.J!.

JV
1375
5730

Purchased by U. S. Dept. of Agric. for Official Use.

if. Transesterification Kinetics of Soybean DiP

Bernard Freedman", Royden O. Butterfield and Everett H. Pryde 2
Northern Regional Research Center, Agricultural Research Service, U.S. Department of Agriculture, Peoria, IL 61604

Transesterification of soybean oil (SBO) and other methyl ricinoleate (7), the glycerolysis of methyl oleate
triglycerides with alcohols, in the presence of a catalyst, (13) and the transesterification of the esters of
yields fatty esters and glycerol. Di- and monoglycerides 9(10)carboxystearic acid (14) and methyl 9(10)car
are intermediates. Reactions are consecutive and boxymethylstearate (15). The present study was
reversible. Rate constants have been determined for undertaken with a commercially important vegetable
each reaction with a computerized kinetic program. The oil, SBO, utilizing a computerized kinetic program (16)
effects of the type of alcohol, I-butanol or methanol to shed additional light on the kinetics of transesterifica
(MeOH); molar ratio of alcohol to SBO; type and amount tion reactions.
of catalyst; and reaction temperature on rate constants
and kinetic order were examined. Forward reactions
appear to be pseudo-first order or second order EXPERIMENTAL
depending upon conditions used. Reverse reactions Materials. Refined SBO was obtained from Central
appear to be second order. At a molar ratio of Soya, Chattanooga, Tennessee. The molecular weight
MeOH/SBO of 6:1, a shunt reaction was observed. was assumed to be 872.4. Methanol (MeOH) and
Energy of activation was determined for all forward and l'butanol (BuOH) were MCB Omnisolv (spectrograde)
reverse reactions under a variety of experimental and were stored over molecular sieves 4A. A solution of
conditions from plots of log k vs liT. Values ranged sodium butoxide (NaOBu) was prepared by the reaction
from 8-20 kcal/mol. of 240 mg of sodium with 10 ml of BuOH at 5999 C ~th
stirring. Sulfuric acid was purchased from B&A, Allied
Chemical, Morristown, New Jersey. Sodium methoxide
Transesterification of vegetable oils with simple (NaOCH3 ) was obtained from Aldrich Chemical Com
alcohols has long been a preferred. method for preparing pany, Milwaukee, Wisconsin.
fatty esters. These esters have good potential as an Reaction conditions employed in kinetic studies. The
alternative or emergency fuel to replace #2 diesel oil in scope of our kinetic studies is outlined in Table l.
farm tractors. Because of this potential, we have focused Experiments were designed to determine kinetic order
our attention on variables that affect 'the yield of these and rate constants using two simple alcohols, two molar
esters (1,2) as well as analysis of intermediates and ratios of alcohol to SBO, acidic vs alkaline catalysis, two
products as determined by an Iatroscan analyzer (3). In weight percentages of NaOBu and two tempera~ure
the present paper we report on a study of the kinetics of ranges each containing five temperatures. All reactIOns
the transesterification of soybean oil (SBO). A capillary studied were conducted at atmospheric pressure.
gas chromatographic (CGC) method used to support this Transesterification reaction and sampling. The first
study has been described (4). One of our objectives was two of five systems shown in Table 1 utilized a 250-ml
to determine how variations in type of alcohol, molar round-bottomed four-necked flask, equipped with a
ratio of alcohol to SBO, catalyst type and reaction mechanical stirrer, thermometer, condenser and/or
temperature affected kinetic order, reaction rates and drying tube and stopper (for sample removal). The
energies of activation. Another objective was to gain a remaining systems in Table 1 employed a 400-ml
more fundamental understanding of the chemistry custom-made reaction flask with five necks. The
involved in the transesterification of vegetable oils. additional neck was equipped with a thermocouple
Various mechanisms have been proposed for both connected to a digital pyrometer. This pyrometer
acid- and alkaline-catalyzed transesterification (5-7). A provided rapid temperature readings that were essential
number of kinetic studies have been reported for the to follow the very fast reactions. The larger size flask
transesterification of simple (non-fatty) esters with enabled us to scale up the reaction and possibly
alcohols (6,8-10) and dimethylterephthalate with permitted more vigorous agitation.
ethylene glycol (11,12). Only a few kinetic studies have The reaction flask was immersed in a Polyscience
dealt with the transesterification of vegetable oil or fatty Series 9000 refrigerated constant temperature bath and
esters. These include the conversion of castor oil to circulator. The operating range of this bath was -15 to
*To whom correspondence should be addressed. +150 C, with a control accuracy of 0.02 C. The
'Presented at the AOCS meeting in Philadelphia, PA, May, 1985. temperature control of the reaction mixture was
'Deceased. generally ca. 0.1 C.

