Polarity of the acid chain of esters and transesterification activity of acid catalysts

Article abstract of DOI:10.1016/j.jcat.2008.11.026

The effect of polarity of the esters on the transesterification reaction rate has been investigated. The effect was studied in homogeneous and heterogeneous catalysts. The polarity of different ethyl alkanoate esters was varied by (i) increasing the number of carbon atoms in the acid alkyl chain of the esters and (ii) introducing Br and hydroxy substituents at the end of the acid chain of ethyl hexanoate. Polarity was determined through the λ_max of the UV-Vis spectrum of the betaine dye dissolved in the investigated esters (E_T(30) scale). The transesterification reaction was carried out with methanol and by using sulfuric acid and a Dowex DR2030 sulfonic resin as homogeneous and heterogeneous catalysts, respectively. It was observed that, in addition to steric hindrance, the polarity of the ester chain has an effect on the reaction rate of the heterogeneous acid catalysts. It is proposed that the positive or negative effect of the polarity is due to repulsive or attractive interactions of the ester chain with the polar groups of the resin and/or with the methanol molecules present in the pores. A very positive effect is found in heterogeneous acid catalysis if H-bonds can stabilize the active intermediate participating in the rate determining step. The attractive or repulsive interactions are absent in the homogeneous case.

Full text of DOI:10.1016/j.jcat.2008.11.026

Journal of Catalysis 262 (2009) 18–26
Contents lists available at ScienceDirect
Journal of Catalysis
www.elsevier.com/locate/jcat
Polarity of the acid chain of esters and transesterification activity of acid catalysts
D. Martín Alonso a, M. López Granados a,∗, R. Mariscal a, A. Douhal b
a
Instituto de Catálisis y Petroleoquímica, CSIC, C/Marie Curie 2, Campus de Cantoblanco, 28049, Madrid, Spain
Departamento de Química Física, Sección de Químicas, Facultad de Ciencias del Medio Ambiente, Universidad de Castilla-La Mancha, Avda. Carlos III, S.N., 45071, Toledo, Spain
b
a r t i c l e i n f o
a b s t r a c t
Article history:
The effect of polarity of the esters on the transesterification reaction rate has been investigated. The
effect was studied in homogeneous and heterogeneous catalysts. The polarity of different ethyl alkanoate
esters was varied by (i) increasing the number of carbon atoms in the acid alkyl chain of the esters and
(ii) introducing Br and hydroxy substituents at the end of the acid chain of ethyl hexanoate. Polarity was
determined through the λmax of the UV–Vis spectrum of the betaine dye dissolved in the investigated
esters (ET(30) scale). The transesterification reaction was carried out with methanol and by using sulfuric
acid and a Dowex DR2030 sulfonic resin as homogeneous and heterogeneous catalysts, respectively. It
was observed that, in addition to steric hindrance, the polarity of the ester chain has an effect on the
reaction rate of the heterogeneous acid catalysts. It is proposed that the positive or negative effect of the
polarity is due to repulsive or attractive interactions of the ester chain with the polar groups of the resin
and/or with the methanol molecules present in the pores. A very positive effect is found in heterogeneous
acid catalysis if H-bonds can stabilize the active intermediate participating in the rate determining step.
The attractive or repulsive interactions are absent in the homogeneous case.
Received 1 October 2008
Accepted 21 November 2008
Available online 7 January 2009
Keywords:
Esterification
Transesterification
Acid catalyst
Biodiesel
Reichardt’s dye
ET(30) polarity scale
© 2008 Elsevier Inc. All rights reserved.
1. Introduction
luminous carboxylic acids or esters, the heterogeneous reactions
proceed at much slower reaction rates and longer reaction times,
and higher temperatures are required to achieve conversions com-
parable to those of homogeneous catalysts [7–9]. An explanation is
required for these effects.
Esterification and transesterification reactions are widely used
for the synthesis of fine chemicals [1], pharmaceuticals [2], plastics
monomers [3], and many other chemical products [4]. Nowadays,
interest in these reactions has increased because they are required
for the production of biodiesel, a mixture of fatty acid methyl es-
ters (FAME). Biodiesel is a non-toxic biodegradable fuel produced
from vegetable oils, and is now widely accepted as environmental
friendly and as a renewable fuel provided by biomass [5,6].
Many processes using esterification and transesterification re-
actions are carried out by using homogeneous acid catalysts. The
use of homogeneous acid catalysts has serious drawbacks: they
are corrosive and they must be neutralized at the end of the
reaction; they cannot be recovered, generating wastewaters and
increasing the cost of the overall process. To cope with the prob-
lems associated with homogeneous processes, solid acid catalysts
have been proposed. However homogeneous reactions always pro-
ceed at higher rates than when heterogeneous catalysts are used.
Moreover when heterogeneous esterification and transesterification
reactions are applied to carboxylic acids and esters with short fatty
acid chains, both reactions proceed at reaction rates that are still
satisfactory. By contrast, when applied to long chain or more vo-
In principle, one can assign the lower activity in solid catalysts
to mass transfer restrictions derived from the lower diffusion co-
efficient of heavier molecules and/or from the porous character of
solid catalysts. However, in the case of esterification of carboxylic
acids by homogeneous catalysis there are no mass transfer restric-
tions, and low reaction rates are also found for longer acid chains.
In such cases, steric effects related with the size of the molecule
and with preferential conformations have been invoked to explain
the lower reaction rate of the esterification of carboxylic acid with
longer chains [7].
The esterification reaction catalyzed by homogeneous acids pro-
ceeds through the following mechanism: in a first step, the car-
bonyl oxygen is attacked by the acid proton that increases the
electrophilicity of the carbon atom. This carbon atom is then more
susceptible to the nucleophilic attack of the hydroxyl group of the
alcohol, which is the rate determining step. The following steps are
the dehydration and deprotonation (esterification) of the interme-
diate with the corresponding formation of the new ester and the
recovery of the catalyst [10,11]. Taking into account that transes-
terification proceeds through a similar mechanism (see Scheme 1),
the same steric and conformational effects can be invoked to ex-
plain the lower activity found in the transesterification of esters
Corresponding author. Fax: +34 915 854 760.
