Stability of sulfated zirconia and the nature of the catalytically active species in the transesterification of triglycerides

Article abstract of DOI:10.1016/j.jcat.2008.02.014

Sulfated zirconia (SZ) exhibits remarkable activity for various hydrocarbon reactions under mild conditions and is of interest for biodiesel synthesis. Nevertheless, to date no detailed study has addressed its activity and stability in liquid polar media such as alcohols, although a number of papers have suggested the possibility for some sulfur leaching. This paper presents an investigation into the activity and stability of a commercial SZ catalyst for the liquid-phase transesterification of triglycerides at 120 °C. The kinetics of tricaprylin (TCP) transesterification with a series of aliphatic alcohols (methanol, ethanol, and n-butanol) were investigated at 120 °C and 6.8 atm in a Parr batch reactor. It was found that the catalytic activity for TCP conversion decreased as the number of carbons in the alkyl chain of alcohol increased, most likely as a result of increased steric hindrance. The SZ catalyst exhibited significant activity loss with subsequent reaction cycles. The characterization of used catalysts after their exposure to various alcohols at 120 °C showed that the SO^2-₄ moieties in SZ were permanently removed. The SO^2-₄ species were leached out, most likely as sulfuric acid, which further reacted with alcohols to form monoalkyl and dialkyl sulfate species, as demonstrated by ¹H NMR studies. This was in essence the main route for catalyst deactivation. Our findings conclusively demonstrate for the first time that in alcoholic-liquid media at higher temperatures, SZ deactivates by leaching of its active sites, most likely leading to significant homogeneous rather than heterogeneous catalysis.

Full text of DOI:10.1016/j.jcat.2008.02.014

Journal of Catalysis 255 (2008) 279–286
www.elsevier.com/locate/jcat
Stability of sulfated zirconia and the nature of the catalytically active species
in the transesterification of triglycerides
Kaewta Suwannakarn, Edgar Lotero, James G. Goodwin Jr. ∗, Changqing Lu
Department of Chemical and Biomolecular Engineering, Clemson University, Clemson, SC 29634, USA
Received 15 November 2007; revised 11 February 2008; accepted 17 February 2008
Abstract
Sulfated zirconia (SZ) exhibits remarkable activity for various hydrocarbon reactions under mild conditions and is of interest for biodiesel
synthesis. Nevertheless, to date no detailed study has addressed its activity and stability in liquid polar media such as alcohols, although a number
of papers have suggested the possibility for some sulfur leaching. This paper presents an investigation into the activity and stability of a commercial
◦
SZ catalyst for the liquid-phase transesterification of triglycerides at 120 C. The kinetics of tricaprylin (TCP) transesterification with a series of
◦
aliphatic alcohols (methanol, ethanol, and n-butanol) were investigated at 120 C and 6.8 atm in a Parr batch reactor. It was found that the catalytic
activity for TCP conversion decreased as the number of carbons in the alkyl chain of alcohol increased, most likely as a result of increased steric
hindrance. The SZ catalyst exhibited significant activity loss with subsequent reaction cycles. The characterization of used catalysts after their
2−
2−
◦
exposure to various alcohols at 120 C showed that the SO moieties in SZ were permanently removed. The SO species were leached out,
4
4
1
most likely as sulfuric acid, which further reacted with alcohols to form monoalkyl and dialkyl sulfate species, as demonstrated by H NMR
studies. This was in essence the main route for catalyst deactivation. Our findings conclusively demonstrate for the first time that in alcoholic-
liquid media at higher temperatures, SZ deactivates by leaching of its active sites, most likely leading to significant homogeneous rather than
heterogeneous catalysis.
© 2008 Elsevier Inc. All rights reserved.
Keywords: Transesterification; Triglycerides; Tricaprylin; Sulfated zirconia; Deactivation; Leaching
1. Introduction
sions. For biodiesel synthesis, transesterification of fats or oils
containing mainly triglycerides is performed to reduce the vis-
cosity, producing a biofuel (fatty alkyl esters) that can substitute
for petroleum-based diesel fuel without the need for engine
modifications.
