Laboratory-scale continuous hydrogenation: Effect of pressure

A laboratory-scale, high-pressure, continuous reactor was used to partially hydrogenate soybean oil with copper catalysts. Effects of pressure on the kinetics of the reaction were studied by conducting experiments in a central composite design. The interaction of pressure (75\s-200 psig) with the other independent variables of temperature (155\s-255 C) and copper concentration (0.15\s-1.85%) was evaluated. Dependent variables studied were linolenate selectivity and formation of trans isomers and conjugated dienes. in addition, effects of pressure up to 500 psig, use of experimental and commercial copper catalysts and comparison of continuous with high-pressure batch rections were investigated. Linolenate selectivity (8\s-10) and trans-isomer formation were not significantly affected by any of the independent variables. Conjugated dienes were eliminated as products of the reaction when pressure was above 200 psig. Experimental copper-silica catalyst gave a 1.6-fold increase in reaction rate over commercial copper catalysts. Presented at ISF-AOCS meeting, New York, April 1980.     

Catalytic behavior of palladium in the hydrogenation of edible oils

Palladium supported on alumina was used to hydrogenate soybean and canola oil. Previous literature reports indicated that palladium forms moretrans isomers than nickel. At 750 psig, 50 ppm palladium, and at 70 C, only 9.4%trans were formed when canola oil was hydrogenated to IV 74. In general, high pressure and low temperature favored lowtrans formation with no appreciable loss in catalyst activity. The effect of pressure, temperature and catalyst concentration on reaction rate,trans formation and selectivity is presented. A survey of various catalyst supports is discussed. Apparent activation energies of 6.3 to 8.9 kcal/mol were obtained; they are in good agreement with values reported in the literature.     

Continuous ultrasonic hydrogenation of soybean oil. II. Operating conditions and oil quality

In previous work we found that ultrasonic energy greatly enhanced the rate of hydrogenation of soybean oil. We have now investigated parameters of ultrasonic hydrogenation and the quality of the resulting products. Refined and bleached soybean oil was hydrogenated continuously with and without ultrasonic energy at different temperatures, pressures and catalyst concentrations. Flavor and oxidative stability of the oils were compared with a commercially hydrogenated soybean oil. The extent of hydrogenation (ΔIV) was not affected by temperature between 245 and 290 C, but was greater at 106 psig than at 65 psig hydrogen pressure. The ΔIV of hydrogenated oils increased linearly with catalyst concentration from 40 ppm to 150 ppm nickel. At the same catalyst concentration the IV drop was significantly increased when ultrasonic energy was used. By reducing the amount of power supplied to the ultrasonic reactor to 40% of full power, the specific power (watts/ΔIV) was lowered by 60%. Linolenate selectivities and specific isomerization (%trans/ΔIV) remained the same, but linoleate selectivities were lower than for batch hydrogenation under varied operating parameters. Flavor scores were not significantly different, initially or after storage eight days at 60 C, for oils continuously hydrogenated with and without ultrasonic energy. Hydrogenation of soybean oil with ultrasonic energy offers a method to produce good quality products at potentially lower cost than present methods.     

Selective hydrogenation of soybean oil: X. Ultra high pressure and low pressure1

Soybean oil was partially hydrogenated with copper-chromite catalyst at 170 C and up to 30,000 psig hydrogen pressure. Catalyst activity increased with increase in pressure up to 15,000 psig. The linolenate selectivity (S_Ln) of the reaction remained essentially unchanged over 50–1000 psig pressure range. A S_Ln of 5.5 to 5.6 was achieved at 15,000 to 30,000 psig pressure range. This value is somewhat lower than the selectivity at 50–1000 psig, but much higher than that obtained with nickel catalysts. Geometric isomerization increased as pressure increased up to 200 psig; above this pressure, the percenttrans remained the same up to 500 psig.trans Isomer content decreased when the pressure was increased to 30,000 psig. cis,trans Isomerization of linoleate was greater at 1000 psig and 15,000 psig than at 50 psig. At 15,000 psig, part of the linoleate in soybean oil was hydrogenated directly without prior conjugation, whereas at low pressures, all of the double bonds first conjugate prior to hydrogenation. This difference in mechanism might explain the lower selectivities obtained at high pressures. Conjugated diene isomers were found in the products up to 200 psig. Above this pressure conjugated diene was not measurable. No significant differences were found in the double bond distribution oftrans monoenes even though the amount oftrans monoene formed decreased as pressure was increased to 30,000 psig. 1 Presented at the AOCS meeting, San Francisco, May 1979.     

