Transesterification of triglycerides using nitrogen-functionalized carbon nanotubes

Article abstract of DOI:10.1002/cssc.200900181

Nitrogen-functionalized carbon nanotubes were synthesized by grafting amino groups to the surface of the nanotubes. The nanotubes exhibited promising results in the base-catalyzed liquid phase transesterification of glyceryl tributyrate with methanol, which is a model reaction for the production of biodiesel. The concentration of the active sites and the reaction parameters, such as temperature and glyceryl tributyrate to methanol ratio, were shown to significantly affect catalytic performance. The grafting technique employed allowed for design and control of the active sites. As a consequence, it was possible to design a nitrogen-functionalized carbon nanotube catalyst with a few strong, basic groups. This might be of interest for carbohydrate conversion reactions where strong basic sites are required but the pH of the solution should remain mild to avoid the degradation of the reactants and/or products.

Full text of DOI:10.1002/cssc.200900181

DOI: 10.1002/cssc.200900181
Transesterification of Triglycerides Using Nitrogen-
Functionalized Carbon Nanotubes
Alberto Villa, Jean-Philippe Tessonnier, Olivier Majoulet, Dang Sheng Su,* and
Robert Schlçgl[a]
Nitrogen-functionalized carbon nanotubes were synthesized
by grafting amino groups to the surface of the nanotubes. The
nanotubes exhibited promising results in the base-catalyzed
liquid phase transesterification of glyceryl tributyrate with
methanol, which is a model reaction for the production of bio-
diesel. The concentration of the active sites and the reaction
parameters, such as temperature and glyceryl tributyrate to
methanol ratio, were shown to significantly affect catalytic per-
formance. The grafting technique employed allowed for
design and control of the active sites. As a consequence, it
was possible to design a nitrogen-functionalized carbon nano-
tube catalyst with a few strong, basic groups. This might be of
interest for carbohydrate conversion reactions where strong
basic sites are required but the pH of the solution should
remain mild to avoid the degradation of the reactants and/or
products.
Introduction
Chemical routes to biodiesel from renewable resources such as
plant oils have attracted much attention over the last years.[1–3]
The conventional production of biodiesel from lipid feedstocks
involves the use of homogenous basic catalysts such as
sodium or potassium hydroxides, or alkoxides, which are very
active, but also present many problems.[4,5] These processes re-
quire an aqueous wash step to remove the dissolved catalyst
and glycerol from the product phase. This step often suffers
from separation problems owing to the formation of emul-
sions, and also leads to waste treatment issues. The substitu-
tion of homogeneous catalysts with heterogeneous catalysts is
crucial to eliminate these problems. A wide variety of solid
bases have been examined for this purpose, for example, alka-
line earth oxides,[6] potassium and lithium-promoted oxides,[7]
calcined hydrotalcites,[8] zeolites,[9] and nano-magnesium
oxide.[10] However, most of these catalysts do not dissolve in
the reaction media. These solids are Lewis bases and in the
presence of polar species, for example methanol and glycerol,
they can form methoxylates and glycerates with the conse-
quence of leaching from the solid catalyst. Therefore, many ef-
forts have been made to develop basic heterogeneous cata-
lysts that show high stability throughout transesterification re-
actions. Liu et al. reported the high catalytic activity of amino-
necessary to study new supports, which are easy to synthesize
and to functionalize. Carbon nanotubes (CNTs) are suitable for
this purpose as they can be chemically modified with heteroa-
toms, which are either inserted in their structure[12–18] or graft-
ed on their surface.[19,20]
We recently reported on the importance of the methodolo-
gy in grafting organic amine molecules on the surface of CNTs
to obtain active and stable catalysts for the transesterification
of glyceryl tributyrate.[21,22] In particular, we developed a syn-
thetic route where CNTs are reacted with an excess of n-butyl-
lithium (nBuLi), followed by addition of bromo amino derivates
to the activated CNTs to form the corresponding nitrogen-
CNTs (N-CNTs).[21,22] The catalysts show high stability during re-
generation/recycling cycles, thus proving that the amino
groups are strongly anchored to the surface of the CNTs.[22] We
also proved that the basicity of the catalyst can be tuned by
grafting amines with different pKa values. Herein, we studied
the effect the amount of grafted amines has on the basicity of
the catalyst. Furthermore, we studied the effect the reaction
parameters, such as temperature and molar ratio of glyceryl
tributyrate to methanol, have on catalytic performance.
