Ph-SO3H-modified mesoporous carbon as an efficient catalyst for the esterification of oleic acid

Article abstract of DOI:10.1016/j.apcata.2012.03.044

Mesoporous carbon materials with thin pore walls (～1.7 nm) were synthesized using low-cost γ-Al₂O₃ as a hard template and in situ polymerized resorcinol-furfural resin as the carbon precursor. Compared with sugar, resin, a widely used carbon precursor, has higher carbon yield and simplifies the synthetic process. Ph-SO₃H modified mesoporous carbon was synthesized by covalent grafting of Ph-SO ₃H groups on mesoporous carbon via the diazonium salt. The resulting materials were characterized by means of nitrogen adsorption analysis, TEM, ¹³C NMR, XRD, FTIR and sulfur elemental analysis. The modified carbons were shown to possess high surface area (～1000 m²/g), a bimodal pore size distribution and high strong acid density (1.86 mmol H ⁺/g). These sulfonated carbons were used as solid acid catalysts in the esterification of oleic acid and methanol, a key reaction in biodiesel production. Compared with the traditional solid acid Amberlyst-15, the optimized carbon catalyst exhibited much higher activity with a rate constant (1.34 h^-1) three times to that of Amberlyt-15 and a turnover frequency (TOF) of 128 h^-1 eight times that of Amberlyst-15. The efficient catalytic ability was attributed to the high surface area and a proper mesopore texture. This carbon catalyst could then be easily separated from the product by filtration. The catalyst was reused six times, and no distinct activity drop was observed after the initial deactivation.

Full text of DOI:10.1016/j.apcata.2012.03.044

Applied Catalysis A: General 427–428 (2012) 137–144
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Ph-SO3H-modified mesoporous carbon as an efficient catalyst for the
esterification of oleic acid
Liang Geng, Gang Yu, Yu Wang, Yuexiang Zhu∗
Beijing National Laboratory for Molecular Sciences, State Key Laboratory for Structural Chemistry of Unstable and Stable Species, College of Chemistry and Molecular Engineering,
Peking University, Beijing 10087, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Mesoporous carbon materials with thin pore walls (∼1.7 nm) were synthesized using low-cost ␥-Al2O3
as a hard template and in situ polymerized resorcinol–furfural resin as the carbon precursor. Compared
with sugar, resin, a widely used carbon precursor, has higher carbon yield and simplifies the synthetic
process. Ph-SO3H modified mesoporous carbon was synthesized by covalent grafting of Ph-SO3H groups
on mesoporous carbon via the diazonium salt. The resulting materials were characterized by means
of nitrogen adsorption analysis, TEM, 13C NMR, XRD, FTIR and sulfur elemental analysis. The modified
carbons were shown to possess high surface area (∼1000 m2/g), a bimodal pore size distribution and
high strong acid density (1.86 mmol H+/g). These sulfonated carbons were used as solid acid catalysts in
the esterification of oleic acid and methanol, a key reaction in biodiesel production. Compared with the
traditional solid acid Amberlyst-15, the optimized carbon catalyst exhibited much higher activity with
a rate constant (1.34 h−1) three times to that of Amberlyt-15 and a turnover frequency (TOF) of 128 h−1
eight times that of Amberlyst-15. The efficient catalytic ability was attributed to the high surface area
and a proper mesopore texture. This carbon catalyst could then be easily separated from the product by
filtration. The catalyst was reused six times, and no distinct activity drop was observed after the initial
deactivation.
Received 21 December 2011
Received in revised form 23 March 2012
Accepted 31 March 2012
Available online 7 April 2012
Keywords:
Mesoporous carbon
Carbon-based solid acid
Alumina template
Diazonium modification
Esterification
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
biomass (sucrose [6], glucose [7,9], starch [10,11] or biochar [12])
as the carbon precursor, the pioneering work in this area synthe-
Due to its renewable feedstock and low CO2 emission, biodiesel
produced from vegetable oil is considered a sustainable fuel to
replace conventional diesel [1]. The production of biodiesel usu-
ally involves the transesterification of triglycerides catalyzed by a
homogeneous base, such as NaOH, KOH or NaOCH3 [1–3]. However,
the free fatty acid (FFA) contained in the crude vegetable oil will
be saponified by the basic catalysts. Therefore, a pre-esterification
process is necessary to convert FFA into the corresponding esters
before transesterification. Conventionally, the catalysts for pre-
esterification are liquid H2SO4, which is toxic, corrosive and costly
to neutralize, separate and recycle. Therefore, the development of
a green and easily separated acid catalyst has received much atten-
tion [4].
sized carbon catalysts through direct carbonization of precursors
followed by sulfonation in concentrated H2SO4 [5–16]. Such cat-
alysts performed well in the esterification of fatty acid, but direct
carbonization led to carbon catalysts with low porosity and low sur-
face area, which is unfavorable for reactant accessibility to active
sites. Moreover, the carbon yield of sugar is low. For instance, the
theoretical carbon yield of sucrose is 40%, while the actual yield is
only approximately 20% [17].
A new type of sulfonated ordered mesoporous carbon (OMC)
has been developed for FFA pre-esterification [18–21]. In this work,
the organic precursors were first impregnated into the pores of
an ordered mesoporous silica template (typically SBA-15) and car-
bonized, and then the template was removed by treatment with
HF solution. The remaining carbon materials were sulfonated by
concentrated H2SO4 [18,19] or 4-benzene-diazoniumsulfonate (4-
BDS) [20,21] to yield OMC-based solid acids with high surface
area and large pores. This improved texture facilitates the dif-
fusion of fatty acid molecules with long carbon chains to the
active site, thereby increasing the catalytic efficiency. In most
cases, silica templates must be prepared beforehand through mul-
tiple steps, which inevitably increase the cost of this type of
catalysts.
