface between t-ZrO2 and WOx species, and
then it is possible that iron may occupy those vacancies.
Fig. 6 shows the Raman spectra of catalysts with 15% W and a
surfactant–zirconia ratio equal to 1 (Sx = 1); all samples present the
typical bands of the tri-dimensional crystalline tungsten oxide at
272, 324, 716 and 808 cm−1 , as well as the band at higher frequency
at 993 cm−1 related to the stretching mode of the W O bond cor-
responding to a mono-oxo species in highly distorted octahedral
coordination [23]. The comparison between samples prepared with
iron (B–D) with the sample without iron (A) shows that the Fe-
promoted catalysts present a higher spectral intensity of the bands
corresponding to WO3 than the unpromoted sample, while this last
catalyst also present slight bands corresponding to the tetragonal
74 M.L. Hernandez-Pichardo et al. / Applied Catalysis A: General 431–432 (2012) 69–78

a 5 zirconia at about 460 and 653 cm−1 , which is an indication of its

A) Fe(0) ZW(10) Sx= 0 lower tungsten surface density, the fact that the intensity of these
B) Fe(0.5) ZW(10) bands diminish with the iron incorporation could reinforce the sug-
4 C) Fe(1.0) ZW(10) gestion of the increase of the segregation of the WOx species in
D) Fe(2.0) ZW(10) samples prepared with this dopant, since the growing of the WO3
bands at about 716 and 808 cm−1 avoids the observation of the
3 bands corresponding to zirconia, which is in agreement with the
1/2

XRD and UV–vis results discussed above.

[F(R00)*hν]

HAADF-STEM, EDS-STEM semi-quantitative analyses as well as

2 a chemical mapping were obtained with a beam probe with size
below one nanometer, in order to determine the dispersion degree
D of Fe, Pt and W in the F1 ZW10 -S0 sample. Fig. 7a shows a HAADF-
C STEM image, bright spots corresponded to elements with high
1 B
A atomic number like Pt and W [20]. Fig. 7b corresponded to the EDS
spectrum obtained from this HAADF-STEM image, in this case the
signal of Fe and Pt is very low, however when the analysis was
0
made in a bright region (punctual EDS analysis), marked by the
1 2 3 4 5 6
circle, the intensity of Fe (6.403 keV) and Pt (9.441 keV) is increased
Photon Energy (eV) as it is shown in Fig. 7c. On the other hand, in Fig. 7d the map-
ping indicates that in this catalyst, with low iron content (1%), the
b 6 iron signals are distributed over the entire surface which indicates
A) Fe(0) ZW(15) Sx= 1
that this component is homogeneously dispersed on the surface,
B) Fe(0.5) ZW(15)
5 C) Fe(1.0) ZW(15)
probably including on WOx surface. The peaks of Cu that appear in
D) Fe(2.0) ZW(15) the EDS spectra are due to the grid used to observe the sample in
the microscope. Furthermore, chemical mapping presents spurious
4 X-ray irradiation [24], however it is clear that the most impor-
1/2

tant emission is coming from the irradiated region of the sample.

[F(R00)*hν]

Several studies on Pt/ZW-Fe catalysts by XANES, EXAFS and Möss-

3
bauer spectroscopy [11–13], where the iron was incorporated by
C B A
impregnation have found that the iron promoter exists as highly
2 dispersed Fe3+ species either bounded to WOx surface or located at
the surface vacant sites of zirconia, as well as ␣-Fe2 O3 clusters on
the surface of catalyst. However, all of these authors found that the
1 D iron incorporation via impregnation generates a higher dispersion
of the WOx phase, whereas in this work, iron was incorporated by
0 coprecipitation and we have observed that the coprecipitation of
1 2 3 4 5 6 iron with the Zr(OH)4 produces a growth of the WOx clusters. This
growth could generate a promotional effect or hinder the activity
Photon Energy (eV)
trough the modification of the WOx nanostructures which gener-
Fig. 5. Diffuse reflectance UV–vis absorption spectra of Fe–WOx –ZrO2 samples at ate the acid sites depending on the tungsten and surfactant content.
different iron contents: (a) samples with 10% W without surfactant and (b) samples Zhou et al. [20] suggest that the incorporation of zirconium cations
with 15% W with surfactant. to the WOx clusters to form active species, should be easier from the
metastable decomposing ZrOx (OH)4−2x than a stable pre-formed
ZrO2 , showing the important differences produced by the synthesis
A) 15% W - 0.0% Fe Sx= 1 method.
808
B) 15% W - 0.5% Fe Fig. 8 shows the IR spectra of adsorbed pyridine (Fig. 8a and
C) 15% W - 1.0% Fe b) and acidity which corresponded to the integrated IR vibration
D) 15% W - 2.0% Fe bands (Fig. 8c and d) of F0 ZW15 -S0 and F1 ZW15 -S0 samples, respec-
Raman Intensity (a.u.)