Empirical modeling of soybean oil hydrogenation

Empirical hydrogenation models were generated from statistically designed laboratory experiments. These models, consisting of a set of polynomial equations, relate the operating variables of soybean oil hydrogenation to properties of the reaction and of the fat produced. These properties include reaction rate,trans-isomer content and melting point. Operating variables included in the models were temperature, hydrogen pressure, catalyst concentration, agitation rate and iodine value. The effects of catalyst concentration and agitation rate were found to be significant in determiningtrans-isomer content, which in turn influences the melting characteristics of the hydrogenated oil. Pressures above 30 psig were found to have little effect ontrans-isomer content, although pressure was very important in determining reaction rate. Reaction temperature was observed as the most important factor in determining thetrans-isomer content for a given iodine value. Generally, 50 to 60%trans isomer content is predicted by the model for the iodine value range and operating conditions used in this study. Thus, these predictive models can assist in scaling up hydrogenation processes and in determining the optimum operating parameters for running commercial hydrogenation. Presented at the AOCS Meeting, Chicago, May 1983.     

Fixed- Bed continuous hydrogenation of soybean oil with palladium- Polymer supported catalysts

To compare a continuous hydrogenation system with batch hydrogenation, soybean oil was treated with Pd and Ni catalysts in a fixed-bed system under conditions that gave trickle flow. The influence of processing variables such as space velocity, pressure, temperature and hydrogen flow on the selectivity, specific isomerization and the activity was investigated. Both the Pd and Ni catalysts gave significantly lower specific isomerization(trans isomer per drop in Iodine Value) when compared to reported values for batch hydrogenation with similar type catalysts. The linolenate and linoleate selectivities were also significantly lower. Heterogenized homogeneous Pd-on-polystyrene catalyst gave lower specific isomerization formation and higher selectivity than carbon-supported Pd catalyst at same conditions. This work indicates that Pd-on-styrene, Pd-on-carbon and extruded Ni catalysts, in fixed-bed continuous hydrogenation can produce soybean oil of desirable composition after further optimization.     

The electrochemical hydrogenation of edible oils in a solid polymer electrolyte reactor. II. Hydrogenation selectivity studies

Soybean oil has been hydrogenated electrochemically in a solid polymer electrolyte (SPE) reactor at 60°C and 1 atm pressure. These experiments focused on identifying cathode designs and reactor operation conditions that improved fatty acid hydrogenation selectivities. Increasing oil mass transfer into and out of the Pd-black cathode catalyst layer (by increasing the porosity of the cathode carbon paper/cloth backing material, increasing the oil feed flow rate, and inserting a turbulence promoter into the oil feed flow channel) decreased the concentrations of stearic acid and linolenic acid in oil products [for example, an iodine value (IV) 98 oil contained 12.2% C_18:0 and 2.3% C_18:3]. When a second metal (Ni, Cd, Zn, Pb, Cr, Fe, Ag, Cu, or Co) was electrodeposited on a Pd-black powder cathode, substantial increases in the linolenate, linoleate, and oleate selectivities were observed. For example, a Pd/Co cathode was used to synthesize an IV 113 soybean oil with 5.3% stearic acid and 2.3% linolenic acid. The trans isomer content of soybean oil products was in the range of 6–9.5% (corresponding to specific isomerization indices of 0.15–0.40, depending on the product IV) and did not increase significantly for high fatty acid hydrogenation selectivity conditions.     

