Complementarity of heterogeneous and homogeneous catalysis for oleic acid esterification with trimethylolpropane over ion-exchange resins

Article abstract of DOI:10.1016/j.catcom.2014.10.033

This work investigates the effect of Amberlyst 36, Purolite CT482 and Purolite CT275DR on the esterification of oleic acid with trimethylolpropane for production of the triester. The results assess that, while the three resins catalyze heterogeneously the

Full text of DOI:10.1016/j.catcom.2014.10.033

Complementarity of Heterogeneous and Homogeneous Catalysis for Oleic
Acid Esterification with Trimethylolpropane Over Ion-Exchange Resins
Maryna Kuzminska, Re´nal Backov, Eric M. Gaigneaux
PII:
DOI:
Reference:
S1566-7367(14)00452-X
doi: 10.1016/j.catcom.2014.10.033
CATCOM 4121
To appear in:
Catalysis Communications
Received date:
Revised date:
Accepted date:
7 October 2014
28 October 2014
30 October 2014
Please cite this article as: Maryna Kuzminska, R´enal Backov, Eric M. Gaigneaux, Com-
plementarity of Heterogeneous and Homogeneous Catalysis for Oleic Acid Esterification
with Trimethylolpropane Over Ion-Exchange Resins, Catalysis Communications (2014),
doi: 10.1016/j.catcom.2014.10.033
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

ACCEPTED MANUSCRIPT
Complementarity of Heterogeneous and Homogeneous Catalysis for Oleic
Acid Esterification with Trimethylolpropane
Over Ion-Exchange Resins
Maryna Kuzminska,1, 2 Rénal Backov2, Eric M. Gaigneaux,*, 1
1 Université catholique de Louvain, Institute of Condensed Matter and Nanosciences - IMCN,
Croix du Sud 2, L7.05.17, B-1348 Louvain-la-Neuve, Belgium
2 Université de Bordeaux, Centre de Recherche Paul Pascal, CRPP-UPR CNRS 8641, 115 av.
du Dr. A. Schweitzer, F-33600 Pessac, France
*eric.gaigneaux@uclouvain.be, maryna.kuzminska@uclouvain.be
Abstract This work investigates the effect of Amberlyst 36, Purolite CT482 and Purolite
CT275DR on the esterification of oleic acid with trimethylolpropane for production of the
triester. The results assess that, while the three resins catalyze heterogeneously the formation
of mono- and diesters, the final triester is generated thanks to the acidity of residual oleic acid.
This cooperative scenario between heterogeneous and homogeneous catalysis is of first
importance, as to complete the reaction over conventional heterogeneous catalyst was not
feasible because of the strong steric hindrance of the intermediate product (diester) and the
final product (triester).
Introduction Esterification of oleic acid (OA) with trimethylolpropane (TMP) is used in the
industry for the production of a biolubricant, namely trimethylolpropane trioleate (TMPTO)
[1]. TMPTO possesses useful physical properties such as high lubricity and viscosity indexes,
high thermal stability, biodegradability, and so forth [2].
To our knowledge only
homogeneous catalysts are being used to synthesize TMPTO at the industrial scale (e.g. p-
toluenesulfonic acid, tin oxalate). As many homogeneous catalytic processes, the latter
reaction requires additional separation, neutralization and purification steps. To minimize the
separation procedure, in the industrial scale the reaction is being performed without additional
solvent through stoichiometric conditions. Additionally, the replacement of homogeneous
catalyst by heterogeneous one could bring even more benefits in the practical, economic and
ecological aspects. This is the reason why extensive research is dedicated toward the finding
of efficient heterogeneous catalyst which could be used in the industry [3-6].
1

