See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/43276316

Hydrogenation of Vegetable Oils Using Mixtures of Supercritical Carbon Dioxide

and Hydrogen

Article  in  Journal of the American Oil Chemists' Society · February 2001

DOI: 10.1007/s11746-001-0229-8 · Source: OAI

CITATIONS READS
106 5,616

4 authors, including:

King J.W Gary R. List

Critical Fluid Symposia University of Illinois, Urbana-Champaign
182 PUBLICATIONS   5,973 CITATIONS    190 PUBLICATIONS   4,579 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Trait modified oils View project

Lipidtechnoogy View project

All content following this page was uploaded by King J.W on 16 May 2014.

The user has requested enhancement of the downloaded file.

 Hydrogenation of Vegetable Oils Using Mixtures
of Supercritical Carbon Dioxide and Hydrogen
Jerry W. King*, Russell L. Holliday, Gary R. List, and Janet M. Snyder
Food Quality and Safety Research Unit, NCAUR, ARS, USDA, Peoria, Illinois 61604

ABSTRACT: Hydrogenation of vegetable oils under supercriti- hancement of hydrogen solubility in the oil phase.
cal conditions can involve a homogeneous one-phase system, or Several investigators have shown the efficacy of conduct-
alternatively two supercritical components in the presence of a ing hydrogenation of oleochemicals where both hydrogen and
condensed phase consisting of oil and a solid catalyst. The for- the fluid of choice (usually supercritical carbon dioxide, SC-
mer operation is usually conducted in flow reactors while the lat- CO2) are technically in the supercritical fluid state. These
ter mode is more amenable to stirred, batch-reactor technology. studies have been predominantly carried out in tubular flow
Although many advantages have been cited for the one-phase hy-
reactors by Harrod and coworkers (1–3) and Tacke et al.
drogenation of oils or oleochemicals using supercritical carbon
(4,5), and have utilized mostly nontraditional catalysts to ac-
dioxide or propane, its ultimate productivity is limited by the oil
solubility in the supercritical fluid phase as well as unconven- celerate the hydrogenation reaction. Recently, Andersson and
tional conditions that affect the hydrogenation. In this study, a King (6) expanded on these initial studies and successfully
dead-end reactor has been utilized in conjunction with a head- coupled a transesterification reaction conducted in SC-CO2
space consisting of either a binary fluid phase consisting of varying with downstream supercritical hydrogenation to produce sat-
amounts of carbon dioxide mixed with hydrogen or neat hydro- urated fatty alcohols directly from vegetable oil feedstocks.
gen for comparison purposes. Reaction pressures up to 2000 psi A variation in using a batch reactor for the hydrogenation of
and temperatures in the range of 120–140°C have been utilized hops (7) has been patented using either the addition of dry ice
with a conventional nickel catalyst to hydrogenate soybean oil. or liquid CO2 to facilitate the synthesis of tetrahydroiso-α-
Depending on the chosen reaction conditions, a wide variety of acids. Likewise, Bertucco and coworkers (8) utilized an in-
end products can be produced having different iodine values,
ternal recycle reactor to hydrogenate unsaturated ketones.
percentage trans fatty acid content, and dropping points or solid
Whereas conventional hydrogenation of vegetable oils is
fat indices. Although addition of carbon dioxide to the fluid phase
containing hydrogen retards the overall reaction rate in most of accomplished in a multiphase, gas–liquid system at low pres-
the studied cases, the majority of products have low trans fatty sures in which the hydrogen is contacted with a nickel-based
acid content, consistent with a nonselective mode of hydrogena- catalyst that is suspended in the vegetable oil, the above su-
tion. percritical hydrogenation systems employ a one-phase sys-
Paper no. J9724 in JAOCS 78, 107–113 (February 2001). tem in which the oil is dissolved in the supercritical fluid at
much higher pressures. In addition, traditional oil hydrogena-
KEY WORDS: Carbon dioxide, hydrogenation, reaction, super- tion processes use batch-stirred autoclave reactors as opposed
critical fluid, vegetable oil. to fixed-bed flow reactors which are conveniently utilized
with the above binary fluid mixtures. Such factors, coupled
Hydrogenation of vegetable oils is a traditional oleochemical with supported palladium, platinum, or possibly ruthenium-
modification process that has undergone limited change since or rhodium-based catalysts when used to demonstrate the
its inception earlier in this century. Recently the concept of benefits of conducting hydrogenation of vegetable oils in
utilizing supercritical fluid conditions for improving hydro- supercritical fluids, are in marked departure from existing in-
genation of fats and oils has been advocated based on several dustrial practice where supported, less expensive nickel cata-
criteria. These include improvement in mass transfer of the lysts yield a wide range of activity and selectivity.
reactants to the surface of the catalyst particles, higher reac- The effect of pressure on the hydrogenation of vegetable
tion rates, reduction in the amount of hydrogen required, and oils has been studied by several research teams (9–12). It is
reduction of undesirable by-products yielding a higher end- found that in commercial dead-end reactors, dissolution of
product quality. Such results are a direct result of the use of hydrogen in the oil is the rate-controlling step (13). Related
supercritical fluids in hydrogenation processes which result to this study are the earlier efforts of Koritala and coworkers
in accelerating heat transfer, decreasing oil viscosity, and en- (11,12,14), who explored soybean oil hydrogenation at pres-
sures up to 30,000 psi using nontraditional copper catalysts
in both batch and flow reactors. Increasing pressure for the
*To whom correspondence should be addressed at Food Quality and Safety hydrogenation of vegetable oils is known to increase the re-
Research Unit, NCAUR, ARS, USDA, 1815 N. University St., Peoria, IL
61604. action rate, the concentration of the hydrogen in the oil and
E-mail: kingjw@mail.ncaur.usda.gov on the catalyst surface, while negating selectivity and trans-

