Bioresource Technology 101 (2010) 245–254

Contents lists available at ScienceDirect

Bioresource Technology
journal homepage: www.elsevier.com/locate/biortech

Lubricants from chemically modified vegetable oils

Alejandrina Campanella a, Eduardo Rustoy b, Alicia Baldessari b, Miguel A. Baltanás a,*
a
Instituto de Desarrollo Tecnológico para la Industria Química (UNL – CONICET), Güemes 3450, S3000GNL Santa Fe, Argentina
b
Departamento de Química Orgánica y UMYMFOR, Facultad de Ciencias Exactas y Naturales (UBA), Piso 3, Pabellón II, Ciudad Universitaria, C1428EGA,
Ciudad Autónoma de Buenos Aires, Argentina

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

Article history: This work reports laboratory results obtained from the production of polyols with branched ether and
Received 4 September 2008 ester compounds from epoxidized vegetable oils pertaining to annual, temperate climate crops (soybean,
Received in revised form 6 August 2009 sunflower and high-oleic sunflower oils), focusing on their possible use as components of lubricant base
Accepted 6 August 2009
stocks. To this end, two different opening reactions of the epoxide ring were studied. The first caused by
Available online 28 August 2009
the attack with glacial acetic acid (exclusively in a single organic phase) and the second using short-chain
aliphatic alcohols, methanol and ethanol, in acid media. Both reactions proceed under mild conditions:
Keywords:
low synthesis temperature and short reaction times and with conversions above 99%. Spectroscopic
Vegetable oils
Oxirane ring-opening
(NMR), thermal (DSC) and rheological techniques were used to characterize the oils, their epoxides
Polyol derivatives and polyols, to assess the impact of the nature of the vegetable oil and the chemical modifications intro-
Lubricants duced, including long-term storage conditions. Several correlations were employed to predict the viscos-
ity of the vegetable oils with temperature, and good agreement with the experimental data was obtained.
Ó 2009 Elsevier Ltd. All rights reserved.

1. Introduction complex structures that improve both the low-temperature and

oxidative stability properties (Erhan and Perez, 2002; Lathi and
New technologies aimed at the development of products from Mattiasson, 2006; Hwang and Erhan, 2001; Hwang et al., 2003;
renewable sources have emerged during the last decade, due to in- Adhvaryu et al., 2005; Sharma et al., 2006, 2008).
creased concerns over the use of petroleum-based products caused Epoxidized vegetable oils are promising intermediates for the
by the progressive depletion of the world reserves of fossil fuels, utilisation of VO given that the epoxide, or oxirane, group can be
but also owing to concerns on their environmental impact. easily functionalized. Erhan and co-workers (Erhan and Perez,
In this regard, vegetable oils (VO) constitute a suitable alterna- 2002; Hwang and Erhan, 2001; Hwang et al., 2003; Erhan et al.,
tive for replacing ‘mineral oils’, as they are wholly biodegradable, 2008) recently worked on the production of lubricant base stocks
non-toxic, and ‘Generally Regarded as Safe’ (GRAS) products (Erhan from acid-catalyzed oxirane ring-opening reactions using epoxi-
and Perez, 2002; Erhan and Asadauskas, 2000). Indeed, VOs possess dized soybean oil, employing several alcohols and later esterifying
most of the desirable lubricity properties, such as good contact the resulting hydroxyl group with an acid anhydride. They also
lubrication, high viscosity index (i.e., minimum changes in viscos- synthesized a biolubricant from di-hydroxylated soybean oil, ob-
ity with temperature), high flash-point and low volatility. They tained by acid hydrolysis of the epoxide group followed by esteri-
also have drawbacks, however, that must be overcome, including fication of said hydroxyl groups (Adhvaryu et al., 2005; Sharma
poor low-temperature properties (such as opacity, precipitation, et al., 2006). More recently, they also produced acyl derivatives
poor flowability and/or solidification at relatively moderate tem- of soybean oil via the opening of the oxirane ring using acid anhy-
perature), their susceptibility to oxidative degradation and their drides and employing BF3 etherate as catalyst. These products ap-
propensity to undergo hydrolysis in acid media. The latter, never- pear to be promising lubricants (Sharma et al., 2008).
theless, can be attenuated with relative ease through the use of In recent years, there has been an increased interested in the
additives (Hwang and Erhan, 2001; Hwang et al., 2003). production of biodiesel (BD) as a more benign fuel. BD is usually
Deliberate modification of the chemical structure of VOs is a produced by the transesterification of vegetable oil triglycerides
sound alternative to allow their direct use as lubricant basestocks. with an aliphatic alcohol (such as, methanol) employing sodium
For instance, unsaturated alkene groups of the fatty acid chains in hydroxide as a catalyst. Fatty acid methyl ester (FAME) are
triglyceride molecules can be deliberately altered to give more obtained as the main product of this reaction. Thus, FAMEs have
become extensively available and are produced with high purity.
* Corresponding author. Tel.: +54 342 455 9175; fax: +54 342 455 0944. This has open new pathways to the synthesis of oleochemical
E-mail address: tderliq@santafe-conicet.gov.ar (M.A. Baltanás). products, such as epoxidized FAMEs (Campanella et al., 2008)