JAOCS, Vol. 63, no. 10 (October 1986)

1376

B. FREEDMAN, R.O. BUTTERFIELD AND E.H. PRYDE

TABLE I
Reaction Conditions in Kinetic Studies
Molar ratio Weight of reactants
of
Alcohol used alcoholfSBO Alcohol (g) SBO (g) Catalyst Temperature (CIa

I-Butanol 30:1 51 20 1% H 2 S0 4 77,87,97,107,117

I-Butanol 30:1 51 20 1% NaOBu 20,30,40,50,60
I-Butanol 30:1 102 40 0.5% NaOBu 20,30,40,50,60
I-Butanol 6:1 51 100 1% NaOBu 20,30,40,50,60
Methanol 6:1 22 100 0.5% NaOCH, 20,30,40,50,60

aTemperature control: 0.1 C.

The reaction flask and its associated equipment and Overall Reaction:
other glass apparatus used for handling starting catalyst
Triglyceride (TG) + 3ROR :;;:.==== 3R'C0 2R + glycerol (GL)
materials were first heated to expel moisture. Calculated
amounts of SBO and alcohol were added to the reaction
flask that was brought to a predetermined temperature Stepwise Reactions:
using the constant temperature bath. A calculated TG-DG(k,)
1. TG + ROR ::;:.===== diglyceride (OG) + R'C0 2R
amount of catalyst was quickly added to the reactants, OG-TG(t.;)
and timing of the reaction was begun.
Samples (0.5 ml each) were removed from the reaction DG-MG(k2)
mixture at various intervals, depending upon reaction 2. OG + ROR ::;:.===== monoglyceride (MG) + R'C0 2R
MG-OG(ts)
conditions. For example, alkali-catalyzed reactions at
60 C were sampled initially at 15-sec intervals and then MG-GL(k3 )
over longer intervals during the course of one hr. For the 3. MG + ROR:;;:.===== glycerol + R'C0 2R
GL-MG(ks )
acid-catalyzed reaction at 117 C, early samples were
taken at 6-min intervals, and later samples were taken
at 30-min intervals during the course of three hr. FIG. 1. Chemistry of the transesterification of vegetable oils.
Samples were immediately quenched in 0.5 ml water,
thus stopping the reaction. Because the oil layer did not
always cleanly separate from the water, salting out and
centrifuging were used to enhance separation. The 100
resulting oil was used for both thin layer chromatog-
raphy (TLC) and CGC analysis. 80
TLC was used initiallY to determine the qualitative
composition of each sample:ofthe reaction mixture and "': 60
to determine which samples were to be quantitatively i
analyzed. This TLC procedure has been described ~ 40
previously (2). The reaction was stopped when TLC u.J