Among heterogeneous catalysts, carbon-based solid acid is ideal
for many reactions owing to such advantages as chemical inertness,
mechanical stability, structural diversity and surface hydropho-
bicity. For the pre-esterification of FFA, much work has focused
on sulfonated carbon as the solid acid [5–16]. Using inexpensive
∗
Corresponding author. Tel.: +86 10 62751703; fax: +86 10 62751708.
E-mail address: zhuyx@pku.edu.cn (Y. Zhu).
0926-860X/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apcata.2012.03.044

138
L. Geng et al. / Applied Catalysis A: General 427–428 (2012) 137–144
Our previous work reported the synthesis of carbon-based solid
The remaining carbon materials, denoted C-Ax-T (Ax denotes the
alumina template, T denotes the carbonization temperature), were
washed repeatedly with water and dried at 110 ◦C.
acid with sucrose as the carbon precursor, commercial ␥-Al2O3 as
the template and 4-BDS as the sulfonating agent [22–25]. Owing
to delicate control of the sucrose loading on the template surface,
such catalysts consist of small and thin carbon sheets. They can
readily disperse in methanol to efficiently catalyze the esterifica-
tion of oleic acid with methanol. The highest turnover frequency
(TOF) observed was seven times that of Amberlyst-15. However,
due to the low carbon yield of sucrose, repeated impregnation and
carbonization was necessary in the synthesis of such catalysts.
Recently, we successfully synthesized mesoporous carbon
materials with thin walls and large surface areas using
resorcinol–furfural resin, which in situ polymerized on an alu-
mina surface, as the carbon precursor [17]. Owing to the carbon
yield of resorcinol–furfural resin being as high as 50%, one
polymerization–carbonization process was sufficient. Moreover,
resorcinol–furfural resin tends to form complete carbon frame-
works that are stable enough to support themselves. In this paper,
Ph-SO3H groups were attached on the surface of these carbon
materials via 4-BDS. N2 adsorption analysis showed these modi-
fied materials still possessed the mesoporous structure and a large
surface area (∼1000 m2/g), and their strong acid density, deter-
mined by sulfur elemental analysis, reached 1.86 mmol H+/g. The
optimal carbon catalyst was identified with a TOF of 128 h−1, eight
times that of Amberlysts-15 and even higher than that of previ-
ously reported sucrose catalysts (109 h−1) [25]. Due to the complete
carbon framework, the catalyst was easy to separate from the liq-
uid reaction system by filtration. The optimal catalyst was reused
several times with no significant drop in activity after the initial
deactivation.
2.1.3. Sulfonation
The sulfonation of C-Ax-Ts was carried out using a previ-
ously reported method [25]. In a typical procedure, 0.22 g of
4-aminobenzene sulfonic acid was dissolved in a 0.05 M HCl aque-
ous solution and cooled to 5 ◦C. Then, 90 mg of NaNO2 was added
to 4-aminobenzene sulfonic acid to produce 4-BDS. The sulfonation
was achieved by mixing the 4-BDS solution prepared above with
0.4 g of C-Ax-T and holding the suspension at 5 ◦C for 3 h. After
filtration and thorough washing with water, dimethylformamide
(DMF) and acetone, the sulfonated mesoporous carbon catalysts
were obtained and denoted SC-Ax-T.
Amberlyst-15 (Acros Organics) was ground into a powder for
use as a control catalyst and was degassed at 100 ◦C for 7 h before
catalytic tests.
2.2. Characterization
Transmission electron microscopy (TEM) images were recorded
on a JEOL JEM-2100 high-resolution microscope (Japan) at an accel-
eration voltage of 200 kV. The N2 sorption analysis was performed
on a Micromeritics ASAP 2010 volumetric adsorption system (US)
at 77 K. All samples were degassed at 150 ◦C prior to measure-
ments. The specific surface area was determined using the BET
method based on the adsorption data in the relative pressure (P/P0)
range 0.05–0.20. Pore size distribution (PSD) was evaluated by the
Barrett–Joyner–Halenda method from the desorption branch of the
isotherm (additional note will be present when evaluated from the
adsorption branch). The total pore volume was estimated from the
amount of N2 adsorbed at a relative pressure of 0.99.
2. Experimental
2.1. Catalyst preparation
Fourier transform infrared (FT-IR) absorbance spectra in the
range of 600–4000 cm−1 were recorded on a Nicolet iN10 FT-IR
Microscope (USA) using pure samples. Powder X-ray diffraction
(XRD) patterns were recorded on a Rigaku D/MAX0200 powder
diffractometer (Japan) using Ni-filtered Cu K˛ radiation at 40 kV
and 100 mA. 13C solid state NMR spectra were recorded on a Bruker
AVANCE III WB Solid-State NMR Spectrometer with a WVT 4 mm
CP/MAS double resonance probe head. The spin rate was 8.0 kHz.
Sulfur and carbon elemental analyses were carried out on an Ele-
mentar Vario Micro Cube (Germany).
2.1.1. Source of alumina template
Three types of alumina were used as template: (a) commercial
alumina SBA150 (Engelhard Corp., USA), denoted AI; (b) alumina
prepared by calcination of pseudoboehmite (Shandong Aluminum
Corp., China) at 550 ◦C for 4 h, denoted AII; and c) alumina pre-
pared by calcination of aluminum hydroxide dried gel (Sinopec
Corp., China) at 550 ◦C for 4 h, denoted AIII. All alumina templates
were calcined at 800 ◦C for 4 h before use.
