Toluenesulfonate-acyl esters of ethylene glycol and other 1,2-diols as industrial antioxidants with cupric ion

Article abstract of DOI:10.1007/s11746-002-0560-0

Ethylene glycol esters of soybean oil FA increased in viscosity much more slowly than methyl or glycerol esters when oxidized at 105°C in the presence of flowing air and colloidal copper. This increased stability was caused by a minor constituent of the e

Full text of DOI:10.1007/s11746-002-0560-0

Toluenesulfonate-Acyl Esters of Ethylene Glycol and Other
1,2-Diols as Industrial Antioxidants with Cupric Ion
Yongyi Julia Jiang and Earl G. Hammond*
Department of Food Science and Human Nutrition and Center for Crops Utilization Research,
Iowa State University, Ames, Iowa 50011
ABSTRACT: Ethylene glycol esters of soybean oil FA increased
in viscosity much more slowly than methyl or glycerol esters
when oxidized at 105°C in the presence of flowing air and col-
loidal copper. This increased stability was caused by a minor con-
stituent of the ethylene glycol esters, which was shown by MS to
be a mixed ethanediol fatty acylate p-toluenesulfonate (EFAT).
The p-toluenesulfonate group came from the catalyst used in the
formation of the ethylene glycol esters. EFAT was quantified by
UV spectrometry, HPLC, or GC of the acyl group that it contains.
EFAT could be synthesized in good yield by reacting ethylene gly-
col, a FA, and p-toluenesulfonic acid (TSA) in a 1:1:1 molar ratio
using a benzene azeotrope to remove water of esterification.
We theorized that the polymerization and viscosity in-
crease of soybean oil resulting from oxidation should be a
function of the number of oxidizable fatty acyl groups per
molecule. It follows that esters of a particular FA composi-
tion should increase in viscosity in the following order when
oxidized under similar conditions: methyl esters < ethylene
glycol esters < glycerol esters. To test this hypothesis we
compared the rates of viscosity increase of these three types
of esters of soybean oil FA when they were oxidized under
the conditions of Ruger et al. (3). We found that the glycerol
esters increased in viscosity faster than the methyl esters, but
EFAT increased the time required for the polymerization of soy- the ethylene glycol esters were unusually stable in the pres-
bean oil by about 27 times but required concentrations of 2–5%
by weight. EFAT made with a variety of FA were active in delay-
ing viscosity increase. Ethyl and decyl p-toluenesulfonate were
inactive. Replacing ethylene glycol by glycerol and 1,2-propyl-
ene glycol but not by 1,3-propylene glycol resulted in active
EFAT. TSA itself delayed the polymerization of soybean oil, espe-
cially in the presence of free ethylene glycol and FA, but this
probably was caused by formation of EFAT during the oxida-
tion test. Colloidal copper could be replaced by cupric ion.
EFAT–copper appeared to act as an antioxidant by destroying hy-
droperoxides without initiation of free radical chains.
ence of colloidal copper and had hardly increased in viscosity
long after the methyl and glycerol esters had increased to
about five times their original viscosity. This paper tells of
our identification of a minor constituent in the ethylene gly-
col esters that was responsible for this antioxidant effect, our
attempts to determine the structural elements necessary for
antioxidant action, the role of copper ions, and information
about a possible mechanism of action.
MATERIALS AND METHODS
Paper no. J10091 in JAOCS 79, 791–796 (August 2002).
Materials. Refined commercial soybean oil (Crisco; Procter
& Gamble, Cincinnati, OH) was purchased locally. Unless
otherwise specified, the solvents and reagents were purchased
from Fisher Scientific (Fair Lawn, NJ) and were certified
grade. Colloidal copper and iron and ethyl p-toluenesulfonate
were purchased from Aldrich (Milwaukee, WI).
Synthesis of esters. Soybean oil was converted to methyl
esters by mixing 500 mL of oil with 55 g of methanol and 15
g of 30% sodium methoxide in methanol (Fluka, Milwaukee,
WI) at room temperature for 15 h. The product was washed
with water and dried with sodium sulfate.
