Preparation of highly active and reusable heterogeneous Al 2O3-Pd catalysts by the sol-gel method using bayberry tannin as stabilizer

Article abstract of DOI:10.1007/s11164-012-0487-z

A novel heterogeneous Al₂O₃-Pd catalyst has been prepared by the sol- gel method; bayberry tannin (BT) was used as stabilizer to prevent the migration and aggregation of Pd species during calcination. According to N₂ adsorption/ desorption determination, Al ₂O₃-Pd has a mesoporous structure and its specific area is as high as 336.5 m²/g. Transmission electron microscopy observation indicated that the size of the Pd particles was greatly reduced by the presence of BT. On the basis of X-ray photoelectron spectra analysis, it was found that the most of Pd nanoparticles were dispersed in the pores, implying that BT can prevent migration of Pd particles from the pores to the outer surface of Al ₂O₃ during calcination. For comparison, Al ₂O₃-Pd* was prepared by the sol-gel method but without use of BT. In the hydrogenation of acrylic acid, Al₂O ₃-Pd had high catalytic activity and excellent reusability compared with commercial and traditionally prepared heterogeneous Pd catalysts. The turnover number of Al₂O₃-Pd is as high as 11,328.0 mol/ mol after recycling seven times, which is much higher than that of a commercial Pd-C catalyst (8048.0 mol/mol). Springer Science+Business Media B.V. 2012.

Full text of DOI:10.1007/s11164-012-0487-z

Res Chem Intermed (2012) 38:1609–1618
DOI 10.1007/s11164-012-0487-z
Preparation of highly active and reusable
heterogeneous Al O –Pd catalysts by the sol–gel
method using bayberry tannin as stabilizer
2
3
Yanping Wang
•
Xin Huang
•
Xuepin Liao Bi Shi
•
Received: 30 September 2011 / Accepted: 10 January 2012 / Published online: 3 February 2012
Ó Springer Science+Business Media B.V. 2012
Abstract A novel heterogeneous Al O –Pd catalyst has been prepared by the sol–
2
3
gel method; bayberry tannin (BT) was used as stabilizer to prevent the migration
and aggregation of Pd species during calcination. According to N adsorption/
2
desorption determination, Al O –Pd has a mesoporous structure and its specific area
3
2
2
is as high as 336.5 m /g. Transmission electron microscopy observation indicated
that the size of the Pd particles was greatly reduced by the presence of BT. On the
basis of X-ray photoelectron spectra analysis, it was found that the most of Pd
nanoparticles were dispersed in the pores, implying that BT can prevent migration
of Pd particles from the pores to the outer surface of Al O during calcination. For
2
3
*
comparison, Al O –Pd was prepared by the sol–gel method but without use of BT.
2
3
In the hydrogenation of acrylic acid, Al O –Pd had high catalytic activity and
2
3
excellent reusability compared with commercial and traditionally prepared hetero-
geneous Pd catalysts. The turnover number of Al O –Pd is as high as 11,328.0 mol/
2
3
mol after recycling seven times, which is much higher than that of a commercial
Pd–C catalyst (8048.0 mol/mol).
Keywords Sol–gel method ꢀ Al O –Pd catalyst ꢀ Bayberry tannin ꢀ
2
3
Reusability
Y. Wang ꢀ X. Huang ꢀ X. Liao (&)
Department of Biomass and Chemical Engineering, Sichuan University,
Chengdu, People’s Republic of China
e-mail: xpliao@scu.edu.cn
B. Shi (&)
National Engineering Laboratory for Clean Technology of Leather Manufacture,
Sichuan University, Chengdu 610065, People’s Republic of China
e-mail: shibi@scu.edu.cn
123