0960-8524/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biortech.2009.08.035
246 A. Campanella et al. / Bioresource Technology 101 (2010) 245–254

era General Deheza S.A., Córdoba, Arg.) and high-oleic sunflower

oil (Ecoop brand, from SEDASA, Buenos Aires, Arg.) were used as re-
ceived. Crude soybean oil (from Oleaginosa Humboldt, Santa Fe,
Arg.) was refined in our laboratory. The main fatty acids composi-
tion of the oils was determined by gas chromatography (GC) after
derivatization of the triglycerides to fatty acid methyl esters, fol-
lowing the AOCS Official Method Ce 1e (Official Methods and Rec-
ommended Practices of the American Oil Chemists’ Society – AOCS,
1997). The iodine value was also determined using the AOCS pro-
tocol, Official Method Da 15-48 (Table 1).
Reagents used in the synthesis and/or preparation steps were
Scheme 1. Oxirane ring-opening reaction of epoxidized vegetable oils. glacial formic and acetic acids (acs, 99.81%), fluoroboric acid,
HBF4 (acs, 50%), p.a. grade methanol, ethanol and benzene and
50 wt.% hydrogen peroxide, all purchased from Fisher Scientific
and, also, lubricants (Moser and Erhan, 2008; Moser et al., 2007a,b; (Pittsburgh, PA). Water was doubly distilled (DD) and deionized.
Moser and Erhan, 2007). Moser et al. (2007, 2008) have recently
been producing branched chain derivatives, partially hydrogenated
and, also, diesters through the chemical modification of FAMEs and 2.2. Preparation of the epoxidized vegetable oils
fatty acids.
In this work we propose different synthesis alternatives that Stocks of epoxidized soybean oil (ESBO), epoxidized sunflower
can be carried out prior to the esterification step (reaction with oil (ESO) and epoxidized high-oleic sunflower oil (EHOSO) were
an acid anhydride). Several epoxidized vegetable oils obtained prepared by reacting each VO with performic acid generated
from temperate climate crops (soybean, sunflower and high-oleic in situ, at low temperature (20–40 °C), using benzene as a diluent
sunflower oil) were used for the production of polyols. Two types of the organic phase to lower viscosity and minimize ring opening
of ‘degradation’ reactions of the epoxide group were studied (Gan et al., 1992). The chosen molar ratio of hydrogen peroxide/
(Scheme 1). In the first approach, we employed glacial acetic acid formic acid/double bonds of the oils was 20/2/1. The VOs, benzene
(i.e., in organic phase exclusively), whereas the second reaction in- and formic acid were jointly placed into a well-stirred, round-bot-
volved the opening of the oxirane ring with a short-chain aliphatic tom glass reactor kept at 20 °C and then dilute H2O2 (30 wt.%) was
alcohol (methanol and ethanol) in aqueous, acid media. Only epox- added dropwise. The reactor temperature was then slowly raised
idized soybean oil was used in the second study. Various thermal to 40 °C, to complete the reaction (20 h). The mixtures were next
(differential scanning calorimetry, DSC) and rheological character- washed with distilled water until complete removal of performic
ization studies were performed with the VOs, their epoxides and and formic acids from the organic phase was achieved, and then
polyol derivatives, to evaluate each natural oil and the impact of with NaCl (5 wt.%). Finally, water and benzene were separated
introducing chemical modifications – the results of which expand from the epoxidized oils in a rotary evaporator. Oxirane and hydro-
the available database on these novel products. xyl values of the oils were analyzed using AOCS recommended
practices, methods Cd 9-57 and Tx 1a-66, respectively.
2. Methods
2.3. Oxirane ring-opening reactions with glacial acetic acid
2.1. Materials
Each of the epoxidized oils was first used to study the epoxide
Well-refined, edible quality (i.e., degummed, neutralized, ring-opening with glacial acetic acid. These reacting systems fea-
bleached and deodorized) sunflower oil (Natura brand, from Aceit- ture a single, organic phase (Zaher et al., 1989; Campanella and

Table 1
Main fatty acid composition and iodine values (IV) of the vegetable oils used.

Oil Fatty acids (wt.%) IV

Total saturated Unsaturated
Oleic Linoleic Linolenic
Soybean oil (SBO) 12.1 21.3 59.8 6.80 130
Sunflower oil (SO) 6.5 23.5 67.5 0.50 128
High-oleic sunflower oil (HOSO) 6.46 86.9 6.64 0.00 84

Table 2
Reaction conditions and final epoxide content in the oxirane ring-opening reaction with glacial acetic acid (base reagent: epoxidized soybean oil).

Run T (°C) [AA]° (M)a Initial AA/epoxide groups molar ratio Total duration (min) Final content of epoxide groupsb (wt.%) Final conversionc
A1 70 12 12.5 1200 0.68 90.08
A2 80 12 12.5 660 0.094 98.63
A3 90 12 12.5 360 0.036 99.46
A4 70 14 16.8 840 0.10 98.53
A5 70 8 4.2 2085 0.086 98.73
a
Initial molarity of acetic acid, [AA]° = mol AA/l of total volume of the reacting system.
b
Final weight percent of the epoxide groups in the organic phase (% Epfinal).
c
The conversion of the epoxide group was calculated as: [(% Epinitial  % Epresidual)/% Epinitial]  100. The initial weight percent of epoxide (% Epinitial) in the organic phase
was 6.78 (This is equivalent to 0.42 mol of epoxide per 100 g of epoxidized soybean oil).
 A. Campanella et al. / Bioresource Technology 101 (2010) 245–254 247