2.1.2. Preparation of mesoporous carbon materials
Mesoporous carbon materials were prepared according to a pre-
viously reported procedure [17]. The alumina templates (AI, AII
or AIII) were mixed and thoroughly ground with resorcinol. The
weight ratio of alumina to resorcinol, mc, was calculated according
to the empirical formula below.
2.3. Catalytic tests
The esterification of oleic acid and methanol was performed in
a stirred 20-ml autoclave at 65 ◦C. Typically, 20 mg of solid acid
catalyst or 1.82 mg of H2SO4 (contain identical H+ amount to SC-AI-
900) was added to 1 g of oleic acid (99%, TCI) and 8 ml of methanol.
At selected reaction times, the autoclave was rapidly cooled in a
room-temperature water bath. The catalyst was then removed by
filtration, and the product was extracted and diluted with n-hexane
to 100 ml. The yield of methyl oleate was analyzed by gas chro-
matography equipped with a FID detector and a BD-5 capillary
column, where methyl heptadecanoate (AccuStandard, 10.0 mg/l
in hexane) was used as the internal standard.
The stability of SC-AI-900 was investigated by performing the
esterification reaction six times with 20 mg of SC-AI-900, 1 g of
oleic acid and 8 ml of methanol at 65 ◦C for 1 h. After each use,
the catalyst was recycled by filtration and washed with methanol.
After recovery, the catalysts were activated in 0.1 M H2SO4 at room
temperature, washed with deionized water and dried at 110 ◦C.
SBET,A × 2 × 12.0
mc = 1.05 × 1.5 ×
(1)
√
3 × ( 3/2) × (1.42 × 10−10
)
2 × 6.02 × 1023
where SBET,A is the Brunauer–Emmett–Teller (BET) surface area
of the alumina template. Typically, 7 g of alumina AI was ground
with 1.12 g of resorcinol for at least 20 min. Then, the mixture was
transferred to a round bottom flask containing 30 ml of mesity-
lene, 2.5 ml of furfural and 3 drops of 1,2-ethylenediamine. After
stirring at 90 ◦C for 4 h, the alumina/resorcinol–furfural resin com-
posite was filtered, washed with ethanol three times and dried in
an oven at 130 ◦C.
To obtain the mesoporous carbon, the composite was car-
bonized under N2 for 3 h at 500 ◦C, 600 ◦C, 700 ◦C, 800 ◦C, 900 ◦C and
1000 ◦C, respectively, and then immersed in a 24% hydrofluoric acid
(HF) solution at room temperature for 3 h to dissolve the alumina.

L. Geng et al. / Applied Catalysis A: General 427–428 (2012) 137–144
139
For kinetic studies, the conversion results were fitted to a
pseudo-first order rate equation to determine the rate constant,
k. TOF was calculated according to the equation below.
n0,OA
TOF = k
(2)
nH
+
where n0,OA is the initial amount of oleic acid and nH+ is the amount
of strong acid sites on the catalyst, as determined by sulfur elemen-
tal analysis.
3. Results and discussion
3.1. Preparation and characterization of sulfonated mesoporous
carbon catalysts
Mesoporous carbon materials were prepared by a previously
published method using low-cost ␥-alumina as the template and
resorcinol–furfural resin as the carbon precursor [17]. Rather than
direct impregnation, the precursor was generated in situ on the
alumina surface. First, resorcinol was mechanically mixed with alu-
mina and ground thoroughly at room temperature, which enabled
resorcinol to uniformly disperse on the alumina surface. Then, after
the introduction of furfural, the polymerization occurred in the
presence of the basic catalyst 1,2-ethylenediamine to form resin
layers uniformly covering the alumina surface. The empirical for-
mula (1) was used to determine the weight ratio of resorcinol
and alumina, mc, on the basis that the carbon materials prepared
according to mc possess an ideal graphene structure and a thick-
ness of 1.5 graphene layers. Our previous work demonstrated that
1.5 nominal graphene layers was thin enough to reserve the meso-
pores in the alumina but also tough enough to support itself as a
stable framework when alumina was removed by HF [17].
The resin/alumina composite was carbonized under N2 in a tem-
perature range from 500 ◦C to 1000 ◦C to investigate the effect of
carbon surface properties on catalytic performance. The resulting
carbon/alumina composite was immersed in HF solution to dissolve
alumina and obtain the carbon material C-Ax-Ts.
C-Ax-Ts displayed a distinct mesoporous structure, thin pore
walls and a high BET surface area. Taking C-AI-900 as an example
(Fig. 1), the sample exhibited a typical type IV adsorption isotherm
with a hysteresis loop and peaks in the mesopore range in the PSD
curve. Furthermore, a PSD peak for C-AI-900 resided close to that
of its alumina template, AI (vertical dotted line in Fig. 1). The differ-
ence between the most probable pore size of C-AI-900 (10.1 nm)
and AI (11.8 nm), 1.7 nm, could be approximately viewed as the
Fig. 1. N2 sorption isotherms and pore size distribution plots (inset) of C-AI-900
and SC-AI-900. The vertical dotted line represents the most probable pore size of
the alumina template AI.
average thickness of the carbon walls. Thin carbon walls could also
be observed in the TEM image of C-AI-900 (Fig. 2a).
