Relation between the hydroxylation state of zirconia, the sulfate promotion method, and the catalytic activity of SO42--ZrO2 catalysts

Article abstract of DOI:10.1006/jcat.1997.1512

Zr(OH)₄ samples were prepared from ZrOCl₂ and ZrCl₄; they were calcined at temperatures between 100 and 620°C and were sulfated with H₂SO₄ solutions. The catalytic activity in n-butane isomerization o

Full text of DOI:10.1006/jcat.1997.1512

JOURNAL OF CATALYSIS 165, 254–262 (1997)
ARTICLE NO. CA971512
Relation between the Hydroxylation State of Zirconia,
the Sulfate Promotion Method, and the Catalytic
Activity of SO24ꢀ ZrO2 Catalysts
–
ꢁ
Carlos R. Vera and Jose´ M. Parera
Instituto de Investigaciones en Cata´lisis y Petroqu´ımica, INCAPE-(FIQ-UNL, CONICET), Santiago del Estero 2654, 3000 Santa Fe, Argentina
Received May 3, 1996; revised October 18, 1996; accepted November 4, 1996
A thorough study of the effects of the calcination of zirco-
nia gels previous to the sulfation with H2SO4 in solution was
undertaken. It comprised the monitoring of common prop-
erties (pore volume, specific surface area, crystalline struc-
ture, crystallite size, sulfate content, and catalytic activity)
and other ones not previously studied (surface OH content
and surface charge in solution) in their evolution with the
calcination temperature. Surface OH content and reactiv-
ity were probed by chlorination, while the surface charge
was measured by “fast” potentiometric titration. The lat-
ter special focus on the surface chemical changes permitted
monitoring of the dehydroxylation/oxolation surface pro-
cess taking place along with the calcination. The surface
oxolation appears irreversible and a chemical inertness at
room temperature was observed.
In order to study the conditions for the attack of the
surface oxo groups, both chlorination and sulfation under
drastic conditions were attempted. In the last case, SO4H2
in the vapor state as the sulfating agent and high tempera-
tures were used. The method, however aggressive, proved
effective in promoting activity in crystalline monoclinic zir-
conia samples. It has a priori one attractive feature: it allows
the resulfation without unloading the catalyst. Neverthe-
less, the conditions must still be tempered and optimized in
order to prevent the deterioration of the porous structure
of the catalyst.
Zr(OH)4 samples were prepared from ZrOCl2 and ZrCl4; they
were calcined at temperatures between 100 and 620ꢂC and were
sulfated with H2SO4 solutions. The catalytic activity in n-butane iso-
merization ofthesesamples(at 300ꢂC, 1atm, and WHSV = 2.5hꢀ1),
previously activated in air at 620ꢂC is not well correlated with their
textural properties or the SO24ꢀ concentration.
Zr(OH)4 samples calcined at temperatures higher than 375ꢂC
are crystalline and adsorb SO24ꢀ groups, but which are not active
catalytically. The catalytic activity decreases with the calcination
temperature in a similar way as the OH groups (determined by
chlorination) are eliminated, forming less reactive surface groups.
When the crystalline samples are sulfated with H2SO4 vapors, cata-
lyticallyactivematerialsareproduced, irrespectiveofthecrystalline
state.
The drastic sulfation overcomes the inertness of the crystalline
surface and catalytic active sites for n-butane isomerization are pro-
duced. This in situ resulfation may also compensate the sulfur losses
c
during the reaction.
ꢃ 1997 Academic Press
INTRODUCTION
Two statements are repeatedly found in the literature
of oxoanion promoted ZrO2 catalysts for acid catalyzed
hydrocarbon reactions, concerning its crystalline structure
and the catalytic activity: (i) The promotion (with SO42ꢀ
or other dopants, e.g., WO3) must be performed on amor-
phous zirconium hydroxyde; promoted crystalline samples
are inactive (1). (ii) For SO24ꢀ ZrO2, the tetragonal struc-
–
METHODS
ture is the dominant crystalline phase present in catalyti-
cally active materials (2). Apart from these and related to
the sulfate content of the catalysts, (iii) there is a growing
Different zirconia samples were prepared using differ-
ent precursors and following different synthesis routes. A
schematic description can be found in Table 1.
concern over the sulfur losses of SO24ꢀ ZrO2 catalysts oc-
–
curring during the activation and reaction, which may ham-
per the long-range activity of the catalyst. Statements (ii)
and (iii) are reexamined in this work and evidence for its
refutation is provided. A practical solution for (iii) is also
devised.
Zirconia Gels
They were prepared by hydrolysis in water of
ZrOCl2 ꢄ 8H2O (Strem, 99.99% ) (sample ZO1) and ZrCl4
(Aldrich, 99.9% ) (sample ZCl), followed by precipitation
with NH4OH (Merck, 37% ) up to pH 10. The gels were
then washed and dried overnight. A detailed description
ꢁ
To whom all correspondence should be addressed. E-mail:
parera@fiqus.unl.edu.ar.
254
0021-9517/97 $25.00
c
Copyright ꢃ 1997 by Academic Press
All rights of reproduction in any form reserved.