Copyright © 2001 by AOCS Press 107 JAOCS, Vol. 78, no. 2 (2001)
108 J.W. KING ET AL.

isomer formation (15). A recent study by Fillion and Morsi nickel on the support. The catalyst charge to oil of 0.02 wt%
(13) has characterized the gas–liquid mass transfer and hy- nickel concentration based on the weight of the oil was used
drodynamic parameters for the hydrogenation of soybean oil in the hydrogenation experiment.
under industrial conditions. Generally, the above factors lead Hydrogenation reactors. The hydrogenation reactions, both
to a higher solids content in the resultant fat and higher, flat- conventional and with the aid of supercritical fluids, were
ter solid fat indices (SFI) curves. performed using a standard Parr 2-L, batch autoclave (Parr In-
In this study the effect of using binary mixtures of SC-CO2 struments, Inc., Moline, IL), equipped with both a pressure
and H2 above the critical point of the mixture was tested using transducer and gauge for monitoring the pressure of the reactor.
a standard industrial nickel catalyst in a batch, stirred reactor. In-line with the above monitors was a relief valve in case of
The liquid phase consisting of oil, suspended catalyst, and over-pressurization as well as a rupture disk for venting the con-
dissolved binary fluid mixtures interfaced with a reactor head- tents of the reactor. The reactor also had type J and K thermo-
space gas phase. The headspace gas phase varied in composi- couples for measuring the temperature of the reactor. A vacuum
tion and applied pressure. Thus, the variation from a traditional exhaust line was used for venting the autoclave along with a
hydrogenation practice (both selective and nonselective) is gas/sampling inlet tube. Termination of the reaction as well as
focused in changing the gas (fluid) phase while maintaining maintenance of the reactor’s heat balance was aided by use of
conventional reactor (stirred, batch) and nickel catalyst usage internal cooling coils. In addition, external air-cooling knives
consistent with standard industrial practice. Additional bene- (Exair Corp., Cincinnati, OH) were used as an aid in cooling the
fits may also accrue when using binary fluid mixtures of CO2 reactor contents down after completion of the reaction.
and H2, such as the direct use of water gas shift reaction-de- Reaction parameters such as temperature, pressure, and agi-
rived CO2 and H2 mixtures without the need for prior gas-gas tation rate were monitored with the aid of a computer utilizing a
separation. Also the prophylatic effect of CO2 addition to H2 National Instruments Inc. (Austin, TX) LabVIEW software pro-
for increasing its flammability limit over that of neat H2, as gram. As shown in Figure 1, an external Isco Model 260D sy-
well as final product color improvement (16), provides extra ringe pump (Isco, Inc., Lincoln, NE) was used for pumping the
incentives for considering the approach described in this hydrogen into the bottom of the reactor vessel to the required
study. For the above reasons, as well as reducing the depar- pressure and volume. Pressurized CO2 was fed into the reactor
tures from traditional processing methodology, it is hoped by a Model AGT 162-52 gas booster pump made by Haskel,
that the results obtained in this study might be considered for Inc. (Burbank, CA). To aid in establishing the pressure-based
adoption by the oleochemical industry. additions of the various gases, a second pressure transducer
(PT), Heise Model ATS 2000 (Dresser Industries, Newtown,
NH) was placed in-line between the delivery pumps and reactor
MATERIALS AND METHODS
vessel (see Fig. 1). External cooling water for the reactor was
Materials and analysis methods. A refined, bleached, and de- from an in-house, deionized water supply, triggered by a sole-
odorized soybean oil (RBD-SBO) from Riceland Industries noid valve on demand, and the shaft of the reactor’s mechanical
(Stuttgart, AR) that had an initial iodine value (IV) of 129 and stirrer cooled with the aid of a Neslab Instruments (model RTE-
a fatty acid composition of 18:0 = 4.2%, 18:1 = 23.1%, 18:2 100, Portsmouth, NH) circulating water bath held at −15°C.
= 52.5%, and 18:3 = 6.2% was used in all experiments. The Reaction conditions. Conditions for establishing a baseline
extent of the hydrogenation reaction was monitored using an for conventional selective hydrogenation were: 170°C, 15 psi
Abbe 3L refractometer from Bausch & Lomb (Rochester, H2, 0.05 wt% Ni catalyst, medium stirring rate (ca. 400 rpm).
NY) thermostated at 60°C. A plot of IV vs. refractive index For nonselective hydrogenations, the conditions were: 120°C,
(RI) was interpolated to measure the extent of the hydrogena- 50 psi H2, 0.02 wt% Ni catalyst, and high stirring (ca. 800 rpm).
tion reaction. For these experiments, 950 g of RBD–SBO were used in the
Fatty acid compositions of the resultant products were de-
termined by a standard procedure using gas chromatography
and fatty acid methyl ester (GC–FAME) derivative formation
(17). End products were also characterized by dropping point
(D.P.) and SFI measurements as described in the Official and
Recommended Methods of the American Oil Chemists’ Society
(18).
Hydrogen was obtained from BOC Gases (Murray Hill,
NJ) and was 99.995% minimum purity grade. The CO2 uti-
lized in the experiments was also from BOC Gases and was
of 99.8% minimum purity. Propane (Suburban Gas Company,
Peoria, IL) was also briefly examined in place of CO2 during
the binary fluid hydrogenations. The nickel catalyst, Calsicat
FIG. 1. High-pressure hydrogenation system for conducting conven-
Ni catalyst droplets, E-479D, was obtained from Mallinck- tional and binary hydrogenations on soybean oil (SBO). PT, pressure
rodt Chemical, Inc. (Erie, PA). This consisted of 0.25 wt% transducer; J and K, thermocouples.

JAOCS, Vol. 78, no. 2 (2001)