indicated high conversion to the ester. Fourteen samples 20

were taken, of which nine were quantitatively analyzed
by the CGC. The CGC method provided a quantitative O'---~~~~~"""""-~~~~-'-'-'--~--'
analysis of the triglyceride (TG), diglyceride (DG), 0.1 0.2 0.5 2 5 10 20
monoglyceride (MG) and ester composition needed for Time, hr
the kinetic study.
FIG. 2. Effect of time and temperature on ester formation.
Preparation of CGC samples and their analysis. BuOH/SBO (30:1). 1 % H 2 S04 , 77-117 C. X. 117 C; 107 C; .... 97 C;
Samples from the oil layers obtained above were 87 C; +. 77 C.
accurately weighed and silylated, internal standard was
added. and the concentration was adjusted and then
analyzed by CGC as previously described (4). A input consists of the concentration of each component
computerized program was used to analyze CGC data. with time. Concentration units were either mmol or
The output from this program gave wt %. ,.anol or mmol, ,.anoUml depending on requirements of the program. The
standard deviation and relative standard deviation for program utilizes these data to produce plots of
each component. concentration vs time, which are shown later. A separate
Computerized kinetic program. A computerized plot is produced for each temperature. Based upon the
kinetic program. described previously (16), was used to kinetic scheme being tested. the program attempts to
determine whether a proposed kinetic order appeared draw a line through the points. A close fit of the lines to
possible. The program requires that a specific kinetic the points suggests that the kinetic scheme employed is
scheme be proposed for each reaction being studied. The adequate. and from this a kinetic order can be proposed.
reactions studied are shown in Figure 1. Additional On the other hand. a poor fit of the lines to the points

JAOCS, Vol. 63, no. 10 (October 1986)

 1377

TRANSESTERIFICATION KINETICS OF SBO

suggests that an incorrect scheme was proposed and/or 60 ,-------------..---::--==-_""'--.=_----,

the data are faulty.
The program also reports two types of errors ..../
...--.--
50 /~
associated with the fit of the lines to the points at each
, ?'
temperature. The first is the "summed squared error,"
which is the sum of all vertical distances squared 40
between points and lines. The greater the coincidence ~
Q) I
I
between points and lines, the smaller the error. The ~ 30
second error reported is the" average error," which is the
o
t; I
square root of the summed squared error divided by the
number of data points (concentrations). By comparing
~
20 '-J I

the total of either the summed squared or average errors 10 I ",,- ~

of the various schemes being tested at all temperatures, I

--&--..:::--"-- - ... ----. ---a

one can determine which scheme gives the best fit. o I/'~.. .
Finally, the computer program reports rate constants o 25 50 75 100 125 150 175 200
for each reaction at each temperature. Time. min
FIG. 3. Change of component concentration with time for a molar
ratio of BuOH/SBO (30:1), 10/0 H 2 S04 , 117 C. , Soybean butyl
RESULTS AND DISCUSSION esters; .... soybean oil; X. diglyceride; monoglyceride.
Chemistry of the transesterification of vegetable oils.
The stoichiometry of the overall transesterification
reaction requires 1 mol of TG for 3 mol of alcohol to give
3 mol of ester and 1 mol of glycerol (GL) (Fig. 1). This
..._a:=:-::-==='==-
100 , - - - - - - - - - - - _ - :
...... ---
~:--.----------.----------

.-/;'..
80 ,/
reversible reaction, which is either acid- or alkaline-
catalyzed, involves stepwise conversions of TG to DG to
MG to GL producing 3 mol of ester in the process. One ~
"0
60
I
objective of this study was to determine the rate E
o
I