SO24ꢀ ZrO2 CATALYSTS
–
255
TABLE 1
Description of Samples and Analysis Performed
Sample
ZO1
Preparation
Analysis
XRD, textural
Zr(OH)4 from ZrOCl2 ꢄ 8H2O calcined at T = 100–
200–300–400–620ꢂC for 3 h
ZCl
Zr(OH)4 from ZrCl4 calcined at T = 100–250–375–
XRD, textural
500–600ꢂC for 3 h
ZO1-S
ZCl-S
ZO1-S-CT
Samples ZO1, sulfated in 1 N H2SO4 for 2 h
Samples ZCl, sulfated in 1 N H2SO4 for 1 h
Samples ZO1-S, after the catalytic test
SO24ꢀ content, catalytic test
SO24ꢀ content, catalytic test
SO24ꢀ content, textural
ZO2-Cl
Zr(OH)4 from ZrOCl2 ꢄ 8H2O calcined at T =
100–200–300–400–500–600ꢂC for 3 h,
chlorinated with SOCl2 (liq), 25ꢂC
Cl content
ZB-Cl
Zr(OH)4 from Zr(n-But-O)4, Cl free, calcined at T =
100–250–400–500–600ꢂC for 3 h, chlorinated
with SOCl2 (liq), 25ꢂC
Cl content
ZB-RCl
Idem ZB, chlorinated under reflux with SOCl2
Cl content
ZO3600-VPC
ZrO2 from ZrOCl2 ꢄ 8H2O calcined at 600ꢂC 3 h, treated
with CCl4 (vap) at 300–375–450–525–600ꢂC.
Idem ZO3, treated with H2SO4 vapors (instead of
CCl4) at different temperatures
XRD, Cl content
ZO4600-VPS
catalytic test, SO24ꢀ content
Note. In every case, the temperature of calcination of the gel is included in the sample name as a superscript, e.g.,
ZCl600. Idem for chlorination and sulfation temperatures (VPC, VPS).
of the method can be found elsewhere, either for samples used. The reactor had a diameter of 1 cm approximately and
ZO1 (3) or ZCl (4). Portions of the gels were calcined in a a length of 45 cm. A 35–80 mesh fraction of the catalyst was
muffle at different temperatures (see Table 1) for 3 h.
loaded for the reaction (0.5 g), and supported over a porous
Samples ZO1 and ZCl were analyzed by XRD in order to glass disk in the middle of the reactor. The effluents of the
determine the crystal structure and the crystal size. The ra- reactor were analyzed in a gas chromatograph connected
diation used was CuKꢀ filtered with Ni. The crystal size was “on line.” The chromatographic column was dimethylsul-
determined with Scherrer’s formula, with Lorentz correc- fonale supported on Chromosorb P, packed in a 6-m long,
tion factor and without corrections for network tensions. 1/8-in. external diameter steel tube. The isomerization of
Textural properties were determined in a Quantachrome n-butane comprised the following steps: (i) activation at
Nova-1000 sortometer, by adsorption of nitrogen at 77.4 K. 620ꢂC in air flow (10–12 ml/min) for 3 h and (ii) a reaction
The specific surface area (Sg) was determined by the BET at 300ꢂC for 3–4 h with an 8 ml/min flow of pure n-butane
method using four points of the adsorption branch of the (Matheson 99.99% ) and WHSV = 2.5 hꢀ1. The sulfate con-
isotherm, at p/p0 values of 0.10, 0.15, 0.20, and 0.25. The tent and textural properties were determined before and
pore volume was determined by nitrogen adsorption at after the catalytic test.
p/p0 = 0.99. The pore distribution was determined by the
Assessment of OH Groups in Precalcined Zirconia Gels
BJH method using the desorption branch of the nitrogen
isotherm.
Experiments of chlorination of samples ZO2 and ZB
were conducted in order to assess the quantity and reac-
tivity of the surface hydroxyls on the samples calcined at
Precalcined Zirconia Gels Sulfated in Solution
Samples ZO1-S and ZCl-S were generated by sulfation different temperatures. Many methods of characterization
of samples ZO1 and ZCl, after calcining at different tem- of surface groups have been used in the past. They com-
peratures (see Table 1). After the calcination and before prise basically two types: spectroscopic (IR, NMR) and
sulfation the samples were allowed to rehydrate overnight chemical. The latter is used extensively for quantitative
in contact with room humidity. The method of sulfation con- determinations. Chlorination, etherification, fluorination,
sisted in the dipping of the samples in 1 N H2SO4 for 1 or 2 h and reactions with organometallic compounds are the most
(samples ZCl-S or ZO1-S, respectively), without agitation. common (5). Chlorination reactions are adopted in this
The sulfated samples were subjected to the catalytic test work, for convenience. The determination of surface Cl in
of isomerization of n-butane. A quartz reactor operating catalysts is reliable and reproducible, and the stoichiometry
under isothermal conditions and atmospheric pressure was of reaction with the hydroxyl groups is expected to be 1 : 1,

256
VERA AND PARERA
a fact not found with other polydentate ligands (notably, sidered that no surface chlorine was present in these cal-
SO24ꢀ).
The reacting agent used, thionyl chloride (SOCl2), is one
cined samples.
The equipment for potentiometric titrations consisted of
of the most commonly used, because it decomposes giving an Altronix SX-II pH-meter coupled to a glass electrode.
only gaseous products:
Acid and base titrating solutions HNO3 and KOH were
added with a microburet to a 0.1 N KNO3 electrolyte solu-
tion containing the sample. The titrations were conducted
in an inert atmosphere of nitrogen. This gas was also bub-
bled in the solution prior to the titration in order to desorb
CO2. The surface charge was determined by substracting
the values of added base or acid between the titration of
the sample and the blank experiment (electrolyte solution
without the solid). More details can be found in Ref. (7).
