A solvent having switchable hydrophilicity

Article abstract of DOI:10.1039/b926885e

A new kind of switchable solvent, a switchable-hydrophilicity solvent, is hydrophobic and has very low miscibility with water when in air but is hydrophilic and has complete miscibility with water when under an atmosphere of CO₂. We report here the first example of such a solvent, N,N,N′-tributylpentanamidine. Solvents such as these could be used for the extraction of low-polarity organic products, such as vegetable oils, followed by the removal of the solvent from the product by carbonated water. Carbonated water is able to extract the solvent from the product because the CO₂ converts the solvent to its hydrophilic form. The solvent can then be separated from the carbonated water upon removal of the CO₂, because this removal triggers the conversion of the solvent back to its hydrophobic, water-immiscible form. Importantly, distillation is not required for removal of the solvent from the product.

Full text of DOI:10.1039/b926885e

View Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/greenchem | Green Chemistry
A solvent having switchable hydrophilicity†
Philip G. Jessop,* Lam Phan, Andrew Carrier, Shona Robinson, Christoph J. Du¨rr and Jitendra R. Harjani
Received 23rd December 2009, Accepted 4th February 2010
First published as an Advance Article on the web 5th March 2010
DOI: 10.1039/b926885e
A new kind of switchable solvent, a switchable-hydrophilicity solvent, is hydrophobic and has very
low miscibility with water when in air but is hydrophilic and has complete miscibility with water
when under an atmosphere of CO2. We report here the first example of such a solvent,
N,N,N¢-tributylpentanamidine. Solvents such as these could be used for the extraction of
low-polarity organic products, such as vegetable oils, followed by the removal of the solvent from
the product by carbonated water. Carbonated water is able to extract the solvent from the product
because the CO2 converts the solvent to its hydrophilic form. The solvent can then be separated
from the carbonated water upon removal of the CO2, because this removal triggers the conversion
of the solvent back to its hydrophobic, water-immiscible form. Importantly, distillation is not
required for removal of the solvent from the product.
Introduction
The removal of solvents by distillation is a common industrial
practice but suffers from two disadvantages that result in
environmental damage. First, distillation requires the use of
a volatile solvent, which leads to significant vapour emission
losses and the resulting contribution to smog formation; in
Canada alone, 4 400 tonnes per year of hexane are emitted to the
atmosphere, one third of that from oilseed processing.1 Second,
distillation requires a large input of energy. Therefore it would be
desirable to find a new non-distillative route for the separation
of solvents from products, so that volatile solvents would not
Fig. 1 The phase behaviour of a mixture of water and an SHS.
have to be used.
This paper describes the successful search for an SHS and
preliminary tests of that solvent for the extraction of soy oil
from soybean flakes.
Solvent removal from a hydrophobic product, without distil-
lation, could be achieved if there were a solvent that could be
reversibly switched from being hydrophobic to hydrophilic. Such
a solvent could be termed a switchable-hydrophilicity solvent
(SHS). Specifically, we have been looking for a solvent that
exhibits the following desired phase behaviour with water: very
little miscibility of the solvent with water in the absence of
CO2 but complete miscibility with water in the presence of CO2
(Fig. 1).2 CO2 is preferred as the trigger for the switching process
because it is non-toxic, benign, inexpensive and easily removed.
Based upon our previous experience with amidine/CO2 and
guanidine/CO2 chemistry,3-6 we believed that there may be
a liquid amidine or guanidine that would exhibit this phase
behaviour. If such a solvent were found, then it could be used as
a substitute for volatile solvents and be removed from organic
products (such as oilseed oils), without distillation, by extraction
of the solvent with carbonated water.
Results and discussion
A series of amidines was tested for the desired phase behaviour:
immiscibility with water under air and miscibility with water in
the presence of 1 bar of CO2. The compounds tested are shown
in Scheme 1. Amidines 1 and 2 were unsatisfactory because they
were soluble in water, even in the absence of CO2. Two amidines
with greater hydrophobicity (3 and 4) exhibited the desired phase
behaviour.