functionalized silica and resin in the transesterification of tria- Results and Discussion
cetin with methanol.[11] They also highlighted a partial leaching
We previously demonstrated that the synthetic procedure for
of the grafted amino functionalities using silica as the support,
whereas with resin this problem was not observed. It appears
the support plays an important role on the stability and activi-
ty of the catalyst. Resin-based catalysts show better stability,
however, several groups observed that approximately 65% of
the active sites are not accessible, the active sites are poorly
dispersed, and their basic character is strongly influenced by
the structure of the resin.[11] In few cases, a strong deactivation
was also observed in the case of the anionic resins. It is thus
grafting N-containing groups onto the surface of multi-walled
CNTs (MWCNTs) has a dramatic role on the subsequent catalyt-
[a] Dr. A. Villa, Dr. J.-P. Tessonnier, O. Majoulet, Dr. D. S. Su, Prof. R. Schlçgl
Department of Inorganic Chemistry
Fritz Haber Institute of the Max Planck Society
Faradayweg 4–6, 14195 Berlin (Germany)
Fax: (+49)30 8413 4401
http://www.fhi-berlin.mpg.de/acnew/welcome.epl
E-mail: dangsheng@fhi-berlin.mpg.de
ChemSusChem 2010, 3, 241 – 245
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
241

D. S. Su et al.
ic activity of the N-CNTs.[21,22] The most common synthetic
route consists of three steps; oxidation of the surface to create
carboxylic acid groups, acylation with thionyl chloride, and the
anchoring of the desired molecule by amidation.[19,20] Although
this technique is efficient, we found that the proximity of
acidic O-containing functional groups to basic amino groups
leads to weakly basic samples, showing poor catalytic activity
and poor stability. This observation is consistent with the re-
sults obtained by several groups when grafting amines on
silica or SBA-15 materials.[11] Therefore, we developed a syn-
thetic route where the desired molecules are directly grafted
to the surface defects by CÀC coupling reactions. Indeed, con-
trary to single-walled CNTs, MWCNTs synthesized by catalytic
chemical vapor deposition (CCVD) exhibit a number of struc-
tural defects, such as vacancies and terminations, where the
carbon atoms are saturated by hydrogen atoms. In our proce-
dure, commercial MWCNTs are reacted with nBuLi to activate
the terminal CÀH bonds near the defects and create nucleo-
philic carbon atoms (CÀLi).[21] In a second step, a bromoalkyla-
mine is added to perform an electrophilic attack on the CÀLi
bonds. As a result, the desired molecule is directly anchored to
the MWCNT surface through the created CÀC bond. Lithium
bromide is formed as a side-product. This synthetic route is an
elegant way to graft various amines, for example triethylamine,
ethylamine, or pyrrolidine, which can be employed as active
sites for base-catalyzed reactions.[22] Herein, we investigated
the role of the concentration of the active sites on the subse-
quent catalytic activity of the N-CNTs. The number of active
sites was controlled by adjusting the concentrations of nBuLi
and 2-bromo-N,N-diethyl-ethylamine used during the synthe-
sis.
Inductively coupled plasma-atomic emission spectroscopy (ICP-
AES) analysis of this sample revealed the presence of lithium
(2.6% of the Li added). Thus, the washing procedure was im-
proved by washing the catalyst overnight, and in this case, no
lithium was detected. To confirm the role that lithium can have
on the transesterification reaction, we performed catalytic tests
using both the fast-washed and long-washed catalysts. The
catalysts were tested in the transesterification of glyceryl tribu-
tyrate with methanol to form methyl butanoate. Figure 1
Figure 1. Glyceryl tributyrate conversion after the fast and improved wash-
ing procedures (catalyst, Et3N-CNTs 3).
shows the observed catalytic activities and, as expected, the
sample containing lithium was more active. Thus, all samples
were thoroughly washed with methanol to exclusively study
the catalytic performances of the grafted amines.
The basicity properties and the number of accessible basic
sites were characterized by acid-base titration with HCl as the
titrant (Table 1). The pH generated by the N-CNTs suspended
Experiments were performed to study the possible effect on
the catalytic activity of a pre-contact between the catalyst and
reagents. In the first test the catalyst was introduced at the
same time as glyceryl tributyrate and methanol, whereas in the
second experiment the catalyst and methanol were stirred for
ten minutes before addition of the oil.
Table 1. Properties of Et3N-CNTs catalysts.