The surface charge was calculated with the expression:
M-OH + SOCl2 → M-Cl + SO2 ↑ + HCl ↑
(M = surface cation).
The HCl adsorbed on the support is removed by thor-
ough degassing. After the chlorination treatment all sam-
ples were heated in vacuo at 100–200ꢂC for 1 h to remove
the adsorbed products. Two kinds of samples were used in
the chlorination experiments. Samples ZO2 were prepared
from ZrOCl2 ꢄ 8H2O in the same way as samples ZO1. In or-
der to avoid the interference of chloride ions retained in the
matrix of the gel, samples ZB were prepared by hydrolysis
of zirconium n-butoxide (Strem, stabilized solution, 70% in
n-butanol). The butoxide was first dissolved in an organic
solvent (n-butanol or hexane) for 2 h under agitation. Then
the dissolved alcoxide was hydrolyzed by adding distilled
water in excess (water/alcoxide > 20–30). The gel was then
filtered and dried overnight. Samples ZO2 and ZB were
first calcined at different temperatures (see Table 1) in a
muffle for 3 h.
After this treatment the samples were allowed to re-
hydrate in contact with the ambient atmosphere. Then
they were chlorinated at room temperature by means of a
solution of SOCl2 dissolved in benzene (1 : 4). The samples
were immersed for 72 h. Then they were filtered, washed
with hexane and dried at 110ꢂC. The samples were always
handled and stored in humidity-free vessels, since the Cl
attached to the surface could be hydrolized. Finally, the
samples were degassed and their specific surface area was
measured. The samples calcined at 100ꢂC were degassed at
100ꢂC for 2 h, and the samples calcined at greater temper-
atures were degassed at 200ꢂC for 1 h.
ꢀ = FV (Cb ꢀ Ca ꢀ (OHꢀ) ꢀ (H+))/S,
where F = Faraday’s constant, V = volume, S = surface
of the sample, Cb = concentration of added base; Ca =
concentration of added acid; (OHꢀ), (H+) = equilibrium
concentration of OHꢀ and H+ ions.
Vapor Phase Treatments
Portionsofthe dried gel were calcined at 600ꢂC for 3h in a
muffle furnace (samples ZO3600 and ZO4600). XRD analysis
was performed in order to determine the crystal structure.
To test the reactivity of the surface, chemical attacks under
aggressive conditionswere used:chlorination with CCl4 and
sulfation with H2SO4. It has been reported (5) that CCl4 at
150–350ꢂC produces an incomplete chlorination (25% ) of
surface groups on silica; however, at 400ꢂC siloxane bridges
are opened; and a maximum chlorination is obtained at
600ꢂC. Mc Daniel (7) reports that chlorination of silica with
CCl4 begins at 300ꢂC and that siloxane bridges are opened
at T > 500ꢂC. The reaction is
M-OH + CCl4 → M-Cl + HCl ↑ + COCl2 ↑ .
Chlorination with thionyl chloride at room temperature The temperatures of chlorination used were 300, 375, 450,
is found to produce a replacement of a fraction of all the 525, and 600ꢂC. Pure CCl4 (Mallinckrodt, 99% ) at a rate of
available hydroxyl groups, while full replacement is said to 3.9 ml/h was injected over a glass wool plug at the inlet of a
be accomplished under reflux conditions (5). Such a treat- quartz reactor. A flow of nitrogen (30 ml/min) carried the
ment was also applied to samples ZB, reflux heating the reactive, which vaporized rapidly in the first inlet portions
solution (70–80ꢂC) for 8 h. The method for the determi- of the reactor; 1-g samples were supported over a fritted
nation of chlorine is thoroughly depicted in Ref. (6). The glass slab in the middle of the reactor. The chlorination was
Cl content was considered equal to the content of surface conducted for 40 min; no reactive was injected during the
reacting OH groups.
cooling and heating ramps; and a flushing with nitrogen for
To assess the influence of the calcination on the adsorp- 15–20 min was performed after the chlorination was fin-
tive properties in solution, potentiometric titrations were ished. The samples were then dried and the specific surface
performed on dried and calcined gels. Since retained an- area was measured.
ions affect the adsorption behavior of surfaces, a gel dried
Samples ZO3600 were also sulfated in the vapor phase.
at 100ꢂC prepared from Zr n-butoxide (ZB100) was taken For this purpose a solution of H2SO4 5 N was injected in a
as the reference for the uncalcined samples. In the case of nitrogen stream in the same way described above (3.9 ml/h,
calcined samples, gels heat treated at 600ꢂC prepared from 40 min). The temperatures of sulfation were 300, 375, and
ZrOCl2 (ZO1600) and ZrCl4 (ZCl600) were used. It was con- 400ꢂC; 1-g samples were also used. The sulfated samples

SO24ꢀ ZrO2 CATALYSTS
257
–
TABLE 3
Crystal and Textural Properties of Samples ZO1
DpW
(Angstrom)
Vp
(ꢁl/g)
Sample
Sg (m2/g)
Structure
ZO1100
ZO1250
ZO1375
ZO1500
ZO1600
262.13
235.44
120.14
66.50
22.711
25.145
46.722
74.850
102.67
298
296
281
249
203
Amorphous
Amorphous
T,M
T+.Mꢀ
39.48
M
Note. Notation as in Table 2.