ACCEPTED MANUSCRIPT
To our knowledge, the available studies related to esterification of OA with TMP over
heterogeneous catalysts are focused mainly on the comparative investigation of the catalyst
activity and selectivity [1, 3, 6, 7]. Even though these reports often claim that the catalysts
have high selectivity towards triester (50-90%), evaluation of real possibilities to use
heterogeneous catalysts in industrial conditions for the synthesis of TMPTO is missing. As
the presence of free fatty acids and mono- and/or diesters negatively affect the properties of
final lubricant, the industrial application of these catalysts is valid only if the reaction is
complete, with nearly 100% of OA conversion (selectivity towards triester being minimum
98% ) [2]. That is also why the esterification of OA with TMP should be performed with the
stoichiometric ratio of reagents to avoid additional separation of OA.
Here we examine for the first time the possibility of acidic macroporous ion-exchanged resins
to complete the tri-esterification of OA with TMP which is challenged by the industrial needs
(the possibility to reach > 98% of triester selectivity).
Materials and methods
Chemicals. Amberlyst 36 (mesh 600-850 µm) (Sigma), Purolite CT482 (425-1200 µm) and
Purolite CT275DR (650 - 900 µm) was kindly provided by Purolite company , oleic acid 90%
(Alfa Aesar), trimethylolpropane 98% (Alfa Aesar), pentadecane 99% (Sigma), N,O-
bis(trimethylsilyl) trifluoroacetamide (BSTFA) (Alfa Aesar). Acetonitrile, hexane and
dimethylformamide (DMF) were of HPLC grade.
Leaching test in DMF (elemental analysis). 0,20 g of resin was placed into 10 ml of DMF and
refluxed at 120°C for 19h. The amount of leached sulfur was measured in the filtered solution
by Inductively Coupled Plasma-Atomic Emission Spectroscopy (ICP AES).
Esterification of OA with TMP. 5.4 g OA and 0.86 g TMP (molar ratio 3:1) were placed into
the batch reactor. The mixture was heated for 1 min to melt TMP, after which the catalyst
(1.6% wt.) was added. The start of the reaction was considered after all components were
mixed and the reactor was placed in an oil bath at 120°C (unless specified otherwise). The
reaction was performed under magnetic agitation at 400 rpm and in an open reactor (unless
specified otherwise). For sampling, 10 µl of the reaction mixture was taken at certain time
intervals, derivatized as described below and followed by gas chromatography (GC). The
separation on the GC system was performed with the method described in [8]. For the
2

ACCEPTED MANUSCRIPT
quantification, internal standard (pentadecane) was used (was added in the beginning of the
esterification). In the chromatogram the evolution of mono-, di- and triesters was followed.
Control tests. Test with H2SO4 was performed under the same conditions as described above
for the esterification. A “blank” test was performed under identical conditions but without the
catalyst.
Derivatization of esterification products. Esterification products were derivatized with
BSTFA following the protocol [8] with some modifications. 10 µl of the mixture was
solubilized with 2 ml of acetonitrile. Then 500 µl of BSTFA was added and the final mixture
was heated at 60°C for with vigorous agitation in a closed vessel. In 30 min the final products
were extracted with 2 ml of hexane. The hexane extract was analyzed by GC. Standard
deviations of the GC measurements were in the range of 0.04-0.20 (mol/l) for the detection of
monoester, 0.03-0.15 for diester and 0.01-0.11 for the detection of triester.
Conversion of OA. The conversion of OA was calculated from the titration of the reaction
medium at certain time intervals with solution of 0.1N KOH, according to the equation:
Conversion, % = (Initial amount of OA - Final amount of OA)×100/Initial amount of OA.
Results and discussions
Three resins were studied in this work: Amberlyst 36 (Amb), Purolite CT482 (Pur482) and
Purolite CT275DR (Pur275). Their physico-chemical properties are summarized in table 1.
All the studied resins belong to the macroporous family of acidic ion-exchange resins with
comparable ion-exchange capacities. The main differences between these resins are their
thermal stability and textural properties (case of Pur275).
Table 1. Properties of the studied resins
Leaching in DMFb,
Max.
Median pore
Ion-exchange
mol % of sulfur
(relatively to the
initial amount of
S)
operating
temperature,
°C
Sulfonation
typec [9, 10]
Resin
Type
diameter (dry capacity, meq
resin), nm
H+/g
Macro-reticular
Macro-reticular
Macro-reticular
150 [11]
190a
O
C
O
27 [9]
5.4
4.3
5.2
1.2
0.5
3.3
Amb
26.8 [10]
40-70a
Pur482
Pur275
180a
a Provided by the supplier;
3

ACCEPTED MANUSCRIPT
b ICP data (see materials and methods)
c Conventionally sulfonated (C), oversulfonated (O)
The catalytic test of OA esterification with TMP with Amb (Fig. 1) demonstrated that Amb
accelerates the production of esters compared to the reaction without any catalyst (“blank-1”
test). However, if compared to the homogeneous test over H2SO4, the concentration of the
triester after 1500 min is 8 times lower when catalyzed by Amb in the identical conditions
(the evolution of esters under H2SO4 is shown in the supplementary information, Fig.S1).
4,5
Max. concentration of triester obtained over H2SO4
4,0
2,0
Diester
1,5
Monoester
1,0
Triester
0,5
Monoester
0,0
0
250
500
750
1000
1250
1500
Reaction time (min)
Figure 1. Evolution of mono-, di- and triesters concentration in the esterification of OA with TMP: ● - over
Amb (test A), ▲ - “Blank-1” test. Reaction conditions: 3:1 molar ratio of OA:TMP, 1.6% wt. of catalyst, 120°C,
400 rpm, open batch reactor. The test with homogeneous H2SO4 is performed in the same conditions (evolution
of the esters with time is shown in the supplementary information). “Blank-1” test is performed under the same
conditions but without catalyst.
The conversion of OA over Amb reaches around 65% after 1500 min of the reaction (Fig. S2)
whereas the obtained conversion of OA under identical conditions is nearly 100% over H2SO4
after the same time. Despite its lower productivity, industrial application of Amb for the
production of biolubricant could be worthy if the final cost of the heterogeneous process is
4

