Organic polymers as pore-regulating agents in TiO2-Al2O3 mixed oxide catalytic supports

Article abstract of DOI:10.1023/A:1004386212175

The synthesis of TiO₂-Al₂O₃ mixed oxide catalytic supports with a TiO₂/(TiO₂+Al₂O₃) molar ratio equal to 0.5, was made by the co-precipitation of the corresponding metallic isopr

Full text of DOI:10.1023/A:1004386212175

JOURNAL OF MATERIALS SCIENCE 33 (1998) 1981—1990
Organic polymers as pore-regulating agents in
TiO2—Al2O3 mixed oxide catalytic supports
T. KLIMOVA, E. CARMONA, J. RAMIREZ
UNICAT, Departamento de Ingenierıa Quımica, Facultad de Quımica, UNAM, Mexico Cd.
Universitaria, Mexico D.F. (04510), Mexico
E-mail: klimova@servidor.unam.mx
The synthesis of TiO2—Al2O3 mixed oxide catalytic supports with a TiO2/(TiO2#Al2O3)
molar ratio equal to 0.5, was made by the co-precipitation of the corresponding metallic
isopropoxides using different organic polymers as pore-regulating agents. The influence of
the preparation parameters (type of polymeric additive, its amount, molecular weight and
method of additive incorporation into the hydroxide precipitate) on the surface area and
pore structure of the final solid were studied. It was found that organic polymers added
during the hydrolysis step, in general, resulted in a significant increase in surface area and
total pore volume. In all cases the pore-size distribution becomes clearly bimodal with two
maxima in the pore-size distribution curve around 3 and 30—50nm. In order to investigate
the action mechanism of the polymeric additives, characterizations by Fourier transform-
infrared, thermogravimetric/differential thermogravimetry and differential thermal analysis,
as well as a TEM study of the hydroxide precipitates before calcination, were carried out. The
results from these characterizations indicate that the textural changes induced by the
polymeric additives can be rationalized in terms of a ‘‘filler’’ effect in the case where the
additive is incorporated during the filtration step, and that when the additive is added during
the hydrolysis step it is the chemistry of the polymer functional groups which determines the
final texture of the solid.  1998 Chapman & Hall
1. Introduction
control of the textural properties of the support is an
important issue.
The role of the oxide support in the performance of
heterogeneous catalytic systems is very important, be-
cause in many cases the activity and stability of the
catalyst is known to depend not only on the inherent
catalytic properties of the active phase, but also on the
textural and physicochemical characteristics of the
support. Among the catalysts in which the support
characteristics plays an important role, are the hydro-
treatment catalysts which are normally supported on
alumina. In the case of these catalysts, in view of the
increased concern over environmental problems, some
new catalytic systems have been studied in order to
obtain more active and selective catalysts. In these
attempts, the use of new supports, such as Ti—Al mixed
oxides, has shown promising results. In this case,
greater catalytic activity has been observed, due to the
effect of the support [1].
Organic polymers have been used with some suc-
cess in the past, as additives to control textural
properties of catalysts supported on single oxides
[3—5], as well as in some mixed oxides systems
such as Al O — SiO [6, 7]. However, the role and
ꢀ ꢁ
ꢀ
the action mechanism of these additives is not yet
clear. Furthermore, it is known that the same
polymer can cause different changes in the textural
properties of the different oxides. For example, it was
noted that with alumina and silica—alumina gels, the
use of polyethylene glycols brings about only a small
increase in the surface area. In contrast, the same
additive used in a silica gel system causes a marked
decrease in the surface area [3]. Therefore, each poly-
meric additive and gel system must be considered
separately.
In general, sol—gel methods have been preferred to
produce such mixed oxide systems. However, the con-
trol of the surface area and porosity of the catalysts
still remains a problem when titanium and aluminium
alkoxides are used as support precursors. In this case,
small pore diameters of the order of 2—3 nm are ob-
tained [2]. This pore dimension is useful only when
light petroleum fractions are to be processed. How-
ever, for the effective processing of heavier petroleum
fractions, larger pore diameters are needed. Therefore,
Here we study the effectiveness of the use of different
organic polymers, added at different stages of the
sol—gel preparation of the support with the aim to
control the porosity and the surface area of
a TiO —Al O mixed oxide. The effect of the type of
ꢀ
ꢀ ꢁ
polymeric additive, its amount, molecular weight and
sequence of addition of the reactants on the surface
area and pore-size distribution in TiO —Al O
ꢀ
ꢀ ꢁ
mixed oxides is reported. Additionally, in order to
understand better the changes occurring in the mixed
0022—2461 ( 1998 Chapman & Hall
1981

oxides, some experiments were also performed with
pure TiO and Al O oxides.
characterized by surface area, S , and porosity using
#2
a Micromeritics ASAP 2000 apparatus. Immediately
ꢀ
ꢀ ꢁ
after a pre-treatment (3 h at 523 K under vacuum),
surface area measurements were performed by nitro-
gen adsorption at 78 K using a five-point BET
method. The Barret-Joiner-Halenda (BJH) method
[8] was applied to determine pore-size distribution.