Thin walls lead to high BET surface area. As shown in Table 1,
entry 4, the specific surface area of C-AI-900 was as high as
1671 m2/g and its pore volume reached 3.60 cm3/g. Additionally,
two peaks were distinguishable in the PSD curve of C-AI-900 (Fig. 1).
In accordance with previous results, the texture of C-AI-900 dupli-
cated the pore size of the alumina templates, corresponding to one
PSD peak close to that of AI. After the template was removed, how-
ever, another type of pore with the size of the alumina particles
emerged, corresponding to the other PSD peak centered at 6.6 nm.
Because the most probable size of the alumina pores and particles
was discrepant, a bimodal PSD was observed for C-AI-900.
The sulfonations of C-Ax-Ts were achieved through reacting the
mesoporous carbon materials with diazonium salt 4-BDS at 5 ◦C.
This method efficiently grafted benzenesulfonic acid groups onto
the carbon surface [20], transforming the carbon materials into
excellent solid acid catalysts. Compared with the traditional sul-
fonation method (i.e., boiling in concentrated H2SO4), diazonium-
salt modification is carried out under relatively mild conditions
[26,27]. In the interest of keeping the mesoporous structure intact,
this method seems to be more suitable for the functionalization of
carbon materials with large pores, thin walls and large BET areas.
As shown in the TEM images (Fig. 2), the morphology of sulfonated
Fig. 2. TEM images of C-AI-900 (a) and SC-AI-900 (b).

140
L. Geng et al. / Applied Catalysis A: General 427–428 (2012) 137–144
Table 1
Textural properties of alumina templates, carbon materials and catalysts and the catalytic performance of the acid catalysts.
Entry
Sample
BET area (m2/g)
Da (nm)
Vb (cm3/g)
Strong acid density (mmol H+/g)
Conv.c (%)
k (h−1
)
TOF (h−1
)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
AI
AII
AIII
127
176
208
15.4
9.0
0.49
0.40
1.02
3.60
2.39
1.75
2.43
2.39
2.20
1.94
1.27
0.59
0.31
19.7
8.63
8.16
5.96
9.77
8.16
7.88
7.50
5.72
3.74
29.1
C-AI-900
SC-AI-900
SC-AII-900
SC-AIII-900
SC-AI-1000
SC-AI-800
SC-AI-700
SC-AI-600
SC-AI-500
Amberlyst-15
H2SO4
1671
1118
1175
996
1170
1114
1036
886
1.86
1.84
1.83
1.88
1.61
1.58
1.47
1.16
5.03
20.4
77.7
66.5
66.1
76.7
72.8
64.6
53.9
34.4
40.5
90.7d
1.34
128
636
43
0.43
2.25
15
214
a
Average pore diameter.
Pore volume.
b
c
Conversion of oleic acid. Reaction conditions: 1 g of oleic acid, 8 ml of methanol, 20 mg of catalyst, and 65 ◦C for 1.25 h.
Catalytic acid amount (H+) of H2SO4 equal to that of SC-AI-900.
d
carbon SC-AI-900 had a similar disordered mesoporous structure
to that of C-AI-900. In addition, the texture of SC-AI-900 preserved
the bimodal PSD observed in C-AI-900. The 13C NMR spectra of
SC-AI-900 (Fig. 3a) showed resonance peak at 128 ppm, 155 ppm
and 171 ppm, which are attributed to aromatic carbon, phenolic-
OH and COOH, respectively [13]. The XRD patterns (Fig. 3b) also
showed that both C-AI-900 and SC-AI-900 were poorly crystallized.
The successful attachment of benzenesulfonic acid groups onto
the carbon matrix was confirmed by FT-IR spectra (Fig. 3c).
Compared with raw carbon material C-AI-900, the S O vibra-
tion bands at 1010 cm−1 (symmetric stretching) and 1034 cm−1
(asymmetric stretching) could be distinguished in the FT-IR
spectra of sulfonated SC-AI-900, proving the presence of SO3H
on the catalyst surface [28]. The vibrational band at 837 cm−1
could be attributed to the out-plane bending of the C H of
1,4-disubstituted benzene, suggesting the Ph-SO3H groups were
attached to the carbon surface para to SO3H. This conformation was
consistent with the molecular structure of the sulfonating agent
4-BDS.
The grafting of benzenesulfonic acid groups could also be con-
firmed by N2 absorption analysis. As seen in Fig. 1 and Table 1
(entries 4 and 5), the BET area, average pore diameter and pore vol-
ume of C-AI-900 dramatically decreased after sulfonation. It was
therefore reasonable to infer that Ph-SO3H groups on the carbon
surface occupied a part of the porous space, leading to an obvious
reduction in porosity. However, the mesopores of carbon materials
were not seriously blocked by these groups. As shown in Table 1,
most SC-Ax-Ts still exhibited mesoporous structures and large BET
areas approximately 1000 m2/g.
The acid density of carbon-based solid acids is a key factor
for determining the catalytic activity. In terms of sulfonated car-
bon catalysts, there are two types of acidic sites: (1) the strong
acid sites (i.e., the introduced SO3H groups) and (2) the weak acid
sites, including COOH and OH from incomplete carbonization of the
organic precursor. Generally, the acid density of carbon-based solid
acids is determined by titration with NaOH or Na2CO3 [29], which
are not able to distinguish the strong acid sites from the weak acid
sites. Given that the esterification of oleic acid is poorly catalyzed
by COOH and OH, only the strong acid sites will affect the perfor-
mance of the carbon catalysts. Therefore, the strong acid densities
of SC-Ax-Ts were calculated from the content of elemental sulfur
and are listed in Table 1, assuming all sulfur atoms were in the form
SO3H. The effectiveness of sulfonation via the diazonium salt was
confirmed by the strong acid density of SC-AI-900 (1.86 mmol/g),
which was higher than any reported carbon catalyst sulfonated by
H2SO4.