Baltanás, 2006). About 63 ml of the epoxides were placed into a 5 ml samples were taken and decanted in a separatory funnel. After
three-necked round-bottom flask, furnished with a reflux con- phase separation, chloroform was added to the organic phase and
denser and a mechanical stirrer (1500 rpm). The flask was heated washed twice with 200 ml of distilled water. The samples were
to the desired working temperature in a thermostatized bath. Next, subsequently washed with 200 ml of NH4OH (5 wt.%) and, finally,
a suitable volume of glacial acetic acid, previously heated at the with a saturated solution of NaCl, after which they were dried
same temperature, was added to the reactor. The reactor tempera- using a rotary evaporator under reduced pressure to eliminate
ture was kept to within ±1 °C. Sample aliquots were periodically traces of water and the aliphatic alcohol.
taken, thoroughly washed to eliminate acid, and dried using a ro-
tary evaporator under reduced pressure prior to their derivatiza- 2.5. Analytical methods
tion and chemical analysis. The experimental program covered a
broad range of temperature and acetic acid concentration, as illus- The temporal evolution of the oxirane ring-opening reactions
trated in Table 2 and Fig. 1 for the case of ESBO. was followed using a conventional volumetric technique, the AOCS
Official Method Cd 9-57 (AOCS, 1997). The products obtained after
2.4. Oxirane ring-opening reactions with short-chain aliphatic alcohols each reaction were analyzed by NMR, according to well-estab-
lished chemical shift assignments (Aerts and Jacobs, 2004; Aerts
Polyol derivatives were prepared via the reaction of ESBO with et al., 2005; Sharmin et al., 2007; La Scala and Wool, 2002; Lathi
either methanol or ethanol at moderate temperature, using a and Mattiasson, 2006; Sharma et al., 2006; Miyake et al., 1998;
homogeneous acid catalyst, HBF4 (Scheme 1b and Table 3), follow- Karmee et al., 2004; Lie Ken Jie and Lam, 1995; Isbell and Mund,
ing Guo et al.’s procedure (Guo et al., 2000). The molar ratio of alco- 1998).
hol to epoxide groups was 4/1. Water and fluoroboric acid were
always added to the mixtures (10 wt.% and 0.4 wt.% with respect 2.5.1. 1H, 13C and 13C DEPT-135 NMR
to the reacting mixtures, respectively). The epoxidized soybean The samples were analyzed in a Bruker AMX 500 unit (Bruker
oil was placed into the mechanically agitated, three-necked BioSpin GmbH, Rheinstetten, Germany) at 25 °C, using 5 mm tubes.
round-bottom flask, which was heated to the desired working tem- For 1H NMR, 60 mg of the samples were dissolved in 0.6 ml of
perature in the thermostatized bath. A suitable volume of the CDCl3. A total of 16 and 5000 scans were used for the 1H and 13C
aqueous solution of the aliphatic alcohol and HBF4, previously analyses, respectively, with a pulse width of 90°. The relaxation de-
thermostatized at the same temperature, was added to the reactor. lay was varied according to the functional level of each VO. For
The reactor temperature was kept to within ±1 °C. Periodically, example, the relaxation delays for SBO and ESBO were 5 s, whereas
the epoxide ring-opening products of ESBO exhibit relaxation de-
lays of 8 s.

2.5.2. Differential scanning calorimetry (DSC)

Experiments were carried out using a Mettler Toledo Star Ther-
mal Analysis System DSC821 (Mettler-Toledo International Inc.,
Greifensee, Switzerland). Prior to measuring, the temperature
accuracy and heat flow precision of the DSM modulus was checked
with pure indium. Approximately 15 mg of each sample was
placed inside a standard aluminum crucible, which was then her-
metically sealed and placed into the DSC modulus together with
an identical, empty (reference) crucible. The heating ramp used
in all studies was 10 °C/min.

2.5.3. Viscosity
Rheological measurements were performed at 25 °C using a
stress-controlled rheometer, Rheostress 80 (Haake Instrument
Inc., Paramus, NJ). A cone and plate geometry (40 mm diameter,
0.04 rad, 1 mm gap, 1 Hz) was used for all the samples. Ascending
and descending shear stress vs. shear rate curves (carried out in the
Fig. 1. Oxirane ring-opening reaction with glacial acetic acid (base reactant:
1–6000 s1 range) were obtained. The effect of temperature on the
epoxidized soybean oil). The operating conditions corresponding to each run are
indicated in Table 2. viscosity of the oils, their epoxides and the polyol derivatives was
also assessed. Samples were equilibrated at 20 °C prior to measure-

Table 3
Reaction conditions and final content of epoxide in the oxirane ring-opening reaction of epoxidized soybean oil with short-chain aliphatic alcohols (using an aqueous mixture of
the alcohol and fluoroboric acid).

Run T Aliphatic Initial alcohol/epoxide group [HBF4]0 [H2O]0 Total duration Final content of epoxide groupsa Conversionb
(°C) alcohol mole ratio (wt.%) (wt.%) (min) (wt.%)
B1 30 Methanol 4 0.4 10 240 0.71 88.98
B2 40 Methanol 4 0.4 10 120 0.78 87.87
B3 50 Methanol 4 0.4 10 60 0.078 98.80
B4 45 Ethanol 4 0.4 10 120 0.081 98.75
B5 50 Ethanol 4 0.4 10 60 0.062 99.04
a
Final weight percent of the epoxide groups in the organic phase (% Epfinal).
b
The conversion of the epoxide group was calculated as: [(% Epinitial  % Epresidual)/% Epinitial]  100. The initial weight percent of epoxide (% Epinitial) in the organic phase
was 6.78 (This is equivalent to 0.42 mol of epoxide per 100 g of epoxidized soybean oil).
248 A. Campanella et al. / Bioresource Technology 101 (2010) 245–254

ment. Experiments were performed by slowly heating samples to

70 °C with further cooling back to 20 °C. Duplicate or triplicate
runs were performed with each sample.