E-mail address: mlgranados@icp.csic.es (M.L. Granados).
*
0021-9517/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2008.11.026

D.M. Alonso et al. / Journal of Catalysis 262 (2009) 18–26
19
Scheme 1.
with longer or more voluminous acid chain by using homogeneous
acid catalysts. In the case of heterogeneous catalysts similar effects
can explain the lower activity, albeit enhanced by their solid na-
ture [7].
pyridinium N-phenolate betaine dye (Reichardt’s dye) [12,13]. The
details of the experimental procedure will be presented in the ex-
perimental section. Basically, it involves measuring the wavelength
of the CT band when the dye is dissolved in the different esters
studied in this work, and the ET(30) magnitude will be determined
from the wavelength. The data of ET(30) for a number of molecules
is tabulated in Refs. [12,13] that facilitates the comparison with the
molecules studied in this work.
Less attention has been paid to the polar character of the acid
chain. The aim of this article is to understand the role of the po-
larity of the acid chain of esters on the transesterification reaction
rate when using homogeneous and, especially, heterogeneous acid
catalysts. In this work, the reaction rate of the methanolysis of dif-
ferent ethyl esters using a sulfonic resin was determined and was
compared with those obtained with sulfuric acid as homogeneous
catalyst. The alkyl chain of the carboxylic fragment was varied (i)
by increasing the number of carbon atoms and (ii) by introducing
polar functionalities at the opposite end of the acid chain in esters
with the same number of carbon atoms. The first strategy accounts
for steric effects as well as for polarity effects, but the second se-
ries should account only for polarity effects.
It will be shown that, when using heterogeneous acid catalysts,
the transesterification reaction rate can be largely improved by in-
creasing the polarity of the chain of the acid group in the ester
(an ester group has two moieties, the acid group and the alco-
hol group), especially when functionalities in the acid chain allow
for the formation of hydrogen bonds between the acid chain and
the surface of the catalysts and/or with liquid methanol molecules
present within the pores of catalyst. In the case of homogeneous
catalyst, the effect of polarity is negligible. The conclusions pre-
sented here can be extended to esterification reactions.
The polarity of a given molecule is defined through its solva-
tion capacity, that is, “through its capacity to establish non-specific
and specific intermolecular interactions with a given solute, ex-
cluding obviously those interactions leading to chemical alterations
of the solute (such as protonation, oxidation, reduction, com-
plex formation, etc.)” [12,13]. The non-specific interactions include
purely electrostatic forces arising from Coulomb interactions and
polarization forces (those between ion and non-polar molecule,
dipolar/non-dipolar molecule and two non-polar molecules). The
specific forces include hydrogen bond between hydrogen bond do-
nating (HBD) and accepting (HBA) molecules and the interactions
between electron pair donor (EPD) and electron pair acceptor (EPA)
molecules.
Among the various methodologies that can be used for the
measurement of polarity, solvatochromic effect has been selected
for this article. Solvatochromism is the change in the position,
intensity and shape of UV/Vis/near IR absorption spectra of a
chemical probe when dissolved in the molecule whose polarity
is to be determined [12,13]. Specifically, the widely applied and
well-established Dimroth–Reichardt ET(30) scale will be used. This
scale is based on the changes in the position of the intramolec-
ular charge transfer (CT) of the π–π∗ absorption band of the
The choice of this methodology tacitly assumes that the inter-
actions established between Reichardt’s probe and the esters are of
the same type as those established between the ester and both the
polymeric chains of the sulfonic resins and the methanol molecule.
The band position of Reichardt’s dye is very sensitive to HBD–HBA
properties and to non-specific interactions with the solvents (the
esters), whereas its sensitivity to interactions like EPD or EPA is
negligible [12,13]. We assume that the former type of interactions
between the acid chain of the ester and both the polymeric chain
and methanol molecule are essential in explaining the effect of
polarity on the transesterification rate, whereas the latter prop-
erties are irrelevant. Therefore, ET(30) appears as a very suitable
scale for our purpose. It will be shown that structural changes in
the sulfonic resin (unfolding, swelling, etc.) driven by the differ-
ent polarity of the esters are minimal in the selected resin, and
therefore the contribution of these changes to the reaction rate
are also negligible. Negative or positive inductive effects of the po-
lar group transmitted through the hydrocarbon chain to the ester
group can be considered negligible because of the long distance
between them. The non-specific and hydrogen bond properties of
the esters may be affected because of their presence within the
nanocavities of the solid, although these variations are expected to
be similar in all the investigated esters.
2. Experimental
2.1. Description of the catalytic resin
All the catalytic experiments were carried out using a Dowex
DR2030 ion exchange resin in its dried form supplied by Dowex
as 1.19–0.42 mm beads (moisture content is less than 3%, mesh
size 16–40). The Dowex DR2030 catalyst is based on divinylben-
zene (DVB) and styrene copolymer and functionalized with sul-
+
g
−1). This poly-
fonic acid groups (weight capacity >4.7 meq H
mer was chosen because of its mesoporosity and relatively large
pore volume (0.33 cm3 g−1). The specific surface was determined
using a Micromeritics ASAP 2000 apparatus to record the N2 ad-
sorption isotherm at 77 K after evacuation at 363 K for 20 h and
−1
applying the BET method. The value was 29.8 0.2 m2 g – con-
sistent with that given by the supplier. The mean pore diameter
obtained by applying the BJH method to the desorption isotherm

20
D.M. Alonso et al. / Journal of Catalysis 262 (2009) 18–26
Table 1
Ethyl esters investigated in this work.