1610
Y. Wang et al.
Introduction
Palladium (Pd) catalysts are among the most important noble metal catalysts, with
remarkable catalytic properties which have been applied in various reactions, for
example the Heck reaction, hydrogenation of alkynes and nitro compounds,
asymmetric hydrogenation reactions, and oxidation of NO [1–6]. Pd catalysts are
generally classified into homogeneous catalysts and heterogeneous catalysts.
Homogeneous Pd catalysts often have high activity and selectivity but separation
from the reaction system is a problem, and, therefore, their practical application is
limited [7]. In contrast, heterogeneous Pd catalysts can be easily recovered by use of
a simple operation (filtration, centrifugation, or natural decantation), which is much
preferred for practical application and economy [8]. However, the activity of
heterogeneous Pd catalysts is often lower than that of homogenous ones and the Pd
particles easily migrate to the outer surface of porous supports during calcination.
Therefore, preparation of heterogeneous Pd catalysts with high activity and
reusability is still a challenge.
Heterogeneous Pd catalysts are conventionally prepared by supporting Pd on
porous oxides with a high specific surface area, and impregnation is a frequently
used method. In a typical synthetic process, Pd precursors dissolved in solvent are
first introduced into solid matrices by impregnation; drying and calcination are then
followed by reduction [9, 10]. Unfortunately, because of weak interactions between
Pd precursors and matrices, the Pd precursors located in pores of supports might
substantially migrate to the outer surface of the support on evaporation of the
solvent during drying and calcination, which further leads to serious aggregation
and/or partial expulsion of Pd particles after reduction [11]. Such heterogeneous Pd
catalysts often have relatively low activity and poor reusability.
The sol–gel method is generally used to prepare porous oxides with a high
specific surface area, for example Al O , SiO and TiO [12, 13]. On the basis of
2
3
2
2
this method, we assumed that if a water-soluble stabilizer which can chelate with
metal ions were added to the sol with the Pd precursors, migration of Pd the species
could be substantially suppressed during the drying and calcination process. In this
way, the Pd precursors should be well distributed in the porous matrix, and the Pd
particles formed from subsequent calcination and reduction would be mainly
constrained in the pores of matrix. Accordingly, aggregation and leakage of Pd
particles can be alleviated during the catalytic reaction process. Herein, we
implemented this new strategy to prepare a novel heterogeneous Al O –Pd catalyst,
2
3
because Al O –Pd is commonly used as a heterogeneous Pd catalyst.
2
3
Plant tannin is a natural polyphenol extracted from roots, leaves, fruits, and bark
of plants [14]. There are numerous phenolic hydroxyl groups in tannin molecules
[
15] which can chelate with many metal ions [16]. Hence, it is believed that tannin
2
?
chelated Pd can be well dispersed in the alumina sol when using the sol–gel
method to prepare the Al O support. Moreover, tannin mainly consists of the
2
3
elements carbon, hydrogen, and oxygen, which can be effectively removed by
thermal treatment. In this manner, the pore structure of Al O can be well preserved,
and the Pd ions can be converted to Pd oxides in the pores of Al O , which could be
2
3
2
3
further reduced to elemental Pd by H reduction. On the basis of these ideas we
2
1
23