To make ethylene glycol esters, 500 mL of oil was saponi-
fied with 130 g potassium hydroxide, 750 mL water, and 25 mL
ethanol and stored at 70°C under nitrogen for 16 h. FFA, re-
covered from the soaps with acid, were dried with sodium sul-
fate. The FFA and ethylene glycol were reacted in a 2.04:1
molar ratio in refluxing benzene for 18 h with 3 g p-toluenesul-
fonic acid (TSA) catalyst. A Dean–Stark trap was used to re-
move the water of esterification. The milliliters of benzene
were set at 2.5 times the weight of the FA plus 30 mL to fill the
leg of the Dean–Stark trap. The reaction mixture was neutral-
ized with 5% aqueous sodium carbonate solution and washed
KEY WORDS: Antioxidant, copper, cupric ion, EFAT, ethyl-
ene glycol toluenesulfonate fatty acylate.
Because of the poor biodegradability of petroleum-based fluids
there is growing interest in using vegetable oils for industrial
applications, such as lubricants and hydraulic fluids, especially
in ecologically sensitive settings. In addition, vegetable oil-
based lubricants are said to have an enhanced lubricity that can
result in significantly decreased gasoline consumption (1,2).
The use of many vegetable oils in these applications is limited
because of their rapid oxidation and polymerization. Ruger
et al. (3) advocated testing the effectiveness of antioxidants for
such applications by their ability to prevent viscosity increase
in soybean oil at 105°C with bubbling air in the presence of
colloidal iron and copper. He concluded that TBHQ was par-
ticularly effective in this application, and its effectiveness was
slightly enhanced by the addition of citric acid.
*To whom correspondence should be addressed at Dept. of Food Science
and Human Nutrition, Food Sciences Bldg., Iowa State University, Ames,
IA 50011. E-mail: hammond@iastate.edu
Copyright © 2002 by AOCS Press
791
JAOCS, Vol. 79, no. 8 (2002)

792
Y.J. JIANG AND E.G. HAMMOND
gently with water five times; the benzene was removed with a and detector temperatures at 250°C; initial column tempera-
rotary evaporator (Buchi, Flawil, Switzerland). Ethylene gly- ture was 170°C for 3 min, programmed at 20°C/min to 225°C,
col monoacyl esters of soybean oil FFA was produced similarly and held 6 min. Carrier flow was 1.5 mL/min, and other gases
using a 3:1 molar ratio of ethylene glycol to FFA.
were at the setting recommended by the manufacturer.
Ethylene glycol fatty acylate p-toluenesulfonate (EFAT)
GC/MS. For EI-MS an HP 5890 Series II instrument with
was synthesized with various FA (lauric, palmitic, stearic, an HP5790 mass detector was used. The column was a Su-
linoleic, and linolenic) using a molar ratio of ethylene gly- pelco SPB-1 30 m × 0.25 mm i.d., 0.25 µm film. Injector and
col/TSA/FFA of 1:1:1 and using benzene and a Dean–Stark detector temperatures were 320°C; initial oven temperature
trap as before. Excess TSA was removed by washing with was 170°C for 3 min, programmed at 10°C/min to 320°C, and
sodium carbonate solution. The same method was used to held 6 min. For atmospheric pressure chemical ionization
make esters using 1,2- and 1,3-propylene glycol in place of (APCI) a Finnigan JSQ 700 (Finnigan Instruments Corp., San
ethylene glycol. To make the decyl ester of TSA, the molar Jose, CA) mass spectrometer fitted with a Finnigan APCI ion
ratio of decanol to TSA 1:1 was used. To substitute glycerol source was used. Samples were injected in the loop injection
for ethylene glycol, the molar ratio of glycerol/TSA/FFA was mode. The vaporizer temperature was maintained at 400°C.
