Liquid-phase catalytic oxidation of unsaturated fatty acids

Article abstract of DOI:10.1007/s11746-999-0236-7

Liquid-phase catalytic oxidation of oleic acid with hydrogen peroxide in the presence of various transition metal/metal oxide catalysts was studied in a batch autoclave reactor. Azelaic and pelargonic acids are the major reaction products. Tungsten and tantalum and their oxides in supported and unsupported forms were used as catalysts. Alumina pellets and Kieselguhr powder were used as supports for the catalysts. Tungsten, tantalum, molybdenum, zirconium, and niobium were also examined as catalysts. Tertiary butanol was used as solvent. Experimental results concluded that tungsten and tungstic oxide are more suitable catalysts in terms of their activity and selectivity. The rate of reaction observed in the case of supported catalysts appears to be comparable or superior to that of unsupported catalysts. In pure form, tungsten, tantalum, and molybdenum showed strong catalytic activity in the oxidation reaction; however, except for tantalum the other two were determined to be economically unfeasible. Zirconium and niobium showed very little catalytic activity. Based on the experimental observations, tungstic oxide supported on silica is the most suitable catalyst for the oxidation of oleic acid with 85% of the starting oleic acid converted to the oxidation products in 60 min of reaction with high selectivity for azelaic acid.

Full text of DOI:10.1007/s11746-999-0236-7

Liquid-Phase Catalytic Oxidation of
Unsaturated Fatty Acids
H. Noureddini* and M. Kanabur
Department of Chemical Engineering, University of Nebraska-Lincoln, Lincoln, Nebraska 68588-0126
ABSTRACT: Liquid-phase catalytic oxidation of oleic acid with several compounds forming a wide range of products. These
hydrogen peroxide in the presence of various transition reactions may involve either the carboxylic group or the car-
metal/metal oxide catalysts was studied in a batch autoclave re-
actor. Azelaic and pelargonic acids are the major reaction prod-
ucts. Tungsten and tantalum and their oxides in supported and
unsupported forms were used as catalysts. Alumina pellets and
Kieselguhr powder were used as supports for the catalysts.
Tungsten, tantalum, molybdenum, zirconium, and niobium
were also examined as catalysts. Tertiary butanol was used as
solvent. Experimental results concluded that tungsten and
tungstic oxide are more suitable catalysts in terms of their activ-
bon-carbon double bond in an unsaturated fatty acid. Reac-
tions such as hydrogenation, epoxidation, or oxidation may
occur at the carbon-carbon double bond. The oxidation of un-
saturated fatty acids results in partial or complete cleavage of
the carbon-carbon double bond. Mono- and dicarboxylic
acids are the main products of the complete oxidation of the
unsaturated fatty acids.
Dicarboxylic acids and their derivatives have many indus-
ity and selectivity. The rate of reaction observed in the case of trial applications (1–3). Diesters and linear polyesters of
supported catalysts appears to be comparable or superior to that straight-chain fatty acids are used as plasticizers for polyvinyl
of unsupported catalysts. In pure form, tungsten, tantalum, and chloride (3–5). Esters of dicarboxylic acids, with a specific
molybdenum showed strong catalytic activity in the oxidation
reaction; however, except for tantalum the other two were de-
termined to be economically unfeasible. Zirconium and nio-
bium showed very little catalytic activity. Based on the experi-
mental observations, tungstic oxide supported on silica is the
most suitable catalyst for the oxidation of oleic acid with 85%
of the starting oleic acid converted to the oxidation products in
amount of branching, are used as lubricants and hydraulic flu-
ids over a wide temperature range (3,6). Applications of di-
carboxylic acids are as polymer intermediates in polyamides,
polyesters, polyurethane coatings, and alkyd resins. For ex-
ample, nylon 1313 produced from brassylic acid, shows en-
hanced properties compared to those of commonly available
nylons, e.g., a lower melting point, lower density, and more
hydrophobicity than nylon-11 and nylon-12 (7), properties
advantageous in specific applications. Further, the reaction
60 min of reaction with high selectivity for azelaic acid.