3.2. Effect of catalyst texture on catalytic activity
Since the BET area, pore size and pore volume of the carbon cat-
alysts are associated with the accessibility of reactant molecules
to the catalytic active sites, it is necessary to investigate the effect
of catalyst texture on catalytic activity to find the optimal cata-
lyst. The texture of Ph-SO3H-modified carbon catalysts was tuned
using three types of alumina templates (AI, AII and AIII) with dif-
ferent porous structures. Fig. 4 displays adsorption isotherms and
PSD curves for the three alumina templates, and the textural prop-
erties are listed in Table 1 (entries 1, 2 and 3). All three templates
were mesoporous materials, but their PSDs were different. AI and
AII exhibited concentrated PSD with the most probable pore sizes of
11.8 nm (AI) and 6.2 nm (AII), respectively. Meanwhile, AIII exhib-
ited a wide PSD peak, and the most probable pore size was 14.9 nm,
indicating AIII had considerably larger pores.
Carbonized at the same temperature (900 ◦C), the carbon cata-
lysts were prepared using the three templates. As shown in Fig. 5
and Table 1 (entries 5, 6 and 7), SC-AI-900, SC-AII-900 and SC-AIII-
900 not only duplicated the PSD of their respective templates but
also exhibited pores from the template particles (bimodal PSD). The
two peaks of SC-AII-900 overlapped each other because the particle
size of AII was close to its pore size.
Different templates led to carbon catalysts with different tex-
tures. First, the three carbon catalysts revealed different pore sizes.
The PSD of SC-AI-900 was in the range of 6–10 nm, SC-AII-900
mainly possessed smaller pores of 4–5 nm and SC-AIII-900 was rich
in pores larger than 10 nm, but small pores approximately 4 nm
were also visible. Second, SC-AII-900 and SC-AIII-900 contained
inkbottle shaped pores, which could be inferred from the presence
of false peaks on the PSD curves. These curves were obtained from
the desorption branch of the isotherm, where N2 desorption takes
place. Thus, the bottle necks of the inkbottle pores will bring about
the so-called tensile strength effect (TSE) and result in a false peak
on the PSD curves at a pore size of 3–4 nm [30]. In PSD curves for
both SC-AII-900 and SC-AIII-900, a sharp peak could be seen at a
pore size of 3.6 nm. To prove they were indeed false peaks, PSD
curves were calculated from the adsorption branch of isotherms
and are shown in Fig. 5c. On these curves, no peak was observed
at a pore size of 3.6 nm. Therefore, the presence of the false peak
was confirmed, suggesting the two carbon catalysts SC-AII-900 and
SC-AIII-900 contained pores shaped like ink bottles.
To investigate the effect of catalyst texture on catalytic activity,
SC-Ax-900s were tested as solid acids in the esterification of oleic
acid and methanol under identical reaction conditions. The cat-
alytic performance was evaluated based on the conversion of oleic

L. Geng et al. / Applied Catalysis A: General 427–428 (2012) 137–144
141
Fig. 4. N2 sorption isotherms and pore size distribution plots (inset) of AI, AII and
AIII.
the absence of inkbottle pores in SC-AI-900 should be viewed as the
primary reason for its superior catalytic performance.
The effect of catalyst pore size on activity is not clear. The length
of an oleic acid molecule is roughly 2 nm, smaller than the pore
size of all three SC-Ax-900s (8.16 nm, 5.96 nm and 9.77 nm, respec-
tively), so diffusion should be easy. Previous work on mesoporous
carbon catalysts reported that a large pore size would enhance cat-
alytic activity [15], but this work focused on carbon catalysts with
2–4 nm pores. It is therefore possible that catalyst frameworks with
too large of pores (e.g., approximately 10 nm in SC-AIII-900) are not
suitable for catalysis, as they reduce the effective BET surface area.
3.3. Effect of carbonization temperature on catalytic activity
Besides texture, carbonization temperature is another impor-
tant concern for carbon-based catalysts. To investigate its influence,
resorcinol–furfural resin was carbonized under N2 gas at 500 ◦C,
600 ◦C, 700 ◦C, 800 ◦C, 900 ◦C and 1000 ◦C using AI as the tem-
plate. First, the carbonization temperature directly determines the
amount of oxygen-containing groups on the carbon matrix. Carbon
materials pyrolyzed from organic precursors always possess these
groups (e.g., OH and COOH). Increasing the carbonization temper-
ature could eliminate these groups to some extent. As shown in
Table 2, the elemental analysis of raw C-AI-Ts demonstrated this
effect. As carbonization temperature increased, the carbon con-
tent of SC-AI-Ts increased, while oxygen and hydrogen content
decreased. The atom ratios of carbon to oxygen also clearly revealed
this trend. When the temperature was higher than 700 ◦C, how-
ever, the reduction of oxygen content was not clear, and the carbon
content of C-AI-900 (92.56%) and C-AI-1000 (92.88%) were almost
identical.
Fig. 3. 13C NMR spectra of SC-AI-900 (a), and XRD patterns (b) and FT-IR spectra (c)
of C-AI-900 and SC-AI-900.