3. Results and discussion

3.1. Oxirane ring-opening reactions

The first reaction studied was the opening of the oxirane ring of
the epoxidized oils caused by the attack of glacial acetic acid (i.e.,
exclusively in the organic phase). For ESBO both the reaction tem-
perature and the molar ratio of epoxide groups (Ep) to acetic acid
(AA) were varied (Table 2 and Fig. 1). It can be readily appreciated
that the final conversion could be achieved at moderate to short
reaction times. For epoxidized sunflower and high-oleic sunflower
oils a single concentration of AA (12 M) and a single temperature
(70 °C) were employed. The kinetics of this reaction has already
been studied (Zaher et al., 1989; Campanella and Baltanás, 2006).
The constitutive equation for the reaction rate in the organic phase Fig. 2. Oxirane ring-opening reaction of epoxidized soybean oil with short-chain
is the following: aliphatic alcohols using aqueous solutions of alcohol and fluoroboric acid as a
catalyst. The operating conditions corresponding to each run are indicated in Table
org
Rorg ¼ kAA ð½AAorg Þ2 ½Eporg V org ð1Þ 3.

org org org

where kAA is the reaction rate constant, while [AA] and [Ep] are
molar concentrations in the organic phase. The activation energy of with the desired ring-opening reaction. Most likely this better
this reaction is 16.5 ± 1.1 kcal mol1 and the kinetic rate constant, reactivity is due to shorter reaction times and lower reaction tem-
org 2 2 1
kAA ¼ 2:81  105 l =ðmol min Þ, at 70 °C (Campanella and Baltanás, perature, since our syntheses at 50 °C require approximately one
2006). hour for completion.
Epoxide ring-opening with acetic acid can be used as an initial Summarizing, both oxirane ring-opening reactions employed in
synthetic step in lubricant production, followed by the reaction this study are put forward here as possible alternatives to those al-
with acetic anhydride (not shown here), to produce di-acetate veg- ready available in the literature (Erhan and Perez, 2002; Hwang
etable oils derivatives. A similar product was developed by Adh- and Erhan, 2001; Hwang et al., 2003; Adhvaryu et al., 2005), pri-
varyu et al. (2005), whom synthesized di-hexanoyl soybean oil in marily for their shorter reaction times, lower reaction temperature,
two steps. The first involved the production of di-hydroxylated high conversion (ffi99%) and selectivity, and desirable lubrication
soybean oil from ESBO at 100 °C (48 h) in the presence of HClO4. properties, as will be discussed in the following sections.
In a second step, the di-hydroxylated soybean oil was reacted with
an anhydride (e.g., acetic or hexanoic anhydrides), in which final
3.2. 1H NMR characterization
products were found to be optimal lubricants after a final esterifi-
cation step. Thus, comparing the first synthesis step in Adhvaryu
The peak assignments in the 1H NMR spectra of refined soybean
et al.’s (2005) work to the reaction described previously, and taking
oil (SBO) and epoxidized soybean oil (ESBO), and in the final prod-
into account that the second synthesis step could be equivalent to
ucts of the oxirane ring-opening reactions of ESBO with methanol
their procedure, using acetic acid can significantly decrease both
(MeOSBO), ethanol (EtOSBO) and glacial acetic acid (AcOSBO) are
working temperature and reaction times. For instance, using
shown in Table 4 (see also Scheme 2). To this end, we also used
[AA]° = 12 M at 90 °C, the reaction is almost complete in about
6 h (conversion ca. 99%).
Scheme 1b illustrates the second reaction that was studied: at- Table 4
tack (opening) on the oxirane-ring with an alcohol, in the presence Assignment of the 1H NMR spectral lines of soybean oil (SBO) and epoxidized soybean
of an acid catalyst. Aqueous solutions of two short-chain aliphatic oil (ESBO) and its oxirane ring-opening products upon using methanol (MeOSBO),
ethanol (EtOSBO) and glacial acetic acid (AcOSBO).
alcohols (methanol and ethanol) were used, at mild temperatures.
Only the impact of temperature on the time of the reaction was H d (ppm)
studied, using a single set of reactants and catalyst concentrations SBO ESBO MeOSBO EtOSBO AcOSBO
(Table 3). Fig. 2 shows that the reaction rate almost doubles with A 5.40–5.21 5.27–5.18 5.25–5.19 5.23–5.16 5.28–5.23
each 10 °C-temperature increase. At 50 °C the conversion is about A0 – 2.98–2.82 4.20–3.20 4.19–3.20 5.21–4.73
99% with both alcohols. Additionally, Fig. 2 indicates that there is 4.20–3.29
practically no difference between both alcohols regarding reaction A00 – 3.18–2.99 4.20–3.20 4.19–3.20 5.21–4.73
4.20–3.29
time (Runs 3 and 5).
B 4.30–4.10 4.32–4.10 4.29–4.06 4.28–4.04 4.31–4.05
Erhan and co-workers have also performed these syntheses, C 2.34–2.26 2.32–2.22 2.34–2.29 2.31–2.16 2.29–2.23
using sulphuric acid as a catalyst (Erhan and Perez, 2002; Hwang D 2.09–1.96 1.52–1.36 2.09–1.68 2.07–1.76 2.07–1.93
and Erhan, 2001). They found that the kinetics were highly depen- E 1.64–1.55 1.62–1.53 1.65–1.46 1.61–1.49 1.72–1.16
dent on the type of alcohol used, where their study focussed on a F 1.38–1.20 1.35–1.16 1.45–1.10 1.47–1.16
G 0.90–0.83 0.90–0.80 0.86–0.78 0.89–0.79 0.97–0.81
wide variety of species: methanol, 2-ethylhexanol, butanol, cyclo- G0 0.96 1.10–0.99 0.86–0.78 0.89–0.79 –
hexanol and decanol. Yet, their reaction times were larger, even H 2.81–2.73 1.80–1.63 2.65–2.35 2.66–2.40 2.37–2.30
though the temperatures they used were higher. For instance, the OCH3 – – 3.40–3.34 – –
reaction with methanol at 70 °C required 5 h to reach completion. OCH2CH3 – – – 3.62–3.48 –
OCH2CH3 – – – 1.18–1.09 –
In deviation with our findings, they systematically observed
OOCCH3 – – – – 2.06–2.00
transesterification of the triglycerides with the alcohol, together
 A. Campanella et al. / Bioresource Technology 101 (2010) 245–254 249