2. Experimental procedure
2.1. Synthesis of TiO2—Al2O3 mixed oxides
The Al—Ti mixed oxide samples with molar ratio
TiO /(TiO #Al O )"0.5 were prepared by co-
The average pore-diameter was calculated as 4»/S
.
ꢀ
ꢀ
ꢀ ꢁ
(&
precipitation of the titanium and aluminium isopropox-
ides as precursors and n-propyl alcohol as solvent. In
the hydrolysis step, water, in an excess of 30 times the
stoichiometric amount required, was used to produce
the formation of the metallic hydroxides. A number of
polymers such as polyethylene glycol (PEG), poly-
propylene glycol (PPG), polyacrylamide (PAA) and
polyvinyl alcohol (PVA) with different degrees of poly-
merization for each type of polymer were used in order
to study the effect of the polymer molecular weight
and its functional groups on the surface area and
pore-size distribution of the final solid. Two series of
samples were prepared. In Series 1 the polymer was
added together with the water to the alkoxide precur-
sors during the hydrolysis step, i.e. the polymer was
present before the formation of the metal hydroxides.
In the second series (Series 2), the polymer was added
into the already formed precipitates, after the hydroly-
sis step.
A potentiometric method was used to determine, by
titrating with n-butylamine, the total number of sur-
face acid centres and the relative acid strength of the
surface sites of the prepared solids [9]. In this method,
a small quantity (0.2 ml) of 0.1 N n-butylamine in
acetonitrile was added to 0.15 g solid and kept under
agitation for 3 h. Later, the suspension was titrated
\ꢂ
adding 0.1 ml min of the same base. The electrode
potential variation was registered on a Metrohm 691
digital pH meter with a combined glass and Ag/AgCl
electrode. The reproducibility of the electrode poten-
tial curve was $5 mV.
3. Results
3.1. Surface area and porosity
3.1.1. Effect of polymer functional group
The results from the textural characterizations show
that, in general, for the first series of samples, when the
polymer is added during the hydrolysis step, the use of
organic polymers results in a large increase of surface
area and total pore volume accompanied by signifi-
cant changes in the pore-size distribution (Table I).
A strong influence of the polymer functional group is
also noted. Thus, for the same type of polymeric addi-
tives — poly-glycols; the change from PEG to PPG of
similar molecular weight does not alter significantly
the specific surface area. On the other hand, it can be
noted that a larger increase in surface area and total
pore volume was observed in the case of PAA and
PVA, using additive amounts of 20 times less than
those used in the case of poly-glycols.
2.2. Characterization
In order to investigate the action mechanism of the
polymers, characterization of the hydroxide precipi-
tates before calcination was carried out by means of
infrared (IR) spectroscopy, thermogravimetric/differ-
ential thermogravimetry (TGA/DTG) and differential
thermal analysis (DTA) and transmission electron
microscopy (TEM). The infrared spectra were re-
corded in a Nicolet 510 FT—IR spectrometer operat-
ing in the transmission mode, using KBr pressed discs.
The differential thermal analysis and thermogravimet-
ric experiments were performed using a DuPont 2000
apparatus in an air atmosphere with a heating pro-
The d»/d log D versus D curves show a bimodal
pore-size distribution for all the samples prepared
with polymers, with a first maximum at about 3 nm
and a second one at around 30—50 nm (Fig. 1). In
contrast, the TiO —Al O mixed support prepared
°
\ꢂ
gramme rate of 20 C min . The TEM observations
were made with a Jeol 100CX electron microscope,
equipped with a side-entry stage. For this character-
ization, the gel samples dispersed in an n-propanol—
water solution (6:1 vol/vol) were deposited on copper
grids coated with a holey carbon film.
ꢀ
ꢀ ꢁ
without additive presents a monomodal pore-size dis-
tribution (curve 1 in Fig. 1).
A detailed analysis of Fig. 1 shows differences in the
behaviour of the different polymers which may be
related to the differences in the type of functional
The mixed oxides formed after drying at 373 K for
24 h and subsequent calcination at 773 K for 24 h were
TABLE I S
and porosity of the Series 1 samples of TiO —Al O mixed oxides
#2
ꢀ
ꢀ ꢁ
Sample
Organic
polymer
Molecular
weight
(g mol\ꢂ)
Amount of polymer
(ml/1.2 g support)
S
Total pore
volume
(cmꢁ g\ꢂ)
Average pore
diameter
(nm)
#2
(mꢀ g\ꢂ)
1
2
3
4
5
—
—
—
247
301
303
361
380
0.210
0.448
0.396
0.471
0.733
3
5
3.8
4.7
6.1
PEG
PPG
PVA
PAA
3 400
3 000
5.0
5.0
0.25
0.25
89 000—98 000
200000
1982

group present in the polymer but also to differences
in the solubility of the polymer. Similar bi-modal
3.1.2. Effect of polymer molecular weight
The textural characteristics of the samples prepared
with PEG, PPG and PVA with different molecular
weights, which are presented in Table III, show that
increasing the polymer molecular weight leads only to
minor changes in the surface area and pore volume,
and that these variations do not follow a definite
pattern of behaviour.