Preparation of highly active and reusable heterogeneous Al
2
O
3
–Pd catalysts
1611
prepared a novel heterogeneous Al O -supported Pd nanoparticle (Al O –Pd)
2 3
2
3
catalyst. The as-prepared catalyst was then characterized by wide-angle X-ray
diffraction (XRD), X-ray photoelectron spectra (XPS), and transmission microscopy
(
TEM). To investigate the catalytic behavior of this novel catalyst, the catalytic
hydrogenation of acrylic acid was used as a probe reaction. The reusability of this
catalyst was also investigated. For comparison, catalytic hydrogenation of acrylic
*
acid was also performed, under the same experimental conditions, with Al O –Pd
2
3
*
*
catalyst prepared by the sol–gel method without use of BT, c-Al O –Pd prepared
2
3
by the impregnation method, and commercial Pd–C.
Experimental
Materials and instrument
Bayberry tannin (BT) was extracted from the bark of Myrica esculenta with
acetone–water solution (1:1, v/v) and then spray-dried. The tannin content of BT is
6
7.3%, and its average molecular weight is 5,000 Da. Aluminum isopropoxide
AIP), PdCl , and other chemicals were all analytical-grade reagents and used as
(
received.
2
Preparation and characterization of catalysts
AIP (10.27 g) was added to a mixture of 50 mL hexamethylene and 5 mL deionized
water. The resulting mixture was supersonically dispersed for 15 min at room
temperature, and then kept under vigorous stirring for 12 h at 335 K. This process
resulted in the formation of well dispersed sol. Subsequently, 50 mL 20 g/L BT
2
?
solution was mixed with 2.5 mL 10 g/L Pd solution, and this mixture was then
added drop-wise to the above sol. After stirring for another 12 h, hydrolysis of AIP
was complete and the sol had changed into a gel. The resulting gel was collected by
centrifugation, dried, calcined (550 °C for 3 h), and reduced (at 200 °C and 2 MPa
H pressure for 2 h). The as-prepared catalyst was denoted ‘‘Al O –Pd’’ and the Pd
2
2 3
*
content was 0.98% (w/w). For comparison, Al O –Pd was prepared by the same
2
procedures in the absence of BT. In addition, c-Al O –Pd prepared by the
3
*
*
2
3
impregnation method and commercial Pd–C (Pd 10%) were used for comparison.
*
*
For preparation of c-Al O –Pd , 1.0 g commercial c-Al O was suspended in
1
2
3
2 3
2
?
0 mL 1 g/L Pd solution and then dried at 80 °C. The collected solid product was
calcined at 550 °C for 3.0 h, followed by reduction with 2.0 MPa H at 200 °C for
2
*
*
2
.0 h. The Pd content of c-Al O –Pd was determined to 0.96% (w/w).
2
3
*
**
The crystalline structures of Al O –Pd, Al O –Pd , and c-Al O –Pd were
2
3
2
3
2 3
characterized by XRD (X’Pert Pro, Philips, The Netherlands). X-ray photoelectron
spectroscopy (XPS, Kratos XSAM-800, UK) analysis was conducted with Mg Ka
radiation (hv = 1253.6 eV) and the pass energy was 31.5 eV. All of the binding
energy peaks of XPS spectra were calibrated by placing the principal C 1 s binding
energy peak at 284.7 eV. The specific surface area and pore structure of the
catalysts were analyzed by N adsorption/desorption by use of a surface area and
2
123

1612
Y. Wang et al.
porosity analyzer (BET, Micromeritics, Tristar 3000, USA). Transmission electron
microscopy (TEM; Jeol 2010, Japan) performed at an acceleration voltage of
2
00 kV was used to observe the morphology of the Pd particles in the catalysts.
Hydrogenation of acrylic acid
Catalytic hydrogenation of acrylic acid was used to assess the activity and
**
*
reusability of Al O –Pd, Al O –Pd , c-Al O –Pd , and Pd–C. All the reactions
2
3
2
3
2 3
were carried out in a stainless steel Parr autoclave equipped with a stirring bar. In a
typical catalytic reaction, 25.0 mL methanol was added into the reactor, and then
specific amounts of catalyst (containing 5 lmol Pd) and acrylic acid (10.0 mmol)
were added. Subsequently, the reactor was heated to 30 °C and kept at 1.0 MPa H
2
under constant stirring. When the reaction was complete, the products were
analyzed by gas chromatography (GC) and the catalysts were collected by
centrifugation, followed by thorough washing with methanol. The turnover
frequency (TOF) and turnover number (TON) of the catalysts were calculated by
use of the equations:
TOF ¼ Substrate hydrogenatedðmol)/[Pd(mol) ꢁ tðh)]
TON ¼ Substrate hydrogenatedðmol)/Pd(mol)
ð1Þ
ð2Þ
Results and discussion
Characterization of catalysts
*
**
Figure 1 shows the XRD patterns of Al O –Pd, Al O –Pd , and c-Al O –Pd .
2
3
2
3
2 3
It can be observed that these three Pd catalysts have similar XRD spectra which
contain the characteristic peaks of c-Al O at 36.7°, 48.7°, and 67.1° [17]. c-Al O
2
3
2 3
is often used as supporting matrix to prepare heterogeneous metal catalysts, because
of its high specific surface area and well defined porosity [18]. N adsorption/
2
desorption was used for analysis of the structural features of the as-prepared Al O –
2
3
*
**
Pd, Al O –Pd , and c-Al O –Pd . The N adsorption/desorption isotherm curves of
Al O –Pd, Al O –Pd , and c-Al O –Pd are shown in Fig. 2a–c; all have type IV
2
3
2
3
2
**
*
2
3
2
3
2 3
or V hysteresis loops, suggesting the mesoporous structures of the catalysts [19].
2
However, the specific surface area of Al O –Pd is as high as 336.5 m /g, which is
2
3
*
much higher than that of Al O –Pd (225.4 m /g). In addition, the average pore
2
2
3
*
diameter of Al O –Pd (5.83 nm) is also larger than that of Al O –Pd (3.66 nm), as
2
3
2 3
shown in Fig. 2d–e. The different structural properties of those two catalysts is
probably because of the use of BT during the sol–gel preparation of the catalysts.
Actually, tannins can interact with each other by hydrophilic and/or hydrophobic
interactions, and thus, supramolecules may be formed by assembly of the tannins.
It can be reasoned that these supramolecules could act as templates during the sol–
gel process, and are subsequently removed by calcination, creating more developed
pores. From the perspective of mass transfer, the Al O –Pd should have much better
2
3
catalytic activity, owing to its higher specific surface area and larger pores.
1
23

