Sol-gel synthesis of ternary phosphate-ZrO2-SiO2 catalysts for alcohol dehydration

Article abstract of DOI:10.1139/v01-108

Phosphate-ZrO₂-Si₂O catalysts were synthesized by sol-gel method using tributyl phosphite, zirconium propoxide, and tetraethyl orthosilicate as precursors. They were characterized by N₂ adsorption, ³¹P CP MAS NMR, and DRIFTS. At lower P content, monomeric phosphates were formed on the surface of the catalysts, which were mainly responsible for the isopropanol dehydration activity. At higher P content, polyphosphates were formed, and thus, the dehydration activity decreased. An optimum P content for dehydration activity was found to be at 10 mol%.

Full text of DOI:10.1139/v01-108

1224
Sol-gel synthesis of ternary phosphate–ZrO2–SiO2
catalysts for alcohol dehydration
Quan Zhuang and Jack M. Miller
Abstract: Phosphate–ZrO2–Si2O catalysts were synthesized by sol-gel method using tributyl phosphite, zirconium
propoxide, and tetraethyl orthosilicate as precursors. They were characterized by N2 adsorption, 31P CP MAS NMR,
and DRIFTS. At lower P content, monomeric phosphates were formed on the surface of the catalysts, which were
mainly responsible for the isopropanol dehydration activity. At higher P content, polyphosphates were formed, and
thus, the dehydration activity decreased. An optimum P content for dehydration activity was found to be at 10 mol%.
Key words: sol-gel synthesis, ternary oxides, phosphate, acid catalyst, alcohol dehydration, 31P CP MAS NMR, N2 ad-
sorption, DRIFTS.
Résumé : On a synthétisé des catalyseurs phosphate–ZrO2–SiO2 par la méthode sol-gel utilisant du phosphite de trié-
thyle, du propylate de zirconium et de l’orthosilicate de tétraéthyle comme précurseurs. On les a caractérisés par ad-
sorption de N2, RMN CP MAS du 31P et DRIFTS. À des teneurs faibles en P, il se forme des phosphates monomères à
la surface des catalyseurs qui se sont avérés responsables pour activité de déshydratation de l’isopropanol. À des te-
neurs en P plus élevées, il se forme des polyphosphates et il en résulte une diminution de l’activité de déshydratation.
On a déterminé que la teneur optimale en P pour l’activité de déshydratation est de 10 mol%.
Mots clés : synthèse sol-gel, oxydes ternaires, phosphate, catalyseur acide, déshydratation des alcools, RMN CP MAS
du 31P, adsorption du N2, DRIFTS.
[Traduit par la Rédaction] Zhuang and Miller 1228
Introduction
method can be used to prepare both catalyst and catalyst
support (9).
Recently, phosphorus-containing oxides have been synthe-
sized by various methods to obtain inorganic materials with
new porous structures and (or) new properties. Among them,
SAPO and AIPO types have been reported as having uni-
form porous structure, showing molecular sieve function, as
well as catalytic activity, for some reactions due to the sur-
face acidity (1, 2). Meanwhile, supported phosphoric acid
catalysts prepared by impregnation have been widely used as
catalysts for catalyzing olefin hydration (3), alkylation (4),
dehydration (5–7), and dimerization (4).
In this paper, we report our attempt to prepare the ternary
phosphate–zirconia–silica catalysts by the sol-gel synthesis
method using tributyl phosphite, zirconium(IV) propoxide,
and tetraethyl orthosilicate as precursors. Its catalytic perfor-
mance was evaluated for isopropanol dehydration, and N2
adsorption, DRIFTS, and 31P CP MAS NMR were employed
to characterize the catalysts.