 HYDROGENATION OF OILS IN BINARY FLUID MIXTURES 109

Parr reactor. using either the specified CO2/H2 or C3H8/H2 reaction condi-
Initially, supercritical reaction conditions using binary mix- tions produced no change in the IV of the product, even for
tures of CO2/H2 were as follows: 2000 psi CO2 + 100 psi H2 reactions run 4–6 h. Therefore this approach was abandoned
(2,100 psi total), 120°C, 0.02 wt% Ni catalyst. In this case, the in favor of using binary fluid mixtures in which the CO2 and
headspace volume of the reactor varied between 1000 and hydrogen pressure were equivalent, utilizing nonselective hy-
1,500 mL for charges of 500 to 950 g of oil. Some brief experi- drogenation conditions. In this case, equivalent gas pressures
ments were also run with a C3H8/H2 system in which 1,000 or ranging from 250 to 1,000 psi were utilized.
1,400 psi of propane would be blended with 100 psi of H2. For Binary supercritical fluid mixtures. In Table 1 the hydro-
the C3H8/H2 system, reaction conditions were the same as for genated oil products, using the above gas compositions, were
the CO2/H2 system above, except only 200 g of oil was utilized. characterized with respect to their IV, percentage trans fatty
These initial conditions for the CO2/H2 and C3H8/H2 produced acid content, and overall fatty acid composition for 2- and
nonselective hydrogenation. 4-h reaction times. Also included in Table 1 are the results
A variation in the respective amounts of the components in from conducting hydrogenation with just H2 at 50 psi for the
the binary fluid system, CO2 and H2, was also made using the same time period. Here the extent of hydrogenation decreases
following conditions: 1,000 psi CO2 + 1,000 psi H2, 500 psi as the overall pressure of the binary fluid system decreases as
CO2 + 500 psi H2, and 250 psi CO2 + 250 psi H2. All of these indicated in the corresponding IV values. The percentage of
reactions were performed at 120°C using 0.02 wt% Ni catalyst, trans fatty acid content at either 2- or 4-h sampling periods
with high stirring in the reactor containing 500 g of oil. also decreased as the overall system pressure was decreased
Hydrogenation conditions using higher hydrogen pressures (10), as did the saturated fatty acid content of the resultant oil
were as follows: 120°C, 0.02 wt% Ni catalyst, and 500 g of (i.e., stearic acid). These two results suggest that nonselective
RBD–SBO. In this case, a pressure ladder of 500, 1,000, and hydrogenation is taking place under these conditions, yield-
1,900 psi of H2 was run using the above conditions. To observe ing oils that have quite different properties from the nonse-
the effect of increasing the reaction temperature to 140°C, two lective hydrogenated product produced at 50 psi.
CO2/H2 systems were used with 2,000 psi CO2 and 100 psi H2 Pure H2 elevated pressure hydrogenations. Additional hy-
or 500 psi CO2 and 500 psi H2, respectively. A reference hy- drogenations were run without the second supercritical fluid
drogenation was also run using only a 500 psi H2 atmosphere. component, i.e., SC-CO2, using a nearly equivalent total pres-
All of these reactions were conducted at 140°C, 0.02% Ni con- sure to that used in the above-described CO2/H2 pressure
centration, 800 rpm stirring rate, with 500 g of soybean oil. ladder. These results are tabulated in Table 2 for 2- and 4-h
A typical reaction sequence consisted of metering the gases sampling intervals so as to compare them against the nonselec-
(fluids) to the required pressures after heating the RBD–SBO tive, lower-pressure hydrogenation results. In this case, the IV
to the desired reaction temperature under vacumn. When of the resultant hydrogenated oils is similar to the IV exhibited
required, cooling water to the reactor coils was applied to coun- by the product from nonselective low-pressure hydrogenation.
teract the effects of the hydrogenation exotherm. After the re- The percentage of trans fatty acid content of the hydrogenated
action was conducted for a specific time, the reactor was shut oils produced using higher hydrogen pressures is significantly
off and cooled to ca. 65°C, before venting the gases utilized in lower compared to the oil hydrogenated at 50 psi, while the
the experiment. Collected oil samples for characterization were stearic acid content increased using a higher hydrogenation
filtered through celite. pressure. These results point to a set of reaction conditions that
can produce low trans fatty acid levels, but that is inherently
nonselective with respect to the mode of hydrogenation.
RESULTS AND DISCUSSION
Effect of reaction temperature. Experiments were also con-
Initial supercritical reaction conditions. Initial reactions ducted to see what effect the reaction temperature would have

TABLE 1
Properties of Soybean Oil Hydrogenated Using Binary Fluid Mixtures of Carbon Dioxide
and Hydrogen
Soybean Nonselective 1,000 psi CO2 500 psi CO2 250 psi CO2
oil (%) 50 psi H2 1,000 psi H2 500 psi H2 250 psi H2
Time (h)
2 4 2 4 2 4 2 4
IVa
105 69 116 82 118 96 122 109
Trans 7.1 23.3 1.9 6.4 1.5 5.0 1.4 3.8
18:0 5.8 16.8 8.3 23.5 7.3 17.0 5.7 10.7
18:1 42.6 61.4 28.2 35.6 27.6 33.2 26.6 31.4
18:2 33.8 6.4 45.2 25.2 46.6 33.3 48.9 40.3
18:3 2.8 0 4.8 2.0 5.0 3.4 5.3 4.2
a
IV, iodine value.