ACCEPTED MANUSCRIPT
lower than of the homogeneous one. An indispensable requisite for this is that triester remains
the main and ideally only product of the reaction.
In this work we thus examine the possibility of ion-exchange resins to complete the tri-
esterification of OA with TMP under conditions close to the industrial ones (stoichiometric
amounts of reagents and continuous removal of water to shift the equilibrium towards the
products). For this purpose we run the reaction with Amb for 450 min after which we
removed the Amb and continued following the evolution of esters with time. The latter test
was compared to the reaction where Amb was present continuously (Fig. 2). The kinetic
curves of the formation of esters after the catalyst was removed (test B) superpose with the
curves where the catalyst was present continuously (test A). This scenario indicates that the
catalyst either lost its activity or was leached in the liquid reaction medium. However, ICP
analysis of the dimethylformamide medium (DMF) after interacting with Amb (leaching test)
revealed the presence of only 1.2 mol% of S-species relatively to the total amount of S within
the resin (Table 1).
To address the possibility of deactivation caused by water that may block the active centers of
the resin, we added fresh Amb to the reaction mixture after 1350 min from the start of the test
(Fig. 2; test B, where the catalyst was removed previously). The results show that even after
insertion of the fresh catalyst the activity does not change compared to the test A. Shortly
speaking, after several hours of the reaction the ester production does not depend on the
presence or absence of the heterogeneous catalyst. This observation means that the lack of
activity after several hours of catalytic test is not due to the poisoning of active centers.
5

ACCEPTED MANUSCRIPT
Test A
Test B
Insertion of fresh Amb (test B)
Removal of Amb (test B)
4,5
4,0
2,0
1,5
1,0
0,5
0,0
Diester
Triester
Monoester
0
250 500 750 1000 1250 1500 1750 2000 2250
Reaction time (min)
Figure 2. Evolution of mono-, di- and triesters concentration in the esterification of OA with TMP: ● Presence
of Amb continuously (test A), □ Removal of Amb after 450 minutes from the beginning of the reaction and
new insertion of fresh Amb after 1320 minutes (test B). Reaction conditions: 3:1 molar ratio of OA:TMP, 1.6%
wt. of catalyst, 120°C, 400 rpm, open batch reactor.
To explore the behavior of the system at higher temperature we performed some tests with
Purolite CT482 (Pur482) and Purolite CT275DR (Pur275) bearing higher thermal stability
than Amb. Moreover, Pur275 has bigger permanent pores that could be beneficial in the
intraparticle mass transfer. For the test with Purolites (at 180°C) we first started the reaction
without catalyst to be able distinguishing the “pure” heterogeneous catalysis from
homogeneous one. Esterification is indeed homogeneously catalyzing thanks to free OA.
After the concentration of OA decreased (to 15% from initial amount) we added resin and
continued following the OA conversion. Figure 3a shows the curves of OA conversion as a
function of time for the “Blank-2” test and the test with Pur482. We notice that the curves for
both tests superimpose, showing that at 180° the presence of Pur482 does not accelerate the
OA conversion. Interestingly, although Pur275 has higher number of active sites and bigger
pores compared to Pur482 (Table 1), the behavior of Pur275 at the end of the reaction does
not differ from the Pur482 (Fig. 3b). The latter suggests that the inefficiency of the resins at
6

ACCEPTED MANUSCRIPT
the end of the reaction does not come from the internal diffusion limitations caused by small
pores. Moreover, it is clear that both Purolites do not catalyze the end of the reaction and that
it is high temperature instead that helps to accelerate the third esterification.
We should mention that when used in the beginning of the reaction, both Pur482 and Pur275
accelerate the formation of the esters as Amb did as well (the evolution of esters with time is
shown in Fig. S3). Pur275 showed to be more active in this case compared to Pur482 and
Amb.
7

ACCEPTED MANUSCRIPT
a)
100
80
60
40
20
0
Insertion of Pur482
0
220 440 660 880 1100 1320 1540 1760 1980 2200
Reaction time, min
b)
100
80
60
40
20
0
Insertion of Pur275
0
220 440 660 880 1100 1320 1540 1760 1980 2200
Reaction time, min
Figure 3. Conversion of OA with TMP. Reaction conditions: 3:1 molar ratio of OA:TMP, 1.6 %wt. of catalyst
(except “blank” test), 180°C, 400 rpm, continuous flow of N2 (closed batch reactor connected to distillation
8