1:2:1 to maximize the yield of glycerol esterified with two For exact mass MS a Kratos MS-50 double-focusing mag-
TSA and one FFA.
netic sector mass spectrometer (Kratos Instruments, Man-
Stability measurement. Oxidative stability was measured chester, United Kingdom) was used, with a solid probe intro-
in duplicate using the apparatus described for the Active Oxy- duction, electron ionization mode, and manual peak matching
gen Method (AOM) in the American Oil Chemists’ Society at 10,000 resolution (10% valley definition) with perfluoro-
Method Cd-12-57 (4). Air, supplied by an aquarium pump, kerosene as the reference compound.
was regulated at 2.33 mL/s by controlling the pressure with a
Quantification of EFAT. The amount of EFAT was deter-
water column. The heating block was maintained at 105ºC mined by dissolving the sample in acetonitrile (HPLC grade)
with a Barnant controller model 689 (Barrington, IL). Sam- and measuring the absorbance of the TSA moiety at 273 nm
ples consisted of 20 mL of oil to which approximately 10 mg with a Hitachi TU-2000 Spectrophotometer (Hitachi Instru-
each of colloidal iron and copper were added. In some exper- ments, Inc., Conroe, TX). EFAT also was quantified by using
iments only colloidal copper was added, or the colloidal cop- GC to determine the amount of acyl group it contained. EFAT
per was replaced with various amounts of an ethanolic solu- preparations were fractionated by TLC, the EFAT band was
tion of cupric acetate. Samples (1 mL) of the soybean oil were recovered, converted to methyl esters with methanolic
withdrawn periodically, and their viscosities were measured sodium methoxide (6), and extracted with hexane. Methyl
using a Brookfield LV-DV-II+ cone and plate viscometer heptadecanoate was added as an internal standard, and the
(Stoughton, MA) operated at 40ºC. After measuring the vis- sample was analyzed by GC.
cosity, the oil samples were returned to their respective AOM
To follow the synthesis of EFAT, 0.05 moles each of ethyl-
tubes. A value of 150 cP was chosen as an arbitrary end point. ene glycol, lauric acid, and TSA were esterified as before.
Initial isolation of EFAT. A column, 2 × 54 cm, was filled Samples of ~0.3 g were taken periodically and washed with
with 120 mL hexane, and 150 g of 80–200 mesh alumina, ac- water to remove unreacted TSA and ethylene glycol. The sol-
tivated at 260°C for 15 h (5), was added and allowed to settle vent was evaporated, and the residue was weighed and dis-
under gravity. Ethylene glycol esters of soybean oil FA (50 g) solved in acetonitrile. Dibenzyl (1 mg/mL) was added as an
dissolved in 30 mL hexane were added, and the column was internal standard. Analysis was accomplished with a Shi-
eluted with hexane, hexane/ethyl acetate (7:5, vol/vol), and madzu LC-600 instrument with an SPD-6 AV UV detector
methanol. The eluates were collected in 20-mL portions, each (Kyoto, Japan), Zorbax SB-C18 4.6 × 25 mm column (Agi-
of which was checked for its composition by TLC.
lent Technologies, Palo Alto, CA), and a mobile phase of ace-
TLC. Mixtures produced by ester syntheses (50 mg) were ap- tonitrile at 1 mL/min. The eluate was monitored at 273 nm,
plied to Adsorbisil-Plus 1 silica thin-layer TLC plates, 20 × 20 and the data were corrected for the relative absorbance ratio
cm, 0.5 mm thickness (Alltech Assoc., Deerfield, IL) with a of lauric EFAT and dibenzyl. The amount of ethylene glycol
streaker (Applied Science Laboratories, Inc., State College, PA). dilaurate was determined by GC, using methyl heptadec-
The plate was developed with hexane/ethyl ether/acetic acid anoate as an internal standard and an SP-2330 column. The
(70:30:1, by vol), and the components were visualized by spray- data were corrected for the relative GC response.
ing lightly with 0.1% (wt/vol) 2′,7′-dichlorofluoresein (Sigma
Proton NMR. Measurements were carried out at room tem-
Chemical Co., St. Louis, MO) in methanol. Silica containing perature using a Varian VXR-300 instrument. CDCl3 (δ =
visible bands was collected, and each band was extracted with 7.26 ppm) was used as solvent and reference standard.
freshly distilled ethyl ether. The ether was evaporated, and the
residues were weighed and used for further analysis.