Hydrogenation of canola oil in the presence of nickel and the methyl benzoate-chrome carbonyl complex

Canola oil was hydrogenated using a mixture of homogeneous methyl benzoate-Cr(CO)₃ and heterogeneous nickel catalysts. The effect of the methyl benzoate-Cr(CO)₃_to-nickel ratio on the activity, specific isomerization index, linoleate and linolenate selectivities, and fatty acid composition was evaluated, and the results compared with those obtained with commercial nickel catalyst and methyl benzoate-Cr(CO)₃ used individually. At higher chromium-to-nickel ratios the activity of nickel was inhibited and the system behaved essentially like the pure chrome complex, while at low chromium-to-nickel ratios the characteristics of the nickel predominated. In a short transition zone relatively high reaction rates were obtained with significantly reducedtrans-isomer levels in the product. In a broader sense, it may be possible to combine a homogeneous and heterogeneous catalyst while retaining the advantages of both. We may thus be able to design catalyst systems for specific applications.     

Homogeneous catalytic hydrogenation of canola oil using a ruthenium catalyst

Dichlorodicarbonylbis (triphenylphosphine) ruthenium (II), RuCl₂ (CO)₂ (PPh₃)₂, was investigated as a catalyst for edible oil hydrogenation in a preliminary screening of potential catalysts for producing partially hydrogenated fats with lowtrans-isomer content. Refined, bleached and deodorized canola oil was hydrogenated using 1.77 × 10⁻⁵ − 6.64 × 10⁻⁴ mol/kg-oil of ruthenium catalyst equivalent to 1.79 × 10⁻⁴ − 6.71 × 10⁻³ wt% Ru. The effects of temperature (50–180 C) and pressure (50–750 psig) on reaction rate,trans-isomer content and fatty acid composition were examined. The activities of RuCl₂ (CO)₂ (PPh₃)₂ and nickel (Nysel HK-4 and AOCS standard nickel catalyst) were compared on a molar basis. At 4.40 × 10⁻⁴ mol/kg-oil (0.0026 wt/Ni or 0.0044 wt% Ru), 140 C and 50 psig, the nickel catalysts were completely inactive, but the ruthenium catalyst produced an IV drop of 40 units in 60 min. At 110 C, 750 psig and 1.34 × 10⁻⁴ mol/kg-oil (1.35 × 10⁻³ wt% Ru), a hydrogenation rate of 0.89 ΔIV/min and a maximumtrans-isomer content of 10.4% (IV=45.0) was obtained with the ruthenium catalyst.     

Catalysts for selective hydrogenation of soybean oil.1 II. Commercial catalysts

A survey of commercial hydrogenation catalysts demonstrated the higher selectivity (S_L= 2.4\s-2.7) of certain platinum, palladium and rhodium catalysts for hydrogenating linolenic components in soybean oil. Nickel catalysts generally showed selectivities below S_L=2.0 although skeletal nickel achieved higher values.Trans-isomers were in the range 7.8\s-15.4% for the above noble metal catalysts. Nickel catalysts provide a lesser degree of isomerization, 5.2\s-7.4% oftrans-isomers for the most selective catalysts. Presented at the AOCS Meeting at Toronto, 1962.     

Deuteration of methyl linoleate with nickel,palladium, platinum and copper-chromite catalysts

