Journal of Energy Chemistry 24 (2015) 463–471

Contents lists available at ScienceDirect

Journal of Energy Chemistry

journal homepage: www.elsevier.com/locate/jechem

Efficient production of biodiesel from both esterification and

transesterification over supported SO4 2− –MoO3 –ZrO2 –Nd2 O3 /SiO2
catalysts
Xiuqin Li, Dongmei Tong∗, Changwei Hu∗
Key Laboratory of Green Chemistry and Technology, Ministry of Education, College of Chemistry, Sichuan University, Chengdu 610064, Sichuan, China

a r t i c l e i n f o a b s t r a c t

Article history: SO4 2 − –MoO3 –ZrO2 –Nd2 O3 /SiO2 (SMZN/SiO2 ) catalysts for the production of biodiesel via both esterifica-
Received 22 January 2015 tion and transesterification were prepared and characterized by N2 adsorption-desorption isotherms, X-ray
Revised 21 March 2015
diffraction (XRD), scanning electron microscopy (SEM), thermogravimetry analysis (TGA), ammonia adsorp-
Accepted 25 May 2015
Available online 10 August 2015
tion Fourier transform infrared spectra (NH3 -FTIR), and ammonia adsorption temperature programmed des-
orption (NH3 -TPD) to reveal the dependence of the stable catalytic activity on calcination time. The reason for
Keywords: catalyst deactivation was also studied. The calcination time remarkably affected the types of active centers
Biodiesel on SMZN/SiO2 , and 4 h was found to be the optimal calcination time. SO4 2− species bonded with small size
SMZN/SiO2 solid acids ZrO2 were found to be the stable active centers, where the leaching of SO4 2− and the deposition of coke were
Stable active centers inhibited. The deposition of coke was easier on large size ZrO2 than on small size ones. Calcination in air flow
Small size ZrO2
could eliminate the deposited coke to recover the deactivated catalysts.
© 2015 Science Press and Dalian Institute of Chemical Physics. All rights reserved.

1. Introduction Later, they paid close attention to the roles of these ions, and suc-
cessfully developed sulfated metal oxide, whereas zirconium oxide
Because of the inevitable future decline of fossil fuel resources, was found to show exceedingly high surface acidity [14]. During the
there was a demand to find an alternative energy source to meet the last years, sulfated zirconia (SO4 2− /ZrO2 ) was widely used in the syn-
increasing world energy needs [1–3]. Biodiesel has attracted consid- thesis of biodiesel for their unique properties, such as strong acidity,
erable attention worldwide as a potential alternative to fossil fuels, high catalytic activity, environmentally benign, readily separable, and
due to its advantages such as renewability, biodegradability and en- simultaneously active for both transesterification and esterification
vironmental benefit [3–5]. It was generally produced by either the [15]. However, they exhibit unsatisfactory stability owing to the de-
transesterification of triglycerides, or the esterification of free fatty crease of active centers [16]. The acidity and stability can be enhanced
acids (FFAs) with short chain alcohols via base or acid catalysis [6,7]. by supporting SO4 2− species on composite oxides of ZrO2 and other
In base catalyzed processes, the catalysts will react with FFAs to form metal oxides [9,17].
soap leading to difficulty of separation [6,8]. Therefore, acid catalysts, Huang et al. reported that the solid acids SO4 2− /MoO3 –ZrO2
such as mineral acid (e.g. H2 SO4 ) and organic sulfonic acid, were exhibited higher catalytic activity than SO4 2− /ZrO2 for the esterifi-
widely used in the synthesis of biodiesel from raw materials with cation of acetic acid with n-butane, because MoO3 had the effects
high acid value [9]. However, these homogenous acid catalysts have to delay the crystallization of ZrO2 and to stabilize SO4 2− [18]. A Nd
some disadvantages, such as serious contamination, corrosion to re- modified solid super acid catalyst SO4 2− /ZrO2 –Nd2 O3 displayed an
actors and inevitable purification steps [10,11]. Therefore, investiga- excellent activity in the synthesis of n-butyllactate from lactic acid
tions have been focused on environmentally benign catalysts for the and n-butyl alcohol [19]. Jiang et al. [20] found that adding Nd into
synthesis of biodiesel [8,12]. SO4 2− /ZrO2 is beneficial to enhancing the activity for esterification,
In 1979, Hino et al. reported that the acidity and activity of Fe2 O3 while the addition of promoter Mo was of great importance in
could be remarkably enhanced by treatment with sulfate ion [13]. enhancing the stability of SO4 2− /ZrO2 . The excellent activity and
stability of SO4 2− /ZrO2 –MoO3 –Nd2 O3 by the co-addition of Nd and
Mo into SO4 2− /ZrO2 were attributed to the firm combination of
∗
Corresponding authors. Tel./Fax: +86 28 85411105. sulfur species with small crystallite ZrO2 , which hinders the leaching
E-mail addresses: [email protected] (D. Tong), [email protected] (C. Hu). of sulfur species from the surface.

http://dx.doi.org/10.1016/j.jechem.2015.06.010
2095-4956/© 2015 Science Press and Dalian Institute of Chemical Physics. All rights reserved.
464 X. Li et al. / Journal of Energy Chemistry 24 (2015) 463–471