Peroxide decomposition. To test EFAT’s effect on hy-
droperoxides, four 20-mL soybean oil samples that were ap-
GC. The GC was a Hewlett-Packard (Avondale, PA) (HP) proximately one millimolar in cupric acetate were oxidized
5890 Series II instrument with an HP3396 Series II integrator in the AOM apparatus at 40°C for 26 h. Lauric EFAT was
and FID detector. A Supelco (Bellefonte, PA) 2330 column, added at 5% by weight to two of the samples, and the other
15 m × 0.25 mm i.d., 0.20 µm film, was operated with injector samples served as controls. The sample and control oils were
JAOCS, Vol. 79, no. 8 (2002)

EFAT ANTIOXIDANTS
793
gassed with nitrogen for the remainder of the experiment. in the AOM test of glycerol, methyl, and ethylene glycol es-
Samples were taken periodically, and their PV were deter- ters of soybean oil that had been purified by chromatography
mined by the Stamm method (7).
on alumina, which was expected to remove tocopherols and
other polar constituents. The relative rates of oxidation of
these esters were more nearly in agreement with our hypothe-
sis. The addition of copper to the ethylene glycol esters had a
RESULTS AND DISCUSSION
We tested the relative rates of oxidation of methyl, ethylene much smaller effect than before (Table 1, line 10). But when
glycol, and glycerol esters of soybean oil FA to test the hy- the materials of Rf 0.3 and 0.2 were added back to ethylene
pothesis that more FA per molecule would increase the rate glycol esters, they greatly enhanced the esters’ stability (Table
of polymerization. Table 1, lines 1–3, shows the AOM times 1, line 11). Similar effects were obtained when these unknown
to 150 cP for glycerol and methyl esters of soybean oil FA. components were added to soybean oil (data not shown).
Ethylene glycol esters oxidized unexpectedly slowly, and the
The component of Rf 0.2 was the monoacyl ester of ethyl-
sample was exhausted before 150 cP was reached. The addi- ene glycol and was the primary component of syntheses that
tion of iron to these esters did not change their oxidation rate had a high ratio of ethylene glycol to FA. By itself this com-
appreciably (data not shown), but in the presence of added ponent had no effect on oil stability. The unknown component
colloidal copper the rate of oxidation of the methyl and ethyl- of Rf 0.3 was treated with methanolic sodium methoxide, and
ene glycol esters decreased while that of the glycerol esters GC on an SPB 2330 column revealed the presence of methyl
remained about the same (Table 1, lines 4–6). The effect of esters of the FA of soybean oil. By substituting palmitic, oleic,
copper on the rate of oxidation of the ethylene glycol esters linoleic, or linolenic acids individually for soybean oil in the
was particularly striking. The slower oxidation of ethylene synthesis, materials analogous to the unknown spot were ob-
glycol esters, especially in the presence of copper, was con- tained but with only one kind of acyl group.
trary to our hypothesis. The addition of copper to lubricants
has been advocated to slow oxidation (8).
The unknown containing linolenate was chosen for further
study by GC/MS. By electron impact GC/MS there was one
To understand the unexpected stability of the ethylene gly- major component with its mass fragments (m/z, %) at 99 (100),
col ester, we subjected it to TLC and discovered three compo- 55 (74), 139 (20), 79 (20), 67 (17), 91 (18), 112 (18), 153 (16),
nents of Rf 0.7, 0.3, and 0.2. The main component was diacyl 86 (9), and 125 (9). The parent compound seemed to have an
ester of Rf 0.7. The ethylene glycol ester mixture was fraction- M.W. of 304, and the spectrum was consistent with a lactone
ated on an alumina column to obtain enough material for our formed from ethylene glycol monolinolenate by splitting out
oxidation tests. Table 1, lines 7–9, shows the relative stability water from the free hydroxy group and a β-hydrogen from the
TABLE 1
Initial Viscosity and Hours to Reach a Viscosity of 150 cP for Soybean Oil and Its Derivatives
Oxidized at 105°C with Air Bubbling Through at 2.33 mL/min
Initial viscosity
(cP)
Hours to viscositya
150 cP
Oil
1. Soybean oil
30.1
3.4
14.3
30.1
3.4
11.6
26.0
3.4
29.6
2. Soybean methyl esters
148
3. Soybean ethylene glycol esters
>83; 31.4 cPb,c
4. Soybean oil + Cu0
25.2
5. Soybean methyl esters + Cu0
248
6. Soybean ethylene glycol esters + Cu0
7. Soybean oil, alumina-treated
>315; 20.8 cPc
20.8
144
8. Methyl esters, alumina-treated
9. Ethylene glycol esters, alumina-treated
10. Ethylene glycol esters, alumina-treated + Cu0
11. Ethylene glycol esters, alumina-treated + Cu0 + unknown #1
12. Soybean oil + Cu0 + 5% EFAT
12.7
12.7
12.7
28.6
45
28.6
28.6
28.6
28.6
28.6
28.6
33.3
42.0
>83; 17.5 cPc,d
>900; 64.9 cPc
164
13. Soybean oil + Cu0 + 5% TSA
14. Soybean oil + Cu0 + ingredients for 2% EFAT
15. Soybean oil + Cu0 + 2% EFAT
16. Soybean oil + 5% EFAT + no Cu
17. Soybean oil + 5% EFAT + 0.01 mM Cu2+
18. Soybean oil + 5% EFAT + 0.1 mM Cu2+
19. Soybean oil + 5% EFAT + 1.1 mM Cu2+
658
>900; 74.5 cPc
384
588
640
>645; 54.4 cPc
aAverage of duplicate determinations. Overall least significant difference (LSD) was 37.7 at P < 0.05. When divided
into groups of <100 cP and >100 cP, the LSD were 8.3 and 61.3, respectively.