Paper no. J8885 in JAOCS 76, 305–312 (March 1999).
KEY WORDS: Azelaic acid, catalytic oxidation, oleic acid, involved in the synthesis of nylon-1313 is much simpler com-
pelargonic acid, silica-supported catalyst, tungsten oxide, unsatu- pared to the reactions associated with nylon-11 and -12.
rated fatty acids.
The conventional process for the production of dicar-
boxylic acid from unsaturated fatty acids involves oxidative
ozonolysis of unsaturated fatty acids, a process used for man-
With the expected depletion of petroleum reservoirs in the ufacturing azelaic acid from oleic acid (8). Nieschlag and co-
next few decades and growing concern for the associated en- workers (9) utilized erucic acid in a continuous pilot-plant
vironmental pollution, much emphasis has been placed on the scale ozonolysis process for brassylic acid production. The
utilization of renewable resources such as agricultural mater- use of ozone as oxidant provides some advantages, namely,
ial as an alternative to the planetary resources. In this context, high reactivity, selectivity, and lack of by-products. However,
vegetable oils show a lot of promise as an alternative feed- owing to the inherent health and safety hazards, high capital
stock. Efforts are underway to utilize the abundantly avail- costs, handling of large volumes of reaction gases and high
able vegetable oils and oil-based products for the production energy requirements associated with the process, efforts are
of mono- and dibasic acids that in turn find major potential being made to find a safer and economically more viable
application as better substitutes for the petroleum-based feed- process for the production of diacids.
stock.
Oxidation of unsaturated fatty acids with a variety of oxi-
Oleic and erucic acids are the major monounsaturated fatty dizing agents such as potassium permanganate, potassium
acid components of soybean and crambe oils, respectively. dichromate, chromic acid, and sodium hypochlorite appears
Unsaturated fatty acids may undergo various reactions with in the literature. Garti and Avni (10) investigated the oxida-
tion of oleic acid with potassium permanganate. The oxida-
*
To whom correspondence should be addressed.
E-mail: hnoureddini@unl.edu
tion of oleic acid in neutral oil in water emulsion resulted in
Copyright © 1999 by AOCS Press
305
JAOCS, Vol. 76, no. 3 (1999)

306
H. NOUREDDINI AND M. KANABUR
the formation of dihydroxy, ketohydroxy, and diketo acids gradually increase the free-radical concentration until propa-
along with azelaic and pelargonic acids. Zaldman et al. (11) gation becomes a dominant process. Sometimes the poisons,
used sodium hypochlorite as an oxidizing agent for the oxi- which get thoroughly adsorbed, can also cause the induction
dation of oleic acid. The results showed that oxidation of oleic period, and are sufficienctly reactive enough to form less
acid in an emulsion system consisting of sodium hypochlo- thoroughly adsorbed products. The free-radical mechanisms
rite, surface-active agents, and ruthenium chloride as cata- can be ascertained by inhibitions of such processes by unsat-
lysts could be achieved with higher conversion and reduced urated monomers or other acceptors of free radicals or initia-
reaction time compared with the nonemulsified (heteroge- tion of such processes by azo compounds. Willms and co-
neous) systems. The reaction of alkaline potassium perman- workers (20) carried out the rate enhancement studies on
ganate with ricinoleic acid was investigated by Hill and m-xylene oxidation in the presence of hydrogen peroxide and
McEwen (12), resulting in about 34% yield of azelaic acid. phenol and found that the hydrogen peroxide decomposes in-
Ayorinde et al. (13) synthesized azelaic and suberic acids stantaneously under the operating conditions of 127–200°C
from oxidation of vernonia oil using nitric acid as oxidant. and 13 bar pressure to produce the free radicals necessary to
The isolated yields of azelaic and suberic acids were about 15 initiate the m-xylene oxidation, thereby almost eliminating
and 11%, respectively. The reaction by-products consisted of the induction period normally associated with the oxidation
many short- and long-chain mono- and dicarboxylic acids. process of m-xylene.