JAOCS, Vol. 78, no. 2 (2001)

110 J.W. KING ET AL.

TABLE 2
Properties of Hydrogenated Soybean Oils Produced Using Pure Hydrogena
Soybean Nonselective
oil (%) 50 psi H2 1,900 psi H2 1,000 psi H2 500 psi H2
Time (h)
2 4 2 4 2 4 2 4
IV
105 69 108 75 108 72 110 76
Trans 7.1 23.3 2.7 7.0 3.1 7.4 3.4 8.6
18:0 5.8 16.8 11.7 27.8 12.6 30.8 11.0 26.9
18:1 43.6 61.4 29.2 35.3 29.1 33.0 29.7 35.7
18:2 33.8 6.4 41.2 21.7 40.9 21.2 41.5 22.2
18:3 2.8 0 3.5 1.3 3.6 1.8 4.3 1.9
a
See Table 1 for abbreviation.

on the results of the hydrogenation (Table 3). In this case hy- genations conducted at 140°C (500 psi CO2/500 psi H2 and
drogenations of soybean oil were performed with a binary 500 psi H2) show a rapid drop in IV with reaction time com-
fluid system of 500 psi for both CO2 and H2, as well as in pure pared to hydrogenations conducted under the above-described
H2 at a 500 psi level, but the reaction temperature was 140°C, conditions. This trend can be partially ascribed to the increase
rather than 120°C. Here we observed a difference from the oil in reaction temperature (11) but requires that enough H2 be
properties achieved previously when comparing the results available to contact with the catalyst/oil (note the result for
from hydrogenating with a binary fluid mixture vs. pure hy- the 2,000 psi CO2/100 psi H2 mixture).
drogen at 500 psi. In this case, both the IV and percentage of Figure 3 shows the relationship between the percentage of
trans fatty acids, as well as the stearic and oleic acids con- trans fatty acid vs. IV value of the resultant soybean oil. In
tents are very close for the two above reactions. this case, the trans fatty acid content–IV relationship is linear
Comparison of results between different reaction condi- and independent of reaction pressure as observed by others
tions. Figure 2 is a plot of reaction time vs. IV value for the (9), the lone exception being the 2,000 psi CO2/100 psi H2
hydrogenated soybean oil products. In Figure 2 the symbol hydrogenation conducted at 140°C which mimics the conven-
codes have been grouped together: the first three representing tional low-pressure hydrogenation at 50 psi. Thus, the binary
hydrogenations conducted at elevated pressures in a pure hy- fluid compositions and pure H2 atmospheres at higher pres-
drogen atmosphere, the next three representing the binary sures yield lower trans fatty acid content having similar IV
fluid mixtures at 120°C followed by a conventional low- values when compared with the traditional low-pressure hy-
pressure hydrogenation at 50 psi, and then hydrogenations drogenation conditions. Also note that for most of the hydro-
done at 140°C. As noted by others (19), an increase in pres- genations in Figure 3, the percentage trans fatty acid content is
sure increases the reaction rate for hydrogenation (steeper IV 30% lower than that usually found in a hydrogenated soybean
vs. time plots) for the CO2/H2 mixtures. Figure 2 leads one to oil with an IV value of 70 (20). Also note that reaction condi-
the conclusion that the binary gas mixtures are retarding the tions which yield the lower trans fatty acid content at similar
hydrogenation reaction relative to pure H2 at 120°C. The use IV continue to head downward in Figure 3, suggesting that
of higher pressures with pure H2 yields no apparent advan- the trans fatty acid content will continue to remain low as the
tage above 500 psi in terms of reaction rate (10) and yields reactions proceed under the described conditions.
results similar to those obtained at 50 psi. Two of the hydro- The percentage of stearic acid content vs. IV (Fig. 4)

TABLE 3
Properties of Soybean Oils Hydrogenated at a Higher Temperature (140¡C)a
Soybean Nonselective 2,000 psi CO2 500 psi CO2
oil (%) 50 psi H2 120°C 100 psi H2 500 psi H2 500 psi H2
Time (h)
2 4 1 3 1 2 1 2
IV
105 69 104 65 88 39 91 48
Trans 7.1 23.3 9.0 25.4 7.2 12.0 6.9 12.3
18:0 5.8 16.8 5.3 16.1 19.3 49.0 19.8 43.8
18:1 43.6 61.4 46.3 69.7 36.7 33.1 33.2 33.9
18:2 33.8 6.4 31.1 2.2 28.0 4.4 30.5 8.9
18:3 2.8 0 2.2 0.2 2.7 0.3 2.9 0.6
a
See Table 1 for abbreviation.