FIG. 1. A dimensional initial conversion (C*) as a function of calcina-
tion temperature: C* = (initial activity/maximum initial activity in a series);
Tcg = temperature of calcination of the gel (before the sulfation); samples
ZO1-S, ; samples ZCl-S, ᭹.
The pore volume diminishes 30% in the case of samples
ZO1 and 50% for samples ZCl. For both samples the spe-
cific surface area is reduced from 260 m2/g to 30–40 m2/g.
The continuous drop in these values is more drastic at tem-
peratures higher than 400ꢂC, in coincidence with the pro-
cess of crystal growth.
were examined by XRD and their catalytic activity in the
reaction of isomerization of n-butane was tested.
Typical BJH pore size distributions (samples ZO1) are
plotted in Fig. 2. Zirconias were essentially mesoporous ma-
terials. At low temperatures of calcination, when no crystal-
RESULTS
˚
lization hastaken place a peak ofmicroporesat 25–30A and
General Properties
˚
a zone of mesopores at 100–400 A can be seen. Micropores
sinter slowly and completely disappear at 600ꢂC, contribut-
ing to increase the mesopore volume. Samples ZCl were
Figure 1 shows the adimensional initial conversion as
a function of the calcination temperature. The maximum
values of conversion for samples ZO1-S and ZCl-S were
30% and 20% , respectively. The active samples deactivated
rapidly, conversion being practically nil after 3 h of time-
on-stream. Isobutane was the main product (85–90% selec-
tivity) while propane was the main secondary one. A first
search for a correlation of the activity pattern with vari-
ations in other properties can be done by inspecting the
results for the textural properties of the calcined gels, their
crystallinity, and the sulfate content of the sulfated samples.
As shown in Tables 2 and 3, the samples began to crystal-
lize below 400ꢂC. At 400ꢂC crystals of 100 nm are present,
with a mixture of structures T and M. Higher temperatures
of calcination produce the growth of the crystals and the sta-
bilization of the monoclinic phase after calcining at 600ꢂC.
˚
more microporous, displaying a peak at 25–30 A. Calcina-
tion at 600ꢂC increases mesoporosity, with a wide maximum
TABLE 2
Crystal and Textural Properties of Samples ZCl
DpW
Vp
R1
R2 Rm
Sample Sg (m2/g) (Angstrom) (ꢁl/g) Structure (nm) (nm) (nm)
ZCl100 259.39
ZCl200 280.89
ZCl300 236.44
ZCl400 110.18
—
—
Amorphous
—
—
—
129
—
—
—
—
—
—
29.3
30.9
48.8
149.9
206 Amorphous
183 Amorphous
139
101
T+,Mꢀ
82 100
ZCl600
33.83
M
1570 1250 1410
Note. Sg, specific surface area; DpW, Wheeler’s pore diameter; Vp, pore
volume; T, tetragonal; M, monoclinic; R1, crystal size calculated over the
peak 22 = 28ꢂ; R2, crystal size calculated over the peak 22 = 34ꢂ; RM:
(R1+R2)/2.
FIG. 2. BJH distribution of samples ZO1 (from oxychloride, calcined
at different temperatures).

258
VERA AND PARERA
FIG. 3. (a) Weight% of sulfate as a function of the calcination temperature (Tcg): ZO1-S (from oxychloride), ; samples ZCl-S (from chloride), ᭹.
(b) Fraction of sulfate monolayer as a function of the calcination temperature (Tcg). (Sulfate monolayer = 4 SO24ꢀ groups/nm2.) ZO1-S (from oxychlo-
ride), ; Sampels ZCl-S (from chloride), ᭹.
˚
at 150–200 A. For both samples, the loss of surface is higher onset of the crystallization process and the drop in conver-
than the loss of pore volume, due to the collapse of pores in sion, but there is no correlation with the evolution of other
the micropore range, which contribute largely to the total textural properties. In the case of surface area, while the ac-
area. A fraction of the micropores merges to form meso- tivity drops to practically null values at high temperatures
pores with no loss of volume.
(Tcg > 400ꢂC), Sg does not.
Table 4 shows the textural properties of samples ZO1-
The sulfate contents for samples ZO1-S and ZCl-S are
S-CT. During the catalytic test samples ZO1-S suffer the plotted in Fig. 3. The samples differ in the time of sulfation
effect of sintering during the activation step at 620ꢂC. Cok- (1 and 2 h) and in the precursor used in their synthesis.
ing during the reaction is not expected to reduce the pore They display a common decrease in the mass content as Tcg
volume, since the coke deposit on SO24ꢀ ZrO2 amounts to increases (Fig. 3(a)). The results of surface sulfate density
–
1% approximately (2). Samples calcined at 100 and 250ꢂC are somewhat different (Fig. 3(b)). For both samples, the
previousto sulfation displaya more resistant area under cal- densityislower than the monolayer value, for Tcg lower than
cination at 620ꢂC, displaying final values of 125–137 m2/g; 600ꢂC. At this temperature of calcination, zirconia samples
for these samples the reduction of pore volume is 20–30% . exhibit an enhanced capacity for the adsorption of anions.