Scheme 2. Peak assignments in the 1H NMR spectra of soybean oil (SBO) and epoxidized soybean oil (ESBO) and its oxirane ring-opening products upon using methanol
(MeOSBO), ethanol (EtOSBO) and glacial acetic acid (AcOSBO).

information extracted from unidimensional 13C and 13C DEPT-135 2002; Lathi and Mattiasson, 2006; Sharma et al., 2006; Miyake
analyses, as well as bidimensional COSY and HSQC, together with et al., 1998; Aerts and Jacobs, 2004; Karmee et al., 2004; Lie Ken
literature data from related compounds (La Scala and Wool, Jie and Lam, 1995; Isbell and Mund, 1998). The 1H NMR informa-
250 A. Campanella et al. / Bioresource Technology 101 (2010) 245–254

tion was used to establish the extent of selectivity to the desired first group of signals is located in the zone of quaternary carboxylic
product(s) and the degree of conversion along each reaction carbons (171.0–170.5 ppm), and the second in the 81.7–72.3 ppm
pathway. range, which corresponds to the A0 and A00 groups. The latter were
observed in the 57.0–54.1 ppm region in the ESBO spectrum.
3.2.1. Epoxidation reaction
A decrease in the relative intensity of the protons, labelled A, 3.2.3. Oxirane ring-opening with methyl alcohol
was observed. In the case of SBO, the signal at d 5.40–5.21 ppm The first confirmation of ring opening was the shift of signals
is attributed to vinyl protons of the fatty acid chain moiety in the corresponding to the A0 and A00 nuclei, which were observed for
oil and to the methyne group of the glyceryl residue. Additionally, ESBO in the 2.98–2.82 and 3.18–2.99 ppm regions, respectively,
two new groups of signals, labelled as A0 and A00 in Table 4, appear but appeared between 4.20 and 3.20 ppm in the MeOSBO spectra.
between 2.98–2.82 and 3.18–2.99 ppm, respectively, which are as- In addition, a new group of signals was observed, located between
signed to the methyne groups of the new oxirane residue. It was 3.40 and 3.34 ppm, assigned to the protons of the methoxyl resi-
also observed that the nuclei designated as D show a displacement due, and a shift to lower fields of the H nuclei (from 1.80–
to 2.09–1.96 ppm in the SBO sample, which is typical of allylic 163 ppm in ESBO to 2.65–2.35 ppm in MeOSBO). The same effect
groups, whereas in the case of ESBO these protons appear between was observed with respect to the D nuclei (Table 4).
1.52 and 1.36 ppm, a characteristic of methylene groups in position From the comparison of the 13C NMR spectra of ESBO and MeO-
a with respect to the oxirane groups. However, the protons of the SBO, it can be noted that the latter shows signals between 72.5 and
bis-allylic methylene groups (signal labelled H in Table 4), under- 82.8 ppm, assigned to the A0 and A00 groups, and another signal at
went a shift in ESBO with respect to SBO, from 1.80–1.63 to 58.1 ppm corresponding to the methoxyl groups. This information
2.81–2.73 ppm. Lastly, a shift of the signal assigned to the methyl confirms the opening of the oxirane ring and formation of ether
groups at the end of the fatty acid chain residues, G and G0 , was links.
also observed (Table 4). From the observation of the SBO and ESBO
13
C NMR spectra, it is evident that vinyl signals present in SBO 3.2.4. Oxirane ring-opening with ethyl alcohol
(131.9–121.1 ppm range) disappear while a new group of signals Analogous to the previous case, the signals were used to con-
emerged in the ESBO spectrum, between 57.0 and 54.1 ppm, which firm the ring-opening reaction corresponding to the A0 , A00 , D and
were assigned to the oxirane group. H groups. In addition, signals were observed in the 3.62–
3.48 ppm and 1.18–1.09 ppm regions, corresponding to the pro-
3.2.2. Oxirane ring-opening with glacial acetic acid tons of the methylene and methyl groups of the ethoxyl residues
The A0 and A00 nuclei were used to analyze he progress of this of the molecules. Likewise, the comparative analysis of the 13C
reaction, both qualitatively and quantitatively (see Scheme 2). NMR spectra showed the presence of the nuclei pertaining to the
Their location is between 5.21 and 4.73 ppm when the R0 or R00 sub- methylene (66.0–65.8 ppm) and methyl (15.8–15.5 ppm) groups
stituents are AcO, and in the 4.20–3.29 ppm range when the R0 or of the ethoxyl residues in the EtOSBO molecule.
R00 substituents are H. The signals corresponding to the D nuclei From the analysis of these 1H, 13C and 13C DEPT-135 spectro-
underwent a shift toward lower fields with respect to their location scopic data, and using related information from the literature
in the ESBO spectrum; that is, from 1.52–1.36 ppm to 2.07– (Lie Ken Jie and Lam, 1995; Karmee et al., 2004), it was concluded
1.93 ppm. These signals are characteristic of those observed in that the structure of the triglycerides is preserved during each of
methylene groups located in a position with respect to CHOR0 (R00 ) the epoxide ring-opening reactions, with either glacial acetic acid
groups. The same effect was observed in the case of the H nuclei or the aliphatic alcohols, whereby the yield was always close to
(Scheme 2). On the other hand, the appearance of a group of signals 99%. The preservation of the triglyceride structure is of the utmost
at 2.06–2.00 ppm, assigned to the protons of the methyl groups in importance with respect to maintaining the biodegradability of
the OOCCH3 residues, allowed us to corroborate the opening of the these products.
oxirane rings, which subsequently led to the formation of ester
bonds. 3.3. DSC analysis
The comparison of the 13C NMR spectra of ESBO and AcOSBO
was used to confirm the opening of the oxirane rings, as new DSC analysis was performed while heating various materials
groups of signals appeared after the attack reaction with AA. The from 100 to 50 °C. The heating rate was 10 °C/min, following pre-