JAOCS, Vol. 78, no. 2 (2001)

 HYDROGENATION OF OILS IN BINARY FLUID MIXTURES 111

FIG. 2. Iodine value (IV) vs. reaction time for the hydrogenation on SBO FIG. 4. IV vs. percentage of stearic acid (% 18:0) for SBO hydrogenated
under various experimental conditions. See Figure 1 for other abbrevia- under various conditions. See Figures 1 and 2 for abbreviations.
tion.

shows similar loci for all the reported reaction conditions ex- a slightly higher IV than that observed for a commercial mar-
cept for the conventional low-pressure hydrogenation and the garine basestock, the percentage of trans fatty acid being a
2,000 psi CO2/100 psi H2 result at 140°C. These two hydro- decade lower than that found in a commercial sample. Table
genations yield a lower percentage of stearic acid in the final 4 also shows that the stearic acid content of both conventional
product that has a similar IV to that obtained under the other margarine basestock and the hydrogenated oils is almost iden-
eight hydrogenation conditions. This corroborates the trends tical.
for trans fatty acid production shown in Figure 3 and indi- It is also possible to obtain a lower trans fatty acid content
cates the nonselective nature of these hydrogenations. in the hydrogenated products compared to that found in a con-
Comparison of experimental products with commercial ventional shortening basestock (D.P. = 45–52°C) having a
products. The results obtained from the described experimen- similar IV range (see Table 4). Table 4 also shows that our ex-
tal hydrogenation runs have potential application in the food perimental hydrogenated products tend to have a slightly ele-
industry since the properties of the resultant oils closely ap- vated level of stearic acid relative to levels found in commer-
proximate the IV, trans fatty acid content, and solid fat con- cial shortening basestocks.
tent (% 18:0) of margarine and shortening basestocks having Both of the above results suggest that the hydrogenation
similar D.P. A conventional margarine basestock will usually conditions described in this study offer considerable versatil-
have a D.P. of 32–39°C and will exhibit the properties listed ity in designing an appropriate basestock for margarine or
in Table 4. The hydrogenated oils obtained in this study have shortening use. Particularly attractive is the lower trans fatty
acid content relative to synthesized basestocks, since lower
trans fatty acid levels in foodstuffs have an appeal to a health-
conscious public (21,22).
SFI were measured for several of the hydrogenated oils
synthesized in this study. Five of these products after 4 h of
hydrogenation time were characterized by their SFI vs. tem-
perature plots (Fig. 5). For the binary fluid mixtures and oil
hydrogenated with pure H2 at 1,900 psi, the temperature

TABLE 4
Comparison of Experimental Hydrogenated Soybean Oilsa vs.
Conventional Margarine and Shortening Basestocks
Margarine basestock Shortening basestock
(D.P. 32–39°C) (D.P. 45–52°C)
Conventional Experimental Conventional Experimental
%18:0 6–9 7–11 11–13 13–24
%Trans 11–30 1–3 15–20 3–8
FIG. 3. IV vs. percentage of trans fatty acid (% trans) for SBO hydro- IV 90–110 108–114 85–90 88–102
genated under various conditions. See Figures 1 and 2 for abbrevia- a
Experimental conditions: 120°C, 0.02 wt% Ni catalyst, 250 psi H2, 250 psi
tions. CO2. D.P. dropping point; see Table 1 for other abbreviations.

JAOCS, Vol. 78, no. 2 (2001)

112 J.W. KING ET AL.

FIG. 5. Solid fat index (SFI) vs. temperature for five SBO hydrogenated FIG. 7. Comparison of the SFI vs. temperature curves for potential low-
using binary fluid mixtures (CO2/H2) and under conventional condi- trans shortening basestocks derived from binary fluid hydrogenations with
tions. See Figure 1 for other abbreviation. low-trans fatty acid-blended oil mixture. % T, trans fatty acid content;
hyd., hydrogenated. See Figure 2 for abbreviation.