constants for all forward and reverse reactions. Rather ~ 40 f-

than use k, to designate the rate constant for the
conversion of the TG to DG, we have chosen to use the
designation TG-DG. Similarly, designations for the
other rate constants are shown in Figure 1. It should be
noted that the transesterification reaction consists of 0'----'----1.---'----'----'---'-'
o 2 4 6 8 10 12
consecutive as well as reversible reactions that must be Time. min
accounted for in any kinetic study. FIG. 4. Change of component concentration with time for a molar
Effect of time and temperature'on ester formation. To ratio of BuOH/SBO (6:2). 1% NaOBu, 30 C. , Soybean butyl esters;
determine the effect of temperature on ester formation ., monoglyceride X 10; .... soybean oil; X. diglyceride.
for the reaction of BriOH with SBO ata molar ratio of
30:1 catalyzed by 1% H 2 S0 4 "five temperatures ranging
from 77-117 C were examiried~'The results of this study
showed that ester formation is essentially complete in tions would be expected to be second order. The fit of the
3 hr at 117 C, compared to 20 hr at 77 C (Fig. 2). It will be . lines to the points was excellent for all four components.
noted that the curves have spacing which indicates a Similarly, good fits were obtained for the other four
distinct rate change for every 10 C temperature rise. The temperatures studied. These findings support our
curves also indicate the reaction begins at a slow rate, kinetic assignment for the forward and reverse
proceeds at a faster rate and then slows again as the reactions. The curves for all four components shown in
reaction nears completion. The shapes of the curves at Figure 3 are characteristic of consecutive reactions
two adjacent temperatures enabled us to predict the and confirm that the transesterification reaction
approximate shapes of the curves for the remaining consists of a series of consecutive reactions. It should be
three temperatures. This information was valuable for noted that both the forward and reverse reactions are
determining both when samples should be taken and consecutive. The rate constants obtained for these
when to stop the reaction. Similarly shaped curves for reactions are discussed later.
the esters were observed for alkaline-catalyzed reac- Kinetics of alkaline-catalyzed transesterification
tions. reactions. Alkaline-catalyzed transesterifications pro-
Kinetics of acid-catalyzed transestenfication reac- ceed at considerably faster rates than acid-catalyzed
tions. We undertook a kinetic study of the reaction of transesterifications (2,17). Partly for this reason and
BuOH with SBO using a 30:1 molar ratio with 1% partly because alkaline catalysts are less corrosive to
H 2 S0 4 at five temperatures. A 30:1 molar ratio was industrial equipment than acid catalysts, most commer-
chosen because at molar ratios of 6:1 and 20:1, cial transesterifications are conducted with alkaline
conversion to ester was unsatisfactory (1). The results catalysts. Sodium alkoxides are among the most
obtained at 117 C are shown in Figure 3. Because we efficient catalysts used for this purpose (5,6,18) and were
used a large excess of alcohol over that required by the therefore used in our study.
stoichiometry, we expected the forward reactions to The molar ratio of alcohol to SBO has an important
follow pseudo-first-order kinetics. The reverse reac- bearing on the kinetic order of the reaction. We have

JAOCS. Vol. 63, no. 10 (October 1986)