Department of Chemistry, Queen’s University, Kingston, Ontario,
Canada K7L 3N6. E-mail: jessop@chem.queensu.ca; Fax: +1 (613)
533-6669; Tel: +1 (613) 533-3212
† Electronic supplementary information (ESI) available: Syntheses
and characterizations of the amidines and guanidines. See DOI:
10.1039/b926885e
Scheme 1 The amidines and guanidines mentioned in this study. The
number in plain font is the predicted log Kow, while the number in bold
font is the compound number.
This journal is
The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 809–814 | 809
©

View Online
Protonation of the amidine by carbonic acid is the cause of
the change in phase behaviour. In the presence of both CO2
and water, amidines 3 and 4 are converted into water-soluble
through the biphasic mixture, the phases merged, and 1H NMR
spectroscopy showed both water and 4, although the peaks of 4
were shifted slightly downfield because of protonation (Fig. 2c).
Heating this homogeneous solution at 80 ◦C for 60 min caused
a phase split, which persisted when the solution was cooled to
room temperature. NMR spectra showed that the upper phase
was 4 (Fig. 2d) while the lower phase still contained some of
the bicarbonate salt of 4 (Fig. 2e). Therefore, the heat treatment
largely reversed the reaction in eqn (1), but not completely. The
amount of amidine that remained in the aqueous phase was
calculated, with the help of an internal standard, to be 11%
of the original amount. A comparison of the spectra in Fig. 2
clearly shows that the residual amidine in the water is in the form
of the bicarbonate. Similar experiments with 3 showed that the
conversion from biphasic to monophasic with CO2 was facile
but that the reverse process, restoring the biphasic mixture by
heating, was rarely successful.
1
bicarbonate salts (eqn (1)). The 13C{ H} NMR spectrum of 4
in CO2-saturated D2O shows inequivalent butyl groups on the
-
–NBu2 nitrogen, and peaks at 160.1 (HCO3 ) and 165.9 ppm
(the sp2 carbon of the protonated amidine). These peaks were
-
assigned by comparison to the reported chemical shift of HCO3
in D2O (160.8 ppm)7 and because the peak at 165.9 ppm couples
to the neighbouring methylene in the HMBC NMR spectrum. In
contrast, the spectrum of 4 in CDCl3 under air shows equivalent
butyl groups on the –NBu2 nitrogen and only a single peak far
=
downfield, at 160.1 ppm, assigned to the C N central carbon.
(1)
1
Similarly, the 13C{ H} NMR spectrum of 3 in CO2-saturated
Conversion of the aqueous solution of the bicarbonate salt of
4 to the biphasic mixture of water and 4 could also be achieved
without heating. Bubbling air slowly through the solution at
room temperature for 5 h caused the mixture to split into the
two phases.
D2O shows inequivalent propyl groups on the –NPr2 nitrogen,
-
and peaks at 160.1 (HCO3 ) and 165.9 ppm (the sp2 carbon
of the protonated amidine). In contrast, the spectrum of 3 in
CDCl3 under air shows equivalent propyl groups on the –NPr2
nitrogen, and only a single peak far downfield at 160.1 ppm,
Several guanidines were also tested for the desired phase
behaviour (Scheme 1). Guanidine 5 was unsatisfactory be-
cause it was miscible with water in the absence of CO2.
Guanidines 6–8 were progressively less miscible with water.
To determine the degree of miscibility of these partially-
=
assigned to the C N central carbon.
Switching of amidine 4 to the bicarbonate was also monitored
1
by H NMR spectroscopy (Fig. 2). At first, 4 and D2O were
shaken together in a 1 : 2 v/v ratio at room temperature, giving
a biphasic mixture. Aliquots of the two phases were removed,
1
miscible guanidines, H NMR spectra of saturated solutions
1
dissolved in CD3OD and analyzed by H NMR spectroscopy,
in D2O were measured. 2-Hexyl-1,1,3,3-tetramethylguanidine
(6), 2-butyl-1,1,3,3-tetraethylguanidine (7) and 2-hexyl-1,1,3,3-
tetraethylguanidine (8) dissolved in D2O to the extent of 84, 38
and 13 mg mL-1, respectively. All three of these liquid guanidines
became completely miscible with water (with a water : guanidine
ratio of 1 : 1 by volume) when treated with CO2. Again,
the miscibility is due to the formation of bicarbonate salts.