The observed changes in the pore-volume distribu-
tion curves, when the polymer molecular weight is
increased, consist mainly in an increase in the volume
of the pores which lie between 20 and 100 nm, and
a displacement of the curve maximum from 35 to
50 nm.
pore-size distributions were obtained for the TiO —
ꢀ
Al O mixed oxides synthesized with PEG, PPG and
ꢀ ꢁ
PVA. In the case of PAA, one can observe that the use
of this polymer leads to an increase in the pore volume
due to pores with diameters between 3 and 80 nm.
Moreover, the two maxima in this pore-size distribu-
tion curve are wider and less defined than in the
samples prepared with the other polymers.
For comparison, the results from similar experi-
ments, carried out with pure Al O and TiO , adding
ꢀ ꢁ
ꢀ
the polymer during the hydrolysis process, are shown
in Table II. Here it is possible to observe that the
behaviour of the alumina is quite different from that of
titania. In the case of alumina, the addition of poly-
mers leads to increased surface area, pore volume and
pore diameter. On the contrary, in the case of titania
the addition of polymers results in lower surface area
and pore volume, but larger diameter pores.
3.1.3. Effect of the amount of the organic
polymer
In the case of PEG with molecular weight of
\ꢂ
600 g mol , an increase in the amount of polymer
leads to an increase in the values of the surface area
and pore volume, and to slight variations in the pore
diameter towards larger diameter pores (Table IV).
These changes are in line with the small variations
observed in the pore-size distribution plot which
shows that the maximum corresponding to pores of
diameters between 40 and 60 nm shifts from 45 nm to
55 nm and that the area under the curve is slightly
increased when the amount of polymer is also
increased.
3.1.4. Effect of addition method
The addition of the polymers to the already formed
precipitates, before drying and calcination of the latter
(Series 2), results in an increase in the average pore
diameter of all samples which is larger than the in-
crease observed when the same polymers are added
during the hydrolysis step (Series 1). The surface area
and pore volume of the Series 2 sample prepared with
PPG (Table V) are larger than those of the corres-
ponding Series 1 sample. In the case of PEG, PAA and
PVA, the effect on the surface area and porosity was
substantially less than that reported in Table I.
Fig. 2a and b show a comparison of the changes
which occur in the pore-size distribution plots when
Figure 1 Pore-size distributions of TiO —Al O mixed oxides
ꢀ
ꢀ ꢁ
(Series 1) prepared (1) without polymer, and with organic polymers
added during hydrolysis: (2) PVA (89 000—98 000 g mol\ꢂ), (3) PPG
(3000 g mol\ꢂ), (4) PEG (3400 g mol\ꢂ), (5) PAA (200 000 g mol\ꢂ).
TABLE II S
and porosity of the Series 1 samples of pure TiO and Al O
#2
ꢀ
ꢀ ꢁ
Sample
Oxide
Organic
polymer
Molecular
weight
S
(m
Total pore
volume
Average pore
diameter
(nm)
#2
ꢀ
\ꢂ
)
g
\ꢂ
ꢁ \ꢂ
(cm g )
(g mol
)
1
2
3
4
5
6
7
8
9
Al O
—
—
180
351
333
365
312
28
14
13
16
29
0.293
0.879
0.951
1.058
0.650
0.067
0.034
0.033
0.057
0.105
4.7
6.9
7.4
7.7
6.2
7.7
8.1
8.5
13.0
10.4
ꢀ
ꢁ
PEG
PPG
PVA
PAA
—
PEG
PPG
PVA
PAA
3400
3000
13000—23 000
200000
—
TiO
ꢀ
3400
3000
13000—23 000
200000
10
1983

TABLE III Effect of the polymer molecular weight on S
#2
and porosity of the Series 1 samples of TiO —Al O mixed oxides
ꢀ
ꢀ ꢁ
Sample
Organic
polymer
Molecular weight
(g mol\ꢂ)
S
Total pore volume
(cmꢁ g\ꢂ)
Average pore
diameter (nm)
#2
(mꢀ g\ꢂ)
1
2
3
4
5
6
7
8
9
10
11
12
13
PEG!
PPG!
PVA"
400
1000
2000
3400
4000
425
1000
2000
3000
361
321
340
301
332
362
335
355
303
302
320
361
305
0.445
0.467
0.410
0.448
0.458
0.549
0.429
0.531
0.396
0.409
0.380
0.471
0.455
4.9
5.0
4.0
5.0
5.2
6.1
4.9
6.2
3.8
4.6
3.9
4.7
4.7
4000
13 000—23000
89 000—98000
124000—186000
Amount of polymer: ! 5 ml/1.2 g support, " 0.25 ml/1.2 g support.