Name
Molecular structure
No. of carbon atoms in the acid chain of the ester
2
Ethyl acetate
Ethyl pentanoate (or valerate)
Ethyl hexanoate (caproate)
Ethyl 6-hydroxyhexanoate
Ethyl 6-bromohexanoate
5
6
6
6
Ethyl octanoate (or caprylate)
Ethyl hexadecanoate (or palmitate)
Glyceryl triacetate
8
16
2a
a
It is not an ethyl ester.
was between 40 and 50 nm. The sulfonic loading was determined
by titration with NaOH solution and by using phenolphthalein as
with the catalyst. In the case of the sulfonic resin, 2 g of solid was
added to the ester, that is, 9.4 meq H . The resin was ground to
+
+
−1
.
indicator, and was found to be 4.7 meq H
g
obtain finer particles with diameter ꢀ < 0.106 mm. The tempera-
ture of the ester–catalyst mixture was set to 333 K by circulating
water through the jacket of the reactor with a re-circulating bath.
The mixture was under vigorous shaking (1000 rpm). The transes-
terification reaction started when methanol (0.005% H2O Riedel de
Häen) preheated to 333 K was added to the mixture through the
dropping funnel. The methanol/ester molar ratio was 2:1. The reac-
tion was carried out at atmospheric pressure for 3 h in the case of
the solid resin. Aliquots at different reaction times were taken and
the reaction was ended by rapidly cooling down to room tempera-
ture. The solid catalyst is then easily separated from the liquids by
filtering.
2.2. Investigated ethyl esters
Table 1 presents the ethyl esters studied in this work. The vari-
ation in the polarity of the acid chain of the esters was achieved
by two means. In a first series, the polarity changes by increas-
ing the length of the carboxylic alkyl chain. Thus, ethyl acetate
(Sigma-Aldrich, purity >99%) ethyl pentanoate (Sigma-Aldrich, pu-
rity >99%), ethyl hexanoate (Alfa Aesar, purity >99%), ethyl oc-
tanoate (Sigma-Aldrich, purity >99%) and ethyl hexadecanoate
(Fluka, purity >95%) esters were selected. Table 1 also indicates
common names used for ethyl pentanoate, hexanoate, octanoate
and hexadecanoate. In a second series, the polarity of the ethyl
hexanoate ester is modified by introducing a hydroxyl or bromo
group at the opposite end of the carboxylic group. Thus ethyl
6-hydroxyhexanoate (Sigma-Aldrich, purity >98%) and ethyl 6-
bromohexanoate (Alfa Aesar, purity >98%) esters were used. Be-
sides these two series, glyceryl triacetate (Sigma-Aldrich, purity
>99%) was also included in the study as a model of short chain
triglyceride. In principle and if it is not mentioned all these esters
were investigated in reaction as received, with no further purifica-
tion.
In the case of the homogeneous catalyst,
a much smaller
amount of sulfuric acid was used and the reaction time was
shorter (1 h) in order to achieve conversions comparable to that
of the sulfonic resin. The same amount of ester was used (114
mmol) and then heated to 333 K, whereupon a sulfuric acid
methanolic solution at 333 K was added so as to achieve the
methanol/ester molar ratio 2:1. The methanol incorporates the sul-
furic acid amounting to 58.4 mg of sulfuric acid, that is, 0.6 meq
+
H
. The analysis of the compositions was carried out immediately
to prevent the progress of the reaction.
The liquid aliquots were analyzed in an Agilent 6890 GC
equipped with a HP INNOwax capillary column (30 m × 320 μm
× 0.50 μm). An aliquot of 250 mg of the liquid sample was added
to 5 mL of acetone. Since the carbon balance was always >95%,
the yield corresponds basically to the conversion of the ethyl ester.
The yield to methyl ester was calculated from the relative areas of
the methyl and ethyl esters using the equations
2.3. Catalytic activity measurements
The transesterification of the esters was carried out in a 125 mL
three-necked jacketed glass reactor. A reflux condenser was con-
nected to one of the necks and a dropping funnel was connected
to another. In all cases, 114 mmol of ester was placed in contact

D.M. Alonso et al. / Journal of Catalysis 262 (2009) 18–26
21
Fig. 1. Molecular structure of Reichardt’s dye.
mol methylester
Yield (%) =
× 100
mol initial ethylester
mol methylester
mol ethylester
=
× 100,
(1)
(2)
mol methylester
mol ethylester
1 +
where
Fig. 2. UV–Vis spectra of betaine dye in some common solvents and in glyceryl
triacetate, ethyl acetate and ethyl octanoate in the region of interest.
area methylester
MW methylester
area ethylester
MW ethylester
mol methylester
mol ethylester
=
.
Table 2
λmax and ET(30) of some solvents and of the different esters investigated.
It is assumed that the chromatographic factors for methyl and
ethyl esters are similar and that no other products in addition
to methyl ester can be formed. Only in the case of the ethyl
6-hydroxyhexanoate were other products formed, as will be dis-
cussed later. The error in the determination of the yield was 5%
of the absolute yield value.
Solvent
Wavenumber
(nm)
ET[30]
(kcal/mol)
THF
Acetonitrile
MeOH
Ethyl acetate
Ethyl pentanoate
Ethyl hexanoate
Ethyl 6-hydroxyhexanoate
Ethyl 6-bromohexanoate
Ethyl octanoate
Ethyl hexadecanoate
Glyceryl triacetate
763
626
517
743
779
780
570
627
790
–
37.5 (37.4)a
45.7 (45.6)a
55.3 (55.4)a
38.5 (38.1)a
36.7
36.6
50.2
45.6
2.4. ET(30) values obtained by UV/Vis spectroscopy measurements of
Reichardt’s dye
36.2
35.9b
The UV/Vis maxima of the absorption bands of Reichardt’s dye
(2,6-diphenyl-4-(2,4,6-triphenylpyridinio) phenolate, Aldrich 90%),
also known as betaine dye, in each ester were recorded in a Varian
(Cary E1) double-beam spectrophotometer to correct ester absorp-
tion on-line. Fig. 1 represents the chemical structure of the betaine.
A few milligrams of Reichardt’s dye were dissolved in each ester
and the spectrum was collected at ambient temperature from 300
to 900 nm with a resolution of 1 nm using a quartz cell of 1 cm
thickness. The esters were dried for two days with a molecular
sieve in order to remove any water present.