The transesterification of triglycerides, often called alco-
holysis, involves the reaction of triesters of glycerol with an
alcohol to form alkyl esters and glycerol. This reaction has
been the subject of extensive research due to the diverse uses
of its products, including in the synthesis of polyester or PET
in the polymer industry [1], synthesis of intermediates for the
pharmaceutical industry [2], curing of resins in the paint indus-
try [3], and synthesis of biodiesel in the alternative fuel industry
[4–7]. The synthesis of biodiesel, for instance, has been a fo-
cus of much recent research on triglyceride transesterification,
because of the need to replace fossil fuel energy sources with
renewable biofuels amid concerns about greenhouse gas emis-
Transesterification can be catalyzed by both bases and acids.
Although the reaction rate of alkali-catalyzed transesterifica-
tion has been reported to be 4000 times faster than that using
acids [5], the use of base catalysts for biodiesel synthesis neces-
sitates refined feedstocks with low content of water (<0.5 wt%)
and fatty acids (<1 wt%), which in the long run increases the
cost of biodiesel production. This is apparent with the use of
virgin vegetable oils (3–6 wt% fatty acids), which forces the
use of higher amounts of the homogeneous base catalyst (some
catalyst is lost in the neutralization of the free fatty acids), pro-
ducing additional waste (soap) and complicating product sep-
aration. Thus, for feedstocks with high amounts of free fatty
acids, acid catalysis is preferable to base catalysis, because it
*
Corresponding author. Fax: +1 864 656 0784.
E-mail address: jgoodwi@clemson.edu (J.G. Goodwin).
0021-9517/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2008.02.014

280
K. Suwannakarn et al. / Journal of Catalysis 255 (2008) 279–286
allows for the simultaneous esterification of free fatty acids and
transesterification of triglycerides under the appropriate reac-
tion conditions without the formation of soap [7–9].
will be required for the use of solid acid catalysts in biodiesel
synthesis.
Industrial processes generally prefer solid catalysts for
chemical transformations because of their easy separation from
any reaction mixture. In addition, solid catalysts have the poten-
tial to be regenerated and are environmentally benign, allowing
their multiple reuse with little waste released to the environ-
ment.
2. Experimental
2.1. Materials
A sulfated-doped zirconium hydroxide precursor (XZO
1249/01) was kindly provided by Magnesium Electron Inc.
(Flemington, NJ). The sulfated zirconia catalyst was prepared
by calcining the parent-doped hydroxide under static air at
600 ◦C for 2 h. After calcination, the solid was kept in a desic-
cator until further use.
Glyceryl trioctanoate (Tricaprylin, with >99% purity as re-
ported by the supplier) was purchased from Sigma–Aldrich.
Anhydrous methanol (99.8%), ethanol (99.5%), and n-butanol
(99.4%) were purchased from Fisher Scientific. Methyl capry-
late (99%, Aldrich), ethyl caprylate (99%, Aldrich), and glyc-
erol (99%, Acros) were used as reference compounds for GC
calibration. All analytical chemicals were chromatographically
pure and used without further purification. Hexane and ethyl
acetate of HPLC grade were obtained from Fisher Scientific.
These were mixed in a volume ratio of 1:1 and used as solvent
for GC analysis.
Among solid acid catalysts, sulfated zirconia (SZ) has re-
ceived considerable attention over the last 20 years due to its
strong acid properties [10–12]. For instance, SZ has demon-
strated very high activity for various hydrocarbon reactions at
mild temperatures (e.g., alkane isomerization), even though it
deactivates rapidly due to coke deposition [11,12]. Much re-
search has been devoted to modifications of SZ to improve its
resistance to deactivation in gas-phase reactions [11]. The na-
ture of the active sites on SZ has been extensively studied for
hydrocarbon transformations [13–15]; for example, Lercher et
al. [16] showed that the covalent sulfate species on SZ respon-
sible for high catalytic activity for n-butane isomerization at
100 ◦C were easily removed by free water at room temperature.
However, only limited information exists about the activation–
deactivation behavior of SZ for reactions in anhydrous polar
liquid media, such as alcohols.
In liquid-phase reactions, SZ has demonstrated significant
activity in the esterification of acetic acid with n-butanol at
75 ◦C [17]. According to those authors, SZ could be completely
regenerated after a simple calcination at 550 ◦C. Similar results
were reported by Kiss et al. [18] for the esterification of do-
decanoic acid with 2-ethyl-hexanol at 160 ◦C. Those authors
showed that catalyst activity dropped to only 90% of its orig-
inal value after five consecutive runs, but SZ activity could be
restored by recalcination at 650 ◦C. In contrast to these previ-
ous reports, Jitputti et al. [19] used SZ for the transesterification
of crude palm kernel oil with methanol at 200 ◦C and obtained
remarkably high yields; however, the spent catalyst was fully
deactivated and could not be reused. Those authors proposed
that catalyst deactivation was due to a combination of catalyst
leaching and the blocking of active sites by reactants and/or
products. The same conclusion was drawn by Ni and Meunier
for the use of SZ in esterification of palmitic acid and methanol
at 60 ◦C [20]; however, they provided no evidence to support
this hypothesis.