dependence of the SFI is a weak function of temperature. This fatty acid content of SBO hydrogenated at various conditions
is similar to trans-suppressive hydrogenation, which yields a along with their respective D.P. vs. temperature curves. For
long plastic range of melting behavior desired for shortenings both the pure hydrogen and CO2/H2 mixtures, the curves for
(23) but contrasts markedly with the SFI vs. temperature percentage of trans fatty acid vs. D.P. are very similar, an en-
curve for the conventional hydrogenation conducted at 50 psi couraging feature since these D.P. can be achieved on oils
H2 pressure. The SFI results in Figure 5 were all determined having a low-trans fatty acid content.
on hydrogenated oils having IV in the range of 60–70; how- By adjustment of the hydrogenation conditions, it is also
ever, the observed differences in their SFI vs. temperature possible to produce an oil having a low-trans fatty acid con-
curves reflect a varying saturated fatty acid content. For the tent (% T) at higher IV that behaves similarly to blending oil
four oils displaying an invariant SFI vs. temperature curve, it mixtures [e.g., a low-trans (6.4%) mixture consisting of
should be possible to stop the hydrogenation reaction earlier canola and hydrogenated corn oils]. Such a comparison is
before the saturated fatty acid level increases to yield differ- made in Figure 7 between two hydrogenated SBO products
ent SFI values. produced using binary fluid mixtures of CO2 and H2 at ele-
The commercial utility of the products obtained by using vated pressures and the above-mentioned blend of oils. The
the above hydrogenation methods is worth noting. Figure 6 two hydrogenated SBO exhibit somewhat similar SFI vs.
illustrates the relationship between the percentage of trans temperature curves as does the oil blend, making them good
substitutes for such margarine basestocks that have low-trans
fatty acid content.
The studies reported using supercritical hydrogenation
conditions are in contrast to the reported increases in reaction
rate noted by other investigators using flow-reactor systems
and precious-metal catalysts (24,25). Hydrogenations run at
higher pressures were found to be nonselective with respect
to their mechanism of reaction, but they also produced oils
having low-trans fatty acid content. The dependence of the
final oil properties on the hydrogenation conditions suggests
some unique opportunities for optimizing reaction conditions
to produce an array of products having different physical and
chemical properties.

ACKNOWLEDGMENTS
Ray Holloway, Angela Neese, and William Neff of the Food Qual-
FIG. 6. Percentage of trans fatty acid content of SBO hydrogenated under ity and Safety Research Unit are thanked for performing the analyti-
various conditions with binary fluid mixtures (CO2/H2) and pure H2 vs. cal and physical property tests. The mechanical assistance of Jeel
dropping point of the resultant product. See Figure 1 for abbreviation. Teel of the High-Pressure Laboratory at NCAUR in setting up the
hydrogenation reactor systems is gratefully acknowledged.

JAOCS, Vol. 78, no. 2 (2001)