Samples taken during deuteration of methyl linoleate with the title catalysts were separated into saturate, monoene and diene fractions. Monoenes were further separated intocis andtrans fractions. A comparison of the double bond distribution in monoenes with those from hydrogenation of alkaliconjugated linoleate indicated that up to 59% of the linoleate was reduced through a conjugated intermediate with nickel catalyst. The respective percentages for palladium and platinum catalysts were 51 and 23. Copper catalysts have previously been shown to reduce linoleate solely through conjugated intermediates. Copper-chromite catalyst showed infinite selectivity for the reduction of linoleate, because stearate did not form. The decreasing order of various catalysts for the selective reduction was copper-chromite>>>Ni at 195 C>Pd>Ni at 100 C>Pt. Computer simulation of platinum reduction indicated that ca. 20% of the linoleate was directly reduced to stearate through a shunt. Geometrical isomers of linoleate were formed during reduction with all catalysts except copper-chromite. Nickel catalyst formed bothtrans,trans- andcis,trans-isomers, as well as nonconjugatable dienes. These isomers were favored at the higher temperature and deuterium was incorporated into them. Palladium and platinum did not isomerize linoleate to nonconjugatable dienes. Because conjugated dienes are more reactive than linoleate, they were not found in appreciable amounts during reduction. Conjugated dienes were the only isomers formed with copper-chromite catalyst. Deuterium was found in these conjugated dienes, which were also extensively isomerized. As a result of isomerization and exchange during reduction of linoleate-as well as further exchange between deuterium and monoenes-a wide distribution of isotopic isomers in monoenes was found with nickel, palladium and platinum catalysts. Since isomerization of monoenes with copper-chromite is negligible, the isotopic distribution of monoenes must be due to exchange of intermediate conjugated dienes followed by addition. Presented at the AOCS Meeting, Ottawa, September 1972. ARS, USDA.     

Behavior of hydrogenation catalysts. I. Hydrogenation of soybean oil with palladium

A statistical method for evaluation of catalysts was used to determine the behavior of palladium catalyst for soybean oil hydrogenation. Empirical models were developed that predict the rate,trans-isomer formation, and selectivity over a range of practical reaction conditions. Two target iodine value (IV) ranges were studied: one range for a liquid salad oil and the other for a margarine basestock. Although palladium has very high activity, it offered no special advantage intrans-isomer formation or selectivity. Palladium can substitute for nickel catalyst, at greatly reduced temperature and catalyst concentrations, for production of salad oil or margarine basestock from soybean oil. Presented at the AOCS meeting, Chicago, May 1983.     

Functional gel-type resin based palladium catalysts: The role of polymer properties in the hydrogenation of the CC bond of maleic and fumaric acids, the isomers of dicarboxylic acids

Functional gel-type resins (OFP) composed of 2-hydroxyethyl methacrylate (HEMA, 20 mol%), styrene (60–77 mol%) and diethylene glycol dimethacrylate, the crosslinking reagent (DEGDMA, 3–20 mol%) are used as supports for palladium catalysts (0.25–2 wt% Pd). The influence of polymer mass expansion on the activity of Pd/OFP catalysts in the hydrogenation of cis- and trans-isomers of unsaturated dicarboxylic acids, maleic (MA) and fumaric (FA), is examined. Starting polymer swells very well in THF and this solvent is used in the hydrogenation experiments. Swelling ability of polymer decreases after insertion of palladium ions due to the crosslinking effect. Reduction of Pd/OFP by NaBH₄ leads to further decrease in swelling ability of polymer caused by the changes in polymer structure observed in FTIR spectroscopy. In the reduced form of the catalysts the nano-particles of Pd (2–3 nm) are formed. Hydrogenation experiments performed in a wide range of operating conditions (Pd loading, catalysts concentration, crosslinking degree of polymer, the size of catalysts grains) show that the swelling of polymer mass during the hydrogenation is a crucial factor for activity of Pd/OFP catalysts. When the catalyst does not swell, the rates of hydrogenation for both isomers are practically the same. They differ only when the catalysts grains are in swollen state. Much higher rate of hydrogenation is observed for MA (the cis-isomer) than for FA (the trans-isomer). This is related to easier penetration of cis-isomer (MA) inside the bulk of the polymer mass which results in more effective utilization of substrates, MA and FA, when present in THF both enhance the expansion of polymer mass but especially cis-isomer, MA. Substrate effect is attributed to the interaction between functional groups present in the reduced catalysts and the hydrogenated reagents, maleic or fumaric acids.     