In the present work, we studied the use of SZ in the acid-
catalyzed alcoholysis of triglycerides using a model triglyceride
compound, tricaprylin (TCP), and three low-molecular-weight
alcohols (methanol, ethanol, and n-butanol). Tricaprylin, which
has the same chemical functionality as other triglyceride mole-
cules, can be obtained in a pure form, unlike the larger triglyc-
erides. In addition, it has been shown that there is little dif-
ference in rates for triglycerides as large as tricaprylin or larger
[21]. By using a pure model compound like tricaprylin, we were
able to gain some fundamental insight into triglyceride transes-
terification. For the first time, special attention has been given
to the issue of catalyst deactivation in alcoholic condensed me-
dia at temperatures above 100 ◦C, conditions that most likely
2.2. Catalyst characterization
The sulfur content of the fresh calcined and spent SZ sam-
ples was analyzed by Galbraith Laboratories, Inc. (Knoxville,
TN). Thermogravimetric analysis (TGA) was carried out using
a Pyris 1 analyzer (Perkin–Elmer) to characterize sulfur con-
tent in the catalyst. Under a nitrogen flow of 20 mL/min, the
temperature was first stabilized at 30 ◦C for 1 min and then
ramped to 1000 ◦C at 10 ◦C/min. The surface area of the cal-
cined SZ catalyst was determined by N2 BET analysis using a
Micromeritics ASAP 2010. The crystallinity of the calcined SZ
powder was analyzed by a Scintag XDS 2000 diffractometer
using CuKα radiation with a wavelength of λ = 1.54 Å. NH3
TPD was used to estimate the acid strength and site concentra-
tion of SZ, as described previously [22].
2.3. Reaction study
The transesterification of tricaprylin (TCP) with methanol
(MeOH), ethanol (EtOH), and n-butanol (BuOH) was carried
out at 120 ◦C in a Parr 4590 batch reactor consisting of a stain-
less steel reactor vessel, a glass liner, a four-bladed pitched
turbine impeller, and a thermocouple. To ensure that most of
the reactants were in the liquid phase, the reactor was initially
pressurized at 6.8 atm (0.68 atm higher than the vapor pressure
of methanol at 120 ◦C). The typical molar ratio of alcohols to
tricaprylin was 12:1. The catalyst concentration was 10 wt%
based on the weight of the tricaprylin. While the amount of
the catalyst and the tricaprylin was kept constant throughout
the study, the total volume of the reaction mixture was changed
depending on the alcohol used. For the reaction startup, the cat-
alyst was initially charged into the reactant mixture at room

K. Suwannakarn et al. / Journal of Catalysis 255 (2008) 279–286
281
temperature, after which the reaction mixture was pressurized
and heated to the desired temperature at 120 ◦C over 7 min. Fi-
nally, the stirrer speed was increased to 2138 rpm; this point
was taken as time zero for the reaction. The sampling method
was described previously [23]. In brief, at particular times of
reaction, 0.15-mL sample aliquots were withdrawn from the
reaction mixture using a microscale syringe with a pressure-
lock button. The reaction sample was immediately mixed with
0.8 mL of solvent (hexane:ethyl acetate = 1:1 v/v) at room tem-
perature, followed by centrifuging to separate out any catalyst
particles. Then 40 µL of homogeneous liquid was withdrawn
and further diluted in 5 mL of solvent containing a known
amount of methyl laurate (an internal standard), followed by
GC analysis. Sample analysis using a Hewlett–Packard 6890
gas chromatograph followed the same procedure as described
previously [23].
Fig. 1. The catalytic activity profiles for TCP transesterification catalyzed by
SZ with (!) MeOH, (P) EtOH, and (1) BuOH at 120 C, 6.8 atm, molar ratio
◦
of alcohol:TCP = 12:1.
