The zirconium sulfate microcrystal structure in relation to their activity in the esterification

Article abstract of DOI:10.1016/j.molcata.2007.03.028

Zirconium sulfate (ZS) varying in crystalline phase from hydrate to anhydrous was prepared at different treatment temperature. The relationship between microcrystal structure at different treatment temperature and activity in esterification was evaluated.

Full text of DOI:10.1016/j.molcata.2007.03.028

Journal of Molecular Catalysis A: Chemical 272 (2007) 91–95
The zirconium sulfate microcrystal structure in relation
to their activity in the esterification
Joon Ching Juana,b, Jingchang Zhanga,∗, Yajie Jianga,
Weiliang Caoa, Mohd Ambar Yarmob
a Institute of Modern Catalysis, Beijing University of Chemical Technology, State Key Laboratory of Chemical Resource Engineering,
Ministry of Education, Beijing 100029, China
b Advanced Catalysis Technology Laboratory, School of Chemical Sciences and Food Technology, Faculty of Science and Technology,
Universiti Kebangsaan Malaysia, Bangi 43600, Malaysia
Received 3 November 2006; received in revised form 5 March 2007; accepted 6 March 2007
Available online 16 March 2007
Abstract
Zirconium sulfate (ZS) varying in crystalline phase from hydrate to anhydrous was prepared at different treatment temperature. The relationship
between microcrystal structure at different treatment temperature and activity in esterification was evaluated. The number of acid groups was found
to increase with increased of pretreatment temperature, except for the samples treated above 330 ◦C. For that sample, it is observed that hydration
of crystalline water to anhydrous phase is significant at temperature of 330–420 ◦C, reducing the surface acidity. The reaction was poisoned by
the exchange of surface protons on ZS with Cs+. Thus, the major effective active sites must be the protons acidity. It was demonstrated that the
microcrystal structure and catalytic activity of ZS was affected by thermal treatment due to changes in crystalline water.
© 2007 Elsevier B.V. All rights reserved.
Keywords: Zirconium sulfate; Thermal treatment; Esterification; Acid catalyst
1. Introduction
contact with four water molecules, and each water molecule is
in contact with one Zr atom. The sulphate oxygen atoms that are
not in contact with Zr atoms are linked to water molecules via
hydrogen bonds. The interlayer distance, gathered from X-ray
diffraction (XRD) of these compounds is around 0.6 nm.
The use of zirconium sulfate (ZS) or disulfatozirconic acid
as catalyst in esterification reaction has received considerable
attention. In general, metal sulfate surface possesses acid sites
capable of taking part in the catalytic process. Several literatures
have studied the influence of treatment temperature [1], ranging
from 550 to 750 ◦C to attain sulfated zirconia [2]. However the
information concerning ZS, especially the influence of treatment
temperature below 400 ◦C to catalytic activity is limited. In gen-
eral, thehydrationofZSfromhigherhydratetoanhydrouscanbe
obtained by thermal-treatment below 400 ◦C. It is reported that
there is inter-relationship between hydrate water and structure
of ZS [3]. The crystalline ZS is an ion exchanger [4], which has
a layered structure (Fig. 1). Within the layers, each zirconium
(Zr) atom is in contact with four sulfate groups and each sulfate
group is in contact with two Zr atom. Also, each Zr atom is in
Recently, ZS has been used as a catalyst for esterification of
carboxylic acids [5] and redox reaction [6], but the catalyst used
wasnotwellcharacterized. SincethepropertiesofZSarevarious
with thermal treatment conditions, such as treatment tempera-
ture, the reported results must be considered to be preliminary
and qualitative. The microcrystal phase of ZS depends on the
crystalline water [3]. Therefore, it is probable that the catalytic
activity and acidity does also. Thus it is necessary to take all of
these factors into account, if meaningful relationships between
crystalline water and catalytic behavior are to be uncovered.
In this context, we have correlated the esterification activ-
ity with a series of ZS hydrate samples. It is observed that upon
heating the ZS samples to 420 ◦C, they exhibit different catalytic
activity. Moreover, the characterization of the samples was per-
formed by mean of XRD, thermogravimetric analysis (TGA),
Fourier transform infrared (FTIR) and ammonia FTIR analysis.
∗
Corresponding author. Tel.: +86 10 64434904; fax: +86 10 64434898.
E-mail addresses: zhangjc1@mail.buct.edu.cn,
zhangjc1@buct.edu.cn (J. Zhang).