Another approach to modify SO4 2− /ZrO2 was increasing the spe- to remove physisorbed water and other molecules. The surface area
cific surface area, through loading active components on porous car- was calculated according to BET equation, while the pore volume and
rier [21–23]. Compared with unsupported SO4 2− /ZrO2 [21], the sup- pore size distribution were determined by BJH equation.
ported sulfated zirconia owned homogeneous dispersion of zirconia, The XRD patterns of the samples were recorded by a DX-1000
higher density of acid sites, and better tolerance of water. diffractometer using nickel-filtered Cu Kα radiation operated at 40 kV
Tetragonal ZrO2 (t-ZrO2 ) was detected over SO4 2− /ZrO2 and and 25 mA. The scanning step size was 0.06° and scanning range was
SO4 2− /ZrO2 -based solid acids, which provided excellent catalytic ac- between 5° and 80°.
tivity for the production of biodiesel [9,15,16]. Leng et al. [24] found The morphologies of SMZN/SiO2 –x h were examined by SEM per-
that t-ZrO2 , monoclinic ZrO2 (m-ZrO2 ) and Zr(SO4 )2 • 4H2 O co-existed formed on a JSM-5900LV with a mode of 20 kV.
in SO4 2− /ZrO2 when calcined at 450 °C, while diffraction peaks The thermal stability and degradation behavior of SMZN/SiO2 –x h
of Zr(SO4 )2 • 4H2 O weakened gradually with the increase of calci- were determined by TGA with a SDT Q600 thermo-analyzer, under
nation temperature. Jiang et al. found that the t-ZrO2 formed in a N2 flow of 60 mL/min with a heating rate of 10 °C/min from 30 to
solid acids SO4 2− /ZrO2 –MoO3 and SO4 2− /ZrO2 –MoO3 –Nd2 O3 , and 900 °C.
Zr(SO4 )2 • 4H2 O existed in SO4 2− /ZrO2 –Nd2 O3 [20]. All these catalysts The NH3 -FTIR of SMZNSiO2 –x h were measured on a Nicolet 5700
provided catalytic activity for esterification of fatty acids to obtain FT-IR with a DTGS detector running at a spectral resolution of 4 cm−1 .
biodiesel. They also reported that Zr(SO4 )2 in SO4 2− /ZrO2 –Nd2 O3 cat- After the samples had been activated at 300 °C under N2 flow for
alyst provided excellent catalytic activity, but the poor stability made 30 min, background spectra were recorded. The adsorption of am-
it difficult to be industrialized. monia was performed at room temperature and then swept by N2
Here, supported sulfated zirconia catalysts were prepared for for 15 min to remove physisorbed NH3 . Finally, infrared spectra were
biodiesel production. The effects of calcination time on the structure, collected at the target temperature of 50 °C.
acidity, stable catalytic activity, stable active centers of the catalyst, The acidity of SMZN/SiO2 –x h was measured by NH3 -TPD with
and the reason for catalyst deactivation were also investigated. on-line multichannel mass spectrometry (MS) HP R-20QIC connected
to the reactor outlet. The samples were swept by a nitrogen flow
of 25 mL/min at 300 °C for 60 min in a stainless steel U-tube, and
2. Experimental
then cooled to 70 °C. Next, the samples were saturated with NH3
(15 mL/min) for 30 min. Physisorbed NH3 was removed in a flow
2.1. Materials
of pure He until the baseline of MS signals stabilized, then NH3 -
TPD was conducted from 70 to 800 °C at a rate of 10 °C/min. The
The reagents were obtained commercially and used without fur-
possible escaped gas molecules (NH3 , H2 O, O2 , SO2 , and SO3 ) were
ther purification. Methyl laurate (Chromatographic grade) was pur-
monitored by recording the MS signals of m/z = 17, 18, 32, 64 and
chased from Beijinghuagong (China), methyl hexanoate (Chromato-
80. (NH4 )2 SO4 was used as standard to calculate the amount of acid
graphic grade) from Sigma-Aldrich Chemic (Germany), lauric acid
sites.
(99.5%) from Alfa Aesar (China), (NH4 )6 Mo7 O24 • 4H2 O (Analytical
grade) from Jinduicheng Molybdenum Group Co., Ltd. (China), and
Nd(NO3 )3 • 6H2 O (Analytical grade) from Tianjin Guangfu Fine Chemi- 2.4. Biodiesel production
cal Research Institute (China). Mesoporous SiO2 (Qingdao Yuminyuan
Silica-Gel Reagent Factory, China) was pretreated for 4 h under air at- The production of biodiesel through the esterification of lauric
mosphere at 650 °C before use. Other analytical grade reagents, such acid with methanol over the solid acid SMZN/SiO2 –x h was carried
as lauric acid, Zr(NO3 )4 • 5H2 O and sulfuric acid were all purchased out in a 25 mL flask equipped with a reflux condenser under con-
from Chengdu Kelong Chemical Co., Ltd. (China). tinuous magnetic stirring. The flask was charged with 1.00 g lauric
acid and heated to 65 °C, then 0.12 g catalysts (12 wt% based on lau-
ric acid) and 1.82 mL methanol (the molar ratio of methanol to lauric
2.2. Catalyst preparation acid was 9/1) were added. After the reaction, the catalyst was sep-
arated by centrifugation, and washed with acetone. The separated
7.70 g Zr(NO3 )4 • 5H2 O and 0.04 g Nd(NO3 )3 • 6H2 O were dissolved solid samples were characterized or reused. The reaction mixture was
respectively in 20 mL and 4 mL de-ionized water. The two solutions analyzed by gas chromatograph fitted with a SE-54 capillary column
were mixed to obtain a solution with Zr/Nd molar ratio of 75/1. 10 g (30 m × 0.25 mm) and a hydrogen flame-ionization detector to deter-
20–40 mesh mesoporous SiO2 was added into the above solution un- mine the molar yield of methyl laurate. Methyl hexanoate was used
der stirring, aged at 30 °C for 24 h, then dried at 80 °C and finally as internal standard in the analysis.
calcined at 600 °C for 4 h to form ZrO2 –Nd2 O3 /SiO2 .
0.31 g (NH4 )6 Mo7 O24 • 4H2 O was dissolved in a 90 mL solution 2.5. Stability test
of H2 SO4 (1 mol/L). ZrO2 –Nd2 O3 /SiO2 was impregnated in the so-
lution for 2 h, then evaporated at 110 °C and calcined at 600 °C The reusability of solid acids for the esterification of lauric acid
for desired time (2, 4, 6, 8 and 12 h) to obtain solid acids SO4 2− – was tested, where the used catalysts were applied in the next con-
MoO3 –ZrO2 –Nd2 O3 /SiO2 – x h (SMZN/SiO2 – x h, x: calcination secutive reaction. The reaction conditions were the same as the first
time). The actual loadings of Zr, Mo and Nd of SMZN/SiO2 –xh were run, that is, 12 wt% catalyst based on lauric acid, methanol to lauric
10.36 wt%, 0.86 wt% and 0.07 wt%, respectively, which were detected acid molar ratio of 9/1, reaction temperature of 65 °C, and reaction
by inductively coupled plasma–atomic emission spectroscopy (ICP– time of 10 h.
AES). The contents of S on the solid acids SMZN/SiO2 were mea-
sured using a FLASH 1112SERIES Element Analyzer, and displayed in 3. Results and discussion
Table 2.
3.1. Structure and surface morphology of the catalysts
2.3. Catalyst characterization
A larger surface area is beneficial towards spreading a large
N2 adsorption-desorption isotherms were recorded with a Mi- amount of zirconia and SO4 2− species, which can improve the cat-
cromeritics Tristar II apparatus. Prior to analysis, each sample was alytic activity [25]. As shown in Table 1, compared with unsupported
degassed at 120 °C for 2 h, and then at 300 °C for 2 h under vacuum sulfated zirconia, the specific surface area, average pore diameter, and
 X. Li et al. / Journal of Energy Chemistry 24 (2015) 463–471 465