tively. The IR spectra of these catalysts contain the characteristic

716
bands of pyridine coordinatively bonded to Lewis acid sites at
about 1450 and 1608 cm−1 , as well as the bands at about 1540 and
272
324
1640 cm−1 that are assigned to protonation of pyridine by Brønsted
993
D sites [25,26]; whereas the band at 1489 cm−1 is a combined band
C originated from pyridine bonded to both Brønsted and Lewis acid
sites [26,27]. The intensity of the IR vibration bands for both Lewis
B
653 and Brønsted acid sites decreased when temperature was increased
A
indicating the strength of the acid sites. After raising the temper-
460 ature up to 400 ◦ C the iron containing sample retained both kind
of sites whereas those of non promoted sample were almost com-
pletely removed at 300 ◦ C, inferring therefore that the acid strength
200 300 400 500 600 700 800 900 1000 1100
of F1 ZW15 -S0 catalyst was stronger. Moreover, although the ini-
Raman Shift (cm-1) tial acidity of the F0 ZW15 -S0 (Fig. 8c) sample is higher than the
acidity of the Fe-doped catalyst (Fig. 8d) due to increased Lewis
Fig. 6. Raman spectra of samples with 15% W and different content of iron syn-
acidity, it appears that at higher temperature the catalyst retains a
thesized with surfactant (Sz = 1): (a) F0 ZW15 -S1 , (b) F0.5 ZW15 -S1 , (c) F1 ZW15 -S1 , (d)
F2 ZW15 -S1 . higher Brønsted and total acidity than the catalyst without iron.
Then, according to the catalytic activity data, the existence of
 M.L. Hernandez-Pichardo et al. / Applied Catalysis A: General 431–432 (2012) 69–78 75

Fig. 7. STEM analysis of the F1 ZW10 -S0 catalyst. (a) HAADF-STEM image, (b) Global EDS analysis of HAADF-STEM image, (c) Punctual EDS analysis obtained in the zone circled,
and (d) Pt, Fe, W and Zr chemical mapping.

a) L b)
F0 ZW15-S0 L L+B F1 ZW15 -S 0
B L
B B L
B L+B

RT RT
Absorbance
Absorbance

100 °C
100 °C

200 °C 200 °C

300 °C 300 °C

400 °C
400 °C

1700 1650 1600 1550 1500 1450 1400 1700 1650 1600 1550 1500 1450 1400
Wavenumbers, cm-1 Wavenumbers, cm-1
c) 120 d) 120
Bronsted Bronsted
100 Lewis 100
Lewis
80 Total Acidity Total Acidity
80
μmolg-1

μmolg-1

60 F0ZW15-S 0 60 F1ZW15-S 0
40 40

20 20

0 0

0 100 200 300 400 500 0 100 200 300 400 500
Temperature, °C Temperature, °C

Fig. 8. FTIR spectra of adsorbed pyridine (a and b) and total acidity and amounts of Brønsted acid and Lewis acid (c and d) of the F0 ZW15 -S0 and F1 ZW15 -S0 samples
respectively.
76 M.L. Hernandez-Pichardo et al. / Applied Catalysis A: General 431–432 (2012) 69–78

a b
ZW-4f ZWFe1%-4f
W 7/2
W 5/2 W 7/2
W 5/2

Zr 4p

Intensity
W 5/2 Zr 4p
Intensity

W 7/2
W 5/2 W 7/2

40.0 37.5 35.0 32.5 30.0 27.5 25.0 40.0 37.5 35.0 32.5 30.0 27.5 25.0
BINDING ENERGY (eV) BINDING ENERGY (eV)
c d
ZWFe2%-4f ZWFe2%-Fe 2p
Fe 2p 3/2
W 5/2
W 7/2

Fe 2p 1/2
Intensity

Intensity

Zr 4p

W 5/2 W 7/2

40.0 37.5 35.0 32.5 30.0 27.5 25.0 740 730 720 710
BINDING ENERGY (eV) BINDING ENERGY (eV)

Fig. 9. (a) XP spectra of the W 4f region of: (A) ZW15 -S0 and (B) F2 ZW15 -S0 catalysts and (b) Fe2p spectra of F2 ZW15 -S0 catalyst.