707
40.4 (40.1)a
a
Values in parentheses are those tabulated in [12,14].
Estimated value (see text for explanation).
b
in more polar molecules are blue shifted (lower wavelength). The
ground state of dye is more stabilized by solvation with polar
molecules than the excited state, and then the energy difference
increases [12]. Among the different molecules presented in the
figure, the betaine band peaks at the lowest wavelength when dis-
solved in methanol and in methyl octanoate at the longest wave-
length. The molecules represented in the figure can be ordered
according to their betaine λmax as follows:
ET(30) values is the energy of the electronic transition of the
intramolecular charge transfer (CT) of the π–π∗ absorption band
−1
of Reichardt’s dye expressed in kcal mol
when measured at
room temperature (298 K) and atmospheric pressure. Once the
λmax (maximum of the band) is determined (the error was always
smaller than 2 nm), ET(30) was calculated as follows [12]:
methanol < acetonitrile < glyceryl triacetate < ethyl acetate
< THF < ethyl octanoate.
ꢀ
ꢁ
kcal
mol
28591
This order is a consequence of the strength of solvation of Re-
ichardt’s dye by the molecules and therefore of their polarity. It
is worth noting that glyceryl triacetate is more polar than ethyl
acetate and the latter more polar than ethyl octanoate.
Table 2 summarizes the λmax and ET(30) values of the dye
when dissolved in the investigated molecules. For some of the
molecules tabulated (THF, acetonitrile, methanol, ethyl acetate and
glyceryl triacetate), the well-accepted values reported in the lit-
erature [12,14] are also included between brackets in the ET(30)
column. The values determined experimentally are in good agree-
ment with reported ET(30) values. In Fig. 3, the experimentally
ET(30)
=
.
(3)
λmax (nm)
3. Results
3.1. Polarity of the esters
Fig. 2 shows the region of interest of the UV–Vis spectra of
betaine dye in some common solvents (methanol included) and
some of the esters investigated in this work. The betaine dye is a
negative solvatochromic compound, and then the CT π–π∗ bands

22
D.M. Alonso et al. / Journal of Catalysis 262 (2009) 18–26
(A)
Fig. 3. ET(30) values vs. number of carbon atoms of the investigated esters: (2) non-
substituted linear carboxylate esters; (!) extrapolation value for ethyl hexade-
canoate and (P) bromo- or hydroxyethyl hexanoate and glyceryl triacetate.
determined ET(30) values are represented vs. the number of carbon
atoms of the acid chain in the ester: in the main figure, the values
for the alkyl chain, without polar groups, are represented by the 2
symbols, whereas in the inset of Fig. 3 the values for the OH and
Br-substituted ethyl hexanoate and for the glyceryl triacetate are
represented by the P symbols. The value for hexadecanoate could
not be determined because Reichardt’s dye is not soluble in it. An
estimation (!) was proposed according to the well-established be-
havior for other series, such as linear 1-alkyl alcohols. In the latter
series, the ET(30) value decreases in a non-linear mode with the
increase of the chain length, and reaches an asymptotic value for
the alcohols with 10–11 carbon atoms. A further increase in the
chain length does not result in a lower ET(30) value [12]. A similar
trend is expected for the ethyl alkanoate esters, whereby a good
approximation for the ET(30) of ethyl hexadecanoate can be 35.8–
(B)
−1
36.0 kcal mol (35.9 was tabulated). The value of the latter is only
−1
Fig. 4. Yield to methyl ester for the non-substituted alkanoate ethyl esters investi-
gated when using (A) sulfuric acid as homogeneous catalyst and (B) sulfonic resin
as heterogeneous catalyst. The insets in the figure are magnifications of the kinetics
during the first 10 min (A) and 35 min (B).
2.6 kcal mol
smaller than that of the ethyl acetate. Larger differ-
ences are found for the Br and OH substituted esters: ET(30) for
−1
the Br ethyl hexanoate is 9 kcal mol
larger than that of its non-
substituted equivalent and this difference is even larger for the
−1
hydroxyl substituted ester (13.6 kcal mol
of glyceryl triacetate is ca. 2 kcal mol
acetate
larger). ET(30) values
stronger than those caused by the –CH2– units or Br group and
therefore result in larger increases in the ET(30) [12,14]. It is well
known that the OH group increases polarity when present in any
molecule.
−1
larger than that of ethyl
The UV–Vis spectrum of betaine dye is very sensitive to non-
specific forces and to the HBD–HBA forces, whereas it is rather
insensitive to EPD–EPA interactions. In the non-substituted ethyl
alkanoate series, the asymptotic trend can be explained as fol-
lows: the ester group would contribute to the same extent to the
solvation energy, and the only variations are expected from the
non-specific forces of the acid chain. The introduction of non-polar
–CH2– units results in an increasingly larger contribution from the
non-specific forces that are less intense. As a result, the polarity
of the molecule decreases and so do the ET(30) values. The distant
–CH2– units have no effect on the probe molecule, and the effect
of the non-specific forces becomes constant as the alkyl chain be-
comes larger than 10–11 carbon atoms.
The presence of a Br group at the end of the hexanoate chain
means that stronger non-specific interactions can be established
between this molecule and Reichardt’s dye since, as is well-
reported, the Br group brings in stronger non-specific type forces
(polarity/polarizability) that are superimposed onto those of the
–CH2– units with less capacity to shift the λmax [12,14]. This is re-
flected in a large ET(30) value of the ethyl bromohexanoate with
respect to that of ethyl hexanoate. In contrast, the presence of a
hydroxyl (OH) group involves a HBD–HBA type interaction with
the betaine dye. These types of interactions have proven to be
3.2. Transesterification reaction rate and length of the alkyl acid chain
Fig. 4 compares the yields to the non-substituted methyl ester
using sulfuric acid with those involving sulfonic resin, respectively.