pore volume are remarkably improved through loading the active

Table 1
BET properties of samples. components on carrier. The calcination time has no obvious effect on
these parameters.
Samples Sg (m2 /g)a Vp (cm3 /g)b Dp (nm)c
The XRD diffraction patterns of the samples are displayed in Fig. 1.
SO4 2 − –MoO3 –ZrO2 –Nd2 O3 59 0.08 5.4 There is no diffraction peaks for Nd and/or Mo compounds in all
SMZN/SiO2 – 2 h 149 0.42 12.3 samples, indicating the low amount of Nd and/or Mo oxides highly
SMZN/SiO2 – 4 h 145 0.43 12.2 dispersed on the carrier [26–28]. Amorphous phase is observed on
SMZN/SiO2 – 6 h 163 0.46 11.0
either SiO2 or ZrO2 –Nd2 O3 /SiO2 . Clearly, the crystalline structure
SMZN/SiO2 – 8 h 153 0.47 12.1
SMZN/SiO2 – 12 h 171 0.51 11.9 of SMZN/SiO2 –x h is strongly influenced by calcination time. For
SMZN/SiO2 –2 h, the observed peaks at 18.4°, 22°, 23.8°, 24.2°, 26.3°
a
Surface area.
b and 30.0° correspond to Zr(SO4 )2 , which may be derived from the dis-
Pore volume.
c
Average pore size. solution of ZrO2 in H2 SO4 for a long time immersion [24]. The diffrac-
tion intensity of Zr(SO4 )2 weakens with calcination time, and the
diffraction peaks of Zr(SO4 )2 disappear thoroughly when the calcina-
tion time increases to 6 h. It can be attributed to the fact that Zr(SO4 )2
decomposes into ZrO2 , which is well dispersed over SiO2 carrier and
cannot be detected by XRD. The ZrO2 grains grow up gradually with
further calcination, then a peak at about 30.1° appears in the XRD pat-
terns of SMZN/SiO2 –8 h which is correspond to (011) plane of m-ZrO2 .
Finally, more obvious diffraction peaks at 30.1°, 34.4°, 35.0°, 50.1°,
50.2° and 60.0° to (011), (002), (110), (112), (020) and (121) diffrac-
tion planes correspond to m-ZrO2 were detected on SMZN/SiO2 –12h.
ZrO2 is only detected on SMZN/SiO2 –8h and SMZN/SiO2 –12h by
XRD. However ZrO2 is observed on SEM photomicrographs of all
solid acids as presented in Fig. 2. Small size ZrO2 grains are ob-
served on solid acid SMZN/SiO2 –2h, while the amount and the size
increase with the calcination time prolonged. As distinctly observed
in Fig. 2(d), m-ZrO2 grains appears on solid acid SMZN/SiO2 –8h. The
crystal aggregation leads to an obvious rise in the crystalline size
with calcination time, and then m-ZrO2 grains become larger on
SMZN/SiO2 –12h compared with SMZN/SiO2 –8h. SEM data confirm
the above deduction from XRD that Zr(SO4 )2 decomposes into ZrO2
and then the latter gradually grows up to m-ZrO2 in the calcination
process.
The thermal stability of SMZN/SiO2 –x h was measured by TGA.
The TGA and DTG curves of fresh solid acid (Fig. 3) are divided into
Fig. 1. XRD patterns of fresh SMZN/SiO2 – x h (x: (1) 2, (2) 4, (3) 6, (4) 8, (5) 12), three stages: below 200 °C, over the range from 200 to 450 °C, and
(6) SiO2 and (7) ZrO2 –Nd2 O3 /SiO2 . 450 to 750 °C. The first stage belongs to the removal of physisorbed