Experimental
On the other hand, there have been extensive studies on
sol-gel synthesis of oxides, because it may provide specific
properties to the oxide that cannot be obtained by the con-
ventional coprecipitation preparation (8). Under appropriate
synthesis conditions, such as metal alkoxide concentration,
pH, the type of complexing agent, the amount of hydrolyz-
ing water, drying technique, the textural properties of the ox-
ide can be controlled. Since the advent of this method, it has
stimulated more and more interest in its application in the
field of catalytic materials research. The sol-gel synthesis
Sol-gel synthesis of the catalysts
n-Butanol was obtained from Caledon Laboratories, all
other chemicals were obtained from Aldrich and used as re-
ceived.
A series of P–ZrO2–SiO2 (Zr:Si molar ratio was fixed at
75:25) catalysts with P contents of up to 25 mol% were syn-
thesized by the sol-gel method. Briefly, tributyl phosphite
(density 0.925 g mL–1), zirconium propoxide (70%, density
1.044 g mL–1), and tetraethyl orthosilicate (98%, density
0.93 g mL–1) were used as the precursors and 2,4-
pentanedione (H-acac) as the complexing agent (8).2 The ap-
propriate amounts of zirconium propoxide and tetraethyl
orthosilicate were desolved in the solvent, n-butanol (density
0.81 g mL–1). The solution was heated to 60°C under stirring
to mix the components thoroughly, then it was cooled down
to room temperature. Tributyl phosphite and H-acac were
added and the solution was stirred for 30 min. Then the
alkoxides in the solution were hydrolyzed over night by add-
ing water. A yellowish transparent gel was obtained, which
Received July 20, 2000. Published on the NRC Research
Press Web site at http://canjchem.nrc.ca on August 22, 2001.
Q. Zhuang and J.M. Miller.1 Department of Chemistry,
Brock University, St. Catharines, ON L2S 3A1, Canada.
1Corresponding author (telephone: (905) 688-5550 ext. 3789;
e-mail: jmiller@brocku.ca).
2Q. Zhuang and J.M. Miller. Submitted.
Can. J. Chem. 79: 1224–1228 (2001)
© 2001 NRC Canada
DOI: 10.1139/cjc-79-8-1224

Zhuang and Miller
Table 1. Reagent amounts (g) used in the catalyst P–ZrO2–SiO2 preparation.
1225
P mol%
0
Reagent
5
10
15
20
12.00
67.31
10.19
25
15.00
63.11
9.56
Tributyl phosphite
Zirconium propoxide
Teraethyl
0.00
84.14
12.74
3.00
79.94
12.10
6.00
9.00
75.73
11.47
71.52
10.83
orthosilicate
n-Butanol
2,4-Pentanedione
H2O
166.63
12.00
47.50
167.82
12.00
47.50
169.04
12.00
47.50
170.60
12.00
47.50
171.43
12.00
47.50
172.61
12.00
47.50
tral window of 36 kHz. Prior to recording the spectra, the
samples were degassed at 100°C for 30 min in a flow of he-
lium and immediately transferred into the zirconia rotors for
the NMR measurements.
Fig. 1. N2 adsorption isotherms of the catalysts (concentrations
based on synthesis).
DRIFTS spectra were acquired using a Spectra Tech
DRIFTS accessory (The Collector) in an ATI Mattson RS-1
FT-IR spectrometer equipped with a Michelson interferome-
ter, a helium–neon laser, a KBr beam splitter, a DTGS
(deuterated triglycerol sulfate) detector with a spectra range
of 4000–400 cm–1, and a standard high-intensity source.
The diffuse-reflectance FT-IR were recorded at 100°C under
vacuum after evacuating the samples at 100°C for 30 min.
The sample chamber was designed to be used in situ on the
collector DRIFTS accessory.
Isopropanol dehydration evaluation
The catalytic activity evaluation of the catalysts for
isopropanol dehydration was performed using a continuous
flow fixed-bed micro reactor made of Pyrex glass.2 The in-
ner diameter of the reactor was 6 mm, its length was 25 cm.