2.4. Catalyst leaching
absence of the catalyst. Negligible activity (<0.6% TCP con-
version after 2 h) was observed in all cases. With the catalyst
in place, at time zero (after the startup period), about 5% TCP
conversion was observed. The evolution of catalytic activity
over time for TCP transesterification with MeOH, EtOH, and
BuOH at 120 ◦C is shown in Fig. 1. All reactions demonstrated
100% selectivity to the corresponding ester products. The reac-
tion using MeOH showed the highest activity, with 84% TCP
conversion at 2 h. Under the same reaction conditions, the con-
versions of TCP using EtOH and BuOH were only 45 and 37%,
respectively, at 2 h TOS. A similar trend also has been re-
ported for the transesterification of rapeseed oil with various
alcohols under supercritical conditions [26], suggesting that the
lower reaction rates obtained with EtOH and BuOH are due
in part to steric hindrance effects of the larger alkyl chains in
these alcohols [27,28]. It should be noted, however, that triglyc-
eride conversion did not decrease proportionately to alkyl chain
length; a greater effect on the relative triglyceride conversion
was observed when the chain length changed from C1 to C2
than when it changed from C2 to C4, although for the latter case
the chain length increase was also twice as great. Similar obser-
vations have been reported for the esterification of carboxylic
acids with different chain lengths with methanol using a homo-
geneous acid catalyst [29]. As suggested for carboxylic acids
with different chain lengths, the tendency toward similar con-
version profiles with increasing alcohol chain length is probably
the result of “conformational leveling” effects wherein large
alkyl moieties assume conformations that counteract the con-
tribution of steric hindrance [29].
Because a possible cause of catalyst deactivation is the
leaching of active catalyst species (especially sulfur) into the
solution, we investigated changes in sulfur content of the cat-
alyst under the experimental conditions described earlier. To
estimate the degree of sulfur leaching, because a preliminary
study had indicated that the alcohol was the facilitator, a sam-
ple of the fresh calcined SZ was contacted with MeOH, EtOH,
and BuOH at 120 ◦C and 6.8 atm under constant stirring. Af-
ter 2 h, the resulting solutions were centrifuged and filtrated to
remove the solid catalyst. Then the solutions were used for re-
action without any catalyst. The pH and the acidity of the filtrate
were measured by titration with 0.05 M NaOH, using a pH me-
ter combined with phenolphthalein as a colorimetric indicator.
3. Results and discussion
3.1. Catalyst characterization
The fresh calcined SZ contained 1.73 wt% of sulfur (ICP
method; Galbraith Laboratories Inc.). This result was consistent
with sulfur content (1.74 wt% of sulfur) determined by TGA
using N2. N2 BET analysis demonstrated a specific surface area
(SBET) of 155 m2/g. Analysis of the N2 adsorption isotherm at
−196 ◦C showed mesopores of about 4 nm and a pore volume
of 0.15 cm3/g. The acid site concentration determined by NH3
TPD was 105 12 µmol/g. X-ray diffraction of the SZ powder
showed exclusively the tetragonal phase of ZrO2.
3.2. Reaction studies
Interestingly, faster reaction rates have been reported by
Freedman et al. [30] for the acid-catalyzed transesterifica-
tion of vegetable oils with heavier alcohols. This may be ex-
plained by alcohol solubility, with short-chain alcohols (espe-
cially methanol) having poor solubility in oils. But under our
reaction conditions, TCP and all of the alcohols used were com-
pletely soluble; therefore, the lower reaction rates obtained with
the heavier alcohols were due in part to steric hindrance, as ex-
pected based on fundamental chemical principles.
All experiments were conducted using a catalyst particle size
of 89–104 µm and a stirrer speed of 2138 rpm. Mass transfer
limitations could be ruled out, as discussed previously [24,25];
as a result, all of the measured reaction rates can be considered
reaction-controlled.
To exclude contributions from potential noncatalytic reac-
tions, blank transesterification reactions of TCP with the three
different alcohols were carried out at 120 ◦C and 6.8 atm in the

282
K. Suwannakarn et al. / Journal of Catalysis 255 (2008) 279–286
Fig. 3. SZ relative deactivation following multiple reaction cycles with TCP
◦
transesterification of MeOH, EtOH, and BuOH at 120 C, 6.8 atm, molar ratio
of alcohol:TCP = 12:1.
hol:TCP), our results are presented in terms of relative initial
catalytic activity in Fig. 3. As can be seen, the degree of catalyst
deactivation was reasonably similar regardless of the alcohol
type. We discuss this issue in more depth in Section 3.4.