Further, Passi et al. (14) carried out studies by subjecting free
Liquid-phase oxidation of oleic acid with hydrogen perox-
or esterified polyunsaturated fatty acids to the oxidative at- ide in the presence of various transition metal/metal oxide
tack by hydroxyl radicals generated by Fenton reaction, radi- catalysts is discussed in this paper. The stoichiometric equa-
ation-induced oxidation, and enzymatic oxidation. They tion for this reaction is as follows:
demonstrated that, upon lipid peroxidation, polyunsaturated
fatty acids originate specific dicarboxylic acids that may be
regarded as a distinctive feature of the particular positions of
the double bonds in the molecules. A hypothetical scheme for
the formation of dicarboxylic acids as a result of the above-
mentioned reactions was presented.
Catalyst
H C(CH ) CH=CH(CH ) COOH + 4H O
2 2
3
2 7
2 7
H C(CH ) COOH + HOOC(CH ) COOH + 4H O
[1]
3
2 7
2 7
2
One of the objectives of this study was to investigate the para-
Liquid-phase oxidation of hydrocarbons to dibasic acids meters involved in the process scale-up; therefore, catalysts in
using heterogenized catalysts has received a lot of attention supported and unsupported forms were both examined.
in recent years. Catalysts such as transition metal oxides, Tungstic oxide in powder form, supported on Kieselguhr (sil-
cation-exchanged zeolites, and transition-metal complexes ica) powder and supported on alumina pellets, was tested. Sim-
with cross-linked polymer are suitable for the liquid-phase ilarly, a comparison study was made between tantalum oxide
hydrocarbon oxidation. Advantages such as high selectivity in powder form and in supported form on alumina. Tungsten,
associated with homogeneous catalysts and ease of recovery tantalum, molybdenum, zirconium, and niobium were also ex-
and regeneration in the case of heterogeneous catalysts can amined as catalysts. The reaction temperature was 130°C, and
be combined in the case of heterogenized homogeneous cata- the hydrogen peroxide concentration was at 100% excess.
lysts. Shen and Weng (15,16), reported liquid-phase oxida- These conditions were optimized in a previous work (21).
tion of cyclohexane and cyclohexanone with supported cobalt
(
15) and cerium (16) catalysts. Reaction mechanism, rate and
EXPERIMENTAL PROCEDURES
conditions, and product distribution are discussed.
Catalytic oxidation of olefins with hydrogen peroxide in Materials. Oleic acid (99%) and boron trifluoride-methanol
the presence of tungstic acid, molybdic acid, and their het- (14%) were obtained from Sigma Chemical Company (St.
eropoly-acids is discussed in a patent by Fujitani and Louis, MO). Oleic acid (90%) and tungsten oxide powder
Nakazawa (17). Nakano and Foglia (18) studied the oxidation (99.99%) were obtained from Aldrich Chemical Company
of unsaturated and hydroxy fatty acids by ruthenium tetrox- (Milwaukee, WI). Hydrogen peroxide (30%) and tertiary bu-
ide and ruthenium oxyanions. At high molar ratio of oxidant tanol (100%) were purchased from J.T. Baker (Phillipsburg,
to double bond (4:1), both catalysts quantitatively cleaved the NJ). The supports for the catalysts, viz. aluminum oxide (alu-
2
double bonds to give nonanoic and azelaic acid. Rushgen and mina) pellets in activated form of 3.2 mm, 90 m per g sur-
co-workers (19) studied the liquid-phase oxidative cleavage face area, 135 Å average pore diameter, Kieselguhr (silica)
2
of unsaturated fatty acids with hydrogen peroxide or peracetic powder of –325 mesh, 30 m per g surface area, 250 Å aver-
acid and transition-metal catalysts. The oxidation of oleic age pore diameter, tantalum oxide powder (99.99%) of pura-
acid with hydrogen peroxide and tungsten acid resulted in tronic grade, tantalum ethoxide (99.99%), and tantalum wire
nonanoic and azelaic aids as the main oxidation products.