Fig. 2. SEM micrographs of fresh SMZN/SiO2 – x h (x: a-2, b-4, c-6, d-8 and e-12).
466 X. Li et al. / Journal of Energy Chemistry 24 (2015) 463–471

Table 2
The chemical properties of catalysts with various calcination time.

Catalyst Weight loss (wt%)a Content (wt%)b Amount (mmol/g)c

Fresh 1st run Fresh 1st run 6th run

ST d SZ e SZS f SF g S1 h S6 i Cj

SMZN/SiO2 –2h 15.5 3.8 2.3 6.2 2.7 1.6 2.1 11.2
SMZN/SiO2 –4h 14.8 5.2 2.2 5.7 2.8 2.4 1.9 11.3
SMZN/SiO2 –6h 10.6 5.2 2.2 4.3 3.3 2.1 2.7 7.1
SMZN/SiO2 –8h 9.8 4.8 2.2 4.0 3.5 2.9 2.3 4.3
SMZN/SiO2 – 12 h 6.5 4.0 1.9 3.3 2.9 2.3 2.2 1.8
a
By TGA.
b
By elemental analysis technique.
c
By NH3 -TPD.
d
Total SO4 2− .
e
SO4 2− bonded with ZrO2 .
f
SO4 2− in amorphous Zr(SO4 )2 .
g
S on fresh catalysts.
h
S on catalysts after the first run.
i
S on catalysts after the sixth run.
j
Deposited coke after the sixth run.

Fig. 3. TGA and DTG curves of fresh SMZN/SiO2 – x h (x: (1) 2, (2) 4, (3) 6, (4) 8, (5) 12).
Fig. 4. NH3 -FTIR properties of fresh SMZN/SiO2 – x h (x: (1) 2, (2) 4, (3) 6, (4) 8, (5) 12).

water, and the second presents no obvious weight loss, whereas the
last one is assigned to the decomposition of SO4 2− species [29].
A shoulder peak appears at about 600 °C on the DTG curve of predominantly lie on the surface of SMZN/SiO2 –x h. The band at
SMZN/SiO2 –2h (Fig. 2a), indicating that there are at least two forms 2345 cm−1 appeared on SMZN/SiO2 –12h may be ascribed to the
of SO4 2− species. The weight loss assigned to the decomposition of stretching vibration of adsorbed N2 O [31].
SO4 2− species is presented in Table 2. It decreases from 15.5% to 6.5% MS was used to distinguish the possible escaped gases in NH3 -
with the calcination time varied from 2 to 12 h, suggesting that the TPD process. Fig. 5(a) depicts the NH3 desorption profile over the
decomposition of SO4 2− species in calcination process gives rise to range from 120 to 650 °C. The peaks appeared over the range from
the gradual decline of the content of SO4 2− species. 120 to 300 °C, 300 to 450 °C and 450 to 650 °C, correspond re-
spectively to weak, strong and super acid sites. The total amount of
3.2. Surface acidity acid sites calculated based on the peak area of NH3 is presented in
Table 2. SMZN/SiO2 –2h and SMZN/SiO2 –4h own an equivalent value
Ammonia was used as a probe molecule to observe the charac- (11.2 and 11.3 mmol/g). The amount of acid sites significantly de-
teristics of surface acidity on SMZN/SiO2 –x h by FTIR. As shown in creases with calcination time, and the value on SMZN/SiO2 –12h is
Fig. 4, the four main spectral features located at 3202, 3027, 2834 just 1/6 of that on SMZN/SiO2 –4h. It confirms the inference from TGA
and 1463 cm−1 , respectively, are assigned to the vibrations of NH4 + that the amount of acid sites decreases with calcination time due to
species with Brönsted acid sites (B-acid sites) [30]. For all the sam- the decomposition of SO4 2− .
ples, there is no obvious band at about 1610 cm−1 assigned to co- Besides NH3 , a large amount of H2 O was also detected (Fig. 5(b)).
ordinated NH3 with Lewis acid sites. It reveals that B-acid sites The profiles of water present a similar trend with that of NH3 , which
 X. Li et al. / Journal of Energy Chemistry 24 (2015) 463–471 467

Fig. 5. Escaped gas profiles of SMZN/SiO2 – x h (x: (1) 2, (2) 4, (3) 6, (4) 8, (5) 12) detected by NH3 -TPD: (a) NH3 , (b) H2 O, (c) SO2 and (d) O2 .