With each reaction, 200 mg of fresh catalyst was put into the
middle of the reactor. A thermocouple was placed just below
the catalyst bed to monitor the reaction temperature, which
was manipulated by a temperature controller. Before each
reaction, the catalyst was heated to the reaction temperature
under a flow of N2. The reactions were performed at the
temperature range of 150–230°C at 1 atm (1 atm =
101.325 kPa). The reactant, isopropanol, was introduced into
the reactor by a N2 bubbler, which was kept at 23°C (11).2
The flow rate of N2 was 100 mL min–1, giving a weight hour
space velocity (WHSV) of the reactant as 1.8 h–1. The prod-
ucts were analyzed by an on-line PerkinElmer Sigma 3B gas
chromatograph, employing a Porapaq R column, 800/100
mesh, 6 ft × 0.125 in, with a thermal conductivity detector
(TCD). To analyze all of the products in the exit stream from
the reactor, the GC column was heated by programming
from 150 to 220°C for 15 min.2 The activities were stablized
in ca. 20 min. After that time, two consecutive analyses were
carried out, which were within 5%. An average is reported
throughout the Discussion section. The main product was
propylene, with trace amounts of diisopropyl ether as by-
product, with selectivity close to 100%.
was dried at 100°C to remove the solvent and then calcined
at 500°C for 3 h to remove the organics. The total concen-
tration of the alkoxides in n-butanol solution was 0.8 M. The
molar ratio of the complexing agent to the alkoxides was
1:2. The molar ratio of hydrolyzing water to alkoxide was
11:1. Table 1 shows the reagent amounts used in the catalyst
preparation.3
Characterization of the catalysts
N2 adsorption analysis was measured by using a Coulter
SA 3100 instrument at –196°C. Samples were degassed un-
der vacuum at 200°C for 1 h immediately prior to N2 ad-
sorption analysis. From the isotherm, the Brunauer–
Emmett–Tellar (BET) surface area, pore volume, and pore
size distribution were calculated by using the Barret, Joyner,
and Halenda (BJH) method (10).
CP MAS NMR experiments were carried out on a Bruker
Avance DPX 300 multinuclear FT NMR instrument. A stan-
dard bore Bruker MAS/CP MAS probe with 4 mm zirconia
rotors was used. 31P NMR spectra were obtained at
121.50 MHz using cross polarization with proton decoupling
during acquisition and referenced to (NH4)3PO4 (chemical
shift = 0 ppm). Samples were spun at 8 kHz. A 1 ms cross-
polarization contact time was used with a 4 s recycle delay
between the pulses. 124 FIDs were accumulated with a spec-
Results and discussion
Figure 1 shows the N2 adsorption isotherms of the cata-
lysts. The catalysts display type IV isotherms (10). By add-
3 The mol% of phosphorous quoted throughout is based on the synthesis concentrations, not analysis of the calcined catalysts.
© 2001 NRC Canada

1226
Can. J. Chem. Vol. 79, 2001
Fig. 2. Pore size distributions of the catalysts (concentrations
Table 2. 31P CP MAS data of the catalyst P–ZrO2–SiO2
based on synthesis).
(P 10 mol%) after calcining at different temperatures.
T (°C)
31P chemical shift (ppm)
100
233
366
500
–4.1
–5.4
–8.2
–11.5
Fig. 4. 31P CP MAS NMR spectra of the catalysts after calcining
at 500°C, P content: 5, 10, 15, 20, and 25 mol% (upward), with
concentrations based on synthesis.
Fig. 3. 31P CP MAS NMR spectra of the catalyst, P 10 mol%,
after treating at 100, 233, 366, and 500°C, respectively (down-
ward).
Table 3. 31P CP MAS data of the catalysts P–ZrO2–SiO2
with different P content after calcining at 500°C.