Preparation of highly active and reusable heterogeneous Al
2
O
3
–Pd catalysts
1613
*
**
–Pd
Fig. 1 X-ray diffraction patterns of Al
2
O
3
–Pd, Al
2
O
3
–Pd , and c-Al
2
O
3
*
**
The physical properties of Al O –Pd, Al O –Pd , c-Al O –Pd , and Pd–C are
2
3
2
3
2 3
summarized in Table 1.
To confirm that use of BT in the preparation of Al O –Pd can suppress the
2
3
migration of Pd particles to the outer surface of Al O , ICP–AES and XPS analysis
2
3
was performed. The Al O –Pd was first digested by nitric acid, and the
2
3
concentration of Pd in the digested solution was then determined by ICP–AES.
On the basis of ICP–AES analysis, the average Pd content of the Al O –Pd was
calculated to be 0.98% (w/w). On the basis of XPS after reduction, the Pd content at
2
3
the outer surface of Al O –Pd was just 0.64%. According to the literature, XPS
2
3
analysis results mainly reflect the surface composition of samples. Thus, it is
believed that the Pd particle of Al O are mainly dispersed in the inner pores of the
2
3
2
Al O because of the chelation and stabilization effect of BT toward Pd , which
?
2
3
2
?
substantially prevents the migration of Pd with the solvent which evaporates
during calcination process.
Figure 3 is the Pd 3d XPS spectrum of Al O –Pd. It contains two pairs of
2
3
doublets, one located at 335.000 and 339.770 eV, which are assigned to elemental
2
?
Pd(0), and another pair located at 336.370 and 341.620 eV, which belong to Pd
.
2
Actually, approximately 40% of Pd has been reduced to the elemental form.
?
*
**
Figure 4 show the TEM images of Al O –Pd, Al O –Pd , c-Al O –Pd , and
2
3
2
3
2 3
Pd–C. It is obvious that distribution of Pd nanoparticles in the Al O –Pd is much
better than for other three Pd catalysts. Moreover, the average particle size of Pd in
2
3
*
the Al O –Pd is 3–5 nm, which is much smaller than that of Al O –Pd
3
2
2 3
*
*
(10–23 nm), c-Al O –Pd (21–42 nm), and Pd–C (19–35 nm) (Fig. 5). When
considering the porous structure of Al O –Pd and Al O –Pd , the Pd nanoparticles
2
3
*
2
3
2 3
in Al O –Pd should be mainly located in the pores of the Al O , whereas a
3
2
2 3
substantial number of Pd particles are located on the outer surface of Al O for
2
3
*
Al O –Pd . In addition, migration and aggregation of Pd species indeed occurred
2
3
*
during the preparation of Al O –Pd when BT was not introduced.
2
3
123