Guanidine 7, for example, showed peaks at 161.2 (guanidinium
showing that the upper phase was 4 with a small amount of
water (Fig. 2a) and the lower phase contained water and a
barely detectable trace of 4 (Fig. 2b). After CO2 was bubbled
-
1
cation) and 160.1 (HCO3 ) in its 13C{ H} NMR spectrum in
D2O after exposure to CO2. The HCl salt of 7 also had a peak
at 161.2 ppm, but no peak at 160.1 ppm. Guanidine 9 was
abandoned because the crude guanidine was immiscible with
water before and after exposure to CO2.
Removal of the CO2 from aqueous guanidinium bicarbonate
solutions was not successful. For example, 1.0 g of 7 was mixed
with 10 mL of distilled water and treated with bubbling CO2 for
2 h. Then, the solution was put on a glass frit and argon passed
up through the frit and the solution, for 18 h; no phase splitting
was observed. A portion of this aqueous solution was then
evaporated in a rotary evaporator at 33 ◦C and the residue then
1
redissolved in D2O. 13C{ H} NMR spectroscopy confirmed that
the peaks at 161.2 and 160.1 ppm were both still present. Treating
another portion of the aqueous solution with 2 M NaOH caused
the guanidine to separate from the water as an organic liquid
layer. ESI-MS confirmed that the layer consisted of compound
7. Thus, the conversion of the guanidinium bicarbonate salt back
to the neutral guanidine by flushing with a non-acidic gas is too
difficult. The same observation was made with guanidines 6 and
8. Further work with guanidines was therefore abandoned.
Thus, only five compounds (amidines 3 and 4, and guanidines
6–8) switched from water-immiscible to water-miscible, and the
Fig. 2 The switchability of 4 in D2O monitored by 1H NMR spec-
troscopy of aliquots withdrawn from observed phases and dissolved in
methanol-d4 with sodium acetate internal standard. B = compound 4.
BD+ = bicarbonate salt of 4.
810 | Green Chem., 2010, 12, 809–814
This journal is
The Royal Society of Chemistry 2010
©

View Online
switching of only one of these (compound 4) could be reliably
reversed. In order to develop a method for predicting which
amidines and guanidines might exhibit this phase behaviour,
the log Kow values (where Kow is the octanol/water partition
coefficient, a quantitative measure of the hydrophobicity) of
each structure was predicted using standard algorithms (see the
Experimental methods section). The results (Scheme 1) suggest
that, for the desired phase behaviour, the log Kow value should
be approximately in the range 3 to 7. From this limited data set,
it is clear that for the switching to be reversible, the solvent must
be an amidine rather than a guanidine and that the log Kow value
should be closer to 7 than 3. Our further work continued with
compound 4 exclusively.
This SHS (4) not only switches its hydrophilicity but also
switches its polarity. While solvents of switchable polarity have
been reported previously,3,6,8-10 the increase in polarity upon
exposure to CO2 has always been moderate. Because compound
4, in conjunction with an equal volume of water, forms an
aqueous solution in the presence of CO2, it has a polarity similar
to that of water. The wavelength of maximum absorbance of Nile
Red dye, which is used as a measure of polarity (Fig. 3), showed
that the water/4/CO2 single-phase mixture had a lmax of 570 nm.
In the absence of CO2, the water is largely expelled from liquid
4, meaning that the polarity of the solvent is that of the amidine
with very little water dissolved therein (lmax = 510 nm). As
shown in Fig. 3, this results in a far larger polarity change than is
observed in the previously reported switchable solvents. The new
solvent also has the advantage over the some of the previously-
reported switchable polarity solvents (B/ROH mixtures, where
B is an amidine or guanidine) in that it is compatible with water:
the B/ROH mixtures, if contaminated with water, give solid
[BH][O2COH] solid salts rather than [BH][O2COR] ionic liquids
upon treatment with CO2.