1378

B. FREEDMAN, R.O. BUTTERFIELD AND E.H. PRYDE

shown that acid catalysis of the forward reaction appearance of ester without the appearance of inter-
employing a 30:1 molar ratio of BuOH to SBO followed mediate DG and MG, a shunt-reaction scheme was
pseudo-first order kinetics. This was expected because tested. This shunt-reaction scheme comprised both the
of the large molar excess of alcohol. Similarly, we second-order reactions and a shunt reaction in which 3
expected alkaline-catalyzed transesterifications using a mol of MeOH directly attacked 1 mol of TG. By using the
30:1 molar ratio of alcohol to SBO to follow pseudo-first shunt-reaction scheme, we obtained much better fits of
order kinetics. We conducted two studies, both using the data. Thus, the kinetics which best describe the
molar ratios of BuOH to SBO of 30:1, at five reaction of MeOH/SBO at 6:1 molar ratio consist of a
temperatures between 20 and 60 e in which the combination of second-order consecutive and fourth-
catalysts were 1% and 0.5% NaOBu by weight. The data order shunt reactions.
obtained from these experiments were used in compu- The shunt-reaction scheme was further supported by
terized plots of concentration vs time as previously comparing certain results obtained from kinetic studies
discussed. These plots strongly supported our assump- of both the shunt and second-order schemes. Two
tion that the NaOBu-catalyzed forward reactions at criteria were used in this comparison. The first was a
30:1 followed pseudo-first order kinetics. comparison of summed squared and average errors. The
We next studied the effect of lowering the molar ratio summed squared errors for the second-order scheme at
from 30:1 to 6:1 with 1% NaOBu as catalyst. It has 20-60 e were 1.5, 304, 2.9, 6.7 and 4.1, all values X 10-5 ,
previously been shown that with alkaline catalysis, a 6:1 for a total of 18.6 X 10-5 The summed squared errors for
molar ratio of alcohol to vegetable oil results in high the shunt-reaction scheme at 20-60 e were 1.5, 004, 1.9,
conversions (93-98%) of vegetable oil to ester, whereas 1.8 and 1.5, all values X 10-5 , for a total of 7.1 X 10-5 The
below the 6:1 ratio conversion decreases (1). A 6:1 molar average errors for the second-order scheme were 0.8,
ratio would therefore probably be the preferred ratio in 1.1, 1.0, 1.6 and 1.3, all values X 10-" for a total of 5.8 X
commercial operations. At this ratio a second-order 10-3 The average errors for the shunt-reaction scheme
reaction appeared more likely than a pseudo-first-order were 0.8, 004, 0.8, 0.8 and 0.8, all values X 10-" for a total
reaction because the excess alcohol was only twice that of 3.6 X 10-3 Thus both the total ofthe summed squared
of theory compared to 10 times that of theory for a 30:1 errors and the average errors were lower for the
ratio. Both schemes, however, were tested using the shunt-reaction scheme compared to the second-order
computerized kinetic program. Again it was assumed scheme. A plot of data for component concentration vs
that the consecutive, reverse reactions were second time for MeOH/SBO at a 6:1 molar ratio with 0.5%
order. This program indicated a better fit of lines to NaOeH 3 at 40 e, assuming the shunt-reaction scheme,
points for the second-order reaction than for the is shown in Figure 5. The close fit of lines to points, at
pseudo-first-order reaction. The summed squared this temperature and the other four temperatures
errors for the second-order reactions at 20-60 e were studied, strongly supports the validity of this scheme.
46.5, 304, 1704, 27.7 and 22.7, for a total of 117.7. The The second criterion was correlation coefficients
summed squared errors for the pseudo-first-order obtained by linear regression from plots of log k vs lIT.
reactions at 20-60 e were 235.2, 7.1, 33.2, 44.2 and 25.6, These plots were used to determine energy of activation,
for a total of 345.3. The average errors for the E a , and are discussed later. The correlation coefficients
second-order reactions were i.3, 004, 0.8, 1.0 and 0.9, for obtained with the shunt-reaction scheme were generally
a total of 404. The average:~rrors for the pseudo-first- higher than those from the second-order scheme. Thus,
order reactions were 2.9, D.5; 1.1, 1.3 and 1.0, for a total the shunt reaction must be used in conjunction with the
of 6.8. Thus both the total of the summed squared errors second-order reaction to fully explain the kinetics for
and the average errors were lower for the second-order the reaction of MeOH with SBO at a 6:1 molar ratio.
reaction compared to the pseudo-first-order reaction. Shunt reactions have been observed in kinetic studies of
These differences appear to be significant and were the the hydrogenation of conjugated diene (19) and SBO and
basis for choosing the second-order reaction. A plot of linseed oils (20).
data for component concentration vs time for a 6:1 molar We next reexamined the reaction of BuOH/SBO at 6:1
ratio of BuOH to SBO with 1 % NaOBu at 30 e,
assuming a second-order reaction, is shown in Figure 4.
0.12,...----------------,
The fit of the lines to the points in this case was
excellent. Good fits were also observed at the other four
temperatures. These results suggest that at a 6:1 molar 0.09
ratio with 1 % NaOBu, the reaction of BuOH with SBO
follows consecutive, second-order kinetics. ],
We also expected the reactions of MeOH with SBO at '" 0.06
'"
"0
a 6:1 molar ratio with 0.5% NaOeH 3 at 20-60 e to be @
second order. MeOH, however, behaved differently from ~
BuOH in this regard. Although a second-order scheme 0.03
was evaluated by our computer program for the reaction
of MeOH/SBO at 6:1, the average error obtained was
somewhat larger than that observed with BuOH/SBO at
6:1. This increase in error is because of the rapid 5 10 15 20 25 30
Time, min
appearance of ester without a corresponding rapid FIG. 5. Change of component concentration with time for a molar
increase and decrease of DG and MG. Because ratio of MeOH/SBO (6:2), 0.5% NaOCH), 40 C. , Soybean methyl
consecutive reactions could not explain the rapid esters; X. diglyceride X 10; _, monoglyceride X 10;.a., soybean oil.