The insets in the figures represent a magnification of the kinetics
during the first minutes of the reaction. The intrinsic reaction rate
of the homogeneous sulfuric acid is much higher than the active
sites of the resin [7,10,11,15] and in order to obtain comparable
yields, the amount of homogeneous catalyst used was almost two
orders of magnitude smaller than that used for the resin. In the
case of the homogeneous reaction (Fig. 4A), the yield rapidly de-
creases as the number of carbon atoms increases up to when the
acid chain of the ester is 5 carbon atoms long (ethyl hexanoate).
Beyond this ester yield to methyl ester for C5, C6, and C8 esters are
quite similar. The yield for the hexadecanoate is slightly smaller
than that of pentadecanoate. By contrast, in the case of the het-
erogeneous catalyst (Fig. 4B) the behavior does not parallel that
of sulfuric acid. Beyond the ethyl pentanoate yield to methyl ester
decreases as the length of the chain become larger. For the larger
ester investigated, ethyl hexadecanoate, the yield to methyl ester

D.M. Alonso et al. / Journal of Catalysis 262 (2009) 18–26
23
of the resin was placed in a sealed bottle and the height reached
was measured. The solvent was then introduced into the flask un-
til 2/3 of the flask volume was reached. After 24 h, the height of
the swollen resin stabilizes and was measured. The increase was
associated with the swelling degree. The degree was 17, 25, 11 and
12% for methanol, ethyl acetate, ethyl octanoate, and ethyl hexade-
canoate respectively. The differences between the ethyl acetate and
the ethyl hexadecanoate were even smaller if methanol was mixed
with them (as occur in the reaction mixture). Then the swelling
degree was 20 and 18%, that indicates that swelling is defined by
the presence of methanol rather that due to the esters. Summariz-
ing it can be concluded that the swelling degree is rather similar
for all the molecules irrespective of the polarity of the solvent
molecules and that the differences in accessibility of the different
esters are therefore expected to be rather small. A second type of
experiments also verified that the order of the addition (methanol
first or ethyl hexadecanoate first) does not change the reaction
rate. In short, the swelling of the resin can be discarded as the
main cause of modification of the reaction rate.
It can also be argued that diffusional limitations can explain
the decrease in the reaction rate in the heterogeneous reaction:
more voluminous esters are more likely to be affected by exter-
nal and inner mass transfer restrictions. These problems can be
discarded because of the following. In order to assess on the exter-
nal mass transfer restrictions, the rate obtained at shorter agitation
speed (600 rpm) for the ethyl hexadecanoate (the larger ester) was
determined and found to be 0.0021 s−1, similar to the value repre-
sented in Fig. 5 (0.0019 s−1) that was obtained at 1000 rpm. This
data precludes the mass transfer distortion in the determination
of the reaction rate of Fig. 5. Moreover the reaction rate for ethyl
hexadecanoate was also determined by using larger resin particle
Fig. 5. TOF values obtained when sulfuric acid or acid resin are used as catalysts in
the transesterification of non-substituted alkanoate ethyl esters: (2, 1) for sulfuric
acid and (", !) for acid resin. The TOF values were calculated for two different
periods of time.
is much lower than that expected for a behavior parallel to that of
homogeneous catalyst.
In the case of Fig. 5 the TOF numbers for the different esters
were represented, both for the homogeneous and the heteroge-
neous cases. They were calculated by assuming that all the acid
groups of the resin can participate in the reaction. For the homo-
geneous case they were calculated by using the yield at 10 min.
of reaction and for the resin at 30 min. However for the sake of
comparison the data at 60 min and 180 min, respectively, were
also included. The latter were similar to that obtained at shorter
reaction times what indicates that they were still valid as differ-
ential reaction rates (except for the case of the ethyl acetate case
that showed very large yield, out of the differential regime). Fig. 5
shows that in the case of homogeneous reaction the trend follows
the well reported behavior: a plateau is reached as the length of
the chain is longer [7]. In the case of the solid resin, the level-
ing off observed in the homogeneous case for ester longer than 5
carbon atoms is no longer visible and the reaction rate keeps on
decreasing as the chain become longer. The symbol ⊗ represents
the hypothetical solid resin value for hexadecanoate case if the be-
havior was parallel to the homogeneous case.
When explaining the behavior shown in Fig. 5 by the sulfonic
resin, one can, in principle, must consider the swelling of the
resins and diffusional restrictions (inner or external). Swelling of
the resins is a well-known event that may occur to a larger or
smaller extent in polymeric resins when contacted with polar sol-
vents: different interactions between the solvent and the groups of
the polymer chain, especially sulfonic groups, result in the solva-
tion of the polymer chains by the polar molecules. The hydrogen
bond predominates, but other forces such as electron pair dona-
tion and non-specific forces may also be involved. Solvation may
be accompanied by the penetration of the polar molecules into
the pores defined by the intertwined polymeric chains that may
stretch out the resin framework [16,17]. The different degree of
swelling caused by the different molecules used in this study may
determine the accessibility of the reactant molecule and therefore
the reaction rate.
size (0.15 < ꢀ < 0.18) than that used for Fig. 5 (ꢀ < 0.106). The
−1
TOF number was found to be 0.0021 s
that indicates that Fig. 5
result is not affected by inner mass transfer restrictions.
The absence of inner diffusional restrictions is not surprising
if one take into account the mean pore size of the resin deduced
from the N2 isotherms which is around 40–50 nm. The immer-
sion of the resin in the esters and/or in methanol scarcely modifies
its textural properties since the swelling degree of the resin was
small. The textural properties determined by N2 isotherms must
therefore be rather similar to that described by the N2 isotherms.
The largest dimension of the ethyl hexadecanoate, the longest es-
ter, was estimated by the Gaussian program [20]. This dimension
corresponds to the fully unfolded ester. Considering the length of
the C–C, C=O and C–O bonds, the largest dimension is around
2.45 nm, more than one order of magnitude smaller than the mean
pore size of the resin. It should be also taken into account that
from the diffusion point of view this ester can be also consid-
ered a cylinder 2.54 nm long whose cylindrical base is defined
by the ester group. Moreover the C16 ester can adopt folded con-
figurations because the chain is flexible. Therefore, the ester can
even diffuse through the pores in more favorable configurations.