Fig. 4 The current industrial process for obtaining soybean oil involves
extracting the oil from soybean flakes (flattened soybeans) with hexane
and then removing the hexane by distillation. The process causes
significant vapour emissions of hexane, which contribute to smog
formation. Hexane is also a neurotoxin.13
neurotoxicity, and the risk of fire and explosion due to the
large-scale use of hexane, the use of a non-volatile solvent (and
therefore a non-distillative route for solvent removal) would be
preferable. We performed a preliminary evaluation of an SHS
for this application.
Switchable hydrophilicity solvents make it possible to extract
soybean oil from soybeans, and then to separate the solvent from
the extracted oil, without using a volatile organic compound.
The proposed process is illustrated in Fig. 5. First, the oil would
be extracted from the flakes using the SHS in its hydrophobic
form. Then, the SHS would be extracted from the oil using
carbonated water, which would convert the solvent to its
hydrophilic form. Flushing air through the SHS/water mixture
would then separate it into the two components, which would
then be re-used. For practical purposes, if a portion of the SHS
remained in the aqueous phase, this would not be considered
a significant problem because the aqueous phase would be re-
used. However, if a portion of the SHS remained in the oil, then
that would be more problematic, because it would represent the
loss of some SHS and the contamination of the product oil.
Fig. 3 The polarities of the low-polarity (green circle) and high-polarity
(red circle) forms of selected switchable-polarity solvents, compared
to the polarity of an equal volume mixture of liquid 4 and water in
the presence of CO2 (red square) and in the absence of CO2 after
the water has been largely expelled (green square). TMBG = 1,1,3,3-
tetramethyl-2-butylguanidine; DBU = 1,8-diazabicyclo[5.4.0]undec-7-
ene; Bn = benzyl.
Fig. 5 The process by which an SHS can be used to extract soybean oil
from soybean flakes without a distillation step. The dashed lines indicate
the recycling of the solvent and the aqueous phase.
Soybean processing is an industrial process that uses distil-
lation to remove a volatile low-polarity solvent from a low-
polarity product. In the process, soybeans are cracked, de-hulled
and flattened into flakes or extruded into collets. Soybean oil is
extracted with hexane, which is then removed by distillation to
leave the pure oil (Fig. 4).11,12 Because of the vapor emissions,
The ability of compound 4, in its hydrophobic form, to extract
soy oil from soybean flakes was compared to the ability of
hexane. The solvent (500 mg) and flakes (100 mg) were stirred
overnight and then filtered. The amount of extracted oil in the
filtrate, determined by H NMR spectroscopy, was the same
(8.8 mg) for both solvents, within experimental error.
1
This journal is
The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 809–814 | 811
©

View Online
To confirm that switching of the solvent can help remove
the solvent from soy oil, we added D2O to a mixture of 4 and
soybean oil (D2O : 4 : oil 2 : 1 : 1 by volume). After thorough
mixing, two liquid phases of equal volume were visible. 1H
NMR spectroscopy of each layer showed that the upper phase
was primarily soybean oil and 4 (Fig. 6, centre spectrum),
while the lower phase was primarily D2O. CO2 was bubbled
through the system for 1.5 h, after which time the top layer
appeared to halve in volume. 1H NMR spectroscopy, using the
peak for 4 at 3.17 ppm, showed that 96% of the amidine had
been removed from the soybean oil (Fig. 6, lower spectrum).