Apparently, the sulfate ion acts as textural stabilizer pre- The sulfate surface density is stabilized at lower values after
serving part of the microporous volume.
the treatment at high temperatures during the catalytic test,
For both ZO1and ZClsamples, the constant C in the BET as shown in Fig. 4. A content of about 0.5 monolayers is
equation increased after calcination. It was 50–60 at 100– retained in samples ZO1-S-CT. The patterns of catalytic
200ꢂC and 80–100 at 600ꢂC. This indicates that the heat of activity and sulfate content are not well correlated. Mass
adsorption of nitrogen increases when the surface is treated percentages of sulfate do not decrease to such an extent
at high temperatures, a fact addressed either to the creation
of new adsorption sites or an increase in the surface energy.
For both samples ZO1 and ZCl, the initial crystal habitat
is a mixture of phases T and M. The thermodynamically
stable M phase is stabilized upon calcination at higher tem-
peratures. There seems to be a coincidence between the
TABLE 4
Textural Properties of Samples ZO1-S-CT
Sg (m2/g)
Dp Wheeler (A)
Vp (ꢁl/g)
˚
Sample
ZO1100-S-CT
ZO1250-S-CT
ZO1375-S-CT
ZO1500-S-CT
ZO1600-S-CT
125.36
137.23
114.95
44.01
33.39
33.81
51.29
88.99
102.25
209
232
284
196
195
FIG. 4. Sulfate content ofsamplesZO1-S, prepared from oxychloride,
calcined at different temperatures and sulfated: before catalytic test,
;
38.13
after catalytic test, ZO1-S-CT, ᭡.

SO24ꢀ ZrO2 CATALYSTS
259
–
which could explain the total loss of activity. Furthermore,
TABLE 5
sulfate surface densities can remain (after activation) at a
half monolayer value or more for any Tcg value.
Specific Surface Area of Samples ZO2-Cl and ZB-Cl (after Chlori-
nation) and Cl Content of Samples ZO2 (before Chlorination)
Sg (m2/g)
(after chlorination)
Assessment of Surface OH Groups
Cl content (% )
(before chlorination)
Sample
Zirconia gel has an isoelectric point of 6–7, typical of am-
photerous transition oxides (6, 9). Chlorine free zirconia,
prepared from n-butoxide, has a ZPC of 7.0. Samples pre-
pared from chlorided precursors always display lower val-
ues of ZPC, due to the presence of chloride ions retained in
the gel matrix. Therefore, for the comparison of adsortive
propertiesin solution, a sample prepared from n-butoxide is
taken as reference material of the dried gel state. Besides,
during the calcination at high temperatures chloride ions
are removed nearly completely (9). Remaining ions after
calcination at 600ꢂC are expected to be in the bulk and not
in the surface due to the enhanced surface ionic diffusion
during the sintering processes.
ZO2100-Cl
ZO2200-Cl
ZO2300-Cl
ZO2400-Cl
ZO2500-Cl
ZO2600-Cl
ZB100-Cl
1.42
1.04
0.96
0.68
0.57
0.31
—
—
—
—
—
206.25
209.21
173.94
147.15
120.73
23.22
44.55
55.60
38.00
20.38
62.60
27.84
32.91
20.37
ZB400-Cl
ZB500-Cl
ZB600-Cl
ZB250-RCl
ZB400-RCl
ZB500-RCl
ZB600-RCl
—
—
—
The results of surface charge are plotted in Fig. 5. A first
effect induced by the calcination is clearly the shift of the
ZPC from a value of 7.0 to a value of 4.0. Many authors
have studied independently the relation between the ZPC
and the previous history of the sample and indicate that
treatments leading to bulk or surface dehydration result in
more acid values of ZPC (8). In the zone of anion adsorp-
tion (pH < PZC) there is a growth in the surface charge in
relation to the noncalcined sample, in this case, for values
of pH < 3. An increase in the surface charge indicates a
corresponding increase in the number of adsorption sites,
or a greater polarizability of the surface sites, induced by
the heat treatment. The increased anion adsorption can be
related to the higher contents of sulfate anion previously
recorded.
chlorine content is detailed in Table 5. As we can see the
samples before being chlorinated with SOCl2 have retained
Cl. This ion is fairly persistent. In principle, the Cl ion can
be incorporated in the bulk or retained on the surface. It
has been argued (9) that incorporation in the bulk is not
likely because the ionic radius of Clꢀ is bigger than the ra-
dius of oxygen. Therefore, the higher fraction would be on
the surface. However, the more persistent chlorides, found
after calcination at 500 and 600ꢂC must be present in the
bulk, since they persist after the crystallization process, in
which the ionic surface diffusion must be high. Then, the
surface chlorine of the blanks was estimated by subtract-
ing the content at 500ꢂC, taken as the initial bulk chlorine
content.
Samples calcined at different temperatures were chlori-
nated in order to measure their OH content. The blanks
before chlorination had a water content of 10–13% . The
The quantity of OH replaced by Cl decreases with calci-
nation (Fig. 6) and is remarkably low on the samples cal-
cined at 400, 500, and 600ꢂC. In the case of samples ZO2-Cl
FIG. 5. Potentiometric titration curves of fresh and calcined zirconia
gels: ᭡, zirconia gel prepared from n-butoxide, dried at 100ꢂC (ZB100);
4, zirconia prepared from oxychloride, calcined at 600ꢂC (ZO2600); ꢃ,
zirconia prepared from chloride, calcined at 600ꢂC (ZCl600).