bThe rate of increase for methyl esters was greater than for ethylene glycol esters.
cViscosity 150 cP was not reached. The average viscosity at the final reading time is given.
dSingle determination because of lack of material.
JAOCS, Vol. 79, no. 8 (2002)

794
Y.J. JIANG AND E.G. HAMMOND
ethylene glycol and FA to 1:1:1 during synthesis. When this
ratio was used, the concentration of EFAT in the reaction
product rose quickly to ~50% and then more slowly rose to
~60% as shown in Figure 1. The formation of diacyl ester was
faster than that of EFAT, but it decreased with reaction time.
Table 1, line 12, shows the effect on the stability of soy-
bean oil of copper plus 5% EFAT in which the acyl groups
were soybean oil FA. The viscosity of the soybean oil with
EFAT slowly rose to about 65 cP in 900 h compared with the
control (line 1) that reached 150 cP in ~30 h. EFAT is much
SCHEME 1
linolenate. But lactones were not polar enough to account for more effective in stabilizing the oil than TBHQ-citric acid,
the TLC properties of the unknown.
which was the best antioxidant reported by Ruger et al. (3).
The APCI-MS of the unknown based on linolenate re- But the ability of EFAT to prevent the polymerization of soy-
vealed peaks at 477 (100), 305 (46), 261 (18), and 243 (4). bean oil requires a fairly high concentration compared with
This is consistent with an M.W. of 476, with a major fragment other antioxidants. EFAT was active at 2% or more but had
at 304. The peaks at 305 and 261 were shown to be fragments no antioxidant activity at 1%.
of the 477 peak. This molecular weight is consistent with
Modifications of EFAT. Various modifications in the struc-
ethanediol p-toluenesulfonate linolenate, and the 304 and 261 ture of EFAT were tried to determine the structural features
peaks probably represent the loss of toluenesulfonic acid and required for antioxidant activity. Purified FA as short as lau-
a linolenyl group, respectively. This structure was confirmed ric acid were substituted for soybean oil FA to alter the acyl
by exact mass measurement of the molecular weight as group in EFAT. All of these acyl group variations were active
476.26088, which was best fit by C27H40O5S. The exact mass in stability tests. When we tried octanoic acid, most of it dis-
spectrum also contained a strong peak at mass 91, indicating tilled with the benzene azeotrope and was lost, so the yield
a toluene ring. We suggest the trivial name EFAT for the com- was too poor for testing.
pound for ethylene glycol, fatty acylate, and p-toluenesul-
fonate (Scheme 1).
Mono alcohol esters of TSA were tested. Neither ethyl p-
toluenesulfonate nor decyl p-toluenesulfonate, made using
We concluded that the concentration of EFAT in our origi- benzene azeotrope distillation, gave antioxidant activity in
nal ethylene glycol esters was about 5%. This yield was im- the AOM test at the 5% level.
proved considerably by increasing the molar ratio of TSA to
EFAT analogs were synthesized using glycerol and 1,2-
FIG. 1. Weight percent yield of ethylene glycol fatty acylate p-toluenesulfonate (EFAT), ethylene glycol dilaurate,
and ethylene glycol mono-p-toluenesulfonic acid (mono TSA) vs. time. Conditions: 1:1:1 molar ratio of TSA/ethyl-
ene glycol/lauric acid in refluxing benzene with water removal by a Dean–Stark trap.