Homogeneous catalytic hydrogenation of unsaturated fats: Cobalt carbonyl

Homogeneous hydrogenation of unsaturated fats by cobalt carbonyl has been compared with the previously reported catalysis by iron carbonyl. Soybean methyl esters, methyl linoleate and linolenate have been hydrogenated at 75–180C, 250–3,000 psi H₂ and 0.02 molar concn of catalyst. The cobalt carbonyl catalyst is more active at lower temp than iron carbonyl. The partially reduced products are similar to those observed with iron carbonyl, but the reaction differs in showing much less accumulation of conjugated dienes, no selectivity toward linolenate, almost complete absence of monoene hydrogenation to saturates, less double bond migration and moretrans isomerization. No evidence was found for a stable complex between cobalt carbonyl and unsaturated fats as previously observed with iron carbonyl. The rates of hydrogenation/double bond were the same for linoleate and linolenate on one hand, and for alkali-conjugated linoleate and nonconjugated linoleate on the other. Presented at AOCS Meeting in Minneapolis, 1963. A laboratory of the No. Utiliz. Res. & Dev. Div., ARS, USDA.     

Hydrogenation of fatty oils with palladium catalysts. V. Products of the tall oil industry

Conditions were found for reducing tall oil distillate to an iodine number of 22 with a sufficiently small amount of palladium catalyst to make the process commereially feasible. The operating conditions were 200°C and 2,600 psi. Tall oil fatty acids were reduced with palladium and the concentration of linoleic acid,cis-oleic acid, saturated acid, andtrans isomers were determined as a function of iodine number. The five-platinum group metals (Pt, Pd, Ir, Rh, Ru) were compared as to activity, selectivity of partial hydrogenation, and tendeney to formtrans-isomers.     

Hydrogenation of methyl oleate in solvents

The hydrogenation of the oleic acid group was investigated with the objective of determining the effect of solvents on the reaction rate and the formation of positional and geometrical isomers. Methyl oleate, either alone or dissolved in one of several solvents, hexane, ethanol,n-butyl ether, or acetic acid, was hydrogenated to an iodine value of about 50 under atmospheric pressure and at 30°C, with palladium-on-carbon and the W-5 form of Raney nickel as catalysts. Hydrogenation with palladium catalyst, with or without solvents, produced 76.6 to 79.1%trans bonds, based on the total double bonds. This is significantly greater than the 67% obtained previously. Hydrogenation products obtained with Raney nickel and solvents contained as little as 20.7%trans bonds at an iodine value of about 50. In two cases thetrans bonds were equal to about one-third the positional isomers. Positional isomers formed extensively when the Raney nickel was used in the absence of solvents and when the palladium catalyst was used. When the Raney nickel and solvents were used large proportions of double bonds were found in the original 9-position. Presented at the 51st Annual Meeting of the American Oil Chemists’ Society, Dallas, Tex., April 4–6, 1960. One of the laboratories of the Southern Utilization Research and Development Division, Agricultural Research Service, U. S. Department of Agriculture.     

Isomerization during hydrogenation of soap: II. Potassium elaidate and potassium linoleate

Potassium elaidate in slightly alkaline solution was hydrogenated for up to 7 hr with 1.5% of Rufert nickel catalyst at 150 C and 20 kg/sq cm pressure. Potassium linoleate was similarly hydrogenated with 1.0% catalyst for 7 hr, and the hydrogenation continued for another 7 hr after addition of 0.5% fresh catalyst. Periodic samples from each were analyzed for component acids. The positional isomers in thecis andtrans monoenes, isolated by preparative argentation thin layer (TLC) or column chromatography, were estimated after oxidation to dicarboxylic acids. Some diene fractions were isolated for further examination. In potassium elaidate hydrogenation,cis monoenes were initially produced in considerable amounts, but to a lesser extent thereafter. Positional isomers were similarly distributed in bothcis andtrans monoenes after prolonged hydrogenation. In the hydrogenation of potassium linoleate, a drop in iodine value (IV) of 60 units occurred in the first hour, and 38% oftrans monoenes (in which the 10- and 11-monoenes constitute 32% each) were formed. The IV then fell only slowly, and up to 38% ofcis monoene (mostly 9- and 12-isomers) was formed. Addition of fresh catalyst caused a major shift ofcis monoenes totrans forms. The diene fraction was mostly nonconjugated material with the first double bond at the 9, 8 and 10-positions. Minor amounts of conjugated dienes were present as well as a dimeric product.     