FIG. 6. Cl content (% ) of samples treated with SOCl2 in liquid phase:
, ZO2-Cl; ᭡, ZB-Cl; ᭹, ZB-RCl.

260
VERA AND PARERA
calcined at these temperatures there is practically no Cl sulfated samples was similar to the corresponding nonsul-
grafted. The heat treatment seems to produce an exten- fated samples, and no new peaks related to bulk sulfates
sive dehydroxylation, accompanied by a surface passivation were detected. The samples were mainly monoclinic. Al-
which makes it inert to the attack by SOCl2.
though they were active, the samples deactivated rapidly.
Surface chlorine densities were also calculated and were The sulfate content of the sample sulfated at 300ꢂC was
related to the OH density. For comparison, a rough esti- 1.72% or 2.68 SO24ꢀ/nm2.
mation of the maximum content of OH groups (ꢂMOH ) was
The obtention of mainly monoclinic active samples in a
done by assigning a number equal to the number of surface way resembles some results obtained by Morterra et al. (12).
cations, theoretically estimated to be of 8 Zr/nm2 (6, 10). They synthesized SO42ꢀ ZrO2 from Zr isopropoxide and
–
The maximum experimental OH densities were 40–45% obtained some samples which were mainly monoclinic af-
ꢂMOH (ZO2100-Cl, 200 m2/g), 95% ꢂOMH (ZB100-Cl, 44.5 m2/g) ter the crystallization. Comparing with other samples, they
and 100% ꢂMOH (ZB250-RCl, 62.6 m2/g). The higher surface concluded that the maximum acidity was obtained for crys-
Cl density in ZB may be addressed to a higher surface OH talline materials irrespective of their crystalline state.
content of zirconias prepared from alcoxides (11).
DISCUSSION
Vapor Phase Treatments
The attack with CCl4 vapors indicated that the inertness
of the surface was broken under this more aggressive con-
ditions (Fig. 7). Samples ZO3-VPC, which were crystalline
and of similar nature than previous samples calcined at
600ꢂC (ZO2600), showed a chlorine intake of 4–7% (2.0–
3.0 Cl/nm2, i.e., 25–37% ꢂOMH ). It must be recalled that the
Cl intake in ZO2400-Cl, ZO2500-Cl, and ZO2600-Cl was prac-
tically zero. Plotted values correspond to the chlorine in-
corporated (the value of the blank has been substracted,
0.1% at 600ꢂC). The amount of chlorine retained can be
considered as the resultant of an equilibrium between the
Cl grafted and that removed by the high temperatures.
The maximum Cl surface density is somewhat lower than
the maximum obtained for ZO2-Cl samples.
The results show that there is a strong influence of the
temperature of calcination of the gels before the sulfation
over the catalytic activity of SO24ꢀ ZrO2, when sulfation
–
is accomplished by the common method of impregnation
with aqueous H2SO4 at room temperature. The calcination
temperature of 400ꢂC establishes a sharp limit between ac-
tive and inactive samples. There is no correlation with the
textural properties modified by the calcination, since no
parameter diminishes to nearly null values as the catalytic
activity does. Crystallization seems to be related to the on-
set of catalytic inactivity, although it does not determine
an accurate sufficient condition. Gels calcined at temper-
atures near the limit (375–400ꢂC), crystallized to a certain
extent and still showed some catalytic activity. Neverthe-
less, crystallization does diminish the number of active sites
produced by sulfation.
Variation of the quantity of surface OH groups, as mea-
sured by chlorination with SOCl2, offers a better explana-
tion. Sulfate anions, in order to become anchored to the sur-
face of zirconia and form sulfo bridges, need the presence of
adjacent hydroxyls. Therefore, the depletion of surface OH
groups hampers the possibilities of sulfo bridge formation.
Calcination is thought to produce a progressive oxolation
in the bulk and on the surface; the process that becomes
accelerated and massive in the temperature interval corre-
sponding to the “glow exotherm,” 420–470ꢂC (13), as de-
tected by DTA-DSC experiments. Oxolation on the surface
produces oxo bridges, which must be rather inert species,
in analogy with other M-O-M groups, such as the siloxane
groups in silica. In the latter, oxo bridges are formed re-
versibly at low temperatures of calcination; at higher tem-
peratures, siloxane groups are permanent. This is due to
the need of a reorientation of the coordination tetrahedron
of Si. Such reorientation does not seem necessary in the
case of zirconia, since Zr has octahedral or near octahedral
coordination in the oxide; therefore, the formation of oxo
bridges seems favored. The oxo bridges on zirconia seemed
to be inert and would not rehydrate under mild conditions,
Vapor-phase sulfation in turn, succeeded in producing
active samples in the catalytic test. The results plotted in
Fig. 7 show that the samples have a maximum initial ac-
tivity when they are activated by calcination at 650–700ꢂC
in the catalytic test. Temperatures higher than 700ꢂC must
produce decomposition of surface sulfate, as is the case with
common SO24ꢀ ZrO2 catalysts. The XRD spectrum of the
–
FIG. 7. Cl content of samples ZO3600-VPC (from oxychloride, cal-
cined at 600ꢂC) treated with CCl4 vapors: , Cl% ; , OH density.