TABLE IV Effect of the amount of the organic polymer (PEG,
presence of the mixed oxide. Also, the C—N vibrational
\ꢂ
\ꢂ
600 g mol ) on S
and porosity of the Series 1 samples of
band of the pure PAA polymer shifts from 1605 cm
#2
TiO —Al O mixed oxides
\ꢂ
to 1628 cm when the polymer is added to the mixed
oxide. These changes in IR vibrational frequencies
could be related to an interaction of the polymer with
the support structure.
ꢀ
ꢀ ꢁ
Sample Amount of
polymer
S
(m
Total pore Average pore
#2
ꢀ
\ꢂ
g )
volume
\ꢂ
)
diameter
(nm)
ꢁ
(cm
(ml/1.2 g
g
support)
For the Series 2 samples, in which the polymer is
added during the filtration step into the already for-
med precipitate, the intensity of the characteristic IR
polymer bands suggests that the quantity of the poly-
mer incorporated into the precipitate was much more
that in the Series 1 samples. However, in the case of
Series 2 samples, a displacement of the polymer IR
vibrational bands was not detected, indicating a lower
interaction of the polymer with the gel than in the case
of Series 1 experiments.
1
2
3
2.5
5.0
7.5
328
353
371
0.421
0.463
0.464
4.4
4.7
5.0
PEG or PAA are added during hydrolysis or during
the filtration step. In the case of PEG (Fig. 2a), the
addition of the polymer during filtration leads to an
increase in the population of pores of 20—100 nm and
a decrease of the small pores in the range 2—4 nm. In
contrast, the standard preparation without polymer
additive only shows the existence of small pores. The
behaviour of the PAA (Fig. 2b), when added to the
precipitate, is similar to that of PEG because the
hydroxide particles have already been formed. How-
ever, when PEG or PAA are added to the hydrolysis
solution, a comparison of the corresponding pore-size
distribution plots (Fig. 2) indicates that the mecha-
nism of action of PAA is completely different from
that of PEG. In the case of PAA it is possible that the
interaction of the functional groups with the hydrox-
ide structures promotes the formation of pores of
different sizes. In agreement with this, as will be shown
later, the FT—IR results show a shift in the frequency
of the PAA C"O and C—N vibrations.
3.3. Sample thermal behaviour
In line with the IR characterization, the results from
the thermal behaviour of the intermediate species (the
precipitates before calcination) confirm the existence
of a polymer—precipitate interaction.
3.3.1. Series 1 samples
Before analysing the effect of adding the polymers on
the thermal behaviour of the samples, the results from
the thermal behaviour of the samples without polymer
is described. The thermogravimetric results obtained
for the titanium—aluminium precipitate prepared
without polymer (Fig. 3a) show two weight-loss peaks
in the temperature range 300—600K. The low temper-
ature peak (¹
at 342 K) can be related to the elim-
ꢀ!ꢁ
ination of solvent and physisorbed water and the
high-temperature peak, centred at 476 K, is most
probably related to the dehydration process leading to
the formation of the TiO —Al O mixed oxide [10]. In
3.2. FT—IR study
The results from FT—IR show that all the Series 1 pre-
cipitates, obtained when the organic polymer is added
during the hydrolysis step, contain the corresponding
polymer. In the case of PAA a displacement of the
characteristic vibrational bands of the polymer func-
tional groups was observed. Thus, the position of the
ꢀ
ꢀ ꢁ
line with these results, two endothermic peaks in this
temperature range are observed in the corresponding
DTA curves for the same sample.
The thermogravimetric results for the sample con-
taining PEG polymer are presented in Fig. 3b. Three
peaks are clearly observed. The low-temperature
\ꢂ
C"O stretching vibration which appears at 1692 cm
\ꢂ
for the pure polymer, shifts to 1661 cm
in the
1984

TABLE V S
and porosity of the Series 2 samples of TiO —Al O mixed oxides
#2
ꢀ
ꢀ ꢁ
Sample
Organic
polymer
Molecular
weight
(g mol\ꢂ)
Amount of polymer
(ml/1.2 g support)
S
Total pore
volume
(cmꢁ g\ꢂ)
Average pore
diameter
(nm)
#2
(mꢀ g\ꢂ)
1
2
3
4
PEG
PPG
PVA
PAA
3400
3000
89 000—98 000
200000
5.0
5.0
0.25
0.25
266
380
280
240
0.438
0.565
0.363
0.346
7.1
5.9
6.1
6.6
Figure 3 Differential TGA curves of TiO —Al O mixed oxides
ꢀ
ꢀ ꢁ
prepared (a) without polymer and (b) with PEG 3400 g mol\ꢂ
added during hydrolysis and (c) of pure PEG 3400 g mol\ꢂ.