P mol%
31P chemical shift (ppm)
5
10
15
20
25
–10.4
–11.5
–12.0
–14.7
–14.7
Note: P concentrations are based on synthesis, not analysis.
ing phosphorus in the form of tributyl phosphite into the
sol-gel synthesis solutions, the N2 adsorption of the resultant
catalysts increases, especially at the intermediate and high
relative pressure range. The pore size distributions of the
catalysts are displayed in Fig. 2. It can be seen that the tex-
tural structure of the catalysts are mesoporous. As the
amount of P is increased in the synthesis solution, the pore
size distribution broadens to the larger pore diameter side,
and the total pore volume of the catalysts increases. From P
0.0 to 25 mol%, the surface area increases from 180 to
250 m2 g–1 and the pore volume from 0.10 to 0.14 cm3 g–1.
To evaluate the evolution process of the ternary P–ZrO2–
SiO2 catalysts, 31P CP MAS NMR spectra of the catalyst
with P 10 mol% calcined at different temperatures were ac-
quired and the results are shown in Table 2 and Fig. 3. As
the calcining temperature is increased, the chemical shift of
31P NMR resonance shifts up field, from –4.1 ppm after be-
ing heat treated at 100°C to –11.5 ppm with the sample
treated at 500°C. The intensities of the 31P NMR peaks de-
crease when increasing the temperature of heat treatment.
The 31P NMR resonance peaks at the above chemical shift
ranges are due to the monomeric phosphates (2, 12). It is
well-known that P(III) can be easily oxidized to P(V) (13).
Thus in our system, after heat treating the gel in air, the
P(III) from tributyl phosphites is oxidized to P(V) in the
form of phosphates.
Figure 4 and Table 3 show the 31P NMR spectra of the
catalysts with different P contents after calcining at 500°C.
It can be seen that when the P content in the catalyst is in-
creased, the intensity of the 31P resonance increases, with an
© 2001 NRC Canada

Zhuang and Miller
1227
Fig. 6. The catalytic activities of the catalysts with different P
Fig. 5. DRIFTS spectra of the catalyst after calcining at different
temperatures (P 10 mol%).
content for isopropanol dehydration (210°C, 1 atm, WHSV =
1.8 h–1) (concentration based on synthesis).
Fig. 7. The activity (as conversion) of P–ZrO2–SiO2, P 0.0 and
10 mol% vs. reaction temperature.
upfield shift (from –10.4 ppm with P 5 mol% to –14.7 ppm
with P 25 mol%). These changes are caused by partial for-
mation of polyphosphates when the P contents in the cata-
lysts are increased.
The DRIFT spectra of the catalyst with P 10 mol% after
calcining at different temperatures were measured, and the
results are shown in Fig. 5 together with the DRIFT spectra
of ZrO2:SiO2 (75:25) calcined at 233 and 500°C, respec-
tively. The absorption peaks at ~980 cm–1 are assigned to
Si—OH stretching and Zr-O-Si modes of ZrO2–SiO2 (14).
Nogami (15) reported that the peaks at 500–600 cm–1 are
due to the polyhedra ZrOx. It can be seen that in our cata-
lysts these ZrOx polyhedra exist. The absorption at
~3730 cm–1 is caused by the vibration of Si—OH. The peaks
at ~3500 cm–1 are ascribed to the hydrogen-bonded hydroxyl
band (16, 17). As shown in the 31P CP MAS NMR results,
after calcining, the phosphorus species in the gel gradually
change to the phosphate, and hence, the new absorption
peaks at ~3650 cm–1 could be reasonably assigned to the OH
groups affiliated to the phosphates, which are acidic. The
peaks at 1540 and 1445 cm–1 appear after increasing the cal-
cining temperature to 233°C. On further increasing the cal-
cining temperature, they disappear. Since we could not
observe these peaks with pure ZrO2–SiO2 system, they must
be related with some specific intermediate organic phospho-
rous species being formed, which then migrate to the surface
and decompose during the evolution of the structure of the
final catalysts on heat treatment at higher temperature (18).
The phosphorus-containing gel is yellow, whereas ZrO2–
SiO2 is nearly white and tributyl phosphite is colorless, indi-
cating that at the gel stage the two materials have different
structures. This can be seen from their DRIFTs spectra after
calcining at 233°C. On heat treating at higher temperature
(500°C) the frameworks of the two materials are similar, be-
cause they show similar DRIFTs spectra. The phosphates
may have only weak interaction with the framework of
ZrO2–SiO2.