JAOCS, Vol. 63, no. 10 (October 1986)

 1379
TRANSESTERIFICATION KINETICS OF SBO

to determine if the shunt-reaction scheme was applic- 0.5% NaOBu- vs 1% NaOBu-catalyzed reactions was
able. Our computer program calculated a zero reaction unexpected. This difference probably resulted because
rate, clearly indicating that the shunt reaction did not the 0.5%- and 1%-catalyzed reactions were run on
exist for the BuOH reaction. Why was the transester- different scales (see Table 1) and with different reactors
ification of SBO with MeOH different from that with (see Experimental). The larger size of the 400ml flask
BuOH at the 6:1 molar ratio? Two possible explanations may have permitted more vigorous agitation, which
can be proposed. The smaller size of the MeOH molecule could have influenced reaction rates. Had the reactions
relative to that of BuOH could facilitate the simultane- been run under identical conditions, rates for the
ous attack of MeOH on all three chains of the TG. Also, 1%'catalyzed reactions should have been higher than
the greater polarity of the methoxy anion compared to the those for the 0.5%catalyzed reactions. A number of
butoxy anion might promote its attack on the ester kinetic studies have shown that reaction rates increase
carbonyl according to established mechanisms. with increasing catalyst concentration (7,8,14).
Rate constants. Rate constants were calculated by our Energy of activation. The relationship between
computer program for all reactions included in our reaction rate constant, k, and temperature is given by
kinetic study. The rate constants determined for the the integrated form of the Arrhenius equation: loglak =
reaction of BuOH with SBO at a 30:1 molar ratio using (- Eal2.303R)/T + C where Eo is the energy of activation,
various catalysts are shown in Table 2. Because the R the gas constant in calories mol' I degree'l, T the
forward and reverse reactions are not of the same kinetic absolute temperature and C a constant. From a plot of
order at a 30:1 molar ratio, a direct comparison of these log k vs lIT, the slope can be determined. This slope is
rate constants cannot be made. The alkaline-catalyzed equal to (- E o /2.303R). Thus, Eo = -4.58 (slope).
reactions had much larger rate constants than the acid- We determined energy of activation in this manner for
catalyzed reactions for both forward and reverse most of the reactions. An example is shown in Figure 6
reactions, as expected. The higher rate constants for the for the acid-catalyzed butanolysis of TG to DG at
77-117 C. Linear regression analysis of these data gave
a slope of -3258.1 with a correlation coefficient of
TABLE 2
Effect of Various Catalyst Systems on Rate Constants
-1.5
BuOH/SBO, 30:1, rate constants X 103 117 C
-1.7 107 C
Rate 1% H 2 SO 4 , 1% NaOBu, 0.5% NaOBu,
designation 77C 60 C 60 C
-1.9 ~97 C
.;;-

~7C
TG-DG 3 3,822 26,626 ""
..E
DG-MG 8 1,215 3,584 -2.1
MG-GL 7 .792 2,373
DG-TG 0.02 121 439
MG-DG 0.05 7 8
-2.3
"~77 C
GL-MG 0.03 11 7
-2.5
0.25 0.26 0.27 0.28 0.29
For the forward reactions units are min-I. For the reverse reactions
l/T x 10'
units are ml mmol-' min-I for the acid-catalyzed reaction, and ml
11Il101-' min-I for the alkali-catalyzed reaction. TG, triglyceride; DG, FIG. 6. Plot of log k vs liT for the butanolysis of triglyceride to
diglyceride; MG, monoglyceride; GL, glycerol diglyceride with 1% H 2 S0 4 at 77-117 C.

TABLE 3
Energy of Activation lE.) Under Studied Conditions
E. (calfmol): reaction conditions

BuOH/SBO, 30:1 BuOH/SBO, 6:1 MeOH/SBO, 6:1

1% H 2 SO 4 1% NaOBu 0.5% NaOBu 1% NaOBu 0.5% NaOCH 3

Rate designation 77-117 C 20-60 C 20-60 C 20-60 C 20-60 C

TG-DG 14,922 15,360 15,662 17,092 16,062

DG-MG 16,435 11,199 13,053 12,137 17,247
MG-GL 15,067 11,621 13.395 10,693
DG-TG 19,895 17,195 15,587 15,925 15,843
MG-DG 16,885 13,336 15,816 13,571
GL-MG 12,196 13,110 8,181
TG-GL 20,022

TG, triglyceride; DG, diglyceride; MG, monoglyceride; GL, glycerol.