Catalyst
pH
Basic site density
[mmolgÀ1
]
Figure 2 shows that the activity is drastically lowered with-
out the catalyst–methanol pre-contact. This decrease in activity
can be attributed to the strong adsorption of glyceryl tributy-
rate to the inner and external surfaces of the carbon nano-
tubes (Figure 3). The strong adsorption of the oil to the active
sites is also the reason for deactivation during recycling
tests.[22] Thus, to have an active system it is necessary to
adsorb methanol onto the active sites before the introduction
of glyceryl tributyrate.
Et3N-CNTs 1
Et3N-CNTs 2
Et3N-CNTs 3
7.8
9.9
10.3
0.11
0.73
1.00
in a 10À3 m KCl solution was 7.8, 9.9, and 10.3, for Et3N-CNTs 1,
Et3N-CNTs 2, and Et3N-CNTs 3, respectively. The total number of
basic groups was calculated by the amount of HCl added and
was found to be 0.11, 0.73, and 1.00 mmolgÀ1 for 1, 2, and 3,
respectively. In all cases the yield of the grafting procedure
ranged from 15–19%.
The effect on catalytic activity of various loadings of the
amino functionality on the CNTs was investigated. The typical
reaction conditions of 608C with a glyceryl tributyrate to meth-
anol ratio of 1:12 were used (Figure 4). The catalyst with the
highest concentration of basic groups (Et3N-CNTs 3) proved to
be most active. In particular, Et3N-CNTs 1 showed a very low
activity and immediately deactivated. The abundance of amino
groups is a key factor for highly active catalysts.
Attention was dedicated to the washing of the catalyst; the
presence of trace amounts of lithium in the catalyst may con-
siderably change its catalytic activity as lithium is known to
catalyze transesterification reactions.[7] To study the effect of
the presence of lithium, the N-CNTs samples were washed
after synthesis following two different procedures. The first
procedure was a fast washing of the catalyst with methanol.
242
www.chemsuschem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2010, 3, 241 – 245

Transesterification of Triglycerides
To compare the grafting route with more classical methods
for nitrogen functionalization, we compared our best catalyst,
3, to a N-CNTs sample prepared by a gas phase procedure.[23]
In this case, the N-CNTs were obtained by previous acid oxida-
tion and further amination with gaseous ammonia. X-ray pho-
toelectron spectroscopy (XPS) and thermogravimetry mass
spectrometry (TG-MS) revealed that these samples mainly con-
tain pyridinic and pyrrolic groups. Hence, they are similar to N-
CNTs synthesized by CCVD.[18] The performance of the catalysts
were compared (Figure 5); the N-CNTs synthesized by gas
phase led to a very low conversion of 3% compared to 77%
for Et3N-CNTs 3.
Figure 2. Influence of the methanol–catalyst pre-contact (temperature,
608C; glyceryl tributyrate/methanol 1:12, 2% catalyst).
Figure 5. Catalytic activities of Et3N-CNTs 3 and gas phase synthesized N-
CNTs for the glyceryl tributyrate transesterification (608C, glyceryl tributy-
rate/methanol 1:12; 2% catalyst).
The effects of the reaction parameters on the catalytic activi-
ties were also investigated. Figure 6 shows the influence of the
reaction temperature. Increasing the reaction temperature leds
not only to a faster reaction but also improved phase miscibili-
Figure 3. TEM image of Et3N-CNTs 3 after transesterification.
Figure 4. Effect of the amount of grafted triethylamine on the activity in the
transesterification of glyceryl tributyrate (temperature, 608C; glyceryl tributy-
rate to methanol ratio, 1:12; 2 wt% catalyst).
Figure 6. Influence of the reaction temperature on the transesterification of
glyceryl tributyrate (glyceryl tributyrate to methanol ratio, 1:12; catalyst,
Et3N-CNTs 3, 2 wt%).
ChemSusChem 2010, 3, 241 – 245
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
243

D. S. Su et al.
ty, in particular when using heavy oils. At 908C the glyceryl
tributyrate was fully converted in 2 h, whereas at 758C it took
6 h to achieve 90% conversion (Figure 7).
operating conditions. High temperatures and high methanol
to glyceryl tributyrate molar ratios improved the catalytic per-
formances. The limiting factors were the vapor pressure of
methanol and the energy required for the separation of the
product from methanol. Optimal reaction parameters were
identified with respect to these limitations. The results may
serve as orientation for other applications of N-CNTs in syn-
thetic chemistry. For practical applications, scalable procedures
need to be identified, as the gas phase ammonolysis or CCVD
techniques do not seem suitable.