1381-1169/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2007.03.028

92
J.C. Juan et al. / Journal of Molecular Catalysis A: Chemical 272 (2007) 91–95
2. Experimental
was treated with 25 ml of 0.05 M CsCl in ethanol solution. After
drying at room temperature, the solid obtained was washed and
dried before calcined at 420 ◦C. The thermally treated sample
was again mixed with 25 ml of 0.05 M CsCl solution, as stated
above. The cesium ion (Cs+) exchanged ZS was tested for its
catalytic activity in esterification of oleic acid with a view to
determine the poisoning effect of Cs+ and to find out the role of
surface hydroxyl groups in the catalysis. It is reported that only
the surface protons were exchanged by this procedure [7]. It is
possible to exchange only the surface protons of ZS with Cs+
because this ion is too large to exchange into the interior of ZS
in the solution.
2.1. Catalyst preparation
Zirconium sulfate (Beijing, Shiji Chem. Co) were treated in
a muffle furnace at 110, 160, 200, 330 and 420 ◦C, respectively.
2.2. Characterization
XRD was performed on a Shimadzu HR 6000X. The condi-
˚
tions used were: Cu K␣ radiation (λ = 1.54056 A) at 40 kV and
200 mA, scanning angle (2θ) from 5 to 60◦ with scan-steps of
0.04◦/s.
FTIR spectra of the samples were recorded on a Shimadzu
FTIR-8400S by the KBr pellet technique. The sample was
ground with spectral grade KBr to form a mixture, which was
pelletized using a hydraulic press. This pellet was used to record
2.3. Esterification of oleic acid with 1-butanol
Esterification reaction was performed in three-neck flask at
atmospheric pressure equipped with a telfon-coated magnetic
stirring bar, a thermometer, Dean-Strak receiver and reflux con-
denser. A typical esterification reactant consists of oleic acid
(31.6 ml, 0.10 mol), 1-butanol (11.0 ml, 0.12 mol) and fresh cat-
alyst (1.41 g). The moles ratio for oleic acid to butanol is 1:1.2
and the weight ratio of the catalyst was 5 wt.%, based on oleic
acid. No solvent was added to promote more environmental
friendly process. The reaction temperature was slowly raised
to 110–120 ◦C and maintained at the desired temperature dur-
ing the specified reaction periods. The heating was stopped
after 4 h to determine the conversion. The product was ana-
lyzed by Hewlett Packard Gas Chromatography (GC) 5890
model with flame ionization detector (FID). The column used
is MXT-1 Crossbond 100% dimethyl polysiloxane (15 m) and
n-tetradecane was added as an internal standard.
the infrared spectra, ranging 4000–400 cm−1
.
TGA was performed by Mettler Toledo thermal analyzer. All
the samples were heated at 10 ◦C min−1 in air up to 900 ◦C.
Catalyst surface acidity was analyzed by ammonia adsorption
followed by FTIR spectroscopy. Finely dried ground catalyst
samples were pressed for 2 min into a self-supporting wafer.
The wafer was preheated by stepwise heating from 25 to 105 ◦C
under extra-pure nitrogen gas at flow-rate of 30 cm3 min−1 for
30 min. Then, the samples were permeated with ammonia for
30 min at 80 ◦C. After adsorption of the ammonia vapors on the
wafer, the physisorbed ammonia was expelled using nitrogen gas
at a flow-rate of 30 cm3 min−1 for 30 min. Sample spectra were
recorded at each stage on a Shimadzu FTIR-8400S equipped
with a Thermo Spectra-Tech high temperature/vacuum chamber
to elucidate the exact acidic nature of the catalyst surfaces. Dif-
ferences between the spectra of ammonia adsorbed on samples
and a reference were obtained by subtraction.
3. Results and discussion
Ion exchanged of ZS sample with cesium chloride (CsCl)
solution was conducted to find out whether the surface hydroxyl
groups played a role in the catalysis. A known amount of ZS
3.1. XRD
Fig. 2 shows the ZS samples after they have received various
thermal treatments. It is obvious that there is phase trans-
formation upon heating the ZS sample to 420 ◦C. The phase
transformation of ZS hydrate to anhydrous upon thermal treat-
ment has also been reported by Bear [8] and Kotsareneko et
al. [9]. It is observed that upon heating the sample to 110 ◦C,
its phase remains as the parent ZS. Thus, ZS retained it crys-
talline water but adsorbed water was removed upon heating up
to 110 ◦C. Further, the tetrahydrate ZS phase was observed upon
heating the sample to 160 ◦C. At 420 ◦C, tetrahydrate was trans-
formed to anhydrous phase. These findings are in conformity
with standard XRD data reported previously [8]. In addition,
upon heating the sample to 750 ◦C, the ZS decomposed to zir-
conia phase. The strong diffraction lines of ZS treated at 750 ◦C
are corresponded to monoclinic zirconia phase. As shown in
Fig. 2, ZS did not decompose to zirconia upon heating up to
420 ◦C because there is no reflection lines corresponded to zir-
conia phase was observed. Therefore, it can be assumed that
upon heating the ZS to 420 ◦C, the phase transformations are
due the hydration of ZS sample.