JAOCS, Vol. 79, no. 8 (2002)

EFAT ANTIOXIDANTS
795
and 1,3-propylene glycol in place of ethylene glycol. The not shown). These results suggest that for antioxidant activity
structures of the latter two products were verified by MS and the EFAT must, for unknown reasons, have its acyl group and
NMR. Chemical shifts for the propylene glycol esters are tosyl groups on adjacent carbons.
given in ppm (δ), and multiplicities are indicated by s (sin-
EFAT was made with sodium dodecylbenzenesulfonate in
glet), d (doublet), t (triplet), q (quartet), qn (quintet), and m place of TSA. The APCI-MS of this product ionized with hy-
(multiplet). 1H NMR (300 MHz, CDCl3, 16 mg/mL) for the drogen gave peaks at m/z 401, 227, and 353. None of these
1,2-propylene glycol product δ: 0.87 (3H, t), 1.30 (22H, s), represented the intact molecule, but the m/z 227 peak repre-
2.16 (2H, m), 2.45 (3H, s), 4.03 (2H, m), 4.91 (1H, m), 7.35 sents the loss of the dodecylbenzenesulfonyl group, and the
(2H, d), 7.79 (2H, d). For the 1,3-propylene glycol product δ: m/z 353 peak represents the loss of the lauryl chain. This ver-
0.88 (3H, t), 1.25 (18H, s), 1.98 (2H, qn), 2.21 (2H, t), 2.45 sion of EFAT was active as an antioxidant in the AOM test at
(3H, s), 4.09 (4H, m), 7.35 (2H, d), 7.79 (2H, d). The two the 5% level (data not shown).
spectra are similar. The 1,3-isomer had a methylene group in
Table 1, line 13, shows that when free TSA was tested in
the middle of the 1,3-propylene glycol backbone that gave a the modified AOM test, it prolonged the time for viscosity in-
quintet at 1.98 ppm, and the two methylene groups adjacent crease to 150 cP from about 30 h to 164 h. But TSA is an in-
to the ester bonds gave two triplets at 4.09 ppm. For the 1,2- teresterification catalyst and may react with soybean oil under
isomer the methyl group at the end of 1,2-propylene glycol AOM conditions to form a glycerol-based version of EFAT.
joined with methylene groups on the FA chain at 1.25 ppm, Table 1, lines 14 and 15, shows that if ethylene glycol and FFA
and the –CH group in the middle of the propylene glycol gave are put into the reaction mixture with TSA, an antioxidant ef-
a multiplet near 4.91 ppm. This multiplet arose because the fect is observed that is better than that of TSA alone but not as
product is a mixture of two compounds that vary in the posi- great as that of 2% preformed EFAT. These observations sug-
tion of tosyl and lauryl groups. The 1,3-propylene glycol gest that EFAT is formed to some extent during the AOM test.
product gave APCI-MS peaks at m/z 413, 241, 104, and 183, If no source of hydroxy groups is present, EFAT formation
using hydrogen as the ionizing gas. The peak at m/z 413 is the may be limited to the amount of MG and DG in the test oil.
expected mass + 1. The m/z 241 fragment is produced by the
The level of copper needed. Cupric acetate could success-
loss of the tosyl group. The m/z 183 fragment probably comes fully replace colloidal copper in the modified AOM stability
from the lauryl group split between its ester oxygen and car- test, so presumably the colloidal copper was furnishing cupric
bonyl group. For the ethylene glycol product no peak repre- ions. The use of cupric acetate made it possible to measure
senting the intact molecule was obtained; only the m/z 241 the amount of copper ion needed. A comparison of Table 1,
peak resulting from the loss of the tosyl group was observed. lines 16 and 12, shows that with no added cupric ion and 5%
The EFAT made with glycerol and 1,2-propylene glycol EFAT, the AOM time was reduced to about 384 h rather than
were active as antioxidants in the AOM test at the 5% level, the >900 h observed with colloidal copper. Increasing levels
but the EFAT made from 1,3-propylene glycol was not (data of copper ion from 0.01 mmol up to about 1.1 mmol or ~76
FIG. 2. Decomposition of soybean oil hydroperoxides in soybean oil that was 1 mmol in Cu2+
vs. time under nitrogen at 40°C. Curves A and B are controls with no EFAT; C and D contain
5% EFAT. For abbreviation see Figure 1.