Figure 6 shows the catalytic activities of the sol-gel syn-
thesized catalysts with different P content for isopropanol
dehydration. As the P content is increased, the activity in-
creases first, then decreases, and then levels off, giving an
optimum content of P at 10 mol%. It is clear that the dehy-
dration activity of the catalysts are due to the acidity of the
phosphates on the surface (3). As the P content in the cata-
lysts is increased, some of the monomeric phosphates
change to polyphosphates as evidenced by the results of 31P
NMR characterization. This may explain the decline of the
activity of the catalysts with higher P content.
Figure 7 shows the activities of the two catalysts without
P and with 10 mol% of P, respectively, under different tem-
peratures. Under the tested reaction temperature range, the
activity of the P containing catalyst is higher than the cata-
lyst without P.
Conclusions
By sol-gel method, phosphate–ZrO2–SiO2 catalysts were
synthesized. The 31P CP MAS NMR spectra revealed that at
lower P content, monomeric phosphates were formed on the
© 2001 NRC Canada

1228
Can. J. Chem. Vol. 79, 2001
9. M. Schneider and A. Biker. Catal. Rev. Sci. Eng. 37, 515
(1995).
10. S.J. Gregg and K.S.W. Sing. Adsorption, surface area and po-
rosity. Academic Press, London. 1982.
11. S. Srinivasan, C.R. Narayanan, A. Biaglow, R. Gorte, and A.K.
Datye. Appl. Catal. A, 132, 271 (1995).
12. E.C. DeCario, J.C. Edwards, T.R. Scalzo, D.A. Storm, and
J.W. Bruno. J. Catal. 132, 498 (1991).
13. F.A. Cotton, G. Wilkinson, C.A. Murillo, and M. Bochmann.
Advanced inorganic chemistry. 6th ed. John Wiley & Sons,
New York. 1999.
14. D.M. Pickup, G. Mountjoy, G.W. Wallidge, R.J. Newport, and
M.E. Smith. Phys. Chem. Chem. Phys. 1, 2527 (1999).
15. M. Nogami. J. Non-Cryst. Solids, 69, 415 (1985).
16. S.W. Lee and R.A. Condrate, Sr. J. Mater. Sci. 23, 2951
(1988).
17. M. Andriarainarivelo, R. Corriu, D. Leclercq, P.H. Mutin, and
A. Vioux. J. Mater. Chem. 6, 1665 (1996).
18. K. Nakamoto. Infrared and Raman spectra of inorganic and co-
ordination compounds. 5th ed. John Wiley & Sons, New York.
1997.
surface of the catalysts, which were mainly responsible for
the isopropanol dehydration activity. When P content was
increased, polyphosphates were formed, causing the de-
crease of the activity. The optimum P content was 10 mol%.
References
1. Q. Gao, B.M. Weckhuysen, and R.A. Schoonheydt. Micropor.
Mesopor. Mat. 27, 75 (1999).
2. J. Quartararo, M. Guelton, M. Rigole, J. Amoureux, C.
Fernandez, and J. Grimblot. J. Mater. Chem. 9, 2637 (1999)
3. C.M. Fougret, M.P. Atkins, and W.F. Holderich. Appl. Catal.
A, 181, 145 (1999).
4. K. Tanabe and W.F. Holderich. Appl. Catal. A, 181, 399
(1999).
5. J.N. Armor. Appl. Catal. A, 78, 141 (1991).
6. W.F. Holderich. Stud. Surf. Sci. Catal. 75, 127 (1993).
7. Y. Kita, K. Kishino, and K. Nakagawa. Nippon Kagaku Kaishi,
3, 269 (1996).
8. J.M. Miller and L.J. Lakshmi. J. Phys. Chem. B, 102, 6465
(1998).
© 2001 NRC Canada