is consistent with the results of NH3 -FTIR, that is, B-acid sites pre- those materials exhibited high catalytic activity for esterification af-
dominantly exist on the surface of SMZN/SiO2 –x h. The formation of ter the modification with SO4 2− species. Therefore, we infer that the
B-acid sites have been taken into account as a result of the residual highly catalytic activity of catalysts should be mainly ascribed to the
water molecules or surface hydroxyl groups [32,33]. In the present presence of SO4 2− species. Among the five solid acids, SMZN/SiO2 –2h
work, all the samples were activated at 300 °C for 1 h before ammo- and SMZN/SiO2 –4h have an equivalent catalytic activity for the ester-
nia adsorption, whereas the liberation of water was still detected be- ification of lauric acid, and get a high value of 95.2% and 94.2% methyl
low 300 °C. It indicates that B-acid sites mainly stem from the surface laurate yield with 1 h reaction time, respectively. The yield of methyl
hydroxyl groups on the surface of SMZN/SiO2 –x h. laurate declines with calcination time suggesting that the calcination
Fig. 5(c) shows that SO2 peaks appear two peaks over the range time influences greatly on the activity of the solid acid SMZN/SiO2 –
from 430 to 800 °C due to the decomposition of SO4 2− species [29]. x h. The decline of activity can be ascribed to the decrease of SO4 2−
The two SO2 peaks prove the speculation from XRD, SEM and TGA species, and Zr(SO4 )2 crystal performs higher catalytic activity than
measurements that there are at least two kinds of SO4 2− species on m-ZrO2 . Although the five solid acids exhibit different initial catalytic
the catalysts. The liberation of O2 was observed to show roughly sim- activity, the yield of methyl laurate on all samples gradually increases
ilar trend with SO2 in Fig. 5(d), and no SO3 is detected in NH3 -TPD with reaction time, and a yield beyond 97.1% is got within 10 h reac-
process. Hino also reported that the decomposition of SO4 2− pro- tion time.
duced SO2 and O2 other than SO3 by using a TGA with a MS apparatus The production of biodiesel was carried out through the trans-
[34]. esterification of triacetin with methanol over SMZN/SiO2 –x h. Al-
though the catalytic activity for transesterification over solid acid
3.3. Biodiesel production SMZN/SiO2 –x h is slightly lower than that for esterification, it ex-
hibits considerable catalytic activity as presented in Fig. 6(b). A yield
Biodiesel was produced by esterification of lauric acid with of methyl acetate beyond 92.1% is got within 10 h reaction time. As
methanol over the above mentioned solid acid catalysts. As shown in expected, the dependence of the catalytic activity of SMZN/SiO2 –x h
Fig. 6(a), lauric acid hardly reacts with methanol without SMZN/SiO2 – for transesterification on calcination time and reaction time is similar
x h, or with ZrO2 /SiO2 , MoO3 /SiO2 , Nd2 O3 /SiO2 and SiO2 . However, to that for esterification.
468 X. Li et al. / Journal of Energy Chemistry 24 (2015) 463–471

Fig. 6. Catalytic performance of SMZN/SiO2 – x h (x: (1) 2, (2) 4, (3) 6, (4) 8, (5) 12) for lauric acid esterification and transesterification of triacetin: (a) activity for esterification,
(b) activity for transesterification, (c) stability for esterification.

The transesterification of jatropha oil with methanol over the solid first run, manifesting that crystal Zr(SO4 )2 leaches easily in the re-
acid SMZN/SiO2 –2 h and SMZN/SiO2 –4 h was also carried out to pro- action owing to the solvation of Zr(SO4 )2 in the by-product of water.
duce biodiesel. A yield of fatty acid methyl ester beyond 75.4% and However, the two solid acids maintain their activity, which reveals
71.2% were obtained at mild conditions (methanol-to-oil molar ratio that the stable active component is not crystal Zr(SO4 )2 . The diffrac-
of 9, 12 wt% catalyst referred to oil, low react temperature of 65 °C tion patterns of the fresh and used SMZN/SiO2 –x h (x: 6, 8 and 12 h)
and 10 h react time) exhibit no significant difference.
From the above results, it is shown that the solid acid SMZN/SiO2 –
x h catalysts were effective for the production of biodiesel. 3.6. The stable active species