Experimental Section
The reactants for the synthesis of catalysts and those used in the
catalytic tests, that is, glyceryl tributyrate (98%), methanol (99.8%),
dimethyl ether, n-butyllithium (2.5m in n-hexane, and 2-bromo-
N,N-diethyl-ethylamine (98%) were purchased from Aldrich. n-
Hexane obtained from Sigma–Aldrich (>99%) and hexyl ether
from Aldrich (97%) were used respectively as the solvent and the
internal standard for gas chromatography (GC) analysis. MWCNTs
were purchased from Nanocyl (Nanocyl 3100, >95% C purity).
Three N-CNT catalysts with different amounts of grafted triethyla-
mino groups were prepared following the published procedure.[21]
Pristine CNTs (2 g) were added to freshly distilled dimethyl ether
(400 mL) and placed in an ultrasonic bath for 20 min at RT. nBuLi
(2.5m solution in n-hexane) was then added and the mixture was
stirred for 5 h under argon at RT. 2-Bromo-N,N-diethyl-ethylamine
was added and the solution was heated at 358C for 90 min. The
samples were washed following two different procedures; for the
fast washing procedure, the N-CNTs were simply filtered on a Bꢁch-
ner funnel and rinsed with methanol (300 mL). For the improved
washing procedure, the N-CNTs were first filtered and rinsed fol-
lowing the fast washing procedure. They were then suspended in
methanol (600 mL) and stirred for overnight. Finally, they were fil-
tered and rinsed with distilled water (500 mL) and methanol
(200 mL). In all cases, the catalyst was dried overnight at 808C. The
different amounts of nBuLi and of 2-bromo-N,N-diethyl-ethylamine
used in the synthesis are shown in Table 2.
Figure 7. Influence of the ratio of glyceryl tributyrate to methanol on the
transesterification of glyceryl tributyrate (temperature, 608C; catalyst, Et3N-
CNTs 3, 2 wt%).
The use of higher temperatures increases energy consump-
tion. Moreover, at higher temperatures a higher pressure is
necessary to keep the methanol in liquid phase. Hence, the
best temperature despite the transport limitation is still 608C,
which is below the boiling point of methanol. The molar ratio
of the oil and methanol is another important factor in the
transesterification of triglycerides. Three moles of methanol
per mole of triglyceride are required stoichiometrically. Howev-
er, as the transesterification is a reversible reaction, a large
excess of methanol is required to displace the equilibrium. In
our study, different glyceryl tributyrate to methanol molar
ratios (1:60, 1:24, 1:12, 1:6) were used, at a fixed temperature
of 608C and a fixed amount of catalyst (2 wt%). By increasing
the amount of methanol the system became more active.
When using molar ratios of 1:60 or 1:24, complete conversion
was achieved after 2 h and 4 h, respectively. With a ratio of
1:12, only 77% conversion was reached after 8 h and in the
case of the ratio 1:6, fast deactivation was observed. Despite
high activity, an optimal molar ratio must balance the cost of
separation of the main products from the excess methanol.
Hence, a ratio of 1:12 can be regarded as the optimal ratio.
Transesterification reactions were performed in a thermostatted
glass reactor (30 mL) with samples periodically withdrawn for anal-
ysis. The reactants were always loaded into the reactor at RT. The
reactor was then purged with nitrogen to expel air before heating
up and pressurized with nitrogen. The catalyst concentration was
2% based on glyceryl tributyrate weight. The quantification of the
products was performed using an Agilent 6890 N GC equipped
with a HP-5 column (30 m length, 0.32 mm internal diameter,
0.25 mm film thickness) and free induction decay (FID) detector.
GC-MS was used for identification of the products.
The Li content was analyzed by ICP-AES using a Jobin Yvon JY24
instrument. Morphologies and microstructures of the catalysts
were characterized by transmission electron microscopy (TEM). The
Conclusions
Three N-CNTs were synthesized following a published proce-
dure.[21] When the amount of grafted triethylamine was in-
creased, the basicity and the catalytic performance in the
transesterification of glyceryl tributyrate increased. An activity
ten times greater than that for N-CNTs synthesized by ammo-
nolysis of oxygen functional groups was observed. The catalyt-
ic performance was significantly improved if the catalyst was
pre-contacted with methanol or through optimization of the
Table 2. Amount of reagents.