This was achieved with only one wash with carbonated water;
higher levels of removal would be expected with further washes
and/or treatment with silica. The removal of 4 from the water
was then performed as described above, demonstrating that this
switchable hydrophilicity solvent can be used for the extraction
of soy oil, that it can be removed from the soy oil with carbonated
water and that it can be recovered from the water for re-use. The
soybean oil was not changed in colour by this overall process.
primary mechanism for hydrolytic degradation.15 Therefore, we
anticipate no significant degradation of the amidine while it is
dissolved in CO2-saturated water. Because it separates out of the
water once the CO2 has been removed, hydrolysis of the amidine
at that stage is also not anticipated to be a problem, although
small losses are inevitable. If this technology were to be adopted
industrially, mechanisms for monitoring amide levels and for
removing amide would be necessary.
If the amidine is stable and has a very high log Kow value,
then persistence and bioaccumulation in the environment are
serious concerns. Fortunately, if some of the amidine were to
escape into surface waters, then it would hydrolytically degrade
because the pH of surface water (typically 6.0 to 8.5)16 is
typically much higher than the pH of CO2-saturated water (3.9).
The two amide products that would form from hydrolysis of
compound 4 are N-butyl pentanamide (predicted log Kow = 2.3)
and N,N-dibutyl pentanamide (3.9). Neither of these amides are
sufficiently hydrophobic to pose serious bioaccumulation risks.
Experimental methods
General
Chemicals were used as received, except as noted. Compounds 1
and 5 were obtained commercially. The synthesis and character-
ization of the other amidines and guanidines are described in the
ESI material.† Compressed gasses were obtained from Praxair:
4.0 grade CO2 (99.99%) and 5.0 grade Ar (99.999%).
TLC was carried out on aluminium-backed silica gel 60 F524
1
from ELD. 1H NMR and 13C{ H} NMR spectra were collected
at 300 K on a Brucker AV-400 spectrometer at 400.3 and
100.7 MHz, respectively. IR spectra were collected on a Nicolet
Avatar 360 FT-IR E.S.P. instrument between potassium bromide
plates.
Prediction of log Kow values
log Kow values were predicted using ALOGPS 2.1 software,17-19
which calculates the log Kow value for a given structure using
nine different algorithms and then averages these values.
Measuring the miscibility of guanidines in water
1
To determine the miscibility of guanidines in water, H NMR
Fig. 6 1H NMR spectra (CDCl3) of pure soybean oil (top), soybean
oil and 4 (middle) and soybean oil after the removal of 4 by carbonated
water (bottom). Peaks corresponding to compound 4 are indicated.
experiments were performed. The guanidine (0.5 mL) and heavy
water (0.75 mL) were stirred vigorously together for 30 min.
The two phases were separated and 200 mL of the aqueous
solution, 10 mL of 1,4-dioxane and approximately 200 mL of
additional D2O were injected into an NMR tube. Solubilities
were calculated from the 1H NMR spectrum by comparing the
signal of the terminal methyl group of the alkyl chain located
at the 2-position of the guanidine with the singlet signal of the
1,4-dioxane protons at d 3.75.
For the soybean application, where the solid materials (the
flakes after the oil has been extracted) are used rather than
discarded, the removal of residual solvent from the solid
materials is also important. For other applications, the solid
materials would not be used. We have not yet addressed this
issue in our research. Another area of future research is a
comparison of the energy cost for this process compared to
that for a conventional distillation-based process.
Evaluating the switching of guanidines
Persistence, bioaccumulation and stability of a solvent are sig-
nificant practical concerns. For industrial application, especially
for repeated re-use, the solvent must be stable. Alkylamidines
have been shown to be stable in acidic aqueous solution14
because hydroxide attack on the amidinium cation is the
CO2 at 1 bar was bubbled through a stirred mixture of guanidine
(0.5 mL) and D2O (0.75 mL) using a cannula. For testing the
reversion to the hydrophobic form, argon was bubbled through
the solution with vigorous stirring using a cannula. As this
method did not cause a phase split, the method was revised
812 | Green Chem., 2010, 12, 809–814
This journal is
The Royal Society of Chemistry 2010
©

View Online
to introduce the argon from below the solution via a glass frit,
again without a phase split being observed.