1614
Y. Wang et al.
1
23

Preparation of highly active and reusable heterogeneous Al
2
O
3
–Pd catalysts
1615
*
**
–Pd , and Pd–C
Table 1 Physical properties of Al
2
O
3
–Pd, Al
2
O
3
–Pd , c-Al
2
O
3
2
Average Pd content (%) Specific area (m /g) Pore size (nm)
Pd catalyst Support
Al
2
O
O
3
3
–Pd
–Pd
Al
2
2
O
O
3
3
0.98
0.98
0.96
336.5
225.4
315.8
998.3
5.83
3.66
5.60
1.72
*
Al
2
Al
*
*
c-Al
2
O
3
–Pd
2 3
c-Al O
Pd–C
Activated carbon 10.0
Fig. 3 Pd 3d XPS spectrum
of the Al –Pd
2 3
O
Hydrogenation of acrylic acid
Hydrogenation of acrylic acid was used to evaluate the catalytic performance of
*
**
Al O –Pd, Al O –Pd , c-Al O –Pd , and Pd–C. Figure 6 shows the catalytic
2
3
2
3
2 3
*
activity and reusability of these four Pd catalysts. Compared with Al O –Pd , the
2
3
Al O –Pd has much greater activity. The TOF of the Al O –Pd is 7792.0 mol/mol/h
2
3
2 3
*
whereas it is only 3044.0 mol/mol/h for Al O –Pd . As shown in Table 1, the
2
3
*
specific surface area of Al O –Pd is approximately 1.5 times that of Al O –Pd .
2
3
2 3
Moreover, the Pd particles in Al O –Pd have a small particle size of 3–5 nm and are
2
3
*
well dispersed in the pores of Al O , whereas the Pd particles in Al O –Pd are of
2
3
2 3
larger size (11–21 nm) and have suffered serious aggregation, as shown in Fig. 5.
According to the literature, the size of metal nanoparticles greatly affects their
catalytic activity [20–22]. In general, metal particles with smaller size have a larger
fraction of surface metal atoms, and will have greater activity than the
corresponding counterparts with larger Pd particle size. Thus, it is understandable
*
that the Al O –Pd with larger Pd particle size has much lower catalytic activity
2
3
than Al O –Pd.
2
3
The reusability of Al O –Pd was excellent; it was reused seven times without
3
2
significant loss of activity. In the 7th run, the TOF of the Al O –Pd was 7384.0 mol/
2
3
mol/h, which was approximately 94.7% of that in the 1st run. In contrast, the
*
reusability of Al O –Pd is quite poor. It was found that its TOF in the 3rd run was
2
3
drastically decreased to 921.6 mol/mol/h. The excellent reusability of the Al O –Pd
2
3
could be because leakage and aggregation of Pd particles were greatly inhibited.
123

1616
Y. Wang et al.
Fig. 4 TEM images and particle size distribution of Pd catalysts
All these results imply that BT indeed plays a crucial role in preventing Pd
migration during the preparation of the catalyst, which further ensures the high
activity and reusability of the catalyst.
*
*
c-Al O –Pd has the lowest activity and its reusability is also quite poor.
2
3
*
*
Compared with Al O –Pd, c-Al O –Pd had similar physical properties (Table 1).
2
3
2 3
However, the TEM images indicated the diversity of Pd particle size and
distribution. The large size and uneven distribution of Pd particles could be the
*
*
main reason for the low catalytic activity and reusability of c-Al O –Pd . The
2
3
commercial Pd–C catalyst has the highest initial catalytic activity of these four Pd
catalysts, but its reusability is also quite poor. The average Pd content of Pd–C was
much higher and the TEM images revealed serious migration of Pd particles. The
high Pd content and serious migration lead to partial expulsion of Pd particles
during the reaction process and caused the sharp decrease of catalytic activity.
To evaluate the efficiency of these four catalysts, their corresponding TON
was calculated, and is in the sequence: Al O –Pd (11328.0 mol/mol) [ Pd–C
2
3
1
23

Preparation of highly active and reusable heterogeneous Al
2
O
3
–Pd catalysts
1617
Fig. 5 Particle size distribution of Pd catalysts
2 3 2 3 2 3
Fig. 6 The activity and reusability of Al O –Pd, Al O –Pd*, c-Al O –Pd**, and Pd–C
123