(99.9%) of 0.25 mm diameter, were obtained from Alfa Aesar
Liquid-phase oxidation reactions generally proceed with a (Ward Hill, MA). Tungsten, tantalum, molybdenum, zirco-
free-radical mechanism (20). The reactions are generally nium, and niobium were obtained from Alfa Aesar. The am-
characterized by an induction period followed by a rapid re- monium metatungstate (99%) was obtained from Strem
action phase. During the induction period, initiation reactions Chemicals (Newburyport, MA).
JAOCS, Vol. 76, no. 3 (1999)

CATALYTIC OXIDATION OF UNSATURATED FATTY ACIDS
307
The standards for heptanoic methyl ester; hexanoic, oc-
Procedures. Experiments were conducted under a batch-
tanoic, pelargonic, succinic, glutaric, pimelic, suberic, aze- wise constant temperature and constant pressure mode. The
laic, stearic, oleic, and behenic acids were obtained from reaction feed consisted of 20 g of oleic acid, 60 mL of hydro-
Sigma Chemical Company, and standards for brassylic and gen peroxide of 30% concentration, and 150 mL of tertiary bu-
undecane diacids were purchased from Aldrich Chemical tanol. The reactor was purged with nitrogen prior to heating.
Company.
Timing was started as soon as the heating procedure began.
Equipment. A bench-top mini reactor (model number Most experiments were carried out for 5 h. In the case of sup-
4562; Parr Instrument Company, Moline, IL) was used for the ported catalysts, one batch of prepared catalysts, as described
oxidation reactions. The reactor assembly was constructed in the previous section, was charged along with the feed mix-
from type 316 stainless steel with a 450-mL bomb. The reac- ture. In the case of experiments with unsupported catalysts,
tor was equipped with a magnetic stirrer, a four-blade, down- about 1.5 g of tantalum oxide, 1.5 g of tungstic oxide, or 1.2 g
ward-thrust impeller, and a 1/12-hp variable speed motor with of pure metal in the form of a wire was used in the experi-
a pulley arrangement to turn the stirrer at speeds from 0 to ments. The reactor contents were agitated continuously with a
8
00 rpm. Heating mantel and internal cooling loop provided stirrer operating at about 500 rpm. Samples were taken peri-
the heating and cooling requirements. A Parr 4843 PID con- odically from the reaction mixture for analysis.
troller was used for controlling and monitoring the reaction
Sample preparation. Samples from oleic acid oxidation
temperature and the impeller speed. Desired temperatures were analyzed by gas chromatography and were expected to
were programmed into the PID controller, which enabled it to contain a variety of mono- and dicarboxylic acids. All sam-
control temperatures as needed. A solenoid valve controlled ples were converted to their respective methyl esters before
the flow of cooling water into the reactor. The temperature of analysis with Metcalfe and co-workers’ (22,23) derivatiza-
the bomb could be monitored to within ±1°C accuracy. A ni- tion procedures. A 14% boron trifluoride solution in methanol
trogen cylinder provided the purge gas for the process. The was used to derivatize fatty acids. To each sample, weighing
purge gas is introduced into the bottom of the reactor bomb. about 500 mg, 4 mL of boron trifluoride–methanol solution
All the preweighed reactants were manually charged into the was added. The mixture was heated to about 80°C for about
reactor.
15 min. Subsequently, the mixture was allowed to cool at
A “U” tube glass apparatus was used for the calcination of room conditions for about 20 min. The esterified product was
supported catalyst. A radiant heater purchased from Omega dissolved in 2 mL of petroleum ether. Then the sample was
Engineering Inc. (Stamford, CT) was used for heating the cal- extracted two times with 4 mL of distilled water. The aque-
cination tube. The rate of heating during calcination was con- ous phase was discarded, and the organic phase containing
trolled by a temperature controller supplied by Watlow (St. mainly esters of product acids was dehydrated by adding a
Louis, MO). The air flow rate required for calcination was small quantity of anhydrous sodium sulfate. Solvent was then
regulated with the help of a rotameter.
removed by passing a steady stream of dry air over the sam-
Catalyst preparation. Different types of unsupported and ple container. Finally, the product was dissolved in exactly 1
supported catalysts were used in this study. Unsupported cat- mL of anhydrous ether solvent and kept under 0°C prior to its
alysts used were tantalum, tungsten, molybdenum, zirconium, analysis.