But even assuming that the ester occupies a volume defined by a
sphere with a diameter of 2.45 nm, diffusion limitations within the
catalyst pores can be considered negligible.
If diffusional restrictions cannot be the origin of the differences
between sulfuric acid and resin catalyst, another explanation is
required. Similar discrepancies between homogeneous and hetero-
geneous catalysts have been found by Goodwin and coworkers [7],
albeit when studying the esterification of linear alkyl carboxylic
acids with different chain length. Considering the close similarities
in the reaction mechanism, the same arguments can, in principle,
be applied to the transesterification reaction. The different be-
havior was explained by the “conformational levelling” effect that
takes place in the homogeneous case, and it is absent in the het-
erogeneous reaction. The explanation, according to Goodwin and
To test whether the swelling has an effect on the transes-
terification rate, two types of experiments were conducted. The
swelling degree of the resin when immersed in methanol, in the
esters with different polarity and in the methanol/ester solution
with the same molar ratio used for the reaction was measured as
the increase in the volume of the resin before and after immersion
in the liquid. The change of volume was determined as follows: 2 g

24
D.M. Alonso et al. / Journal of Catalysis 262 (2009) 18–26
(A)
Fig. 6. TOF values obtained when sulfuric acid or acid resin are used as catalysts in
the transesterification of hexanoate ethyl ester and bromo and hydroxy hexanoate
ethyl esters.
coworkers is as follows [7]: a first aspect to consider is the rate
determining step of the reaction. This is the nucleophilic attack of
the alcohol on the carbon atom of the protonated carbonyl group
(in the case of esterification, the carbonyl group belongs to a car-
boxylic group, and to an ester in transesterification). A requirement
of this step is for methanol to approach the activated carbonyl
group. The second aspect to consider is the steric hindrance that
in principle is influenced by the size of the chain. Thus, a larger
alkyl chain results in larger constraints of the alcohol molecule ap-
proaching the activated ester group. In the case of homogeneous
reaction, the steric effect is also modulated by conformational ef-
fects [7,21,22] that minimize these constraints. Longer chains are
flexible for adopting other preferential conformations. The confor-
mational minimization is negligible when compared with steric
hindrance for a chain with fewer than 5–6 carbon atoms, but when
the length of the chain becomes longer it starts to be significant.
In other words, the acid chain with more than 6–7 carbon atoms
can adopt conformations that allow the methanol to reach the pro-
tonated carbonyl group as easily as with a shorter chain. As a
result of this conformational flexibility, a leveling is reached in the
reaction rate. This is possible in the case of homogeneous catal-
ysis. However, it is absent in the case of heterogeneous catalyst
because following Goodwin et al.’s suggestion for esterification re-
actions [7], the capacity to adopt another preferential conformation
is restricted. The activated carbonyl group is now anchored onto
the acid sites of the resin and surrounded by polymer chains. The
freedom of the long acid chains of esters to adopt other confor-
mations that facilitate the approach of methanol to the activated
carbonyl is limited, and therefore steric hindrance remains a key
factor in inhibiting the reaction. As a result, the heterogeneous re-
action rate keeps on decreasing with the length of the acid chain,
whereas it remains constant for the homogeneous case. The next
section shows that this hypothesis must be reconsidered and/or
completed by considering the possibility of hydrophobic repulsion
between the activated ester chain and the sulfonic groups of the
resin.
(B)
Fig. 7. (A) TOF values vs. ET(30) values of the different ethyl esters investigated:
(2, 1) acetate; (", !) pentanoate; (Q, P) hexanoate; (a, e) octanoate; (F, E)
hexadecanoate. (B) TOF values vs. ET(30) values of the hexanoate, bromohexanoate
and hydroxyhexanoate ethyl esters. White symbols indicate the reaction performed
with sulfuric homogeneous catalyst and open symbols with heterogeneous sulfonic
resin.
the ester molecule does not result in a noticeable increase in the
reaction rate. In the case of the solid resin, the presence of the hy-
droxyl group results in a remarkable improvement of the reaction
rate, whereas the Br group has only a small effect.
In the case of ethyl hydroxyhexanoate, it should be noted
that other products besides methyl 6-hydroxyhexanoate were also
formed. Those are the products derived from the transesterifica-
tion of ethyl hydroxyhexanoate with itself and from the transes-
terification of methyl hydroxyhexanoate with itself and with ethyl
hydroxyhexanoate. These reactions result in the formation of two
hydroxydiesters with 13–14 C atoms that were detected by GC.
If these products are considered, then the relative conversion of
ethyl hydroxyhexanoate with respect to that of ethyl hexanoate
was slightly higher with both homogeneous and heterogeneous
catalyst. The contribution of these compounds was of the order
of 5%, but in the heterogeneous case the differences are more in-
tense. So the conclusions derived from Fig. 6 are still valid even if
the formation of other products is taken into account.
3.3. Transesterification reaction rate and polar groups in the acid chain
Fig. 7 compares the reaction rate with the ET(30) values. It can
be observed that in the case of linear non-substituted ethyl esters
(Fig. 7A), conversion decreases upon the decrease of the polarity
of the acid chain of the ester for both homogeneous and heteroge-
neous cases. But for the homogeneous the dependence is defined
by the conformational leveling effect and not by the polarity of
the ester: beyond the C5 ester the trend reaches a plateau and rate
Fig. 6 shows the intrinsic reaction rate of ethyl bromohexanoate
and ethyl hydroxyhexanoate when compared with that of ethyl
hexanoate. White columns represent the values obtained when us-
ing sulfuric acid and darker columns when using the solid resin as
catalysts. From the results in Fig. 6, it can be concluded that, in
the case of homogeneous catalysis, an increase in the polarity of

D.M. Alonso et al. / Journal of Catalysis 262 (2009) 18–26
25
of C16 ester is similar to that of C5 although there is a difference
of more than 1 kcal/mol in ET(30) value. In the case of the solid
resin, there is a logarithmic dependence between the reaction rate
and the polarity and the TOF number decreases upon the ET(30)
value of the ester. We will return to this point later to discuss it
deeper.
long chain of activated ester and the surrounding species (either at
the surface of the resin or liquid within the pores). The repulsion
is the result of the preferred chemisorption of methanol over the
resin or the preferred presence of methanol in the pores. For the
long ester molecule to be an activated complex it must displace
the methanol out of the surface sites of the resin surface breaking
the very strong HBD and HBA interactions established between the
sulfonic groups and the methanol molecules and between the liq-
uid methanol molecules themselves present within the pores. This
repulsion depletes the surface concentration of the activated com-
plex involved in the rate determining step for esters with larger
acid chains with respect to those with shorter chains. This effect
must also be taken into account, besides steric constraints, when
explaining the lower reaction rate of long alkyl acid chain esters.