cracking leading to a slight amount of light products as well as
skeletal isomerization to form 2,2-DMB and 2,3-DMB, suggests
the participation of strong acid sites. However, the acidity of the
F1 ZW15 -S0 catalyst was apparently enhanced with respect to that
of the non promoted F0 ZW15 -S0 sample, which is confirmed with
the higher production of the bi-ramified isomers, as well as light
products (Table 3).
The different performance of catalysts with diverse Fe contents
could be explained by the tendency of Fe to be incorporated into
the zirconia surface [14,28]; this could produce a distortion of WOx
lattice that can modify the strain in the W O W bonds, increasing
the WOx reducibility and thus the formation of active Brønsted acid
sites. However at high Fe concentrations (2%) the iron species are
segregated on the zirconia surface forming dispersed iron oxides

Scheme 1. (a) Promotional effect of Fe at low contents trough the modification of the WOx lattice, and (b) Segregation of iron oxides and WO3 crystals at high Fe contents.
 M.L. Hernandez-Pichardo et al. / Applied Catalysis A: General 431–432 (2012) 69–78 77

Table 4
Influence of the surfactant content (Sz ) on the textural properties and surface tungsten density ((W) of catalysts with 20% of tungsten.

Sample SZ (CTAB/ZrO2 )a Sg BET (m2 /g) Vpb (cc/g) ´

Dpb (Å) (W (W ats nm−2 )

F0.5 ZW20 -S0 0 32.7 0.070 59.93 20.0

F0.5 ZW20 -S1 1 60.9 0.182 88.87 10.7
a
Molar ratio.
b
BJH desorption.

(lower than the XRD detection limit), facilitating the growth of
the WO3 crystals which present a lower activity than the dis-
persed WOx polytungstates, and hindering the promotional effect
of iron (Scheme 1). In order to support this model XPS analyses
were performed on the F0 ZW15 -S0 , F1 ZW15 -S0 , F2 ZW15 -S0 cata-
lysts. Although the XRD results do not show any evidence about the
formation of iron oxides, due to their small domain size (lower than
3 nm), it was found for all samples that W4f (Fig. 9) signal appeared
with three states for W, the first one was about 35.2 eV which corre-
sponded to W6+ characteristic of WO3 (NIST), and the second one at
29.3 eV corresponding to a highly reduced W, and finally an inter-
mediately reduced specie at 34.6 eV. This last W specie increased
its intensity along with the amount of Fe. The Fe 2p signal is com-
prised of two signals at about 714.8 eV corresponding to highly
oxidized Fe3,4+ specie and the second one belongs to a satellite sig-
nal characteristic of oxidized Fe species (718.5 eV). These shifted
bands can be indexed to Fe2p3/2 and Fe2p1/2 , which are character-
istic of Fe3+ in Fe2 O3 [29]. This result suggested that the Fe is an
electron donor, and it is possible that is transferring charge to the
W species. Then, the XPS results suggest the formation of hematite-
like particles in both samples but with a higher formation of Fe2 O3
in the sample with 2% Fe, supporting the characterization results
and the suggested model.
Some authors [30] found that in sulfated zirconia catalysts
doped with low iron concentrations (1%), the entire metal amount
is incorporated into zirconia, whereas the addition of higher
amounts of iron produces the segregation of hematite. Whereas
in tungstated zirconia catalysts also doped with 1% of Fe, Carrier
et al. [11] found by XANES, Mössbauer and EXAFS that iron is in the
FeIII oxidation state and it is located at the surface of the support
with a limited formation, about 10%, of iron oxide particles, while
FeIII predominantly forms a solid solution with tetragonal zirconia,
in good agreement with our results al low iron concentrations.
3.3.2. Effect of the surfactant
As we have previously reported, the use of the surfactant in the
synthesis of tungstated zirconia catalysts produces an increment in
the porosity and surface area of these materials [31]. As an example,
Table 4 presents the texture properties for catalysts with 20%W and
0.5% Fe, prepared with different surfactant contents (Sz = 0 and 1). It
is observed that the surfactant addition improves the surface area,
pore volume, and pore diameter compared to the surfactant-free
sample. These materials present a type IV isotherm (not shown)
with a hysteresis loop characteristic of mesoporous materials with
open pores systems. The catalyst prepared with surfactant exhibits
a wider pore size distribution than the surfactant-free sample in
the mesoporous range.
The study of the surfactant effect on the crystalline structure was
also performed by XRD. Fig. 10 shows the XRD patterns of samples
synthesized with surfactant (Sx = 2), it is observed that these sam-
ples also present the reflections corresponding to the tetragonal
zirconia and monoclinic tungsten oxide; however it is observed
a higher intensity of the reflections corresponding to WO3 . The
higher porosity, produced by the use of a surfactant-assisted syn-
thesis favored the tungsten ions diffusion and the expel of the
WOx species from the bulk producing modifications on the dis-
persion and domain size of these species before they form the
active sites on the zirconia surface. Finally, Fig. 11 presents the FTIR
spectra of adsorbed pyridine (9a) and total acidity and amounts of
Brønsted acid and Lewis acid (9b) of the F1 ZW15 -S1 sample. It is
observed that this sample present the same trend in the nature of
the acid sites than samples without surfactant but the total acid-
ity of this catalyst was enhanced mainly due to the increase in the
number of Lewis acidic sites. Then, the main role of the surfac-
tant is that it leads to the formation of a higher surface area and
the increase of the porosity of these catalysts, which modifies the
formation and domain size of the WOx species, and therefore the