JAOCS, Vol. 79, no. 8 (2002)

796
Y.J. JIANG AND E.G. HAMMOND
ppm increased the AOM time (Table 1, lines 17–19). The
copper content of refined soybean oil has been reported to be
~0.02 to 0.06 ppm (9), which would be about 0.001 mmol. It
may be that if the level of copper in refined soybean oil could
be reduced another order of magnitude, EFAT would have no
effect by itself in stabilizing the oil. EFAT at 2% in the oil is
~41 mmol, which is ~37.5 times greater than the moles of
copper necessary to give good stability.
Effect of EFAT on hydroperoxides. Figure 2 shows that the
addition of EFAT to oxidized copper-containing soybean oil
samples held under nitrogen at 40°C caused a much more
rapid decrease of PV than controls. These results agree with
our observation that samples of soybean oil containing EFAT
accumulated much less peroxide during AOM testing than
controls. These results suggest that EFAT exerts its antioxi-
dant action by destroying peroxides and thus reducing the rate
of free radical initiation. Normally the thermal decomposi-
tion of hydroperoxides in the presence of oxygen leads to the
formation of free radicals that results in further accumulation
of hydroperoxides. It is not clear how EFAT prevents this.
REFERENCES
1. Johnson, D., Update on the Soybean Engine Oil Project, paper
presented at Lubricants and Fluids Technical Advisory Panel,
Omnitech, Midland, MI (1998).
2. Lambert, J.W., and D.L. Johnson, Vegetable Oil Lubricants for
Internal-Combustion Engines and Total Loss Lubrication, U.S.
Patent 5,888,947 (1999).
3. Ruger, C.W., E.J. Klinker, and E.G. Hammond, Abilities of Some
Antioxidants to Stabilize Soybean Oil in Industrial Use Condi-
tions, J. Am. Oil Chem. Soc. 79, 733–736 (2002).
4. American Oil Chemists’ Society, Official Methods and Recom-
mended Practices of the American Oil Chemists’ Society, 4th
edn., AOCS Press, Champaign, 1993, Method Cd12-57.
5. Jensen, R.G., T.A. Marks, J. Sampugna, J.G. Quinn, and D.L.
Carpenter, Purification of Triglycerides with an Alumina Col-
umn, Lipids 1:451–452 (1966).
6. Hammond, E.G., Rapid Analysis of Lipids in Many Individual
Plants, in Modern Methods of Plant Analysis, edited by H.F.
Liskens and J.F. Jackson, Springer-Verlag, New York, 1991,
New Series, Vol. 12, pp. 321–330.
7. Hamm, D.G., E.G. Hammond, V. Parvanah, and H.E. Snyder,
The Determination of Peroxides by the Stamm Method, J. Am.
Oil Chem. Soc. 42:920–922 (1965).
8. Garmier, W.W., Biodegradable Lubricant Composition from
Triglycerides and Oil-Soluble Copper, PCT Int. Appl. WO
9743351 (1977).
9. List, G.R., Special Processing for Off Specification Oil, in Hand-
book of Soy Oil Processing and Utilization, edited by D.R. Er-
ickson, E.H. Pryde, O.L. Brekke, T.L. Mounts, and R.A. Falb,
American Oil Chemists’ Society, Champaign, 1980, pp.
355–376.
ACKNOWLEDGMENTS
To Eric Klinker for mass spectra of the EFAT made with propylene
glycols and with dodecylbenzene sulfonic acid and the Hatch Act,
State of Iowa and to the United Soybean Board for financial support.
Journal Paper No. 19517 of the Iowa Agriculture and Home Eco-
nomics Experiment Station, Ames, Iowa, Project No. 3702.
[Received September 17, 2001; accepted June 14, 2002]
JAOCS, Vol. 79, no. 8 (2002)