As has been reported previously [11,12,31], carbon deposi-
tion can be a leading cause of catalyst deactivation with SZ.
Consequently, in an attempt to eliminate carbon deposits that
may have formed during the first reaction cycle, the recov-
ered solid was dried overnight at 100 ◦C and recalcined under
20 mL/min of flowing air at 315 ◦C for 4 h, as was shown to
be effective by Suwannakarn et al. [22]. Activities of the re-
calcined catalyst were comparable to activities obtained for the
second reaction cycle where the catalyst was used without treat-
ment, suggesting that formation of carbonaceous deposits was
not the cause of catalyst deactivation in this case.
3.4. Catalyst leaching and deactivation
As reported previously, ionic sulfur species supported on
the SZ catalyst surface can be modified and successively trans-
formed into H2SO4, HSO4−, and SO2− by the presence of free
water in the liquid phase [31], leadin4g to the loss of active sites
from the solid surface. Omota et al. [32], for instance, reported
that after contacting fresh SZ catalyst samples with water, a
rapid drop in the pH of the solution occurred, indicating that
acid species likely were being leached out into solution. Other
authors also have documented water’s capability of leaching out
the active catalytic species in SZ [16,32–34]. To the best of our
knowledge, however, no one has yet addressed the impact of
alcohol on SZ deactivation by site leaching.
To study the effect of alcohols on SZ catalyst deactivation,
fresh calcined SZ samples were immersed in the three differ-
ent alcohols (MeOH, EtOH, and BuOH) at 120 ◦C and 6.8 atm,
under continuous stirring for 2 h. Afterward, the alcoholic so-
lutions were centrifuged and filtrated to remove the solid cat-
alysts. The recovered catalysts were dried overnight and used
for the reaction. Fig. 4 shows the activities for TCP transes-
terification of the recovered catalysts from the three alcohol
washes and the activities for TCP transesterification in the ab-
Fig. 2. Reusability of SZ during three reaction cycles of TCP transesterification
with (a) MeOH, (b) EtOH, and (c) BuOH using a 12:1 molar ratio of alco-
◦
hol-to-TCP. ((!) first cycle, (P) second cycle, and (1) third cycle.) T = 120 C
and 10 wt% catalyst.
3.3. Catalyst recycling
One of the main advantages of heterogeneous acid catalysts
over liquid acids is that the former can be easily recovered
from the reaction mixture and potentially can be regenerated
and reused. In the present study, catalyst recycling studies were
carried out by recovering the used catalysts after 2 h of reaction
and reusing them (without pretreatment) with fresh reagents in
a subsequent reaction cycle. Fig. 2 presents the results for three
successive 2-h reaction cycles of TCP alcoholysis with MeOH,
EtOH, and BuOH, demonstrating a continuous activity loss for
all reactions. To account for the effect of the different TCP and
alcohol concentrations used in each case due to the different
alcohol volumes used to maintain a molar ratio of 12:1 (alco-

K. Suwannakarn et al. / Journal of Catalysis 255 (2008) 279–286
283
Table 1
Sulfur content of SZ catalyst samples after washing with fresh alcohol batches
◦
multiple times at 120 C, 6.8 atm, and with continuous stirring
Alcohol
Sulfur content (wt%)
after the first wash
determined
Sulfur content (wt%)
determined by N TGA
b,c
2
After 1
After 2
washes
After 3
washes
a
by ICP method
d
d
d
wash
MeOH
EtOH
BuOH
1.47
1.42
1.48
1.44
1.48
1.42
1.27
1.33
1.30
1.20
1.24
1.26
a
Determined by Galbraith Laboratories, Inc.
Sulfur content was determined assuming sulfate moieties desorbed as SO
b
3
gas.
c
Experimental error 0.01 wt%.