and niobium in the form of metal wire, and tungstic and tan-
Analysis. Gas chromatography with flame-ionization de-
talum oxides. Supported catalysts used were tantalum oxide tectors has been widely used for the separation and quantifi-
on alumina and tungstic oxide on alumina and silica support. cation of methyl esters of fatty acids (24,25). A Hewlett-
Supported tantalum oxide catalyst was prepared by im- Packard HP 6890 series GC system equipped with HP IN-
pregnation of dried alumina pellets with tantalum ethoxide NOWax capillary column (30 m × 0.25 mm i.d. × 0.25 mm),
precursor dissolved in ethanol. Each batch of catalyst con- HP Series II GC Electronic Pressure Control (EPC) and split-
sisted of about 6 g of support and 3 g of tantalum ethoxide. splitless inlet (Palo Alto, CA) was used for the analysis of
The impregnation was carried out in a beaker. Then impreg- products. Helium was used as a carrier gas with a velocity of
nated alumina pellets were dried on a hotplate. The dried ma- 42 cm/s and a pressure of 24 psi with EPC at constant flow
terial was subjected to calcination at 450°C for about 3 h in mode (1.9 mL/min). A split ratio of 20:1 was maintained. The
the calcination tube. The temperature of the calcination cham- hydrogen flow and air flow rates were kept at 30 and 275
ber was increased at the rate of 5°C/min. The airflow rate was mL/min, respectively. The front inlet temperature was main-
maintained at about 7–8 mL/min through the calcination tained at 220°C. The oven temperature was maintained at
process.
50°C for 1 min, prior to injection and after injection. Then a
Impregnation of tungstic oxide on alumina was similar to the ramp rate of 10°C/min was used until oven temperature
above procedure except that 2 g of ammonium metatungstate reached 160°C. After a hold-up time of 1 min at 160°C, the
precursor was used in each batch, and distilled water was the oven temperature was further increased at the rate of 5°C/min
solvent. The impregnated support was then dried and calcined. to a final temperature of 260°C. The duration of sample run
Calcination temperature was 400°C for about 2 h. A similar was 35 min. An exact amount of 1 µL of sample was used
procedure was used to impregnate tungstic oxide on silica each time. The detector temperature was maintained at
support.
280°C. The quantification of the acids in the samples was car-
JAOCS, Vol. 76, no. 3 (1999)

308
H. NOUREDDINI AND M. KANABUR
ried out by the HP Chemstation software, which was provided concentration was reached in 1 h or less (see Fig. 2). All other
with the calibration runs. Behenic acid was the internal stan- cases showed a maximum azelaic acid concentration after
dard for the analysis.
about 2–3 h (see Figs. 1–6). In all cases studied, nonanoic
acid appears to be relatively more stable compared to azelaic
acid, which seems to be degrading faster with prolonged heat-
ing in the presence of catalyst.
RESULTS AND DISCUSSION
Catalytic oxidation of an unsaturated fatty acid with hydro-
Transition metal catalysts. In the presence of hydrogen
gen peroxide and transition-metal catalyst results in the for- peroxide, transition metals may potentially oxidize to their
mation of mono- and dicarboxylic acids (17–19). A typical respective oxides. This oxide is believed to be responsible for
profile of various products for the oxidation of oleic acid with catalyzing the oxidation of unsaturated fatty acids. From a
unsupported tungstic oxide catalyst is presented in Figure 1. scale-up perspective, this path could be advantageous if the
This profile summarizes the reaction kinetics after the first amount of metal oxidation is low and the consequential rate
hour of the reaction. As this figure indicates, azelaic and of formation of the products is comparable with the direct use
nonanoic acids are the main reaction products. Numerous of oxides. Tungsten, tantalum, molybdenum, zirconium, and
shorter-chain mono- and dicarboxylic acids of C to C length niobium in the form of a wire were examined in the initial
5
8
are also formed during the course of oxidation, indicating that screening. Molybdenum and tungsten were significantly oxi-
chain degradation also takes place during the course of oxi- dized (total oxidation for molybdenum and 60 wt% for tung-
dation. The reaction results for other cases studied are sum- sten) and, although the rate of oxidation of oleic acid was
marized in Figures 2–6. The formation of the chain degrada- comparable with direct oxidation with tungstic oxide, these
tion products consistently appeared in all cases under investi- catalysts were labeled as economically unfeasible. Zirconium
gation. For the purpose of simplicity, only nonanoic, azelaic, and niobium were very resistant to oxidation with hydrogen
octanoic, suberic, and oleic acids are shown in Figures 2–6. peroxide. The reaction products with these two catalysts were
As these figures indicate, the unsaturation bond in oleic acid similar to the oxidation in the absence of catalyst. Tantalum
is effectively cleaved, yielding mainly the corresponding was the only metal that showed a relatively slow oxidation in
mono- and dicarboxylic acids.