Through affecting oxygen content of the carbon matrix, car-
bonization temperature influences the acid density of carbon
catalysts. As seen in Table 1 (entries 5 and 8–12), the strong acid
densities of SC-AI-Ts increased with carbonization temperature.
acid at 1.25 h, and the results are listed in Table 1 (entries 5, 6 and 7).
Among the three carbon catalysts, SC-AI-900 displayed the highest
activity with an oleic acid conversion of 78%. Meanwhile, the activ-
ities on SC-AII-900 and SC-AIII-900 were similar to each other and
relatively low (∼66%). Because the strong acid density of the three
carbon catalysts was almost the same (∼1.8 mmol H+/g), the higher
activity on SC-AI-900 was attributed to the catalyst texture.
The SC-AI-900 PSD curve has a much smaller false peak than the
other two catalysts (Fig. 5b), suggesting this sulfonated carbon cat-
alyst did not possess significant inkbottle pores. The bottleneck of
such pores would prevent oleic acid from diffusing into the “ink bot-
tle” to reach the active sites. Therefore, the presence of such pores
is unfavorable for catalytic reactions involving large molecules, and
Table 2
Elemental content of C-AI-Ts.
Sample
Elemental content (wt.%)
C/O (at.%)
C
H
O
C-AI-1000
C-AI-900
C-AI-800
C-AI-700
C-AI-600
C-AI-500
92.88
92.56
91.64
90.39
86.63
81.22
0.66
0.76
0.98
1.85
2.52
3.14
6.46
6.68
7.38
7.76
10.85
15.64
19.2
18.5
16.6
15.5
10.6
6.9

142
L. Geng et al. / Applied Catalysis A: General 427–428 (2012) 137–144
carbonization temperature fell below 600 ◦C, the whole network
tended to collapse to denser structures characterized by lower BET
areas and smaller pores. Notably, the high false peak in the PSD
curves indicated the emergence of many inkbottle pores.
To investigate the effect of carbonization temperature on cat-
alytic activity, SC-AI-Ts were tested in the esterification of oleic
acid under identical conditions. The activity was compared based
on oleic acid conversion at 1.25 h. As seen in Table 1 (entries 5 and
8–12), catalysts carbonized at higher temperatures displayed better
catalytic activities. For example, the activities of SC-AI-900 and SC-
AI-1000 were close, suggesting limited improvement of catalytic
performance through elevating the carbonization temperature.
Among the SC-AI-Ts, SC-AI-900 stands out as the best catalyst.
The first factor likely contributing to this activity was the high car-
bonization temperature, which tends to increase the strong acid
density of a carbon catalyst. It is well known that the acid den-
sity of sulfonated carbon is the key factor in determining catalytic
activity. The more SO3H attached to the carbon surface, the bet-
ter performance the catalyst will show. The strong acid density of
SC-AI-900 was as high as 1.86 mmol H+/g, supporting the highest
observed activity among the SC-AI-Ts. Second, a high carboniza-
tion temperature helps stabilize the carbon framework to maintain
the optimal mesoporous texture. When the carbonization tem-
perature was higher than 700 ◦C, inkbottle pores were negligible.
Structural shrinkage and a large quantity of inkbottle pores are
unfavorable for reactant transportation and give lower catalytic
activity. Third, a high carbonization temperature promotes the sur-
face hydrophobility of the catalysts. For carbon-based solid acids,
fewer oxygen-containing groups are related to higher hydropho-
bility, which is favorable for water, a product of esterification, to
desorb from the active site and promote reaction. Hence, compared
with SC-AI-Ts carbonized at lower temperatures, the hydrophobil-
ity of SC-AI-900 should also be one of the reasons for its excellent
activity.
3.4. Catalytic performance of SC-AI-900
As discussed above, through adjusting template texture and car-
bonization temperature, SC-AI-900 was developed as the optimal
catalyst. Intensive study of its catalytic activity and stability was
thus subsequently carried out. The catalytic performance of SC-AI-
900 in the esterification of oleic acid with methanol is shown in
Fig. 7. The equilibrium conversion of oleic acid after 5 h of reaction
was close to 100%. Because the amount of methanol was 66 times
that of oleic acid, the reaction could be viewed as following pseudo
first order kinetics, in which reaction rate depends solely on the
amount of oleic acid. Such a first-order kinetic plot was fit (Fig. 7)
and the rate constant k was determined to be 1.34 h−1. Based on the
strong acid density of SC-AI-900 (1.86 mmol H+/g, Table 1, entry 5),
Fig. 5. N2 sorption isotherms (a) and pore size distribution plots of SC-Ax-900s
calculated from both the desorption branch (b) and the adsorption branch (c) of
isotherms.
the TOF was calculated using formula (2) and found to be 128 h−1
.
As shown in Fig. 7, sulfuric acid, a homogeneous catalyst, has the
highest activity, with the highest k (2.25 h−1) and TOF (214 h−1). But
the kinetic plot of SC-AI-900 is quite close to that of sulfuric acid,
showing the good activity of this solid acid. Compared with the
commercial solid acid Amberlyst-15, SC-AI-900 exhibited higher
catalytic activity and efficiency. Amberlyst-15 was tested in this
esterification reaction under identical conditions. The time course
and kinetic plot are shown in Fig. 7, and k and TOF were determined
to be 0.43 h−1 and 15 h−1, respectively (Table 1 entry 13). Although
the strong acid density of SC-AI-900 (1.86 mmol H+/g) was much
lower than Amberlyst-15 (5.03 mmol H+/g), the rate constant on
SC-AI-900 was three times that of Amberlyst-15, underlining the
excellent catalytic ability of Ph-SO3H-modified mesoporous car-
bon. Notably, the TOF on SC-AI-900 reached eight times that of
Amberlyst-15, further demonstrating the superior catalytic effi-
ciency of the sulfonated carbon catalyst.