 HYDROGENATION OF OILS IN BINARY FLUID MIXTURES 113

REFERENCES drodynamic Parameters in a Soybean Oil Hydrogenation Process

Under Industrial Conditions, Ind. Eng. Chem. Res. 39:2157–2168
1. Van den Hark, S., M. Harrod, and P. Moller, Hydrogenation of (2000).
Fatty Acid Methyl Esters to Fatty Alcohols at Supercritical Con- 14. Koritala, S., K.J. Moulton, Sr., J.P. Friedrich, E.N. Frankel, and
ditions, J. Am. Oil Chem. Soc. 76:1363–1370 (1999). W.F. Kwolek, Continuous Slurry Hydrogenation of Soybean Oil
2. Harrod, M., M.-B. Macher, J. Hogberg, and P. Moller, Hydro- with Copper–Chromite Catalyst at High Pressure, J. Am. Oil
genation of Lipids at Supercritical Conditions, in Proceedings Chem. Soc. 61:909–913 (1984).
of the 4th Italian Conference on Supercritical Fluids and Their 15. Hastert, R.C., Hydrogenation, in Proceedings of World Confer-
Applications, edited by E. Reverchon, Capri, Italy, 1997, pp. ence on Emerging Technologies in the Fats and Oils Industry,
319–326. edited by A.R. Baldwin, American Oil Chemists’ Society,
3. Harrod, M., and P. Moller, Hydrogenation of a Substrate and Champaign, 1986, pp. 1–29.
Products Manufactured According to the Process, International 16. Temelli, F., J.W. King, and G.R. List, Conversion of Oils to
Patent WO 96/01304 (1996). Monoglycerides by Glycerolysis in Supercritical Carbon Diox-
4. Tacke, T., S. Wieland, and P. Panster, Hardening of Fats and ide Media, J. Am. Oil Chem. Soc. 73:699–706 (1996).
Oils in Supercritical CO2, in High-Pressure Chemical Engineer- 17. House, S.D., Determination of Total, Saturated, and Monoun-
ing, edited by P.R. von Rohr and C. Trepp, Elsevier, Amster- saturated Fats in Foodstuffs by Hydrolytic Extraction and Gas
dam, 1996, pp. 17–21. Chromatographic Quantitation: Collaborative Study, J. Assoc.
5. Tacke, T., S. Wieland, and P. Panster, Selective and Complete Off. Anal. Chemists Int. 80:555–563 (1997).
Hardening of Edible Oils and Free Fatty Acids in Supercritical 18. Firestone, D. (ed.) Official and Recommended Methods of the
Fluids, in Proceedings of the 4th International Symposium on American Oil Chemists’ Society, Methods Cc-18-80, Cd-10-57,
Supercritical Fluid, Sendai, Japan, 1997, pp. 511–514. American Oil Chemists’ Society, Champaign, 1989.
6. Andersson, M., and J.W. King, Synthetic Modification of Oleo- 19. Susu, A.A., and A.F. Ogunye, Nickel-Catalyzed Hydogenation
chemicals in Critical Fluids, in Proceedings of the 6th Meeting of Soybean Oil: I. Kinetic, Equilibrium and Mass Transfer De-