However, this peak is shifted to higher temperature in
comparison to the peak of the pure polymer which
decomposes at 554 K (Fig. 3c). This shift could be due
to some structural changes and interactions between
the polymer chain and the functional groups of the
Ti—Al hydroxide precipitate. Similar shifts to higher
temperatures are observed for all polymers used in
this investigation (Table VI).
Figure 2 Pore-size distributions of TiO —Al O mixed oxides for
ꢀ
ꢀ ꢁ
preparations (1) without polymer, (2) polymer added during
hydrolysis, (3) polymer added during filtration; (a) PEG
(3400 g mol\ꢂ), (b) PAA (200 000 g mol\ꢂ).
3.3.2. Series 2 samples
peaks (333 and 478 K) have identical positions with
the peaks observed for the sample without polymer
and are obviously related to the dehydration process.
The third peak is observed at 576 K and can be as-
signed to the thermal decomposition of the polymer.
In this case, in which the polymer was added into the
already formed precipitate after the stage of hydroly-
sis, the weight loss corresponding to the polymer de-
composition process is larger, and the corresponding
¹
shift is less pronounced. For example, for PAA
ꢀ!ꢁ
1985

TABLE VI ¹
of the decomposition process of polymers
ꢀ!ꢁ
Sample
Polymer
Molecular weight
\ꢂ
¹
, pure polymer
¹
, Series 1 samples
ꢀ!ꢁ
ꢀ!ꢁ
(g mol
)
(K)
(K)
1
2
3
4
PEG
PPG
PVA
PAA
3400
3000
89000—98 000
554
525
559
576
597
587
200000
569, 682, 874
645, 813, 933
incorporated into the Ti—Al precipitate during the
filtration step, three peaks are observed at 563, 685
and 863 K. The positions of these peaks are similar to
those observed for the pure PAA polymer (Table VI).
preserve the pore structure through the ‘‘filling-re-
moval’’ mechanism. In contrast, polymers with a re-
active functional group will form chemical bonds with
the metal cations and will therefore change the hy-
drous gel structure and, consequently, the dehydra-
tion process. In general, these types of polymer will
produce radical changes in the pore structure of the
final solid. However, the results from the above study
[3], show that polyglycols, which have side functional
groups, behave more as non-reactive polymer addi-
tives. In contrast, other work shows that the addition
of glycols before precipitation is more effective in the
modification of the alumina structure, in part due to
the adsorption of the polymer on the particle surface
as it is formed [5]. Clearly, the behaviour of some
polymer additives with low reactivity groups will
show the characteristics of the two main groups, i.e.
reactive and non-reactive polymers.
3.4. Sample morphology
The typical morphology of the unmodified tita-
nium—aluminium precipitate prepared by the sol—gel
method is shown in Fig. 4a. It can be observed that the
sample is composed of irregular aggregates with size
from 20—200 nm. In view of the regular pore-size dis-
tribution with a maximum about 3 nm of this sample
(Fig. 1, curve 1), it is concluded that these aggregates
are formed of very fine primary particles of regular
size, which create mesopore voids between them.
The morphology of the different samples obtained
with organic polymers seems to be sensitive to the
type of the polymeric additive (Fig. 4b—e). It should be
noted that their size and form are quite different,
depending on the polymer used in the synthesis.
As has been shown previously [4], other effects can
also play a role in determining the pore structure of
solids when polymers are used as additives to regulate
the pore structure. Steric stabilization would be a re-
sult of the affinity by the solvent and by the particle of
the lyophilic and lyophobic part of the polymer mol-
ecule, respectively. If the medium is a good solvent for
the lyophilic polymer tail, the particle—solvent affinity
will increase and the particle would be stabilized. If the
medium is not a good solvent for the polymer tails,
flocculation will occur. According to this, the final
pore structure, which depends on the size of the pri-
mary particles and aggregates, will be the result of the
many possible interactions between solvent, polymer
and gel. At the same time, the organic tail of the
polymer may act as a ‘‘filler’’, leading to an increase in
mesoporosity the importance of which will depend on
the size and concentration of the polymer chains.
Clearly, if the analysis of the interaction of polymers
with a single type of gel is not easy due to the complex-
ity of the different and possible concurrent phe-
nomena taking place at the same time, the study of
mixed oxide systems becomes even more difficult.
Nevertheless, here we will try to discuss the results
obtained for the Al O —TiO mixed oxide synthesis
3.5. Acidity measurements
The results from the characterization of acid proper-
ties of TiO —Al O mixed oxide samples after calcina-
ꢀ
ꢀ ꢁ
tion are reported in the Table VII. It can be observed
that all the polymers used in this study produced only
minor changes in the total number of acid sites. How-
ever, the measured maximum acid strength changed in
the case of the samples prepared using PVA and PAA.