SO24ꢀ ZrO2 CATALYSTS
261
–
as was revealed by the chlorination of calcined samples with complete depletion of surface OH groups. The oxo bridges
thionyl chloride. formed are inert and do not react under mild conditions;
The necessity of OH groups for the grafting of sulfate they do not rehydrate in the presence of water and do not
and the formation of acid sites has been previously hinted react with sulfate anions to form sulfo bridges. At temper-
by Chen et al. (14), although no surface reactions were de- atures of calcination near but lower than 400ꢂC the content
picted. For zirconia sulfated in the form of Zr(OH)4, they of OH groups is low but enough to produce a minimum
posed a two-step process in which the sulfate first displaces amount of active sulfates, which contribute substantially to
two OH groups of the framework and then upon calcination the conversion of n-butane. Samples with a small fraction
at T > 400ꢂC acid sites are formed upon oxolation.
of sulfate monolayer are found to be as much active as
The presence of oxo bridges on the surface of zirconia samples with a half monolayer or more (19). At tempera-
has already been mentioned by Bianchi et al. (15). Thiel tures greater than 400ꢂC the dehydroxylation is completed
and Madey (16) have highlighted the fact that many ox- and practically no active sites are formed by sulfation in
ide systems (titania, tungsten bronzes, PbO, NiO) do not solution. However, chemically inert, crystalline zirconia
rehydroxylate after they have crystallized and only display calcined at 600ꢂC seems to have similar or enhanced prop-
dissociative chemisorption of water in the presence of de- erties of adsorption. The observed contents of retained sul-
fects. Orlando et al. (17), by means of an ab initio quantum fate and the variation in the value of the C constant of
mechanical model of tetragonal zirconia, have shown that the BET equation confirm this statement. Although calci-
associative and dissociative chemisorption of water on crys- nation is generally thought to diminish the heterogeneity
talline zirconia involve similar energies. They estimated a of gels and the number of adsorption sites, a more ener-
rather high value for the activation energy between the two getic surface may be produced if a new phase is produced.
states of water. It is also well known that at sufficiently high Holmes et al. (20) have found that the appearance of the
temperatures, dehydroxylation produces Lewis acid–base monoclinic habitat upon calcination of zirconia results in a
pairs at the surface which act as catalytic active sites in many material with much higher surface energy; a difference in
reactions. Therefore, we can envisage two kinds of groups the heat of immersion of 300 Erg/cm2 was found. Ardizzone
appearing at high temperatures (see below). The stability and Bassi (21), while performing potentiometric titrations
or the appearance of (a) or (b) depends on the temperature of zirconia, found that zirconia heat treated at 800ꢂC dif-
of activation. The sustained drop of the OH density during fered from samples calcined at 200ꢂC in the mechanisms of
our chlorination studies with SOCl2, may indicate that (a) charge formation, the magnitude of the surface charge, and
proceeds from the early stages of calcination:
the sensibility to the concentration of electrolyte.
The surface inertness of crystalline calcined zirconia was
found to be overcome when more aggressive conditions of
reactions were used. Both chlorination and sulfation when
performed in the gas phase, at high temperatures, seemed
to produce the reopening of the oxo bridges, incorporating
chlorine in one case and yielding active sulfate species in
the other. The latter can be supposed to be sulfo bridges, to
OH OH
Zr + Zr ꢀꢀꢀꢀꢀ→ Zr¡ LZr
O
|
|
ꢀ
(a)
(b)
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
Oꢀ
|
+
ꢀꢀ→ Zr
Zr
which the appearance of catalytic activity in SO24ꢀ ZrO2 is
–
Lewis sites (b) require rather high temperatures at which
practically all the water is eliminated from the surface (18).
Both (a) and (b) can be depicted by the schematic reaction:
customarily related:
O
Cl Cl
2 OHꢀ ꢀ→ H 2 O + O 2 ꢀ
.
|
|
Zr¡ LZr + CCl4(vap) ꢀꢀ→ Zr Zr + COCl2
In the comprehensive review of Parks (8) on the ZPC
(zero point of charge) of many oxide systems, the ratio
(O2ꢀ/OHꢀ) is said to have influence on the position of the
ZPC. Higher ratios produce the shift of the ZPC to acid
values of pH. Surface oxygen would produce a strong po-
larization of the free surface groups involved in ampho-
teric reactions with the solvent. This shift in ZPC for cal-
cined samples was certainly observed in our potentiometric
titrations, with a difference of three points with respect to
noncalcined samples.
O
O
LL ¡¡
S
O
O¡ L
O
Zr¡ LZr + H2SO4(vap) ꢀꢀ→ Zr¡
Zr + H2O.
L
The evolution of the activity of the crystalline samples
sulfated in vapor phase, when activated at high tempera-
–
Now we have a hypothetical scheme to explain the ob- tures in air, is similar to that observed in SO24ꢀ ZrO2 cata-
served pattern of catalytic activity. Calcination at temper- lysts sulfated in the traditional way, having a maximum at
atures higher than the “glow exotherm” produces a nearly 650–700ꢂC (Fig. 8). Higher temperatures produce a steep

262
VERA AND PARERA
in solution at room temperature or chlorination in vapor
phase.