the presence of hydrogen peroxide (less than 1 wt%) and the
Reactivity. In general, during the course of reaction, the oxidation of oleic acid with tantalum was significant, when
concentration of azelaic and nonanoic acids increased to a compared to the direct use of tantalum oxide. The reaction re-
maximum and then started to decrease. The exact time during sults for the oxidation of oleic acid with tantalum and tanta-
which maximal concentration of azelaic and nonanoic acids lum wire are presented in Figure 6.
was reached depended on the type of catalyst used. In the case
Conversion. Tungstic oxide supported on silica showed
of tungstic oxide supported on silica, maximal azelaic acid the highest conversion (see Fig. 2). The concentration of oleic
acid in the product was about 13 mol% after1 h. This corre-
sponds to about 79% conversion of the initial oleic acid to the
FIG. 1. Typical profile of various products formed during the catalytic
oxidation of oleic acid with unsupported tungstic oxide: (◆), nonanoic
acid; (■), azelaic acid; (▲), octanoic acid; (+), succinic acid, (❋),
suberic acid, (■), stearic acid; (◆), heptanoic acid; (×), hexanoic acid;
FIG. 2. Reaction results for the oxidation of oleic acid, subject to silica-
supported tungstic oxide catalyst: (◆), nonanoic acid; (■), azelaic acid;
(▲), octanoic acid; (●), oleic acid; and (❋), suberic acid.
(●), sebacic acid; (●), oleic acid; and (▲), pimelic acid.
JAOCS, Vol. 76, no. 3 (1999)

CATALYTIC OXIDATION OF UNSATURATED FATTY ACIDS
309
FIG. 3. Reaction results for the oxidation of oleic acid, subject to alu-
mina-supported tungstic oxide catalyst: (◆), nonanoic acid; (■), azelaic
acid; (▲), octanoic acid; (●), oleic acid; and (❋), suberic acid.
FIG. 4. Reaction results for the oxidation of oleic acid, subject to alu-
mina-supported tantalum oxide catalyst: (◆), nonanoic acid; (■), aze-
laic acid; (▲), octanoic acid; (●), oleic acid; and (❋), suberic acid.
oxidation products. Conversion of oleic acid was 96% in 2 h
and about 98% after 3 h and longer. When unsupported degradation and low selectivity of catalyst for the oxidation
tungstic oxide was used the concentration of oleic acid in the of unsaturated bond. The unsupported tantalum oxide resulted
product was about 29% and the conversion of oleic acid was in a composition of oleic acid in the product of 11 and 2 mol%
at about 56% after 1 h (see Fig. 1), which was significantly after 1 and 3 h, respectively (see Fig. 5). The corresponding
lower when compared with supported tungstic oxide. How- conversion of oleic acid was about 80 and 96% after 1 and 4
ever, the performance of unsupported tungstic oxide was h, respectively. This appeared to be higher than conversion
comparable with the supported catalyst after 2 h of reaction. with tungstic oxide. However, the concentration of the desired
The composition of oleic acid was almost the same in both products, i.e., nonanoic and azelaic acids, were much lower
cases after 2 h of reaction time (see Figs. 1, 2).
for tantalum oxide.