Hydrogenation of alkali-conjugated linoleate: Isomeric monoene profiles obtained with nickel,palladium and platinum catalysts

Alkali-conjugated linoleate (cis-9,trans-11- andtrans-10,cis-12-octadecadienoate) was hydrogenated with nickel, palladium and platinum catalysts. Thetrans andcis monoenes formed during reduction were isolated, and their double bond distribution was determined by reductive ozonolysis and gas liquid chromatography. About 44–69% of the monoenes were composed of δ¹⁰ and δ¹¹ trans monoene isomers, whereas the δ⁹ and δ¹² cis monoenes amounted to 20–26%. With nickel catalyst, composition of monoene isomers remained the same, even when the hydrogenation temperature was increased. The monoene isomer profiles between nickel and palladium catalysts were indistinguishable. Isomerization of monoenes with platinum catalyst was suppressed at 80 psi. The position of the double bonds in unreacted conjugated diene was always retained, except with nickel at both temperatures and with platinum at 150 C when a slight migration occurred. Geometrical isomerization totrans,trans-conjugated diene was observed in the unreacted diene with nickel (ca. 15% of diene) at both 100 C and 195 C, and with platinum (ca. 7% of diene) at 150 C. ARS, USDA.     

Lipids in margarines and margarine-like foods

The lipid composition of margarines from stores in selected locations in the U.S. is reported. The lipids determined include the fatty acids, tocopherols and major plant sterols. Data are included for isomeric octadecenoic fatty acids (14 isomers or groups of isomers) and four isomeric octadecadienoic fatty acids common in partially hydrogenated vegetable fats, insofar as these are separable by capillary gas chromatography. Amounts of individual lipids found in vegetable oil margarines, spreads, imitation and diet margarines were: palmitate, 8.49 to 13.17% (normalized weight percent, calculated as triglyceride); stearate, 4.78 to 9.53%; linoleate, 6.06 to 46.39%; linolenate, 0.18 to 3.57%; α-tocopherol, 0.3 to 24.3 mg/100g; γ-tocopherol, 3.0 to 55.0 mg/100g; δ-tocopherol, 0.5 to 18.9 mg/100g; campesterol, 10.6 to 106.3 mg/100g; stigmasterol, 13.1 to 60.9 mg/100g; sitosterol, 42.3 to 412.9 mg/100g. Amounts oftrans-unsaturated octadecenoic fatty acids in these margarines varied from 10.74 to 30.06%. Small amounts of thetrans,trans, trans,cis andcis,trans isomers of linoleate also are reported.     

Hydrogenation and deuteration of conjugated isomers of linoleate and linolenate with palladium,platinum, nickel and lindlar catalysts

Conjugated isomers of methyl linoleate and linolenate were reduced with palladium, platinum, nickel and Lindlar catalysts at atmospheric hydrogen or deuterium pressure. After the products were separated, positions of their double bonds were determined by ozonolysis. Palladium and platinum catalysts reduced β-eleostearate directly to monoene. Nickel reduced β-eleostearate to dienes chiefly by 1,2-addition and to a lesser extent by 1,4- and 1,6-addition, whereas Lindlar catalyst reduced by 1,2-and 1,6-addition only. All catalysts reduced conjugated linoleate isomers by both 1,2- and 1,4-addition, with nickel being somewhat preferential for 1,2-addition. Selectivity for the catalytic reduction of dienes to monoenes decreased in the order: nickel, palladium and platinum. Lindlar catalyst did not isomerize or reduce monoenes that formed during reduction. Palladium and platinum did not isomerize conjugated dienes and trienes during their reduction, whereas nickel and Lindlar catalysts isomerized them slightly. Some deuterium was found in unreacted conjugated diene and triene with nickel and Lindlar catalysts, but none with palladium or platinum. Deuterated products contained a wide range of isotopic isomers with some products having up to 31 deuterium atoms. This wide deuterium distribution resulted from (a) exchange followed by addition, (b) addition followed by exchange and (c) exchange-addition-exchange reactions. Presented at the AOCS Meeting, Atlantic City, October 1971. ARS, USDA.