Table 5
Melting and crystallization temperatures of soybean, sunflower and high-oleic sunflower oils, their epoxides and some of the corresponding products after the ring-opening
reactions using acetic acid or aliphatic alcohols.

Source oil Melting temperatures (°C) Crystallization temperature (°C)

Peak 1 Peak 2 Peak 3
Soybean oil SBO 38 25 8 65
ESBO 14 6  35
ESBOa 17 3  25
AcOSBO 27 9 3 –
MeOSBO 25 5 2
MeOSBOa 26 6 3 –
Sunflower oil SO 33 24 9 68
ESO 11 0 – 42
AcOSO 27 4 1.5 
High-oleic sunflower oil HOSO 3 – – 29
EHOSO 36 – – 4
AcOHOSO 35 11 – –
a
Guo et al. (2000).
 A. Campanella et al. / Bioresource Technology 101 (2010) 245–254 251

vious work that suggest this as the optimum value to obtain well- which correspond to the different crystalline polymorphisms (a, b0
resolved DSC features (Erhan and Perez, 2002; Adhvaryu et al., and b) typical of vegetable oil products (Erhan and Perez, 2002;
2003; Guo et al., 2000; Moser and Erhan, 2008). Adhvaryu et al., 2003; Guo et al., 2000).
The melting and crystallization temperatures of the oils and As expected, the higher the degree of unsaturation for each veg-
their epoxides, together with those of the different ring-opening etable oil, the lower its melting point, i.e., highly unsaturated oils
products are detailed in Table 5. Three peaks appear in most cases, exhibit better fluidity properties at lower temperatures (compar-
ing HOSO, SO and SBO, for example). The crystallization tempera-
tures of SBO and SO were quite similar. The epoxidized oils lose
fluidity, but their ring-opened derivatives again show partially re-
stored (often times desirable) fluidity. Our data, obtained using
soybean oil from Argentine cultivars, are very similar to those of
Guo et al. (2000).

3.4. Viscosity analysis

Detailed knowledge of the viscometric properties of natural and

chemically modified vegetable oils is of particular importance
whenever lubrication is their intended usage. The viscosity vs.
temperature data of the SBO, SO and HOSO oils, along with their
epoxides and the polyol derivatives employed in this work, is
shown in Fig. 3. As expected, the viscosity exhibits an exponential
decrease with temperature, as illustrated by the data, with the
exception of EHOSO (which is solid at room temperature). Both
the VO and their epoxides could in principle be used as lubricant
stocks at or above ambient temperature if only the viscosity were
taken into account (and with the aid of antioxidant/antihydrolysis/
antirust packages). Conversely, none of the products obtained after
reaction with either acetic acid or the aliphatic alcohols could be
used as lube oils as such, because of their high viscosity that is well
above the usual Society of Automotive Engineers (SAE) specifica-
tions. Nevertheless, encouraging literature data indicate that after
a second synthesis step (esterification) all of these products can
achieve the desired flowability properties (Erhan and Perez,
2002; Hwang and Erhan, 2001; Adhvaryu et al., 2005; Erhan
et al., 2008).
We also studied the stability of select products obtained after
reacting the epoxidized oils with acetic acid and stored in a refrig-
erator at 4–8 °C. Their viscosity vs. temperature behaviour was
studied one year after their synthesis (Fig. 4) revealing in each case
that the viscosity marginally increased (most likely owing to oligo-
mer formation). Further studies will be needed to investigate the
stability, or ‘shelf life’, of these lube intermediates during storage.