1618
Y. Wang et al.
*
**
(
mol). Therefore, Al O –Pd prepared by our novel method has the highest efficiency.
8048.0 mol/mol) [ Al O –Pd (3074.8 mol/mol) [ c-Al O –Pd
(2155.5 mol/
2
3
2 3
2
3
Conclusions
In this study, we demonstrated that BT can significantly suppress Pd migration and
aggregation during the preparation of the catalyst because of its chelation and
stabilization effects. The Pd particles were reduced in size and well dispersed in the
as-prepared heterogeneous Al O –Pd compared with the catalyst produced without
2
3
use of BT. Therefore, Al O –Pd had high activity and excellent reusability in the
3
2
catalytic hydrogenation of acrylic acid.
Acknowledgments The authors would like to acknowledge the financial support provided by
foundation for the author of National High Technology R&D Program (2011AA06A108) and the
National Natural Science Foundation of China (21176161). Special thanks to the Test Center of Sichuan
University.
References
1
2
3
4
5
6
7
8
. A.A. Vedyagin, A.M. Volodin, V.O. Stoyanovskii, Appl. Catal. B 103, 397 (2011)
. L. Zhang, M. Xue, Y.C. Cui, J. Appl. Polym. Sci. 115, 2523 (2010)
. Z.B. Wu, Z.Y. Sheng, Y. Liu, J. Hazard. Mater. 164, 542 (2009)
. R.K. Edvinsson, A.M. Holmgren, S. Irandoust, Ind. Eng. Chem. Res. 34, 92 (1995)
. H. Wu, L.M. Zhuo, Q. He, Appl. Catal. A 366, 44 (2009)
. A. Marinetti, L. Ricard, Organomet. Chem. 13, 3956 (1994)
. Q.L. Zhang, I.G.D. Lana, Chem. Eng. Sci. 52, 4187 (1997)
. K.G. Georgogianni, A.K. Katsoulidis, P.J. Pomonis, G. Manos, M.G. Kontominas, Fuel Process.
Technol. 90, 1016 (2009)
9. X. Liu, J.G. Khinast, B.J. Glasser, Chem. Eng. Sci. 63, 4517 (2008)
10. P. Handa, K. Wikander, K. Holmberg, Microporous Mesoporous Mater. 117, 170 (2009)
11. S.J. Kowalski, Chem. Eng. Sci. 55, 1289 (2000)
1
2. M. Houmard, D. Riassetto, F. Roussel, A. Bourgeois, G. Berthom e´ , J.C. Joud, M. Langlet, Surf. Sci.
6
02, 3364 (2008)
3. N.E. Bogdanchikova, S. Fuentes, M. Avalos-Borja, M.H. Far ´ı as, A. Boronin, G. D ´ı az, Appl. Catal. B
7, 221 (1998)
1
1
1
1
1
1
1
1
2
2
2
4. P.J. Hernes, J.I. Hedges, Geochim. Cosmochim. Acta 68, 1293 (2004)
5. P. Schofield, D.M. Mbugua, A.N. Pell, Anim. Feed. Sci. Tech. 91, 21 (2001)
6. E. Haslam, Phytochemistry 68, 2713 (2007)
7. M.R. Othman, I.S. Sahadan, Microporous Mesoporous Mater. 91, 145 (2006)
8. C. Agrafiotis, A. Tsetsekou, J. Eur. Ceram. Soc. 22, 423 (2002)
9. C.G.V. Burgess, D.H. Everett, S. Nuttall, Pure Appl. Chem. 61, 1845 (1989)
0. J.P. den Breejen, J.R.A. Sietsma, H. Friedrich, J.H. Bitter, K.P. de Jong, J. Catal. 270, 146 (2010)
1. F. Maillard, M. Martin, F. Gloaguen, J.M. L e´ ger, Electrochim. Acta 47, 3431 (2002)
2. R. van Hardeveld, F. Hartog, Adv. Catal. 22, 75 (1972)
1
23