Washed for 2 h.
d
To further explore these phenomena, SZ samples were
washed multiple times using fresh alcohol. The sulfur content
of the samples pretreated in this manner was then measured
using TGA and ICP, the latter carried out by Galbraith Labora-
tories, Inc. (Table 1). Elemental sulfur analyses by ICP and
TGA were in good agreement within an experimental error
of 3%. As shown, after the first alcohol washing pretreat-
ment, SZ samples retained 85% of the original sulfur content
(1.73 wt%) regardless of the alcohol used, suggesting that un-
der our conditions, sulfur species were leached out to the same
extent regardless of alcohol characteristics (e.g., polarity, nu-
cleophilicity, alkyl chain length). It should be noted that after
the catalyst was washed once in pure distilled water under our
reaction conditions, it retained 1.30 wt% of sulfur (75% of its
original sulfur content). Also note that the amount of sulfur re-
tained in the catalyst after it was washed three times with fresh
alcohol was almost identical to that in the catalyst after only
a single water wash. As would be expected, water more effi-
ciently removes sulfate species from the catalyst surface, likely
due to its higher polarity and stronger hydrogen bonding ca-
pacity. Thus, if one assumes that under our conditions, water
can remove all of the leachable sulfur, then a single 2-h alcohol
wash or (based on the reaction cycle experiments) a single 2-h
reaction cycle with a mixture of alcohol and triglyceride effec-
tively removes almost 70% of the leachable sulfate species on
the catalyst surface.
The leaching process is fast during the first reaction cycle.
As determined by N2 TGA results, the sulfur content of SZ
was 1.47 wt% after washing with methanol for only 15 min
under reaction conditions, which is comparable to the result
obtained from methanol washing for 2 h under reaction con-
ditions. This finding is further supported by the fact that reac-
tions carried out using the methanol solution recovered after
the 15-min washing step exhibited the same reaction profile as
reactions carried out using the methanol solution after 2 h of
catalyst washing (Fig. 5). Similar results were found for EtOH;
however, a longer contacting time (at least 1 h) was required
for BuOH. Thus, it seems that alcohols with larger alkyl chain
lengths have somewhat slower kinetics for leaching the active
sites in SZ.
Fig. 4. Activities of alcohol precontacted SZ catalysts and activities of alcohol
solutions used in the catalyst washing steps compared to catalyst recycling ex-
periments as shown in Fig. 2 ((a) MeOH, (b) EtOH, and (c) BuOH) in TCP
transesterification at 120 C, 6.8 atm, molar ratio of alcohol:TCP = 12:1.
◦
sence of catalyst using the alcohol solutions (filtrates) obtained
after the washing experiments. As can be seen, the alcohol-
precontacted catalysts showed activities close to those of the
second cycle from the cycling experiments. The small varia-
tions in reaction activity can be attributed to differences arising
from the previous step (i.e., precontact with the reaction mix-
ture vs. pure alcohol) or to experimental error. On the other
hand, the filtrates (i.e., alcohol solutions used in the alcohol
precontact pretreatment) showed activities comparable to those
of the first reaction cycle, indicating that in-solution (i.e., ho-
mogeneous, not heterogeneous) catalytic species are apparently
primarily responsible for catalytic activity during the initial re-
action cycle at 120 ◦C.
Another implication of our results is that at temperatures
above 100 ◦C, the ability of sulfate ions to leach from SZ may

284
K. Suwannakarn et al. / Journal of Catalysis 255 (2008) 279–286
Fig. 5. Activity for methanol alcoholysis of TCP of the alcohol used in washing
the catalyst for various lengths of time compared to the first reaction cycle with
◦
SZ (120 C, 6.8 atm).
be related to the presence of the –OH function in the alcohols,
imposing a chemical limitation on using SZ for reactions that
use or produce compounds with this functionality. Certainly,
more research is needed to investigate this issue from a general
standpoint in light of the conflicting reports [18].
3.5. Active species in alcohol solution
In an attempt to determine whether sulfate species leached
out as H2SO4, H2SO4 solutions containing concentrations close
to those measured in the alcohol washing solutions and catalyst
deactivation studies were prepared and their catalytic activity in
the alcoholysis of TCP was investigated (Fig. 6). The catalytic
activity obtained from low-concentration H2SO4 solutions was
similar to that obtained from reactions using SZ, suggesting that
sulfate ions indeed may have leached from the catalyst surface
as H2SO4.
To further support this observation, sulfuric acid solutions
and filtrates from alcohol washing experiments were titrated us-
ing 0.05 mol of NaOH. Our findings were not consistent with
the expected molar ratio of 2H+:S for H2SO4; for instance,
a solution of 113 ppm H2SO4 in BuOH exhibited a H+:S ratio
of 1.3 value. Similar results (H+:S ratio = 0.92) were obtained
for filtrate from BuOH washing experiments. However, as has
been reported for H2SO4 in aliphatic alcohols [35], H2SO4
can undergo esterification with the alcohol to produce mono
alkyl-hydrogen and dialkyl sulfate; the latter of which is widely
used as an alkylating agent in organic synthesis and chemi-
cally induced mutagenesis [36,37]. In fact, dialkyl sulfates are
produced commercially by the direct reaction of alcohol and
sulfuric acid [36]. In addition, it is known that sulfuric acid
can readily react with alcohols to yield dialkyl sulfates even
in the gas-phase conditions of the planet’s atmosphere [38,39].