When resin was carbonized at low temperature, a large number
of oxygen-containing groups remained on the edge of the carbon
planes, occupying positions that would otherwise be attached to
Ph-SO3H and hindering the diazonium modification of the carbon
surface.
Furthermore, the carbonization temperature also affects cata-
lyst texture. Because redundant oxygen-containing groups would
soften the carbon matrix, the higher the carbonization tempera-
ture, the tougher the carbon framework formed and the less liable
the porous structure is to collapse. As seen in Fig. 6 and Table 1
(entries 5 and 8–12), when resin was pyrolyzed at temperatures
higher than 700 ◦C, the carbon framework was rigid enough to
support itself and displayed a bimodal PSD. In these cases, the
BET surface areas were all above 1000 m2/g. However, when the

L. Geng et al. / Applied Catalysis A: General 427–428 (2012) 137–144
143
Fig. 6. N2 sorption isotherms and pore size distribution plots of SC-AI-Ts.
SC-AI-900 outperformed the traditional catalyst Amberlyst-15
to a large extent because of its unique texture. Amberlyst-15, an
ion exchange resin, is composed of macropores with low BET sur-
face area (43 m2/g) and high acid density (5.03 mmol H+/g) (Table 1,
entry 13). Such a texture indicates that most of the acid sites
of Amberlyst-15 are located inside the resin network, which is
not capable of swelling well enough to be penetrated by large
molecules such as oleic acid. On the other hand, the texture of
SC-AI-900 features mesopores with thin carbon walls, a high BET
surface area (1118 m2/g, 26 times that of Amberlyst-15) and rel-
atively low acid density (1.86 mmol H+/g). These characteristics
suggest that most acid sites reside on the catalyst surface and are
readily accessible for large substrate molecules. Therefore, SC-AI-
900 shows a much higher catalytic efficiency in the esterification
of oleic acid than Amberlyst-15.
Previously, we reported a sucrose-carbon catalyst showing high
efficiency in the esterification of oleic acid and methanol (TOF
109 h−1). This catalyst consisted of thin and small carbon sheets
able to disperse in methanol, promoting the catalytic process [25].
However, SC-AI-900 from resorcinol–furfural resin exhibited an
even higher catalytic efficiency (TOF 128 h−1). Moreover, unlike
the sucrose-carbon catalyst, the monolithic carbon framework of
SC-AI-900 facilitated catalyst separation and recycling.
Due to its superior reactivity, the stability of SC-AI-900 was
evaluated by running the esterification reaction six times with
the same catalyst. As shown in Fig. 8, the conversion of oleic acid
gradually decreased in the first four cycles, but no distinct drop
in catalytic ability was observed after three cycles (conversion of
Fig. 8. Recycling of SC-AI-900 for the esterification of oleic acid (1 g) with methanol
(8 ml) at 65 ◦C. The conversions of oleic acid at 1 h are shown.
oleic acid near 60%). To investigate this deactivation, the strong
acid density of SC-AI-900 after six reaction cycles was determined
for comparison with freshly prepared SC-AI-900. The results of the
sulfur element analysis showed that after six reaction cycles, the
strong acid density of SC-AI-900 was reduced from 1.86 mmol H+/g
to 1.34 mmol H+/g. Since the catalytic ability of sulfonated car-
bon largely depends on the acid density, the loss of acid sites was
likely the main reason for catalyst deactivation. As described in the
literature, the leaching of polycyclic aromatic hydrocarbons bear-
ing SO3H groups was observed on sulfonated carbon catalysts and
led to deactivation [31,32]. However, the remaining active sites,
approximately 72% of the freshly prepared SC-AI-900, still attached
to the carbon surface after six reaction cycles and were able to give
a high conversion of oleic acid. This observation suggested that
the Ph-SO3H-modified mesoporous carbon was a stable solid acid
catalyst.
4. Conclusion
Ph-SO3H-modified mesoporous carbon was prepared using
␥-Al2O3 as the hard template, resorcinol–furfural resin in situ
polymerized on an alumina surface as the carbon precursor and
4-BDS as the sulfonating agent via diazonium chemistry. Using
inexpensive ␥-Al2O3 reduced the catalyst cost, resorcinol–furfural
resin, instead of sucrose, simplified the catalyst preparation pro-
cess, and the diazonium modification replaced H2SO4 sulfonation
while increasing strong acid density and keeping the carbon meso-
porous network intact. The Ph-SO3H-modified carbon materials
had a distinct mesoporous structure with bimodal PSD, thin pore
Fig. 7. Esterification of oleic acid with methanol over H2SO4, SC-AI-900 and
Amberlyst-15. Kinetic plot was fitted to first order reaction kinetics. Reaction con-
ditions: 1 g of oleic acid, 8 ml of methanol, 20 mg of solid catalyst or 1.82 mg H2SO4,
and 65 ◦C.