JAOCS, Vol. 63, no. 10 (October 1986)

1380

B. FREEDMAN, R.O. BUTTERFIELD AND E.H. PRYDE

-0.9964. By using this slope in the equation Eo = -4.58 Proceedings of the International Conference on Plant and
(slope), an Eo of 14,922 cal/mol was calculated for this Vegetable Oils as Fuels, American Society of Agricultural
reaction. Correlation coefficients for the other five Engineers, St. Joseph, MI, 1982, p. 117.
3. Freedman, B., E.H. Pryde and W.F. Kwolek, J. Amer. Oil
reactions of BuOH/SBO, 30:1 catalyzed by 1 % H 2 S04 Chem. Soc. 61:1215 (1984).
were -0.9918, -0.9965, -0.8918, -0.9795 and 4. Freedman, B., W.F. Kwolek and E.H. Pryde. Ibid. 62:663
-0.9905. (1985).
Energies of activation determined for other reactions 5. Markley. KS., in Fatty Acids, Their Chemistry, Properties,
in our study ranged from 8,000-20,000 cal/mol (Table 3). Production and Uses, edited by KS. Markley, 2nd ed.,
Other investigators have reported Eo values within this Interscience Publishers, New York, 1961, Part 2. p. 862.
6. Sridharan, R., and 1.M. Mathai.J. Sci.Ind. Res. 33:183 (1974).
range for other transesterification reactions (6,8,9-12). 7. Cotes, A.C., and L.A. Wanzel, Lat. Am. J. Chem. Eng. AppL
Eo for the shunt reaction TG-GL had a value of 20 Chem. 4:126 (1974).
kcal/mol. The spread of values seen in Table 3 for Eo is 8. Davies. B., and G.V. Jeffreys, Trans. Inst. Chern. Eng. 51:271
due partly to experimental error. In addition, the (1973).
simplified model used in the equation log 10K = 9. Farkas, L., O. Schachter and B.H. Vromen, J. Am. Chem. Soc.
(-Eal2.30R)/T + C may not be adequate to account for 71:1991 (1949).
10. Baker, G.E., and R.R. White, Chem. Eng. Prog. Symp. Ser. No.
all the variables involved. This might also explain why 448:75 (1952).
Eo for some reverse reactions are greater than those of 11. Ravindranath, K., and R.A. Mashelkar, J. Polym. Sci. Polyrn.
the corresponding forward reactions. Further research is Chem. Ed. 20:3447 (1982).
needed to develop a model that might provide a narrower 12. Tomita, K, and H. Ida, Polymer 14:55 (1973).
range of Eo values. 13. Pore, J., J.P. Houis and 1. Rasori, Rev. Tech. Ind. Cuir 73:164
(1981).
14. Dufek. E.J., R.O. Butterfield and E.N. Frankel, J. Amer. Oil
ACKNOWLEDGMENTS Chern. Soc. 49:302 (1972).
15. Dufek, E.J., Ibid. 55:337 (1978).
Haifa Khoury assisted in the laboratory and in manuscript 16. Butterfield, R.O., Ibid. 46:429 (1969).
preparation; Wilma Bailey gave statistical assistance. 17. Reid. E.E., Am. Chern. J. 45:506 (1911).
18. Formo. M.W., J. Amer. Oil Chem. Soc. 31:548 (1954).
19. Koritala, S. R.O. Butterfield and H.J. Dutton. Ibid. 50:317
REFERENCES (1973).
20. Vigeron, P.Y.. S. Koritala, R.O. Butterfield and H.J. Dutton,
1. Freedman, B., E.H. Pryde and T.L. Mounts, J. Amer. Oil Ibid. 49:371 (1972).
Chem. Soc. 61:1638 (1984).
2. Freedman, B., and E.H. Pryde, in Vegetable Oil Fuels, [Received December 17, 1985]

JAOCS. Vol. 63. no. 10 (October 1986)