TGA was also used to explore the reason for the deactivation
3.4. Stability test
of solid acids. Fig. 8(a) and (b) depict the TGA and DTG curves of
SMZN/SiO2 –x h after the first and the sixth run. Compared to the
The stability of SMZN/SiO2 –x h for the esterification of lauric acid
DTG curves of fresh SMZN/SiO2 –2h, the shoulder peak appears more
was also investigated. As shown in Fig. 6(c), the five solid acids all per-
clearly after being used once. It implies that the overlapped peaks are
form good activity in the second run. It is clearly that the calcination
divided into two clear weight loss peaks (over the range from 450 to
time plays an important role in keeping the catalytic activity. Because
640 °C and 640 to 740 °C) after the leaching of crystal Zr(SO4 )2 . The
SMZN/SiO2 –2 h and SMZN/SiO2 –4 h nearly keep their activity till the
other solid acids present the similar characteristics. It confirms that at
sixth run, especially, a high yield of methyl laurate (93.7%) is got over
least two kinds of SO4 2− species exist on SMZN/SiO2 –x h, and crystal
SMZN/SiO2 –4 h. However, only 71.4%, 75.6%, and 54.7% methyl laurate
Zr(SO4 )2 easily leaches in the reaction. Moreover, the high tempera-
yield are obtained in the sixth run over SMZN/SiO2 –6h, SMZN/SiO2 –
ture peak over the range from 640 to 740 °C disappears thoroughly
8 h and SMZN/SiO2 –12h, respectively. From what has been discussed,
after the sixth run (Fig. 8b), whereas the low temperature peak still
the calcination time affects significantly not only the activity but also
exists. It can be deduced that the leaching of Zr(SO4 )2 leads to the dis-
the stability of solid acid, while SMZN/SiO2 –4 h achieves the highest
appearance of the above mentioned peaks. Therefore, the peak over
activity and the best stability compared to the other solid acid cata-
the range from 640 to 740 °C can be assigned to the decomposition
lysts. So among the five solid acids, SMZN/SiO2 –4h is the most effec-
of Zr(SO4 )2 . Additionally, it is worth to note that a new peak appears
tive catalyst for the production of biodiesel at the low temperature
over the range from 250 to 450 °C on SMZN/SiO2 –x h after the sixth
with the lower energy consumption and waste of raw material.
run.
To assign the types of weight loss peaks, we used SMZN/SiO2 –12h
3.5. Leaching of SO4 2− species after the sixth run as well as Zr(SO4 )2 to suffer the temperature pro-
grammed decomposition (TPD) over the same temperature range in
To seek for the reason for the obvious deactivation of the He flow, and recorded the evolved gases using MS. Fig. 9(a) shows
long-time (6–12 h) calcined solid acids, XRD characterization of that Zr(SO4 )2 decomposes into SO2 and O2 over the range from 600
SMZN/SiO2 –x h after the first and the sixth run were conducted. to 800 °C. Hino reported that a large weight decrease of Zr(SO4 )2
As displayed in Fig. 7(a) and (b), Zr(SO4 )2 crystal existed on fresh decomposition was observed over relatively high temperature range
SMZN/SiO2 –2h and SMZN/SiO2 –4h disappear thoroughly after the other than that for the surface SO4 2− species in TGA [34]. It confirms
 X. Li et al. / Journal of Energy Chemistry 24 (2015) 463–471 469

Fig. 7. XRD patterns of used SMZN/SiO2 -xh (x: (1) 2, (2) 4, (3) 6, (4) 8, (5) 12): (a) used once and (b) after sixth run.

Fig. 8. TGA and DTG curves of used SMZN/SiO2 -xh (x: (1) 2, (2) 4, (3) 6, (4) 8, (5) 12): (a) used once and (b) after sixth run.

that the weight loss of SMZN/SiO2 –x h used once over the range from acids, and finally disappear thoroughly after the sixth run. This phe-
640 to 740 °C belongs to the decomposition of Zr(SO4 )2 . Therefore, it nomenon indicates that crystal Zr(SO4 )2 leaches thoroughly after the
indicates that the weight loss over the range from 450 to 640 °C of first run, whereas amorphous Zr(SO4 )2 still exists, and then amor-
SMZN/SiO2 –x h used once is corresponded to the decomposition of phous Zr(SO4 )2 also leaches in multiple recycles.
SO4 2− bonded with ZrO2 . Although Zr(SO4 )2 crystal diffraction peaks The weight loss of SMZN/SiO2 –x h after the sixth run over the
of XRD profiles on fresh SMZN/SiO2 –2h and SMZN/SiO2 –4h disappear range from 250 to 450 °C belongs to the release of water, because
thoroughly just after the first run, in the process of TGA, the weight water desorption is observed on SMZN/SiO2 –12h after the sixth run
loss peaks corresponding to Zr(SO4 )2 are still observed on all solid above 200 °C in TPD process (Fig. 9b). The water may be derived from
470 X. Li et al. / Journal of Energy Chemistry 24 (2015) 463–471

Fig. 9. Escaped gas profiles of samples detected by TPD: (a) Zr(SO4 )2 , (b) SMZN/SiO2 – 12 h after the sixth run in the He flow, (c) SMZN/SiO2 – 12 h after the sixth run in the air flow.