Sz = 2
: Tetragonal zirconia

WO3

20% W - 2%Fe
intensity

15% W - 2%Fe
10% W - 2%Fe
20% W - 1%Fe
15% W - 1%Fe
10% W - 1%Fe
20% W - 0.5 %Fe
15% W - 0.5 %Fe
10% W - 0.5 Fe
20% W - 0%Fe
15% W - 0%Fe
10% W - 0%Fe

15 20 25 30 35 40 45 50 55 60 65 70 75 80
2θ

Fig. 10. XRD patterns of the Fe–WOx –ZrO2 catalysts calcined at 800 ◦ C, at different tungsten and iron contents, synthesized with a surfactant molar ratio of 2 (Sz = 2).
78 M.L. Hernandez-Pichardo et al. / Applied Catalysis A: General 431–432 (2012) 69–78
a)
F1ZW1 5-S1
L+B
L
B 160
RT B Total Acidity
Absorbance
100 °C
200 °C
300 °C
400 °C
1700 1650 1600 1550 1500
Wavenumbers, cm-1
Fig. 11. FTIR spectra of adsorbed pyridine (a) and total acidity and amounts of Brønsted acid and Lewis acid (b) of the F1 ZW15 -S1 sample.
reducibility of the polytungstate domains that generates active acid
sites.
In summary, it is observed that depending on the Fe, W and sur-
factant concentrations, WOx species with different domain sizes
are generated, which yield different interactions between WOx , Fe
and zirconia. These interactions could modify the reducibility of
tungsten nanostructures and thus the catalytic activity of Pt/Fe–WZ
materials. Therefore, the differences in the catalytic performance
could then be explained by the differences in the domain size and
connectivity of the WOx species produced by the variation on tung-
sten and iron concentrations, as well as the sample texture.
4. Conclusions
The catalytic evaluation results in the n-hexane isomerization
over Pt/Fe-ZW samples prepared with different iron and tungsten
contents showed that the conversion and selectivity depend on the
Fe/W ratio. The optimum Fe/W ratio was found at about 0.05–0.07
regardless of the tungsten or surfactant content. The results also
show that the tungsten content can be optimized by the promotion
with iron, since the samples with low tungsten content (10% W)
promoted with 0.5% Fe showed the highest catalytic activity even
superior to catalysts with 15% W.
It was also observed a positive influence of the surfactant on
the catalytic performance for all the tungsten concentrations; the
optimum content was a surfactant/zirconia molar ratio of 1. This
component improves the textural properties of these catalysts and
generates a higher porosity that allows a better diffusion of zirco-
nium and tungsten ions into tungstated zirconia catalysts. It was
also found that the acidity of iron-doped samples is slightly higher
than those samples without this dopant. Therefore, the improve-
ment in the catalytic performance of Pt/FZW could be explained
by the differences in the domain size and connectivity of the WOx
species produced by the iron incorporation at low concentration
(0.5–1% Fe) that modifies the acid sites produced on the zirconia
surface and therefore higher catalytic activity.
Acknowledgments
M.L.H.P. would like to acknowledge to the IPN and the Conacyt
for the financial support received. The authors acknowledge the
facilities of the Combinatorial Chemistry Laboratory and the Elec-
tron Microscopy Laboratory of the Instituto Mexicano del Petroleo,
b) 200