The reaction of H2SO4 and alcohols has been reported to occur
through the following reaction mechanism:
Fig. 6. Sulfuric acid catalyzed TCP transesterification with (a) methanol,
(b) ethanol, and (c) butanol at similar sulfur concentrations as in the leachate al-
◦
cohol and compared to the first of reaction with SZ (T = 120 C, P = 6.8 atm,
molar ratio of alcohol:TCP = 12:1).
To corroborate the presence of alkyl-sulfate compounds in our
reaction mixtures, residual alcohol solutions after SZ alcohol
washing under reaction conditions were analyzed by 1H NMR
1
(JEOL ECX 300, 300.5 MHz). Fig. 7a shows the H NMR
spectrum of a methanol solution obtained from washing SZ at
120 ◦C and 6.8 atm, with peaks at 3.739 and 3.636 ppm corre-
sponding to monomethyl hydrogen sulfate and dimethyl sulfate,
1
respectively [39]. In addition, control H NMR experiments
conducted using trace amounts of dimethyl sulfate in the al-
cohols showed a strong 1H NMR peak at 3.636 ppm (Fig. 7b),
confirming the presence of this species in the leachate solutions.
Consequently, a deactivation pathway for SZ in alcohols at tem-
peratures above 100 ◦C can be proposed (see Fig. 8).
H2SO4 + ROH ꢀ RSO3H + H2O,
(1)
(2)
ROSO3H + ROH ꢀ (RO)2SO4 + H2O.

K. Suwannakarn et al. / Journal of Catalysis 255 (2008) 279–286
285
◦
Fig. 7. NMR spectra of (a) the methanol filtrate after washing SZ at 120 C for 2 h, and (b) methanol solution with a representative concentration of dimethyl sulfate.
4. Conclusion
We investigated the activity and stability of SZ for the liquid-
phase transesterification of TCP using a series of aliphatic alco-
hols (methanol, ethanol, and n-butanol) at 120 ◦C and 6.8 atm.
The highest catalytic activity was observed in methanolysis,
followed by ethanolysis and butanolysis, respectively. The de-
creased catalytic activity with alcohol size likely was due to
increased steric hindrance. The SZ catalyst deactivated with
subsequent reaction cycles in all cases due to the leaching of
sulfate ion species, most likely as sulfuric acid. Under the reac-
tion conditions used, almost all catalytic activity in the first re-
action cycle appeared to be due to homogeneous rather than het-
erogeneous catalysis, as a result of sulfur leaching. The degree
Fig. 8. Schematic representation of possible active site leaching mechanism
for SZ.

286
K. Suwannakarn et al. / Journal of Catalysis 255 (2008) 279–286
of sulfur removal depended on the alcohol size and contacting
time. The proposed catalyst deactivation pathway includes the
removal of sulfate ions from the catalyst surface as sulfuric acid,
which subsequently reacts with alcohol to form monoalkyl hy-
drogen sulfate and dialkyl sulfate in solution, as indicated by
1H NMR studies.
[13] A. Clearfield, G.P.D. Serrette, A.H. Khazi-Syed, Catal. Today 20 (1994)
295.
[14] C. Morterra, G. Cerrato, F. Pinna, M. Signoretto, J. Phys. Chem. 98 (1994)
12373.
[15] F. Babou, G. Coudurier, J.C. Vedrine, J. Catal. 152 (1995) 341.
[16] X.B. Li, K. Nagaoka, J.A. Lercher, J. Catal. 227 (2004) 130.
[17] T.A. Peters, N.E. Benes, A. Holmen, J.T.F. Keurentjes, Appl. Catal. A 297
(2006) 182.
[18] A.A. Kiss, A.C. Dimian, G. Rothenberg, Adv. Synth. Catal. 348 (2006)
75.