the decomposition of coke which is assigned to the strong adsorp- The co-existed SO4 2− species, amorphous and crystal Zr(SO4 )2 ,
tion of mixture reactants on the acid sites of SMZN/SiO2 –x h [35]. and SO4 2− species bonded with small size ZrO2 and large size ZrO2 ,
So the release of water over the range from 250 to 450 °C implies all contribute to the considerable initial catalytic activity. Zr(SO4 )2
that there is a large amount of coke on the surface of SMZN/SiO2 – leaches easily, while the SO4 2− species bonded with ZrO2 exist rel-
x h. As displayed in Fig. 9(b), there is a similar desorption trend for atively stably. SMZN/SiO2 –2 h exhibits slightly lower stability than
both CO2 and SO2 over the range from 400 to 650 °C, because coke is SMZN/SiO2 –4 h in the sixth run, which is ascribed to the lower
oxidized by O2 derived from the decomposition of SO4 2− to produce amount of S (SO4 2− species bonded with ZrO2 ) stably existed on the
CO2 . It confirms the deduction that there is a large amount of coke surface of catalysts. SMZN/SiO2 –8 h owns the larger content of S, but
on the surface of SMZN/SiO2 –x h. SMZN/SiO2 –12 h after the sixth run SMZN/SiO2 –4 h exhibits higher stability. It can be attributed to the
was also subjected to TPD process in air flow. Different from the re- fact that the amount of coke on SMZN/SiO2 –4 h is much lower than
sult from He flow, more content of CO2 is observed over the range that on SMZN/SiO2 –8 h, due to the easy formation of coke on the
from 200 to 650 °C (Fig. 9c). In this process, a larger amount of water large size ZrO2 . Therefore, the considerable stability of SMZN/SiO2 –
is also detected ((not displayed for brevity)). It once again confirms 4 h is ascribed to the large amount of SO4 2− species firmly bonded
the inference that a large amount of coke deposits on the surface of with small size ZrO2 , and the least amount of coke deposited on the
SMZN/SiO2 –x h. Therefore, the weight loss of SMZN/SiO2 –x h after surface.
the sixth run over the range from 250 to 450 °C corresponding to the
decomposition of coke, and the weight loss over the range from 450 3.7. The model of stable active species
to 650 °C is assigned to the overlap of the decomposition of coke and
SO4 2− species bonded with ZrO2 . Many reports had focused on the identification of acid site models.
The weight loss of SO4 2− species on SMZN/SiO2 –x h used once is Usually, Lewis acid sites and B-acid sites co-existed on the surface of
presented in Table 2. All solid acids used once show nearly the same sulfated zirconia [25,29]. A structure for acid sites on sulfated zirconia
weight loss of amorphous Zr(SO4 )2 . Different from the weight loss of to be chelated bidentate complexes, where the central ion acted as a
amorphous Zr(SO4 )2 , the content of SO4 2− species bonded with ZrO2 Lewis acid site, was proposed by Tanabe and co-workers [37]. Wang
exhibits an increase with the calcination time and then decreases et al. found that the addition of alumina caused the decrease of the
above 6 h calcination. amount of Lewis acid sites, while B-acid sites became dominant [25].
Table 2 also presents the contents of deposited C on the surface The model structure for the acid sites was suggested through
of SMZN/SiO2 –x h after the sixth run. The contents of deposited C summarizing the experimental results as following. In the present
on SMZN/SiO2 –2 h and SMZN/SiO2 –4 h are obviously lower com- work, the high catalytic activity of catalysts should be mainly as-
pared to those on other solid acids, which is consistent with the vari- cribed to the presence of SO4 2− species from the activity test. The
ation trend of stability. It indicates that a large amount of coke on the molar amount of SO4 2− species is much greater than those of Mo
surface of SMZN/SiO2 –x h leading to the coverage of the active cen- and Nd, therefore, the majority of SO4 2− species may interact with
ter should be considered as another reason for catalyst deactivation ZrO2 . On the other hand, the performance of the catalyst varied with
[36]. To confirm this point, SMZN/SiO2 –12 h after the sixth run was the variation of the crystal form of Zr species. Thus the SO4 2− species
reused after being calcined in a 60 mL/min oxygen–argon (1:1) flow bonded with small size ZrO2 might be the stable active centers. NH3 -
at 400 °C for 1 h and approximately 82% methyl laurate yield was FTIR spectra indicate that B-acid sites predominantly exist on the sur-
obtained. Similarly, the regenerated solid acid SO4 2− /ZrO2 –Yb2 O3 – face of SMZN/SiO2 –x h, which is derived from the weakening of the
Al2 O3 in air stream at 500 °C for 3 h showed that an approximately O–H bond by neighboring sulfate group [38,39]. The infrared spectra
20% increment was achieved [16]. Additionally, it seems that the coke of SMZN/SiO2 –4h and SMZN/SiO2 –12h were obtained (not displayed
formation occurs more easily on the large size ZrO2 than on the small for brevity). The IR bands between 400 and 700 cm−1 of SMZN/SiO2 –
ones. 4h and SMZN/SiO2 –12h can be assigned to crystal ZrO2 and Zr(SO4 )2
 X. Li et al. / Journal of Energy Chemistry 24 (2015) 463–471 471

[4] F. Su, Y. Guo, Green Chem. 16 (6) (2014) 2934.