on Supercritical Fluids, edited by M. Poliakoff, M.W. George, terminations, J. Am. Oil Chem. Soc. 58:657–661 (1981).
and S.M. Howdle, International Society for the Advancement of 20. Hasman, J.M., Trans Suppression in Hydrogenated Oils, inform
Supercritical Fluids, Nottingham, England, 1999, pp. 195–203. 6:1206–1213 (1995).
7. Ting, P.L., J.R. Refling, and H. Goldstein, Hydrogenation of 21. Hunter, J.E., Nutritional Aspects of Hydrogenated Fats, in Hy-
Hop Soft Resins Using CO2, U.S. Patent 6,020,019 (2000). drogenation:Proceedings of an AOCS Colloquium, edited by R.
8. Bertucco, A., P. Canu, L. Devetta, and A.G. Zwahlen, Catalytic Hastert, American Oil Chemists’ Society, Champaign, 1987, pp.
Hydrogenation in Supercritical CO2: Kinetic Measurements in a 227–246.
Gradientless Internal-Recycle Reactor, Ind. Eng. Chem. Res. 22. List, G.R., K.R. Steidley, and W.E. Neff, Formulation, Struc-
36:2626–2633 (1997). ture, and Properties of Commercial Spreads, a 1999 Survey, in
9. Albright, L.F., Partial Hydrogenation of Triglycerides, in Hy- 49th Oilseed Conference—Surviving in a Changing Global
drogenation; Proceedings of an AOCS Colloquium, edited by Economy, New Orleans, LA, 1999, pp. 1–7.
R. Hastert, American Oil Chemists’ Society, Champaign, 1987, 23. Patterson, H.B.W., Hydrogenation of Fats and Oils: Theory and
p. 12. Practice, American Oil Chemists’ Society, Champaign, 1994,
10. Albright, L.F., and J. Wisniak, Selectivity and Isomerization pp. 39–40.
During Partial Hydrogenation of Cottonseed Oil and Methyl 24. Tacke, T., S. Wieland, P. Panster, M. Blankmann, R. Brand, and
Oleate: Effect of Operating Variables, J. Am. Oil Chem. Soc. H. Magerlein, Hardening of Unsaturated Fatty Acids, Fatty
39:14–19 (1962). Acids, or Fatty Acid Esters, U.S. Patent 5,734,070 (1998).
11. Mounts, T.L., S. Koritala, J.P. Friedrich, and H.J. Dutton, Se- 25. Harrod, M., and P. Moller, Hydrogenation of Fats and Oils at
lective Hydrogenation of Soybean Oil. IX. Effect of Pressure in Supercritical Conditions, in High-Pressure Chemical Engineer-
Copper Catalysts, Ibid. 55:402–406 (1978). ing, edited by P.R. von Rohr and C. Trepp, Elsevier, Amster-
12. Koritala, S., J.P. Friedrich, and T.L. Mounts, Selective Hydro- dam, 1996, pp. 43–48.
genation of Soybean Oil. X. Ultra High-Pressure and Low-Pres-
sure, Ibid. 57:1–5 (1980).
13. Fillion, B., and B.I. Morsi, Gas–Liquid Mass Transfer and Hy- [Received August 3, 2000; accepted November 4, 2000]

JAOCS, Vol. 78, no. 2 (2001)

View publication stats