Acknowledgments
[19] J. Jitputti, B. Kitiyanan, P. Rangsunvigit, K. Bunyakiat, L. Attanatho,
P. Jenvanitpanjakul, Chem. Eng. J. 116 (2006) 61.
[20] J. Ni, F.C. Meunier, Appl. Catal. A 333 (2007) 122.
[21] D.E. Lopez, J.G. Goodwin Jr., D.A. Bruce, J. Catal. 245 (2007) 381.
[22] K. Suwannakarn, E. Lotero, J.G. Goodwin Jr., Catal. Lett. 114 (2007)
122.
[23] Y. Liu, E. Lotero, J.G. Goodwin Jr., Appl. Catal. A 331 (2007) 138.
[24] Y. Liu, E. Lotero, J.G. Goodwin Jr., J. Catal. 246 (2007) 428.
[25] Y. Liu, E. Lotero, J.G. Goodwin Jr., J. Catal. 242 (2006) 278.
[26] Y. Warabi, D. Kusdiana, S. Saka, Appl. Biochem. Biotechnol. 113 (2004)
793.
This research was funded by the Animal Co-Products Re-
search & Education Center. The authors thank Magnesium
Electron for providing the SZ catalysts and Dr. Donald Van
Derveer of Clemson University for his valuable help with the
XRD measurements.
References
[1] U. Meyer, W.F. Hoelderich, Appl. Catal. A 178 (1999) 159.
[2] D. Duran, N. Wu, B. Mao, J. Xu, J. Liq. Chromatogr. Related Technol. 29
(2006) 661.
[27] K. Suwannakarn, E. Lotero, J.G. Goodwin Jr., Ind. Eng. Chem. Res. 46
(2007) 7050.
[28] B.R. Jermy, A. Pandurangan, Appl. Catal. A 288 (2005) 25.
[29] Y. Liu, E. Lotero, J.G. Goodwin Jr., J. Catal. 243 (2006) 221.
[30] B. Freedman, E.H. Pryde, T.L. Mounts, J. Am. Oil Chem. Soc. 61 (1984)
1638.
[3] J. Barrault, Y. Pouilloux, J.M. Clacens, C. Vanhove, S. Bancquart, Catal.
Today 75 (2002) 177.
[4] M. Di Serio, M. Ledda, M. Cozzolino, G. Minutillo, R. Tesser, E. Santace-
saria, Ind. Eng. Chem. Res. 45 (2006) 3009.
[31] A. Corma, H. Garcia, Catal. Today 38 (1997) 257.
[32] F. Omota, A.C. Dimian, A. Bliek, Chem. Eng. Sci. 58 (2003) 3175.
[33] T. Okuhara, M. Kimura, T. Nakato, Appl. Catal. A 155 (1997) L9.
[34] M. Kimura, T. Nakato, T. Okuhara, Appl. Catal. A 165 (1997) 227.
[35] M. Liler, Reaction Mechanisms in Sulfuric Acid, Academic Press, New
York, 1971, pp. 150–152.
[5] H. Fukuda, A. Kondo, H. Noda, J. Biosci. Bioeng. 92 (2001) 405.
[6] F.R. Ma, M.A. Hanna, Bioresour. Technol. 70 (1999) 1.
[7] E. Lotero, Y. Liu, D.E. Lopez, K. Suwannakarn, D.A. Bruce, J.G. Good-
win Jr., Ind. Eng. Chem. Res. 44 (2005) 5353.
[8] Y. Zhang, M.A. Dube, D.D. McLean, M. Kates, Bioresour. Technol. 89
(2003) 1.
[36] S. Theodore, P.S.T. Sai, Can. J. Chem. Eng. 79 (2001) 54.
[37] R. Wolfenden, Y. Yuan, Proc. Natl. Acad. Sci. USA 104 (2007) 83.
[38] N.P. Levitt, J. Zhao, R.Y. Zhang, J. Phys. Chem. 110 (2006) 13215.
[39] Y. Suzuki, M. Kawakami, K. Akasaka, Environ. Sci. Technol. 35 (2001)
2656.
[9] Y. Zhang, M.A. Dube, D.D. McLean, M. Kates, Bioresour. Technol. 90
(2003) 229.
[10] M. Hino, K. Arata, J. Chem. Soc. Chem. Commun. (1980) 851.
[11] G.D. Yadav, J.J. Nair, Microporous Mesoporous Mater. 33 (1999) 1.
[12] X.M. Song, A. Sayari, Catal. Rev. Sci. Eng. 38 (1996) 329.