[5] R. Luque, J.C. Lovett, B. Datta, J. Clancy, J.M. Campelo, A.A. Romero, Energy Environ.
Sci 3 (11) (2010) 1706.
[6] R. Song, D. Tong, J. Tang, C. Hu, Energy Fuels 25 (6) (2011) 2679.
[7] Q. Yi, J. Zhang, X. Zhang, J. Feng, W. Li, Fuel 143 (2015) 390.
[8] C.A. Deshmane, M.W. Wright, A. Lachgar, M. Rohlfing, Z. Liu, J. Le, B.E. Hanson,
Bioresour. Technol 147 (2013) 597.
Fig. 10. The model of the stable active species over solid acids SMZN/SiO2 – x h. [9] Y. Kuwahara, W. Kaburagi, K. Nemoto, T. Fujitani, Appl. Catal. A 476 (2014) 186.
[10] G. Chen, R. Shan, J. Shi, B. Yan, Bioresour. Technol 171 (2014) 428.
[11] S.Y. Chen, S. Lao-Ubol, T. Mochizuki, Y. Abe, M. Toba, Y. Yoshimura, Bioresour. Tech-
nol 157 (2014) 346.
[40], and the bands at 700–1500 cm−1 are assigned to bidentate sul- [12] T. Witoon, S. Bumrungsalee, P. Vathavanichkul, S. Palitsakun, M. Saisriyoot,
fate ions coordinated to the ZrO2 [21,38]. Therefore we tentatively K. Faungnawakij, Bioresour. Technol 156 (2014) 329.
[13] M. Hino, K. Arata, Chem. Lett 8 (1979) 1259.
suggest a model for the stable sulfated zirconia catalysts SMZN/SiO2 – [14] M. Hino, S. Kobayashi, K. Arata, J. Am. Chem. Soc 101 (1979) 6439.
x h as displayed in Fig. 10 [38]. [15] Y. Song, J. Tian, Y. Ye, Y. Jin, X. Zhou, J.A. Wang, L. Xu, Catal. Today 212 (2013) 108.
[16] G.X. Yu, X.L. Zhou, C.L. Li, L.F. Chen, J.A. Wang, Catal. Today 148 (1–2) (2009) 169.
[17] N.H.R. Annuar, A.A. Jalil, S. Triwahyono, N.a.A. Fatah, L.P. Teh, C.R. Mamat, Appl.
4. Conclusions Catal. A 475 (2014) 487.
[18] B. Huang, Z. Huang, J. Mol. Catal. 13 (5) (1999) 383.
Stable sulfated zirconia solid acids for the production of biodiesel [19] B. Deng, L.P. Chen, Q. Yu, China Cond. 33 (8) (2008) 848.
[20] K. Jiang, D. Tong, J. Tang, R. Song, C. Hu, Appl. Catal. A 389 (2010) 46.
are successfully prepared, obtaining a stable methyl laurate yield of
[21] X. Chen, Y. Ju, C. Mou, J. Phys. Chem. C 111 (2007) 18731.
93.7% over SMZN/SiO2 with 4 h calcination time in the sixth run. [22] H. Song, K. Shah, E. Doroodchi, T. Wall, B. Moghtaderi, Energy Fuels 28 (2) (2014)
The calcination time remarkably affected the types of active centers 1284.
[23] J. Li, J. Liu, L. Ren, Q. Liu, Z. Zhao, Y. Chen, P. Zhu, Y. Wei, A. Duan, G. Jiang, J. Energy
on SMZN/SiO2 , and SO4 2− species bonded with small size ZrO2 were
Chem 23 (5) (2014) 609.
found to be the stable active centers. Coke occurs more easily on the [24] Y. Leng, Y. Zhang, C. Huang, X. Liu, Y. Wu, Bull. Korean Chem. Soc. 34 (4) (2013)
surface of large size ZrO2 than on that of small size ones, and can be 1160.
eliminated by calcination in air to recover the deactivated catalysts. [25] J.H. Wang, C.Y. Mou, Appl. Catal. A 286 (1) (2005) 128.
[26] W. Li, H. Zhang, X. He, H. Liu, J. Energy Chem 22 (3) (2013) 512.
[27] Z. Li, Y. Tian, J. He, B. Wang, X. Ma, J. Energy Chem 23 (5) (2014) 625.
Acknowledgments [28] Q. Liu, F. Gu, J. Gao, H. Li, G. Xu, F. Su, J. Energy Chem. 23 (6) (2014) 761.
[29] J.H. Wang, C.Y. Mou, Microporous Mesoporous Mater. 110 (2–3) (2008) 260.
[30] D. Liu, P. Yuan, H. Liu, J. Cai, Z. Qin, M. Fan, Q. Tao, H. He, J. Zhu, Acta Miner. Sin. 30
This work is financially supported by the National Nature Science (1) (2010) 33.
Foundation of China (no. 21106089). The characterization of catalysts [31] B.E. Michael, W.T.G. David, R.W. Stephen, S. L. Paul, Anal. Chem 72 (2000) 206.
from Analytical and Testing Center of Sichuan University is greatly [32] B. Devassy, F. Lefebvre, S. Halligudi, J. Catal 231 (1) (2005) 1.
[33] Z. Ye, H. Chen, X. Cui, J. Zhou, J. Shi, Mater. Lett 63 (27) (2009) 2303.
appreciated. [34] M. Hino, M. Kurashige, H. Matsuhashi, K. Arata, Thermochim. Acta 441 (1) (2006)
35.
References [35] M. Gulsnet, P. Magnoux, Catal. Today 36 (1997) 477.
[36] K. Suwannakarn, E. Lotero, J. Goodwinjr, C. Lu, J. Catal 255 (2) (2008) 279.
[1] S. Lian, H. Li, J. Tang, D. Tong, C. Hu, Appl. Energy 98 (2012) 540. [37] J. Tuo, Y. Tsutomu, T. Kozo, J. Phys. Chem 90 (1986) 4194.
[2] P.D. Luu, H.T. Truong, B.V. Luu, L.N. Pham, K. Imamura, N. Takenaka, Y. Maeda, [38] J.R. Sohn, S.H. Lee, J.S. Lim, Catal. Today 116 (2) (2006) 143.
Bioresour. Technol 173 (2014) 309. [39] H. Wu, Y. Liu, J. Zhang, G. Li, Bioresour. Technol 174 (2014) 182.
[3] Y. Li, S. Lian, D. Tong, R. Song, W. Yang, Y. Fan, R. Qing, C. Hu, Appl. Energy 88 (10) [40] G.D. Yadav, A.D. Murkute, J. Catal 224 (1) (2004) 218.
(2011) 3313.