Fig. 3. Viscosity vs. temperature plots for (a) the vegetable oils, (b) the epoxidized Fig. 4. Comparison of viscosity vs. temperature changes with aging (one year)
vegetable oils and (c) various compounds resulting from the ring-opening reactions. between products of the oxirane ring-opening with glacial acetic acid of epoxidized
SBO = soybean oil; SO = sunflower oil; HOSO = high-oleic sunflower oil. soybean oil (SBO), sunflower oil (SO) and high-oleic sunflower oil (HOSO).
252 A. Campanella et al. / Bioresource Technology 101 (2010) 245–254

Vegetable oils are Newtonian fluids, i.e., there is a linear rela- from this curve using the slope from fits of the experimental shear
tionship between shear rate and shear stress (Fig. 5a). This relation stress–shear rate data.
is in agreement with Newton’s law of viscosity, as given by the fol- The epoxidized counterparts of the vegetable oils, and the prod-
lowing equation: ucts derived therein via the opening of the oxirane ring, are thixo-
tropic (Fig. 5b and c), as revealed by the hysteresis loop that
r ¼ lc_ ð2Þ appears on the r vs. c_ plots. Lower shear rates result upon decreas-
ing the applied shear rate as compared with the monotonic in-
where r is the shear stress (Pa), c_ is shear rate (s1), and l is the
viscosity (Pa.s). The viscosity at each temperature can be obtained

Fig. 5. Shear stress vs. shear rate plots of soybean oil (SBO), sunflower oil (SO) and Fig. 6. Comparison of experimental results of viscosity vs. temperature plots for (a)
high-oleic sunflower oil (HOSO), their epoxidized counterparts and some of the soybean oil, SBO, (b) sunflower oil, SO and (c) high-oleic sunflower oil, HOSO.
products of the oxirane ring-opening reactions with glacial acetic acid, methanol or WFL = modified Williams–Landel–Ferry model, PW = Power-Law model; FA = fatty
ethanol (T = 25 °C). acid distribution model.
 A. Campanella et al. / Bioresource Technology 101 (2010) 245–254 253

crease. The epoxidized high-oleic sunflower oil, a solid at room Acknowledgements

temperature, was not analyzed in this study.
The dependency of viscosity with temperature for vegetable oils The authors thank CONICET, ANPCyT and UNL (CAID 2006 grant
can be estimated using the following, contemporary models (Fasi- 29-161) for their financial support. Thanks are given to Pablo
na et al., 2006), such as the modified Williams–Landel–Ferry (WLF) Vicentini and Adolfo Larese for their skilled assistance in the prep-
model, aration of epoxidized vegetable oils and to Daniel De Pianti for his
technical assistance to use the rheometer. A.B. thanks UBA (X089)
aT and ANPCyT (PICT 2005, 32735) for partial financial support. AB
ln l ¼ ð3Þ
ðb þ TÞ and MAB are Research Members of CONICET.