Another hypothesis that can be proposed is that the larger in-
trinsic reaction rates of acid homogeneous catalysts is related with
the fact that the acid species are not concentrated on a polar sur-
face but dispersed in the liquid and therefore the approach of the
acid catalytic species to the ester group is not inhibited by any re-
pulsion effect.
In the case of the bromohexanoate ester, the presence of a Br
group at the end of the hexanoate chain results in an increase
−1
of ca. 9 kcal mol
with respect to that of ethyl hexanoate (see
Fig. 7B). This increase in polarity does not result in a remarkable
increase in conversion either in the homogeneous reaction or in
the heterogeneous reaction. By contrast, the presence of a hydroxy
−1
−1
group increases ca. 14 kcal mol
its ET(30) value (ca. 5 kcal mol
more than the Br substituted ester) and this results in a ca. 7-fold
increase in the reaction rate of methyl ester formation when using
the sulfonic resin and a very small increase when using sulfuric
acid (only ca. 2-fold).
Summarizing, in the case of acid solid resin there is a depen-
dence of the intrinsic reaction rate with the polarity and with the
presence of polar groups with HBD–HBA properties whereas there
is not such correlation for the homogeneous acid catalyst.
Fig. 7B showed that the presence of polar groups in ester chain
with capacity of building up HBD and HBA interactions (like OH
group) remarkably increases the reaction rate. On the contrary
when the polar group does not show that capacity the reaction
rate is scarcely affected. We propose that HBD and HBA interac-
tions between the OH group of the ester chain and other groups
present close to it stabilize the species involved in the rate de-
termining step of the reaction. Scheme 2 describes some of the
configurations of the chemisorbed activated complex derived from
the ethyl 6-hydroxyhexanoate ester stabilized through HBD and
HBA interactions with other species in its close vicinity. Others
species can be proposed but for the sake of simplicity they are
not included because the main guidelines of our interpretation are
not altered. Species I describes the H-bonds that can be formed
with the methanol molecules present in the liquid medium filling
the pores of the solid. A similar picture can be drawn, albeit with
the hydroxyl group of the other hydroxyl-ester molecule present
in the liquid. Species II depicts the H-bonds that can be formed
between the OH-ending of two supposedly adjacent activated es-
ter molecules, although those species should be less likely. Species
III and species IV describe the H-bonding of the OH-ending of
the activated complex with the adjacent sulfonic group and with
the self-activating sulfonic group. These species summarizes other
configurations that can be proposed and that display HBA–HBD in-
teractions with polar groups present at the surface of the resin or
with chemisorbed alcohol (methanol or ethanol) molecules. Since
the species I and II can be also present in the homogeneous catal-
ysis, we propose that the interactions described by species III and
IV, or others alike, are more responsible of the stabilization of the
activated complex involved in the heterogeneous rate determin-
ing step. In other words it is the HBD–HBA interactions of the
ester chain of activated complex with methanol in the pores and
with sulfonic groups at the surface of the resin that notably im-
proves the reaction rate. The increase is much more modest if the
acid chain cannot establish these H-bonding interactions like for
instance the Br-substituted ester. We cannot discard that the tran-
sition state of the rate determining step may also be affected by
any of these interactions, but this hypothesis cannot be assessed
with the experimental evidence provided in this work.
4. Discussion
When looking for hypotheses that explain the effect of the po-
larity on the reaction rate we consider that the reaction mecha-
nism must be taken into account. The reaction mechanism for the
homogeneous catalysts is well known. However the heterogeneous
case has been recently unveiled. It has been recently reported that
the reaction mechanism of esterification and transesterification by
heterogeneous catalyst follows the same steps as the homogeneous
case [7,15,18,19,23]). In Scheme 1, this now widely accepted mech-
anism of transesterification of ethyl esters by methanol in the pres-
ence of a sulfonic resin is presented (the steps for the esterification
reaction with methanol would be similar, but involving carboxylic
acids). The principal peculiarities between heterogeneous and ho-
mogeneous mechanisms refer to the following two aspects: firstly,
the activation of the carbonyl group of the ester comes with the
chemisorption of the ester molecule on the Brønsted acid site. This
chemisorption takes place by the acid attack on the O atom of
carbonyl group forming the protonated carbocation intermediate.
The second aspect refers to the Eley–Rideal model to account for
the nucleophilic attack of the methanol molecule on the carboca-
tion of the activated ester. As in the case of homogeneous system,
this is the rate determining step [7,15,18,19,23]. Although methanol
can also be chemisorbed on the Brønsted acid sites by the sulfonic
attack on the O atom of the methanol molecule, the methanol in-
volved in the rate determining step is not chemisorbed, but in the
liquid medium present in the pores of the solid catalyst. These are
the key peculiarities of the heterogeneous mechanism. The rest of
the steps are similar to the homogeneous reaction, none of them
are the rate determining step: inner rearrangement of protons and,
finally, the deprotonation of the complex with the correspond-
ing formation of the new ester and the recovery of the catalytic
species.
Considering this mechanism, we hypothesized that any varia-
tion in the polarity of the acid chain of the ester that helps to
stabilize the chemisorbed activated intermediate participating in
the rate determining step will favor the reaction: the rate de-
termining step is faster as the concentration of the chemisorbed
species is larger.