CO2 was bubbled through the system for 1.5 h, after which
time the top layer appeared to have halved in volume. 1H NMR
spectra were again acquired for both layers in methanol-d4 and
for the top layer in CDCl3.
Evaluating the switching of amidine 4
A 1.0 mL portion of the bottom layer was withdrawn and
transferred to a new 4.0 mL vial. A magnetic stirrer was
added and the vial stirred at 80 ◦C for 1 h. Samples were
withdrawn from both layers for 1H NMR spectroscopy analysis
in methanol-d4 with reference to the dioxane internal standard.
A methanol-d4 solution was prepared containing a sodium
acetate internal standard (48.8 mM) for use as the NMR solvent.
Two 4 mL glass vials containing 1.0 mL D2O and 0.5 mL 4
were prepared and shaken. A 50 mL sample was withdrawn from
each layer of one vial and combined with 0.50 mL of the sodium
acetate standard solution in each of two NMR tubes: the top
layer (4) in tube A; the bottom layer (aqueous) in tube B.
CO2 was bubbled through the unsampled vial until 4 had
completely converted to the corresponding bicarbonate, as
evidenced by the disappearance of the top layer. The pH of
the solution was measured using pH paper to be approximately
8–9. A 50 mL sample was withdrawn from the aqueous solution
and added to 0.50 mL of the sodium acetate standard solution
in NMR tube C.
The 4.0 mL vial, from which sample C was withdrawn, was
then heated at 80 ◦C for 1 h, stirred by a magnetic stirrer. Bubbles
of CO2 were observed escaping from the solution as a top layer, of
hydrophobic amidine, appeared. After cooling the vial to room
temperature, a 50 mL sample was withdrawn from each layer
and combined with 0.50 mL of the sodium acetate standard in
two NMR tubes: the top layer (4) in tube D; the bottom layer
(aqueous) in tube E.
Conclusions
A switchable-hydrophilicity solvent (SHS), meaning a solvent
that can reversibly switch from having poor miscibility with
water to having very good miscibility with water, has been
identified. CO2 at atmospheric pressure, is used to switch the
solvent to its hydrophilic form, and air and/or heat can be used
to switch it back again. This solvent can be used to extract and
isolate organic materials, such as soybean oil, without the need
for a distillation step. Because distillation is not required, there
is no longer a need to use a volatile organic solvent.
Future work, already under way, includes a search for more
examples, a more detailed phase behaviour study and thereby
the development of solvent design principles for maximizing the
quality of the separations.
The NMR spectra of all five samples are shown in Fig. 3.
Acknowledgements
The authors gratefully acknowledge the Natural Sciences and
Engineering Research Council (NSERC) and the Canada Re-
search Chairs program for funding. We also thank Tom Peterson
(now of Dow Chemical), Allan Hodgson (of Bunge) and James
White (of PNNL) for bringing to our attention the significant
need for a new solvent for soybean oil extraction, and for their
contribution to the evaluation of switchable-polarity solvents
for that application.2
Nile red measurements
To measure the lmax of the solvatochromic dye Nile Red dissolved
in water-saturated 4, 1 mL of 4 was added to a 1 dram vial
containing 1 mL of distilled water. The contents were stirred
at room temperature for 1 h under air. The top phase (water-
saturated 4) was pipetted into a quartz cuvette and 1 mg of
Nile Red was added. The colour at this stage was bright orange-
magenta (lmax = 510 nm). Distilled water (1 mL) was then added
to the cuvette and CO2 slowly bubbled into the mixture for
1 h, after which the homogenous mixture turned purple (lmax
570 nm).
=
References
1 National Pollutant Release Inventory: http://www.ec.gc.ca/inrp-
npri/ (accessed 31 October 2009).
2 L. Phan, H. Brown, T. Peterson, J. White, A. Hodgson and P. G.
Jessop, Green Chem., 2009, 11, 53–59.
3 P. G. Jessop, D. J. Heldebrant, L. Xiaowang, C. A. Eckert and C. L.
Liotta, Nature, 2005, 436, 1102.