or the power-law model,

Appendix A. Supplementary data
l ¼ kðT  T ref Þn ð4Þ
Fasina et al. (2006) also found that either the mass fractions of Supplementary data associated with this article can be found, in
monounsaturated or polyunsaturated fatty acids, but not the mass the online version, at doi:10.1016/j.biortech.2009.08.035.
fraction of saturated or total unsaturated fatty acids of vegetable
oils, could be correlated with their viscosity as a function of tem- References
perature (5–95 °C). They developed the fatty acid (FA) distribution
model, using data from 12 different oils and non-linear Adhvaryu, A., Erhan, S.Z., Perez, J.M., 2003. Wax appearance temperatures of
vegetable oils determined by differential scanning calorimetry: effect of
regressions: triacylglycerol structure and its modification. Thermochim. Acta 395, 191–200.
    Adhvaryu, A., Liu, Z.S., Erhan, S.Z., 2005. Synthesis of novel alkoxylated
B C triacylglycerols and their lubricant base oil properties. Ind. Crops Prod. 21,
l ¼ A exp þ y exp ð5Þ 113–119.
RT RT
Aerts, H.A.J., Jacobs, P.A., 2004. Epoxide yield determination of oils and fatty acid
methyl esters using 1H NMR. J. Am. Oil Chem. Soc. 81, 841–846.
where y stands for the mass fraction of the fatty-acid group consid- Aerts, H.A.J., Sels, B.F.E., Jacobs, P.A., 2005. The use of 1H NMR for yield
ered (monounsaturated or PUFA), R is the universal gas constant determination in the regioselective epoxidation of squalene. J. Am. Oil Chem.
and T is the absolute temperature (Kelvin). A, B and C are constants Soc. 82, 409–413.
AOCS, 1997. Official and Recommended Practices of the American Oil Chemists’
provided in their work. Society, fifth ed. AOCS Press, Champaign, IL.
To apply these models, the a, b, k and n constants indicated in Campanella, A., Baltanás, M.A., 2006. Degradation of the oxirane ring of epoxidized
Eqs. (2) and (3) (for soybean and sunflower oils) were taken from vegetable oils in liquid–liquid heterogeneous reaction systems. Chem. Eng. J.
118, 141–152.
Fasina et al.’s work. For SBO, SO and HOSO the FA model was also Campanella, A., Fontanini, C., Baltanás, M.A., 2008. High yield epoxidation of fatty
used, employing the A, B and C constants found by Fasina et al. acid methyl esters with performic acid generated in-situ. Chem. Eng. J. 144,
(2006) corresponding to the y mass fractions of either monounsat- 466–475.
Erhan, S.Z., Asadauskas, S., 2000. Lubricant basestocks from vegetable oils. Ind.
urated fatty acids (MUFA) or polyunsaturated fatty acids (PUFA). Crops Prod. 11, 277–282.
Fig. 6 reveals that for both SBO and SO the power-law model Erhan, S.Z., Perez, J.M. (Eds.), 2002. Biobased Industrial Fluids and Lubricants. AOCS
gives a poor estimate of viscosity vs. temperature, while the FA Press, Champaign, IL.
Erhan, S.Z., Sharma, B.K., Liu, Z., Adhvaryu, A., 2008. Lubricant base stock potential of
model-MUFA yields better estimates than the FA model-PUFA. chemically modified vegetable oils. J. Agric. Food Chem. 56, 8919–8925.
The FA model was modified and employed to estimate the viscosity Fasina, O.O., Hallman, H., Craig-Schmidt, M., Clements, C., 2006. Predicting
of the VOs, using both sets of Fasina et al.’s (2006) correlation con- temperature-dependence viscosity of vegetable oils from fatty acid
composition. J. Am. Oil Chem. Soc. 83, 899–903.
stants for both MUFA and PUFA:
Gan, L.H., Goh, S.H., Ooi, K.S., 1992. Kinetic studies of epoxidation and oxirane
     cleavage of palm olein methyl esters. J. Am. Oil Chem. Soc. 69, 347–351.
BMUFA C MUFA Guo, A., Cho, Y., Petrović, Z.S., 2000. Structure and properties of halogenated and
l ¼ yMUFA AMUFA exp þ yMUFA exp
RT RT nonhalogenated soy-based polyols. J. Polym. Sci. Part A: Polym. Chem. 38,
     3900–3910.
BPUFA C PUFA Hwang, H.S., Erhan, S.Z., 2001. Modification of epoxidized soybean oil for lubricant
þ yPUFA APUFA exp þ yPUFA exp ð6Þ formulations with improved oxidative stability and low pour point. J. Am. Oil
RT RT
Chem. Soc. 78, 1179–1184.
Thus, the best estimates of viscosity vs. temperature behaviour Hwang, H.S., Adhvaryu, A., Erhan, S.Z., 2003. Preparation and properties of lubricant
basestocks from epoxidized soybean oil and 2-ethylhexanol. J. Am. Oil Chem.
were obtained, in the complete temperature range, for each of the Soc. 80, 811–815.
vegetable oils in this study (Fig. 6). Isbell, T.A., Mund, M.S., 1998. Synthesis of secondary ethers derived from
meadowfoam oil. J. Am. Oil Chem. Soc. 75, 1021–1029.
Karmee, S.K., Mahesh, P., Ravi, R., Chadha, A., 2004. Kinetic study of the base-
4. Conclusions catalyzed transesterification of monoglycerides from pongamia oil. J. Am. Oil
Chem. Soc. 81, 425–430.
La Scala, J., Wool, R.P., 2002. The effect of fatty acid composition on the acrylation
Two new synthetic routes to polyol-derived intermediates, pro- kinetics of epoxidized triacylglycerols. J. Am. Oil Chem. Soc. 79, 59–63.
posed as an alternative way for the production of lubricants from Lathi, P.S., Mattiasson, B., 2006. Green approach for the preparation of
biodegradable lubricant base stock from epoxidized vegetable oil. Appl. Catal.
vegetable oils obtained from temperate climate crops (soybean B: Environ. 69, 207–212.
and sunflower seeds), are proposed in this work. The oils are first Lie Ken Jie, M.S.F., Lam, C.C., 1995. 1H-Nuclear magnetic resonance spectroscopic
epoxidized, and then the oxirane ring is opened with either acetic studies of saturated, acetylenic and ethylene triacylglycerols. Chem. Phys.
Lipids. 77, 155–171.
acid or a low chain aliphatic alcohol (methanol or ethanol). These Miyake, Y., Yokomizo, K., Matsuzaki, N., 1998. Rapid determination of iodine value
routes are promising in that they demand short reaction times by 1H nuclear magnetic resonance spectroscopy. J. Am. Oil Chem. Soc. 75, 15–
and low reaction temperatures, while ca. 99% conversion is 19.
Moser, B.R., Erhan, S.Z., 2007. Preparation and evaluation of a series of alpha-
achieved, with high selectivities. The product identities were
hydroxy ethers from 9,10-epoxystearates. Eur. J. Lipid Sci. Technol., 206–213.
determined by 1H and 13C NMR spectroscopy. Considering only Moser, B.R., Erhan, S.Z., 2008. Branched chain derivatives of alkyl oleates:
the viscosity of the products, they are all viable lubricant base- tribological, rheological, oxidation, and low temperature properties. Fuel 87,
stocks (or grease basestocks in the case of high-oleic sunflower 2253–2257.
Moser, B.R., Sharma, B.K., Doll, K.M., Erhan, S.Z., 2007a. Diesters from oleic acid:
oil HOSO), in which the intermediates can be further esterified to synthesis, low temperature properties, and oxidation stability. J. Am. Oil Chem.
obtain low-temperature lubricants. Soc. 84, 675–680.
254 A. Campanella et al. / Bioresource Technology 101 (2010) 245–254

Moser, B.R., Haas, M.J., Winkler, J.K., Jackson, M.A., Erhan, S.Z., List, G.R., 2007b. Sharmin, E., Ashraf, S.M., Ahmad, S., 2007. Epoxidation, hydroxylation, acrylation
Evaluation of partially hydrogenated methyl esters of soybean oil as biodiesel. and urethanation of Linum usitatissimum seed oil and its derivatives. Eur. J. Lipid
Eur. J. Lipid Sci. Technol. 109, 17–24. Sci. Technol. 109, 134–140.
Sharma, B.K., Adhvaryu, A., Liu, Z., Erhan, S.Z., 2006. Chemical modification of Zaher, F.A., El-Mallah, M.H., El-Hefnawy, M.M., 1989. Kinetics of oxirane cleavage in
vegetable oils for lubricant applications. J. Am. Oil Chem. Soc. 83, 129–136. epoxidized soybean oil. J. Am. Oil Chem. Soc. 66, 698–700.
Sharma, B.K., Liu, Z., Adhvaryu, A., Erhan, S.Z., 2008. One-pot synthesis of chemically
modified vegetable oils. J. Agric. Food Chem. 56, 3049–3056.