4. Discussion
The use of water-soluble organic polymers as addi-
tives during the synthesis of inorganic ceramic mater-
ials by soft chemistry methods, as was reported earlier
[3, 6], are able to modify the drying process or rad-
ically change both the hydrogel structure and there-
fore the subsequent dehydration process, and thus
alter the porosity and specific surface area of the final
calcined solid. However, the mechanism of action of
these additives may be completely different depending
on the nature of the polymer. It has been mentioned
that the action of the polymers will depend on whether
or not they possess reactive functional side groups [3].
According to this, polymers which do not have react-
ive functional side groups will mainly alter the drying
process. In this case, the function of the polymer will
be to prevent the collapse of the gel structure, either by
reducing the surface tension or acting as ‘‘separator’’
in the gel network. Some of these polymers will also
ꢀ ꢁ
ꢀ
when PEG and PPG, non-reactive polymers accord-
ing to Basmadjian et al. [3], and PVA and PAA
(reactive polymers) were used to modify the textural
properties of the solids.
In order to discern if the different polymers were
affecting the structure of the gel, TEM observations of
the modified gels were compared to those of the un-
modified gel. Apparently, from the conclusions of pre-
vious reports [3], one would expect a great similarity
1986

Figure 4 Transmission electron micrographs of a Ti—Al precipi-
tates (Series 1 samples): (a) sample prepared without polymer,
(b) sample prepared with PEG 3400 g mol\ꢂ, (c) sample pre-
pared with PPG 3000 g mol\ꢂ, (d) sample prepared with
PVA 89 000—98000 g mol\ꢂ and (e) sample prepared with PAA
200000 g mol\ꢂ.
TABLE VII Surface acidity measurements of the Series 1 samples
Sample Polymer Molecular
Maximum
Total number
weight
acid strength of acid sites
(g mol\ꢂ)
(mV)
(meq base/g
solid)
1
2
3
4
5
6
7
8
9
—
—
!48
!52
!49
!57
!43
!49
!43
!45
1.40
1.43
1.20
1.37
1.47
1.35
1.47
1.35
1.39
1.33
1.34
PEG
200
600
1000
4000
1000
3000
4000
PPG
between the structure of the unmodified gel and those
obtained with PEG and PPG. However, the micro-
graphs shown in Fig. 4a—e, which were taken at the
same magnification, indicate that the gel morphology
of the polymer-modified gels is quite different from
that of the unmodified sample. This change of mor-
phology is most probably related to changes in the gel
structure. According to this, all the polymers used here
interact with the gel, leading to similar pore-size distri-
butions (see Fig. 1), although the amount of polymer
used in the case of PVA and PAA was 20 times less
than in the case of PEG and PPG. This may be due
either to the reactivity of the functional groups or to
the amount of groups present in each polymer chain.
In the case of PEG and PPG there are only two
terminal OH groups which can be considered as ca-
PVA
PAA
89 000—98 000 !66
124000—186 000 !78
200000 !26
10
11
pable of bridging two particles and therefore contrib-
ute to the formation of larger aggregates [5]. In the
case of PVA and PAA, the greater number of func-
tional groups present all along the polymer molecule
will lead to a greater probability of interaction be-
tween two particles. The formation of aggregates
will then need a smaller amount of polymer than
in the case of PEG or PPG. This is in line with our
1987

experimental results. The case of PAA modification
presents different features which we will discuss later.
A close analysis of Fig. 1 reveals that the polymer-
induced bimodal pore-size distribution curves of the
supports prepared using these three polymers (PEG,
PPG and PVA) are very similar (curves 2—4 in Fig. 1)
and that the first maximum at about 3 nm observed in
the d»/dlogD versus D curves could be attributed to
inter-particle pores. The population of the interpar-
ticle pores of this diameter, promoted by the use of
polymers, may be related to several effects, for
example, to changes in the surface tension of the liquid
within the pores [11, 12]. However, earlier it was
demonstrated that organic polymers have only a small
effect on the surface tension in comparison to
surfactants; moreover, surfactants used as additives
have little effect on the pore structure [3]. It is con-
cluded, therefore, that the surface tension is not the
principal cause of the textural changes in the present
circumstances. The other mechanism that could be
proposed to explain the increase in the first maximum
in Fig. 1 is that in the early stages of the drying and
calcination processes the adsorbed polymer molecules
probably can also act as wall ‘‘separators’’ in the gel
network, up to the temperature required to set the gel.
Above the flash point of the polymer, the gel-structure
supporting agent will oxidize out leaving the pure
calcined gel [6]. This action mechanism is similar to
what has been called the ‘‘filling-removal’’ mechanism
[5].