Catalyst
nBuLi
2-Bromo-N,N-diethyl-ethylamine
[mmol]
[mmol]
Et3N-CNTs 1
Et3N-CNTs 2
Et3N-CNTs 3
2.75
27.5
55
2.75
27.5
50
244
www.chemsuschem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2010, 3, 241 – 245

Transesterification of Triglycerides
[9] E. Leclercq, A. Finiels, C. Moreau, J. Am. Oil Chem. Soc. 2001, 78, 1161.
[10] L. Wang, J. Yang, Fuel 2007, 86, 328.
[11] Y. Liu, E. Lotero, J. G. Goodwin Jr. , C. Lu, J. Catal 2007, 246, 428.
[12] M. Terrones, R. Kamalakaran, T. Seeger, M. Rꢁhle, Chem. Commun. 2000,
2335.
powder samples of the catalysts were ultrasonically dispersed in
chloroform and mounted onto holey carbon coated copper grids.
A Philips CM200 FEG electron microscope, operating at 200 kV and
equipped with a Gatan imaging filter, GIF100, was used for TEM
observation. A Mettler Toledo titrator equipped with a DGi 114-SC
electrode was used for acid-base titration measurements
[13] C. Tang, D. Goldberg, Y. Bando, F. Xu and B. Liu, Chem. Commun. 2003,
3050.
[14] S. H. Lim, H. I. Elim, X. Y. Gao, A. T. S. Wee, W. Ji, J. Y. Lee, J. Lin, Phys. Rev.
B 2006, 73, 45402.
[15] R. Czerw, M. Terrones, J. C. Charlier, X. Blasꢂ, B. Foley, R. Kamalakaran, N.
Grobert, H. Terrones, D. Tekleab, P. M. Ajayan, W. Blau, M. Rꢁhle, D. L.
Carroll, Nano Lett. 2001, 1, 457.
[16] S. Souto, M. Pickholz, M. C. dos Santos, F. Alvarez, Phys. Rev. B 1998, 57,
2536.
[17] H. Sjçstrçm, S. Stafstrom, M. Boman, J. E. Sundgren, Phys Rev Lett. 1995,
75, 1336.
Acknowledgements
The project was partly conducted under the EnerChem Project of
Max Planck Society. DAAD is gratefully acknowledged for finan-
cial support.
[18] S. van Dommele, K. P. de Jong, J. H. Bitter, Chem. Comm. 2006, 4859.
[19] D. Tasis, N. Tagmatarchis, A. Bianco, M. Prato, Chem. Rev. 2006, 106,
1105.
Keywords: basicity
transesterification
· biodiesel · catalysis · nanotubes ·
[20] Y. P. Sun, K. Fu, Y. Lin, W. Huang, Acc. Chem. Res. 2002, 35, 1096.
[21] J.-P. Tessonnier, A. Villa, O. Majoulet, D. Su, R. Schlçgl Angew. Chem.
2009, 121, 6665; Angew. Chem. Int. Ed. 2009, 48, 6543; Angew. Chem.
Int. Ed. 2009, 48, 6543.
[1] E. Lotero, J. G. Goodwin Jr., D. A. Bruce, K. Suwannakarn, Y. Liu, D. E.
Lopez, in Catalysis (Ed.: J. Spivey), Royal Society of Chemistry, London,
2006.
[22] A. Villa, J.-P. Tessonnier, O. Majoulet, D. S. Su, R. Schlçgl, Chem. Comm.
2009, 4405.
[23] R. Arrigo, M. Hꢃvecker, R. Schlçgl, D. S. Su, Chem. Commun. 2008, 4891.
[2] F.R Ma, M. A. Hanna, Bioresour. Technol. 1999, 70, 1.
[3] L. C. Meher, D. V. Sagar, S. N. Naik, Renew. Sust. Ener. Rev. 2006, 10, 248.
[4] J. Van Gerpen, Fuel Process. Technol. 2005, 86, 1097.
[5] K. G. Georgogianni, M. G. Kontaminas, E. Tegou, D. Avlonitis, V. Gergis,
Energy Fuels 2007, 21, 3023.
[6] G. R. Peterson, W. P. Scarrah, J. Am. Oil Chem. Soc. 1984, 61, 1593.
[7] W. Xie, H. Peng, L. Chen, Appl. Catal. A 2006, 300, 67.
[8] D. G. Cantrell, L. J. Gillie, A. F. Lee, K. Wilson, Appl. Catal. A 2005, 287,
183.
Received: July 29, 2009
Published online on November 11, 2009
ChemSusChem 2010, 3, 241 – 245
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
245