The surface chemical inertness or zirconia after calcina-
tion at T > 400ꢂC, is rationalized in terms of the appearance
of oxo bridges. This inertness can be overcome by the attack
of some reactives, e.g., CCl4 and H2SO4, in vapor phase and
at high temperatures. The more aggressive conditions are
thought to produce the reopening of the oxo bridges.
Sulfation of zirconia with sulfuric acid in the vapor phase
is posed as an alternative, however drastic, method of pro-
motion of crystalline materials. An attractive feature of the
method is that it enables the in situ resulfation of catalysts
between reaction periods, thus allowing for a compensation
of sulfur losses.
FIG. 8. Catalytic activity for the isomerization of n-butane of sam-
Predominant monoclinic zirconia can be promoted by
vapor phase sulfation and turned active in isomerization of
n-butane, suggesting that active catalysts need no specific
crystalline phase.
ples ZO4600 sulfated in the vapor phase as a function of the activation
temperature in the catalytic test. Activity: ᭹, ZO4600-VPS300 (calcined at
600ꢂC, sulfated at 300ꢂC); , ZO4600-VPS375; ᭡, ZO4600-VPS400. Selectivity
to i-butane: ꢃ, ZO4600-VPS300
.
REFERENCES
activity drop that is surely related to the decomposition of
surface sulfur species.
1. Arata, K., and Hino, M., in “Proceedings, 9th International Congress
on Catalysis, Calgary, 1988” (M. J. Phillip and M. Ternan, Eds.),
Vol. 4, p. 1727. Chem. Institute of Canada, Ottawa, 1988.
2. Comelli, R. A., Vera, C. R., and Parera, J. M., J. Catal. 151, 96 (1995).
3. Parera, J. M., Catal. Today 15, 481 (1992).
4. Comelli, R. A., Vera, C. R., and Parera, J. M., Latin Am. Appl. Res.
24, 227 (1994).
5. Boehm, H. P., and Knozinger, H., in “Catalysis Science and Technol-
ogy” (J. R. Anderson and M. Boudart, Eds.), Vol. 4, p. 39. Springer-
Verlag, Berlin, 1983.
6. Vera, C. R., Doctoral thesis, Universidad Nacional del Litoral, Santa
Fe, Argentina, 1995.
7. Mc Daniel, M. P., J. Phys. Chem. 77, 2070 (1973).
8. Parks, G. A., Chem. Rev. 65, 177 (1965).
9. Valigi, M., Cimino, A., Gazzoli, D., and Minelli, G., Solid State Ionics
32/33, 698 (1989).
CONCLUSIONS
The variation of the catalytic activity of SO24ꢀ ZrO2 with
–
the temperature of calcination of zirconia gels before the
sulfation in solution, does not correlate well with any tex-
tural or crystal property, or the sulfate content.
The adverse effect of calcination at T > 400ꢂC of gels
before sulfation, is not directly related to the crystallinity
of zirconia. In fact, under certain conditions of preparation,
crystalline materialscan be effectivelypromoted to produce
active catalysts in isomerization of n-butane.
The cause of the ineffective promotion by the customary
sulfation in solution, when applied to crystalline materials,
is explained by the disappearance of surface OH groups,
which is practically complete at 500ꢂC when the material
has crystallized. OH groups would condense forming oxo
bridges that would be inert to the attack by sulfuric acid in
solution under mild conditions.
Sulfate contents cannot be related directly to the activity.
Samples calcined at T > 500ꢂC have an adsorption strength
for sulfate anions that is equal to or higher than amorphous
zirconia gels.
10. Vissers, D. R., J. Phys. Chem. 72, 3236 (1968).
11. Ward, D. A., and Ko, E. I., J. Catal. 157, 321 (1995).
12. Morterra, C., Cerrato, G., Emanuel, C., and Bolis, V., J. Catal. 142, 349
(1993).
13. Mercera, P. D. L., Van Ommem, J. G., Doesburg, E. B. M., Burggraaf,
and Ross, J. R. H., Appl. Catal. 57(1), 127 (1990).
14. Chen, F. R., Coudurier, G., Joly, J.-F., and Vedrine, J. C., J. Catal. 143,
616 (1993).
15. Bianchi, D., Chafik, T., Khalfallah, M., and Teichner, S. J., Appl. Catal.
A 123, 89 (1995).
16. Thiel, P. A., and Madey, T. E., Surf. Sci. Rep. 7(6–8), 211 (1987).
17. Orlando, R., Pisani, C., Ruiz, E., and Sautet, P., Surf. Sci. 275, 482
(1992).
18. Morterra, C., and Marchese, L., in “Acid-Base Catalysis” (K. Tanabe,
H. Hattori, T. Yamaguchi, and T. Tanaka, Eds.), p. 197. Kodansha,
Tokyo, 1989.
19. Yori, J. C., and Parera, J. M., Appl. Catal. A: General 129, L151 (1995).
20. Holmes, H. F., Fuller Jr., E. L., and Gammage, R. B., J. Phys. Chem.
76, 1497 (1972).
The density of surface OH groups and/or their reactivity
seems to depend on the precursor used in the preparation
of the gel. In any case, the density of chlorinated surface
groups is always lower or equal to the surface cationic den-
sity (8 Zr/nm2). Samples prepared from oxychloride show
a maximum of 3–3.7 OH/nm2 titratable by chlorination
21. Ardizzone, S., and Bassi, G., Mater. Chem. Phys. 25, 417 (1990).