We propose that the dependence of the rate for solid resin with
polarity described in Fig. 7A can be explained, besides by steric
constraints, by repulsion effects. We propose that if long alkyl acid
chains are involved, the stabilization of the activated complex is
less favored because of the increase of the repulsion between the
The information derived from this investigation can be extrap-
olated to other situation like of the effect of the polarity of the
carboxylic acid chains on the esterification reaction and the ef-
fect of the polarity of the alcohol moiety of the ester on the
transesterification reaction. The same type of attractive and repul-
sive forces and H-bonding interactions can be proposed in such
cases. However, extrapolation to the polarity effect of the reacting

26
D.M. Alonso et al. / Journal of Catalysis 262 (2009) 18–26
Scheme 2.
alcohol molecule in the transesterification and esterification reac-
tions cannot be done. In principle, according to the Eley–Rideal
mechanism, the chemisorption of the alcohol molecule is not re-
quired and the effects described here refer to the stabilization of
a chemisorbed molecule. The authors are unaware of any research
conducted comparing the effect of the length and of the polar-
ity of the alcohol chain on the reaction rate in both homogeneous
and heterogeneous reaction rates, and therefore cannot comment
on this aspect.
Acknowledgments
Financial support from the Spanish Ministry of Science and
Innovation through the projects BIOGLICAT (ENE-2006-15116-C04-
01) and CTQ-2005-00114/BQU is grateful acknowledged. D.M.A.
thanks the Comunidad de Madrid for a pre-doctoral fellowship. Ms.
Laura Barrio Pliego is also thanked for her help in determining the
size of the esters.
References
5. Conclusions
[1] J. Barrault, Y. Pouilloux, J.M. Clacens, C. Vanhove, S. Bancquart, Catal. Today 75
(2002) 177.
The reaction rate of transesterification catalyzed by solid acids
decreases upon increasing the length of the alkyl chain whereas for
homogeneous acids reaches a plateau for esters with acid chains
larger than 5 atoms. We propose that this trend in heterogeneous
catalyst is explained not only because steric hindrances but also
because of the repulsion between the non-polar chain of the ester
with the polar acid groups on the catalyst surface and/or with the
liquid methanol molecules present in the catalyst pores.
Repulsion or H-bonds effects are of no importance in the case
of homogeneous acid catalyst because the active intermediate is
not attached to a polar surface and this explains the larger intrinsic
reaction rates of homogeneous acid catalysts when compared to
those of solid acids.
[2] G.D. Yadav, H.B. Kulkarni, React. Funct. Polym. 44 (2000) 153.
[3] J. Devaux, P. Godard, J.P. Mercier, Polym. Eng. Sci. 22 (1982) 229.
[4] G.D. Yadav, M.S. Krishnan, Org. Process Res. Dev. 2 (1998) 86.
[5] J. Ni, F.C. Meunier, Appl. Catal. A Gen. 333 (2007) 132.
[6] J. Van Gerpen, Fuel Process. Technol. 86 (2005) 1097.
[7] Y.J. Liu, E. Lotero, J.G. Goodwin, J. Catal. 243 (2006) 221.
[8] D.E. Lopez, J.G. Goodwin, D.A. Bruce, E. Lotero, Appl. Catal. A Gen. 295 (2005)
97.
[9] D.E. Lopez, J.G. Goodwin, D.A. Bruce, J. Catal. 245 (2007) 381.
[10] Y.J. Liu, E. Lotero, J.G. Goodwin, J. Mol. Catal. A Chem. 245 (2006) 132.
[11] E. Lotero, Y.J. Liu, D.E. Lopez, K. Suwannakarn, D.A. Bruce, J.G. Goodwin, Ind.
Eng. Chem. Res. 44 (2005) 5353.
[12] C. Reichardt, Chem. Rev. 94 (1994) 2319.
[13] C. Reichardt, Solvents and Solvent Effects in Organic Chemistry, 3rd edition,
Wiley–VCH, Weinheim, 2003.
The presence of functionalities in the ester chain with capacity
to build up H-bonds, for instance hydroxy substituents, remark-
ably increases the transesterification reaction rate when a hetero-
geneous acid catalyst is used. We propose that the chemisorbed
intermediate participating in the rate determining step of the re-
action is stabilized by the formation of H-bonds with the acid
groups present at the surface of the catalyst and/or with the liq-
uid methanol molecules in the pores. This leads to a larger surface
concentration of the active intermediate and therefore to an im-
provement of the reaction rate. The presence of other polar func-
tionalities but without HBA–HBD properties does not have this
positive effect.
[14] Y. Marcus, Chem. Soc. Rev. 22 (1993) 409.
[15] J. Lilja, D.Y. Murzin, T. Salmi, J. Aumo, P.M. Arvela, M. Sundell, J. Mol. Catal. A
Chem. 182 (2002) 555.
[16] B. Corain, M. Zecca, K. Jerˇábek, J. Mol. Catal. A Chem. 177 (2001) 3.
[17] Y. Pouilloux, S. Abro, C. Vanhove, J. Barrault, J. Mol. Catal. A Chem. 149 (1999)
243.
[18] Y.J. Liu, E. Lotero, J.G. Goodwin, J. Catal. 242 (2006) 278.
[19] S.H. Ali, A. Tarakmah, S.Q. Merchant, T. Al-Sahhaf, Chem. Eng. Sci. 62 (1907)
3197.
[20] M.J. Frisch, et al., Gaussian 03, Revision C.02, Gaussian, Inc., Wallingford, CT,
2004.
[21] D. Datta, D. Majumdar, J. Phys. Org. Chem. 4 (1991) 611.
[22] H. Fujimoto, Y. Mizutani, J. Endo, Y. Jinbu, J. Org. Chem. 54 (1989) 2568.
[23] S.Z.M. Shamshuddin, N. Nagaraju, J. Mol. Catal. A Chem. 273 (2007) 55.