Fig. 1. Idealized structure of zirconium sulfate (ZS) tetrahydrate.

J.C. Juan et al. / Journal of Molecular Catalysis A: Chemical 272 (2007) 91–95
93
ther, the weight loss began at 600–730 ◦C is due to evolution of
SO3 decomposed from zirconium sulfate, following the equa-
tion: Zr(SO4)2 → ZrO2 + 2SO3. This could be observed from
Fig. 3(g) that no weight loss could be observed upon heating to
750 ◦C.
UponheatingtheZSsampleto110, 160, 200, 330and420 ◦C,
the calculated weight loss based on TGA curves below 250 ◦C
is corresponded to the removal of 5, 4, 2, 0.5 and 0 mol of crys-
talline water, respectively. It is noteworthy that the hydration
state of ZS at various thermal treatments is parallel to those
observed from XRD. For instance, upon heating the ZS sample
to 160 ◦C, 4 mol of crystalline water was calculated from TGA
analysis. In addition, upon heating the ZS sample up to 420 ◦C,
a weight loss at 600 ◦C was still detected. This is clearly indi-
cated that ZS did not decompose to zirconia upon heating up to
420 ◦C.
Fig. 2. XRD patterns of ZS. (a) Before heating, and after heating at (b) 110 ◦C,
(c) 160 ◦C, (d) 200 ◦C, (e) 330 ◦C, (f) 420 ◦C and (g) 750 ◦C.
3.3. Ammonia adsorption Fourier transform infrared
Fig. 4 shows the FTIR spectra of ammonia adsorbed on ZS
sample treated at 110, 160, 200, 330, and 420 ◦C. The band at
1525 cm−1 is the characteristics peaks of ammonium ion, which
is formed on the Bronsted acid sites and the adsorption peak
at 1700 cm−1 is contributed to ammonia coordinately bonded
Lewis acid sites, indicating the presence of both Bronsted and
Lewis acid sites on the surface of ZS. The Bronsted acidity is
due to the Zr–OH groups, while the Lewis acidity is from Zr4+.
The peak intensity increased with increasing thermal treatment,
meaning that amount of acidity increase with thermal treatment
temperature.
With the increasing thermal treatment temperature, the total
amount acidity was gradually increased. However, upon heating
the ZS sample to above 330 ◦C, the acidity was slightly reduced.
It is observed that ZS treated at 200 ◦C exhibits higher amount of
acidity than other treatment temperatures. Upon heating ZS sam-
ple up to 420 ◦C, the anhydrous ZS still exhibit certain amounts
of Bronsted acidity. It is reported that the interaction of Zr4+ with
3.2. Thermogravimetric analysis
Weight loss curves for the ZS samples are shown in Fig. 3. It
is noteworthy that the TGA curves at temperature below 250 ◦C
became shallow upon heating the ZS samples to 420 ◦C. This
would mean that ZS began losing it crystalline water upon
heating the ZS to 420 ◦C. As shown in Fig. 3(a), the parent
ZS lose it crystalline water in three steps, around 80, 165 and
220 ◦C to anhydrous phase before decomposed to zirconia. The
weight loss up to 220 ◦C is corresponded to the removal of
5 moles of crystalline water. Although the ZS brought was
labeled as tetrahydrate, it is surprising to observe more than 4
moles of water was removed. The ZS seems to be rather hygro-
scopic and it is concluded that the starting reactant is ZS with
5 mol of hydrate water attached to it. These finding are also
in accordance with those reported by Strydom et al. [10]. Fur-
Fig. 3. TGA curves of ZS: (a) Before heating and after heating at (b) 110 ◦C,
(c) 160 ◦C, (d) 200 ◦C, (e) 330 ◦C and (f) 420 ◦C.
Fig. 4. Infrared spectra of NH3 adsorbed on ZS: (a) before heating and after
heating at (b) 110 ◦C, (c) 160 ◦C, (d) 200 ◦C, (e) 330 ◦C and (f) 420 ◦C.

94
J.C. Juan et al. / Journal of Molecular Catalysis A: Chemical 272 (2007) 91–95
sulfate ions in highly active catalysts have a strong tendency to
adsorb basic molecules [11], such as water [12]. Perhaps, this is
likely to form Bronsted acid sites on anhydrous ZS.