144
L. Geng et al. / Applied Catalysis A: General 427–428 (2012) 137–144
walls, ultra-high BET area and high strong acid density. These
materials showed superior catalytic performance in the esterifi-
cation of oleic acid and methanol. The optimal catalyst, SC-AI-900,
displayed much higher activity than the traditional solid acid cat-
alyst Amberlyst-15 with a rate constant of 1.34 h−1, three times
that of Amberlyt-15, and a TOF of 128 h−1, eight times that of
Amberlyst-15. In addition, SC-AI-900 could be easily separated
from the reaction system by filtration and reused. There was no
distinct activity drop during recycling after the initial deactivation,
suggesting Ph-SO3H-modified mesoporous carbon is a stable and
efficient solid acid catalyst. Such carbon-based solid acids could also
be useful for many other acid-catalyzed reactions involving large
reactant molecules.
[8] W.Y. Lou, M.H. Zong, Bioresour. Technol. 99 (2008) 8752–8758.
[9] J.A. Maciá-Agulló, M. Sevilla, M.A. Diez, A.B. Fuertes, ChemSusChem 3 (2010)
1352–1354.
[10] V. Budarin, J.H. Clark, J.J.E. Hardy, R. Luque, K. Milkowski, S.J. Tavener, A.J. Wil-
son, Angew. Chem. Int. Ed. 45 (2006) 3782–3786.
[11] V.L. Budarin, J.H. Clark, R. Luque, D.J. Macquarrie, Chem. Commun. 6 (2007)
634–636.
[12] A.M. Dehkhoda, A.H. West, N. Ellis, Appl. Catal. A: Gen. 382 (2010)
197–204.
[13] K. Nakajima, M. Hara, J. Am. Ceram. Soc. 90 (2007) 3725–3734.
[14] M. Hara, ChemSusChem 2 (2009) 129–135.
[15] X.H. Mo, E. Lotero, C.Q. Lu, Y.J. Liu, J.G. Goodwin, Catal. Lett. 123 (2008) 1–6.
[16] B.V.S.K. Rao, K. Chandra Mouli, N. Rambabu, A.K. Dalai, R.B.N. Prasad, Chem.
Commun. 14 (2011) 20–26.
[17] Y. Wang, H., Dong, L., Geng, G., Yu, Y.X., Zhu, Y.C. Xie, Acta Phys. Chim. Sin., in
press.
[18] L. Peng, A. Philippaerts, X.X. Ke, J. Van Noyen, F. De Clippel, G. Van Tendeloo,
P.A. Jacobs, B.F. Sels, Catal. Today 150 (2010) 140–146.
[19] J. Janaun, N. Ellis, Appl. Catal. A: Gen. 394 (2011) 25–31.
[20] X.Q. Wang, R. Liu, M.M. Waje, Z.W. Chen, Y.S. Yan, K.N. Bozhilov, P.Y. Feng,
Chem. Mater. 19 (2007) 2395–2397.
Acknowledgments
The authors are grateful for financial support from the National
Science Foundation of China (20773004) and the Major State Basic
Research Development Program (Grant No. 2011CB808702).
[21] R. Liu, X.Q. Wang, X. Zhao, P.Y. Feng, Carbon 46 (2008) 1664–1669.
[22] L. Lin, W. Lin, Y.X. Zhu, B.Y. Zhao, Y.C. Xie, G.Q. Jia, Langmuir 21 (2005)
5040–5046.
[23] L. Lin, P. Wang, S.R. Wang, Y.X. Zhu, B.Y. Zhao, Y.C. Xie, Carbon 44 (2006)
3120–3123.
References
[24] Y. Wang, G. Yu, B. Cai, Y.X. Zhu, Y.C. Xie, Acta Phys. Chim. Sin. (2011) 729–
735.
[25] L. Geng, Y. Wang, G. Yu, Y.X. Zhu, Catal. Commun. 13 (2011) 26–30.
[26] A. Stein, Z.Y. Wang, M.A. Fierke, Adv. Mater. 21 (2009) 265–293.
[27] F. Barrière, A.J. Downard, J. Solid State Electrochem. 12 (2008) 1231–1244.
[28] W.L. Zhang, H.X. Tao, B.H. Zhang, J.W. Ren, G.Z. Lu, Y.Q. Wang, Carbon 49 (2011)
1811–1820.
[29] H.P. Boehm, Carbon 40 (2002) 145–149.
[30] J.C. Groen, L.A.A. Peffer, J. Perez-Ramirez, Microporous Mesoporous Mater. 60
(2003) 1–17.
[1] R. Jothiramalingam, M.K. Wang, Ind. Eng. Chem. Res. 48 (2009) 6162–6172.
[2] R.A. Sold, A.R.S. Oliveira, L.P. Ramos, M.A.F. César-Oliveira, Appl. Catal. A: Gen.
361 (2009) 42–48.
[3] M. Verziu, M. Florea, S. Simon, V. Simon, P. Filip, V.I. Parvulescu, C. Hardacre, J.
Catal. 263 (2009) 56–66.
[4] J.A. Melero, J. Iglesias, G. Morales, Green Chem. 11 (2009) 1285–1308.
[5] M. Hara, T. Yoshida, A. Takagaki, Angew. Chem. Int. Ed. 43 (2004) 2955–2958.
[6] M. Toda, A. Takagaki, M. Okamura, J.N. Kondo, S. Hayashi, K. Domen, M. Hara,
Nature 438 (2005) 178.
[7] M.H. Zong, Z.Q. Duan, W.Y. Lou, T.J. Smith, H. Wu, Green Chem. 9 (2007)
434–437.
[31] X.H. Mo, D.E. López, K. Suwannakarn, Y. Liu, E. Lotero, J.G. Goodwin, C.Q. Lu, J.
Catal. 254 (2008) 332–338.
[32] G. Chen, B. Fang, Bioresour. Technol. 102 (2011) 2635–2640.