In line with the above, the TGA/DTA characteriza-
tion results of the precipitates obtained when the poly-
meric additive was added during the hydrolysis step
(Table VI and Fig. 3), show that the polymer elimina-
tion occurs at temperatures higher than those required
for the dehydration process. Therefore, the incorpor-
ated polymer can prevent the collapse of the porous
structure during the drying and the first stages of the
calcination processes. As a result of this, the structure
of the calcined solid is not so different from that of the
hydrous gel and this allows solid materials to be
obtained with higher surface area and porosity. At the
same time, the organic polymer chain might act as
a ‘‘filler’’ to increase mesoporosity in the higher range
of pore diameters (at about 30—50 nm) [5]. In agree-
ment with this, mesopores in this range of pore dia-
For the same polymer, a regular increase in me-
soporosity should be expected as the molecular weight
of the polymer or its amount increase. Such increase in
mesoporosity would not be expected to be linear,
because coiling of the organic chain can occur [13]. In
general, these predicted trends are consistent with our
experimental results.
For the glycol-type polymers, a change in the mo-
lecular weight results only in a slight decrease in the
surface area of the supports (Table III). This effect can
be easily understood because in these experiments the
same volumetric amount of polymer was used and
therefore the small change in surface area must be due
to the decrease in the concentration of terminal func-
tional groups present. Also, if the same volumetric
amount of polymer was used, no significant variations
in the total pore volume of the final sample would be
expected.
When the amount of the polymer added is varied
(Table IV), a small increase in the surface area and
total pore of the sample is observed. However, a close
analysis of the pore-size distribution curve shows that
increasing the amount of polymer additive leads to
a slight decrease in the population of the small-dia-
meter pores in the range of 3 nm, while the population
of pores in the range 10—100 nm tends to increase also
slightly. These two effects result in a small increase of
the average pore diameter.
In the case of the samples modified by the use of
PVA, the results show that an increase in the molecu-
lar weight of the polymer leads to a clearer ‘‘filler’’ type
effect. In particular, an increase of ten times in the
polymer molecular weight results in a change in the
average pore diameter from 3.9 nm to 4.7 nm (Table
III). In agreement with these results, the pore-size
distribution curves show a slight shift of the second
maximum, corresponding to the wider pores, from 45
nm to 50 nm, and an increase of about double in the
contribution of these pores to the total pore volume.
An analysis of the results from Table I and of the
pore-size distribution curves obtained with PVA and
PAA (Fig. 1), indicate that the action mechanisms of
these two polymers are quite different because the use
of PVA generates a pore-size distribution curve sim-
ilar to that obtained with polymers of the glycol type,
whereas the use of PAA increases the proportion of
pores of larger diameter leading also to a bimodal
pore-size distribution curve but much less defined
than those obtained with the other polymers. This
result could be related to the factors that distinguish
the PAA from the other organic polymers used in this
study, i.e.
meters were consistently observed in all the Al O —
ꢀ ꢁ
TiO mixed oxide samples prepared with polymeric
ꢀ
additives. Moreover, pores in this range of diameters
were not observed in the standard sample prepared
without polymer (see Fig. 1, curve 1), hence its appear-
ance is due to the incorporation of polymer into the
hydrous gel. In line with this, it is also observed that
the pore volume of pores between 30 and 50 nm dia-
meter and the position of the second maxima in the
pore-distribution curves shown in Fig. 1, depend on
the amount and type of polymer used as additive (see
Fig. 1, curves 2—5). These results show clearly that the
‘‘filler effect’’, which is evident in the case of PEG and
PPG and absent in the case of PVA, results from the
different amounts of polymer used in each case. The
case of PAA modification presents different features,
which will be discussed later.
(i) the more pronounced polarity of the functional
groups of the PAA with respect to the other polymers
leads to a stronger interaction of the polymeric addi-
tive with the surface of the Al—Ti gel. The formation of
hydrogen bonds between di-alkyl substituted PAA
and the surface of gels of alumina and silica—alumina
has been reported previously [6]. Our infrared results
from the intermediate species also support this propo-
sal since a shift in the bands corresponding to the
characteristic vibrations of the C"O and C—N bonds
in PAA was observed when the polymer was in the
1988

presence of the Al—Ti gel, indicating the existence of an
interaction of these groups with the Al—Ti gel;
(ii) the low solubility of PAA in the water—alcohol
mixture which exists during the hydrolysis step, might
give rise to the formation of polymer aggregates of
different sizes on the surface of which small particles of
the solid gel might adhere, thus creating a secondary
porosity in all the pore-diameter size intervals [4]. To
corroborate this point, an additional series of experi-
ments using PAA, which is the polymer with the
lowest solubility, were made by increasing the amount
of water to make the polymer more soluble in the
reacting medium. The results from these experiments
show that as the amount of water is increased, the
resulting pore-size distribution curve becomes better
defined with two maxima at 3.3 and 60 nm and bears
a greater resemblance to those of the more soluble
polymers (see Fig. 5). Increasing the amount of water
from 150 ml to 300 ml does not produce further
significant changes in the texture of the final solid,
because under these conditions, we are well below
the solubility limit of the PAA in the water—alcohol
mixture.