ACCEPTED MANUSCRIPT
apparatus) .a) ▲“Blank-2” test (no catalyst, the same conditions); ○ Reaction started as “blank” (no catalyst),
then after 340 min Pur482 was added; b) ▲“Blank-2”; ○ Reaction started as “blank-2” (then after 340 min
Pur275 was added).
Thus, neither Amb at 120°C nor Purolites at higher temperature are able to complete the tri-
esterification whereas they are able to convert OA to monoester and monoester to diester.
Likely because of steric hindrance of these intermediate products the last step of the process -
conversion of the diester to the triester - is not possible over conventional heterogeneous
catalyst. The results also indicate that the formation of the triester is occurring due to the
acidity of the residual OA. When OA is being consumed and its concentration decreases, the
production of the triester is no longer persistent regarding the low concentration of
homogeneous OA. Heterogeneous catalyst in this case becomes inefficient due to the
difficulties in adsorption/desorption and diffusion processes of bulky components (di- and
triesters) on its surface and/or in its structure.
Conclusions
Amberlyst 36, Purolite CT482 and Purolite CT275DR were studied in the industrial process
for the tri-esterification of OA with TMP. Our studies showed that Amb, Pur482 and Pur275
catalyze only the conversion of OA to monoester and of monoester to diester and do not
catalyze the conversion of diester to triester. The reason for latter is steric hindrance of the di-
and tri-esters that complicates their adsorption/desorption and diffusion on the catalyst surface
and in its structure. On the one hand, these penalties make impossible to complete the
esterification of OA with TMP over heterogeneous catalyst. On the other hand we showed
that final diester to triester conversion is induced exclusively through homogeneous catalysis.
This complementarity between homogeneous and heterogeneous catalysis is of primary
importance as it contributes first toward the understanding of complex sequential catalytic
reactions. Secondly, it provides key information for the design of advanced heterogeneous
catalysts capable, for instance, bearing endeavor for partial solubilization/reprecipitation
where homogeneous catalysis would proceed.
Acknowledgements We acknowledge the financial support from the “DGTRE, Région Wallonne” for the
project “CATHESTOL” (Convention number 1117464) and “Mosselman s.a.” for constructive interaction. We
thank the company Purolite® for the Purolite CT resins.
The work was performed in the framework of IDS-FunMat program supported by Erasmus Mundus of EU-FP7.
9

ACCEPTED MANUSCRIPT
References
[1]
155.
[2]
Y. Tao, B. Chen, L. Liu, T. Tan, Journal of Molecular Catalysis B: Enzymatic. 74 (2012) 151-
J.C.J. Bart, E. Gucciardi, S. Cavallaro, in: J.C.J. Bart, E. Gucciardi, S. Cavallaro (Eds.),
Biolubricants, Woodhead Publishing, 2013, pp. 473-526.
[3] C.O. Åkerman, Y. Gaber, N.A. Ghani, M. Lämsä, R. Hatti-Kaul, Journal of Molecular
Catalysis B: Enzymatic. 72 (2011) 263-269.
[4]
[5]
P. Varhadi, M. Kotwal, D. Srinivas, Applied Catalysis A: General. 462–463 (2013) 129-136.
J. Oh, S. Yang, C. Kim, I. Choi, J.H. Kim, H. Lee, Applied Catalysis A: General. 455 (2013)
164-171.
[6]
[7]
R.-J. Li, L. Chen, Z.-C. Yan, J Am Oil Chem Soc. 89 (2012) 705-711.
C.O. Åkerman, A.E.V. Hagström, M.A. Mollaahmad, S. Karlsson, R. Hatti-Kaul, Process
Biochemistry. 46 (2011) 2225-2231.
[8]
[9]
R. Yunus, O.T. Lye, A. Fakhru'l-Razi, S. Basri, JAOCS. 79 (2002) 1075-1080.
C. Casas, R. Bringué, E. Ramírez, M. Iborra, J. Tejero, Applied Catalysis A: General. 396
(2011) 129-139.
[10]
J. Guilera, PhD thesis, Ethyl octyl ether synthesis from 1-octanol and ethanol
or diethyl carbonate on acidic ion-exchange resins, University of Barcelona, 19 July 2013,
http://hdl.handle.net/2445/45985.
[11]
I. Dosuna-Rodríguez, E.M. Gaigneaux, Catalysis Today. 195 (2012) 14-21.
10

ACCEPTED MANUSCRIPT
Graphical abstract
11

ACCEPTED MANUSCRIPT
Highlights
Ion-exchange resins were studied in the tri-esterification of OA with TMP.
The final triester is homogeneously generated thanks to the residual oleic acid.
Bulky di- and triesters hinder adsorption/desorption onto resin surface.
12