3.4. FTIR spectroscopy
The FTIR spectra of ZS sample treated at different tem-
perature are given in Fig. 5. ZS sample treated up to 420 ◦C
showed absorption bands in the S–O stretching region [12],
ranging from 1300 to 850 cm−1 provides proof for the pres-
ence of several families of surface monosulfates and bidentate
sulfate ion coordinated to Zr4+ [13]. It is proven that sulfate
species was preserved upon heating up to 420 ◦C. The adsorption
bands of lower treatment temperature sample are vague, due to
some remaining molecular water complicated the spectrum [12].
Theabsorptionbandsaround3300 cm−1 systematicallychanged
with thermal treatment. In general, the molecular water in the
vapor state found at 3756 and 3652 cm−1, which assigned to OH
stretching frequencies, are lowered when the water is bound in
a crystal lattice. In addition, stronger hydrogen bonds are char-
acterized by shorter Ow–O distances and lower OH stretching
frequencies. Thus, the differences spectra in Fig. 5 are due to the
presence of differences bonded water molecules in the present
hydrates. It is observed that an absorption maximum occurs near
3200 cm−1 in the spectra of ZS samples treated below 160 ◦C,
which due to coordinated water molecules. It is noticeable that
upon heating the ZS sample, the OH stretching position shifted
to higher frequencies. This would mean that the Ow–O distance
became shorter from higher hydrate to anhydrous. Further, it
was reported that Ow–O distance is 0.269 and 0.271 nm for ZS
tetrahydrate [4] and monohydrate [14], respectively. Therefore,
this help us to explain the OH position of ZS sample shifted to
higher frequencies is due to transformation of ZS from hydrate
to anhydrous. For the ZS sample treated at 420 ◦C, the maxi-
mum absorptions shifted to higher frequencies, substantiates the
presence of almost adsorbed water molecules in this compound.
Fig. 6. Catalytic activities of ZS for esterification of oleic acid as a function of
treatment temperature (ꢀ) before heating and after heating at (᭹) 110 ◦C, (ꢁ)
160 ◦C, (ꢂ) 200 ◦C, (ꢀ) 330 ◦C and (×) 420 ◦C.
3.5. Catalyst testing for esterification of oleic acid
The conversion of oleic acid with ZS samples of different
treatment temperature is shown in Fig. 6. The reaction con-
dition is mentioned above for esterification of oleic acid with
n-butanol. It is interesting to observe that the catalytic activity
of ZS is influenced by the thermal treatment. The catalytic activ-
ity of ZS increased with the temperature but slightly fell after
heating above 330 ◦C. It is presumed that the difference in their
activities is closely related to their crystalline water in ZS at ele-
vated temperatures. The catalytic activity of ZS sample treated
at 200 ◦C is most active compared to other samples. During
the first hour, a significant catalytic activity change at different
treatment temperature was observed. The catalytic activity of
ZS sample treated at 200 ◦C increased rapidly compared with
other samples. The conversion of ZS treated at 200 ◦C is 80%
during first hour, which is 32 % more than that of ZS sample
treated at 420 ◦C. The preliminary result from TGA shows that
dihydrate is present upon heating the ZS sample to 200 ◦C. The
catalytic activity data show that the entire ZS sample treated
at various thermal treatments is selective toward the esterifica-
tion. The selectivity for esterification of oleic acid to butyl oleate
is found to be more than 99%. Very strong acid sites are nor-
mally oxidized oleic acid, which are present on sulfuric acid.
Our results indicate that this situation these sites are not present
on ZS because ZS only exhibit moderate acidity. It is noteworthy
that catalytic activity began to decrease upon heating to 330 ◦C.
The decreased of activity for ZS at higher thermal treatment
is attributed to the fact that most of its hydroxyl group, which
contributed to Bronsted acid sites from hydrate water are con-
densed out at 420 ◦C. However, it is seen that ZS sample still
having certain amount of acidity after treated at 420 ◦C.
Therefore, we are interested to determine whether the surface
hydroxyl groups were solely responsible for the esterification
activity. It may assume that the replacement of the protons
by Cs+ should eliminate this activity. Fig. 7 gives the activ-
ity for cesium exchanged ZS samples, from which it is seen
that Cs+ ion exchanged inhibits the reaction considerably. For
Fig. 5. Infrared spectra of ZS: (a) before heating and after heating at (b) 110 ◦C,
(c) 160 ◦C, (d) 200 ◦C, (e) 330 ◦C and (f) 420 ◦C.