L 180 Bronsted
Lewis
140
μmolg-1
120
F1ZW1 5-S1
100
80
60
40
20
0
1450 1400 0 100 200 300 400 500
Temperature, °C
as well as the experimental support of the CNMC-IPN for this work.
Moreover, the authors also acknowledge to Prof. Ricardo Macias
Salinas for the revision of the manuscript.
References
[1] G. Busca, Chem. Rev. 107 (2007) 5366–5410.
[2] T. Buchholz, U. Wild, M. Muhler, G. Resofsky, Z. Paál, Appl. Catal., A 189 (1999)
225–236.
[3] B.R. Jermy, M. Khurshid, M.A. Al-Daous, H. Hattori, S.S. Al-Khattaf, Catal. Today
164 (2011) 148–153.
[4] S.G. Ryu, B.C. Gates, Ind. Eng. Chem. Res. 37 (1998) 1786–1792.
[5] M. Scheithauer, R.E. Jentoft, B.C. Gates, H. Knözinger, J. Catal. 191 (2000)
271–274.
[6] A. Jatia, C. Chang, J.D. Mac Leod, T. Okubo, M.E. Davis, Catal. Lett. 25 (1994)
21–26.
[7] S. Praserthdam, P. Wongmaneenil, B. Jongsomjit, J. Ind. Eng. Chem. 16 (2010)
935–940.
[8] M.L. Hernandez-Pichardo, J.A. Montoya, P. del Angel, A. Vargas, J. Navarrete,
Appl. Catal., A 345 (2008) 233–240.
[9] J.G. Santiesteban, D.C. Calabro, C.D. Chang, J.C. Vartuli, T.J. Fiebig, R.D. Bastian,
J. Catal. 202 (2001) 25–33.
[10] P. Lukinskas, S. Kuba, B. Spliethoff, R.K. Grasselli, B. Tesche, H. Knözinger, Top.
Catal. 23 (2003) 163–173.
[11] X. Carrier, P. Lukinskas, S. Kuba, L. Stievano, F.E. Wagner, M. Che, H. Knözinger,
Chem. Phys. Chem. 5 (2004) 1191–1199.
[12] S. Wong, T. Li, S. Cheng, J. Lee, C. Mou, Appl. Catal., A 296 (2005) 90–99.
[13] J.M.M. Millet, M. Signoretto, P. Bonville, Catal. Lett. 64 (2000) 135–140.
[14] E. Carreto, J.A. Montoya, J. Morgado, J. Am. Ceram. Soc. 87 (2004) 612–616.
[15] H.W. Turner, A.F. Volpe Jr., W.H. Weinberg, Surf. Sci. 603 (2009) 1763–1769.
[16] A. Corma, J.M. Serra, A. Chica, Catal. Today 81 (2003) 495–506.
[17] C.A. Emeis, J. Catal. 141 (1993) 347–354.
[18] Y. Ono, Catal. Today 81 (2003) 3–16.
[19] A. Martínez, G. Prieto, M.A. Arribas, P. Concepción, J.F. Sánchez-Royo, J. Catal.
248 (2007) 288–302.
[20] W. Zhou, E.I. Ross-Medgaarden, W.V. Knowles, M.S. Wong, I.E. Wachs, C.J. Kiely,
Nat. Chem. 1 (2009) 722–728.
[21] R.S. Weber, J. Catal. 151 (1995) 470–474.
[22] E.I. Ross-Medgaarden, I.E. Wachs, J. Phys. Chem. C 111 (2007) 15089–15099.
[23] S.R. Vaudagna, S.A. Canavese, R.A. Comelli, N.S. Fígoli, Appl. Catal. 168 (1998)
93–111.
[24] D.B. Williams, C.B. Carter, Transmission Electron Microscopy, Springer, New
York, 2009.
[25] A. Barrera, J.A. Montoya, M. Viniegra, J. Navarrete, G. Espinosa, A. Vargas, P. del
Angel, G. Pérez, Appl. Catal., A 290 (2005) 97–109.
[26] G. Sunita, B.M. Devassy, A. Vinu, D.P. Sawant, V.V. Balasubramanian, S.B. Hal-
ligudi, Catal. Commun. 9 (2008) 696–702.
[27] B.T. Loveless, A. Gyanani, D.S. Muggli, Appl. Catal., B 84 (2008) 591–597.
[28] J.A. Navío, M.C. Hidalgo, G. Colón, S.G. Botta, M.I. Litter, Langmuir 17 (2001)
202–210.
[29] W.J. Liu, F.X. Zeng, H. Jiang, X.S. Zhang, W.W. Li, J. Chem. Eng. 180 (2012) 9–18.
[30] P. Reyes, M. Oportus, M. do Carmo-Rangel, Appl. Catal., A 334 (2008) 187–198.
[31] M.L. Hernández, J.A. Montoya, I. Hernández, M. Viniegra, M.E. Llanos, V. Garibay,
P. Del Angel, Microporous Mesoporous Mater. 89 (2006) 186–195.