4 Y. Liu, P. G. Jessop, M. Cunningham, C. A. Eckert and C. L. Liotta,
Science, 2006, 313, 958–960.
5 D. J. Heldebrant, C. R. Yonker, P. G. Jessop and L. Phan, Energy
Env. Sci., 2008, 1, 487–493.
6 L. Phan, X. Li, D. J. Heldebrant, R. Wang, D. Chiu, E. John, H.
Huttenhower, P. Pollet, C. A. Eckert, C. L. Liotta and P. G. Jessop,
Ind. Eng. Chem. Res., 2008, 47, 539–545.
7 F. Mani, M. Peruzzini and P. Stoppioni, Green Chem., 2006, 8, 995–
1000.
8 T. Yamada, P. J. Lukac, M. George and R. G. Weiss, Chem. Mater.,
2007, 19, 967–969.
9 T. Yamada, P. J. Lukac, T. Yu and R. G. Weiss, Chem. Mater., 2007,
19, 4761–4768.
10 L. Phan, J. R. Andreatta, L. K. Horvey, C. F. Edie, A.-L. Luco, A.
Mirchandi, D. J. Darensbourg and P. G. Jessop, J. Org. Chem., 2008,
73, 127–132.
Extraction of soybean oil from soybean flakes
Soy flakes (100 mg) were stirred for 18 h in a sealed 3 dram vial
containing 500 mg solvent (either hexane or compound 4). The
mixture was filtered and 0.1 mL of the filtrate was analyzed by 1H
NMR spectroscopy with 0.1 mL ethanol as internal standard.
Separation of soybean oil from SHS
A 4 mL vial was prepared containing 1.0 mL D2O, 0.5 mL 4 and
0.5 mL soybean oil. The vial was shaken thoroughly and allowed
to settle, showing a 1.0 mL upper layer (primarily soybean oil
and 4) and a 1.0 mL lower layer, D2O. 1H NMR samples (50 mL)
were withdrawn from each layer and mixed with methanol-d4
(0.5 mL) containing a dioxane internal standard (51.5 mM).
Another sample was withdrawn from the upper layer for analysis
by 1H NMR spectroscopy in CDCl3, because soybean oil is not
miscible with methanol. The same volume was discarded from
the bottom layer to maintain the initial ratio.
11 G. C. Mustakas, in Handbook of Soy Oil Processing and Utilization,
ed. D. R. Erickson, E. H. Pryde, O. L. Brekke, T. L. Mounts
This journal is
The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 809–814 | 813
©

View Online
and R. A. Falb, American Soybean Association and American Oil
Chemists Society, St. Louis, 1980, pp. 49–65.
12 W. J. Wolf, Soybeans and Other Oilseeds, in Kirk-Othmer Ency-
clopedia of Chemical Technology, Wiley InterScience, 2007, online
version.
16 J. Kalff, Limnology, Prentice Hall, Upper Saddle River, NJ, 2002, ch.
14.
17 I. V. Tetko and V. Y. Tanchuk, J. Chem. Inf. Comput. Sci., 2002, 42,
1136–1145.
18 I. V. Tetko, J. Gasteiger, R. Todeschini, A. Mauri, D. Livingstone, P.
Ertl, V. A. Palyulin, E. V. Radchenko, N. S. Zefirov, A. S. Makarenko,
V. Y. Tanchuk and V. V. Prokopenko, J. Comput.-Aided Mol. Des.,
2005, 19, 453–463.
19 VCCLAB, Virtual Computational Chemistry Laboratory: http://
www.vcclab.org (accessed 22 October 2009).
13 O. F. Sendur, Y. Turan, S. Bal and A. Gurgan, Inhalation Toxicol.,
2009, 21, 210–214.
14 K. Sharavanan, H. Komber, D. Fischer and F. Bo¨hme, Polymer, 2004,
45, 2127–2132.
15 D. R. Robinson, J. Am. Chem. Soc., 1970, 92, 3138–3146.
814 | Green Chem., 2010, 12, 809–814
This journal is
The Royal Society of Chemistry 2010
©