J.C. Juan et al. / Journal of Molecular Catalysis A: Chemical 272 (2007) 91–95
95
are considered to be the active centers for esterification of fatty
acid. Preliminary results show that the esterification may take
place through a ligand exchange reaction with fatty acid on the
Bronsted acid sites, although we do not have decisive evidence.
4. Conclusions
This paper has shown that there is a correlation between
catalytic activity and thermal treatment of ZS for esterifica-
tion reaction. A close relationship between acidity and catalytic
activity could be observed. We can correlate the higher activity
exhibited by the ZS at various thermal treatment to its acid-
ity. The different hydrate or phase transformation was obtained
upon heating the ZS sample. They give rise to different XRD
patterns, FTIR spectra, acidity and catalytic activity. The results
show that ZS treated at 200 ◦C is the most active due to higher
amount of acidity. It is demonstrated that upon heating the ZS
sample to 420 ◦C, the crystalline phase and acidity changes are
due to hydration of ZS. One of the active sites, which get poi-
soned by Cs+ ion exchange, should be related to the hydroxyl
groups (Bronsted acid sites). This fact leads us to assume that
Bronsted acid sites should mainly responsible for activity of ZS.
Fig. 7. Effect of Cs+ ion exchanged on the activity of ZS with no Cs+ treated
at 420 ◦C (A), after first treatment with CsCl (B), after second treatment with
CsCl (C), and without catalyst (D). The reaction was carried out under the same
conditions as stated above for comparison purposes.
comparison purposes, the reaction condition is the same as
shown in Fig. 6. The conversion was 52% and second treatment
with CsCl yielded a still lower conversion of 44%. Although,
the poisoning experiments (Fig. 7) suggest the participation of
protons in the chemical reaction, the fact that the catalyst does
not get totally indicates that the catalytic activity of ZS is gov-
erned not only by the number of protons or acid sites but also
certain unknown inherent characteristics or surface defects cre-
ated by structural modification at elevated temperature. In other
words, the present observations suggest the presence of at least
two kinds of sites on ZS as mentioned above. One of the sites,
which get poisoned by Cs+ ion exchange, should be related to the
hydroxyl group (Bronsted acid site). This fact leads us to assume
that the second kind of site which does not get poisoned by Cs+
ion or by thermal treatment should be responsible for the resid-
ual activity of ZS. This second kind of site should be associated
either with the phase transformation to zirconium–oxygen bond
stretching or Lewis-type centers formed by removal of protons.
Preliminary results show that Lewis-type centers formation is
more pronounced. Besides that, it may also assume that Bronsted
acid sites are mainly responsible for activity of ZS. The Bron-
sted acid sites result from the weakening of the –OH bond by the
sulfate groups attached to Zr. Meanwhile, the Lewis acid sites
are shorts of electrons, at Zr4+ centers as the result of electron
withdrawing nature of the sulfate groups. These active species
Acknowledgments
The authors are thankful to the National Natural Science
Foundation of China (No. 20076004) and National Malaysia
Ministry of Science and Technology (IRPA grant 09-02-
02-0033) for the financial support. We are also grateful to
UNESCO/China Great Wall Co-Sponsored program for a fel-
lowship to J.C. Juan.
References
[1] Y. Huang, B. Zhao, Y. Xie, Appl. Catal. A: Gen. 173 (1998) 27.
[2] K. Arata, M. Hino, N. Yamagata, Bull. Chem. Soc. Jpn. 63 (1990) 244.
[3] I.J. Bear, W.G. Mumme, Acta Cryst. B 26 (1970) 1140.
[4] J. Singer, D.T. Cromer, Acta Cryst. 12 (1959) 719.
[5] W. Cao, J. Zhang, Q. Lv, J. Beijing Univ. Chem. Technol. 26 (1999) 72.
[6] W. Cao, J. Zhang, Q. Lv, J. Petrochem. Univ. 13 (2000) 41.
[7] G. Alberti, M.G. Bernasconi, M. Casciola, U. Costantino, Ann. Chim. 68
(1978) 265.
[8] I.J. Bear, Aust. J. Chem. 20 (1967) 415.
[9] N.S. Kotsareneko, V.P. Shmachkova, Kinet. Catal. 43 (2002) 280.
[10] C.A. Strydom, G. Pretorius, Thermochim. Acta 223 (1993) 223.
[11] C. Miao, W. Hua, J. Chen, Z. Gao, Catal. Lett. 37 (1996) 187.
[12] J.R. Sohn, D.H. Seo, Catal. Today 87 (2003) 219.
[13] E.E. Platero, M.P. Mentruit, C.O. Arean, A. Zecchina, J. Catal. 162 (1996)
268.
[14] I.J. Bear, W.G. Mumme, Acta Cryst. B 26 (1970) 1125.