Figure 5 Pore-size distributions of TiO —Al O mixed oxides
(Series 1) prepared with PAA (200 000 g mol\ꢂ) using different
amounts of water: (1) 50 ml, (2) 150 ml and (3) 300 ml.
ꢀ
ꢀ ꢁ
A comparison of Series 1 and Series 2 samples
demonstrates that, in all cases, the formation of pores
of larger diameter (30—50 nm) is promoted by the
addition of polymer. It appears that the increase in the
population of these pores in Series 2 samples is a result
of the increased local polymer concentration, which
results from the incorporation of the polymeric addi-
tive during the filtration process. This result can be
rationalized by assuming that it is the agglomeration
of the polymer in concentrated solutions which pro-
duces a ‘‘filler’’ effect responsible for the creation of the
larger diameter pores. In contrast, when the polymer
is added during the hydrolysis step, the system is
highly diluted and under these conditions, the small
particles of precipitate will be surrounded by a thin
layer of adsorbed polymer which will change the mor-
phology of the hydrogel. This leads to solids con-
stituted by aggregates of small particles and with high
surface areas and small-diameter pores.
Clearly, the process of interaction between the poly-
mers and the gels is complex and there must be other
effects besides those analysed here. For example, the
change in the acid properties of the final solids with
the nature of the polymeric additives, indicates
a change in the composition or structure of the sur-
face, possibly originated by differences in the titanium
and aluminium alkoxides hydrolysis rates due to the
presence of the polymer functional groups which alter
the properties of the reactive medium.
5. Conclusions
1. The surface area and pore volume of TiO —
ꢀ
Al O mixed oxides prepared by the methods used
ꢀ ꢁ
here are increased by the addition of polymers such as
PEG, PPG, PVA and PAA.
2. The incorporation of the polymers allows the
formation of bimodal pore-size distributions in con-
trast to the monomodal distribution obtained in the
absence of polymers.
3. The addition of the polymers during the hydroly-
sis step leads to greater surface areas but the pore-size
distribution shifts towards the smaller pores com-
pared with adding the additive during filtration.
4. The effect of the type of polymer functional
groups is observed clearly only when the polymer is in
a diluted state, as in the case of addition during the
hydrolysis step. On the contrary, the filler effect is
dominant when the polymer is added to the precipi-
tate during filtration and is therefore highly concen-
trated.
Acknowledgements
Financial support for this work by DGAPA-UNAM,
CONACyT and the IMP FIES programme (Mexico)
is gratefully acknowledged. One of us (E. C.) thanks
PEMEX-Refinacion for the grant received. We are
grateful to Mr R. Hernandez and Mr I. Puente for
obtaining the transmission electron micrographs.
The results of this study show that the action mech-
anisms of the different polymeric additives are quite
similar when added to the precipitate during the filtra-
tion process, and may be explained by a ‘‘filler’’ effect,
whereas in the diluted condition, when the polymer is
added during the hydrolysis step, it is the chemistry of
the polymer functional groups which is determinant
for the final texture of the solid. In this latter case, the
‘‘filler’’ effect will be only of minor importance unless
the solubility of the polymer in the solution is quite
low.
References
1. M. BREYSSE, J. L. PORTEFAIX and M. VRINAT, Catal.
¹oday 10 (1991) 489.
2. J. RAMIREZ, L. RUIZ-RAMIREZ, L. CEDENꢀ O, V.
HARLE, M. VRINAT and M. BREYSSE, Appl. Catal. 93
(1993) 163.
1989

3. D. BASMADJIAN, G. N. FULFORD, B. I. PARSONS and
D. S. MONTGOMERY, J. Catal. 1 (1962) 547.
4. D. L. TRIMM and A. STANISLAUS, Appl. Catal. 21 (1986)
215.
5. A. WHITE, A. WALPOLE, Y. HUANG and D. L. TRIMM,
ibid. 56 (1989) 187.
11. C. J. BRINKER and G. W. SCHERER, ‘‘Sol—Gel Science. The
Physics and Chemistry of Sol—Gel Processing’’ (Academic
Press, New York, 1990).
12. T. KLIMOVA and J. RAMIREZ, Mater. Sci. Forum. 152-153
(1994) 309.
13. T. SATO and R. RUCH, ‘‘Stabilization of Colloidal Disper-
sions by Polymer Adsorption’’ (Marcel Dekker, New York,
1980).
6. L. MASCIA, ¹rends Polym. Sci. 3 (1995) 61.
7. R. SNEL, Appl. Catal. 12 (1984) 347.
8. E. P. BARRET, L. G. JOYNER and P. P. HALENDA,
J. Amer. Chem. Soc. 73 (1951) 373.
9. R. CID and G. PECHI, Appl. Catal. 14 (1985) 15.
10. E. RODENAS, T. YAMAGUCHI, H. HATTORI and
K. TANABE, J. Catal. 69 (1981) 434.
Received 7 April
and accepted 5 December 1997
1990