Tungstic oxide supported on alumina showed a similar pat-
In the case of the oxidation of oleic acid with unsupported
tern to the silica-supported catalyst as shown in Figure 3. The metallic tantalum, the concentration of oleic acid in the prod-
rate of reaction for alumina-supported catalyst was slightly
lower than for silica supported catalyst. The smaller average
pore diameter of alumina (135 Å) compared to the silica (250
Å) may have been responsible for the lower rate of reaction
for the alumina-supported catalyst. Also alumina pellets tend
to settle down in the bottom of the reactor vessel and create
diffusion resistances which may result in slower reaction.
Using a stirred-contained solids reactor where the catalyst
basket is mounted around the paddle or agitator may increase
the reactivity of this catalyst. This ensures better contact of
reactants with the active sites of the catalyst.
The reaction results for supported and unsupported tanta-
lum oxide are presented in Figures 4 and 5, respectively.
When tantalum oxide supported on alumina was used, the
performance of the catalyst was relatively poor. Oleic acid
concentration in the product was about 18 mol%, which
amounts to about 70% conversion after about 3 h of reaction.
The maximal conversion that could be attained in 5 h was
about 77%. Moreover, the relative concentrations of nonanoic
and azelaic acids with respect to conversion of oleic acid were
lower than unsupported tantalum oxide and much lower than
FIG. 5. Reaction results for the oxidation of oleic acid, subject to un-
supported tantalum oxide catalyst: (◆), nonanoic acid; (■), azelaic acid;
supported tungstic oxide, indicative of a high rate of chain (▲), octanoic acid; (●), oleic acid; and (❋), suberic acid.
JAOCS, Vol. 76, no. 3 (1999)

310
H. NOUREDDINI AND M. KANABUR
uct was about 36 mole% after 1 h and 2% after 4 h as shown diacids early in the reaction media (1 h into the reaction) is
in Figure 6. The corresponding conversion of oleic acid was indicative of low selectivity of tantalum oxide for the oxida-
about 52 and 96% after 1 and 4 h, respectively. Formation of tion of unsaturation bond of the oleic acid (Fig. 5).
nonanoic and azelaic acids was comparable with tungstic
In all cases for tungstic oxide- and tantalum oxide-based
oxide between 2–5 h of reaction. However, the rate of oxida- catalysts, products were either about equally distributed in the
tion for this catalyst was relatively low during the first hour mole percentage of nonanoic and azelaic acids or were more
of the reaction. This may have been caused by the slow rate concentrated in nonanoic acid than azelaic acid except for un-
of oxidation of tantalum by hydrogen peroxide to tantalum supported tantalum that appeared to be more selective toward
oxide, which then catalyzes the oxidation of oleic acid.
the formation of azelaic acid (see Fig. 6). This may be due to
Selectivity. The selectivity of unsupported tungstic oxide the lower azelaic acid chain degradation rate in the presence
and tungstic oxide supported on silica and alumina were com- of transition metal catalysts.
parable and were marginally different in terms of types and
amounts of by-products formed during the reaction (Figs. grade 90 mole percentage oleic acid was used. The main im-
–3). Octanoic acid, which was detected as the most concen- purity in this acid was determined to be stearic acid, a C18
Effect of impurities. In the entire case studies, technical-
1
-
trated by-product, appeared at a relatively uniform concentra- saturated monocarboxylic acid. Stearic acid appeared to be
tion. In all cases, the concentration of pimelic acid was the very stable under the reaction conditions. Its composition re-
lowest at about 2 mol%. Although numerous mono- and di- mained unchanged throughout the reaction period. To study
carboxylic acids were also formed during the course of the the effect of impurities during the course of the reaction, oleic
reaction, tungstic oxide appeared to be very selective in the acid of 99% purity was used to establish the baseline for the
oxidation reaction, and for most of the systems the concentra- products. With pure oleic acid, the formation of the products
tion of the chain degradation products was under 5 mol%.
and their quantities were comparable with that of technical-
The selectivities of the unsupported tantalum oxide cata- grade oleic acid. Stearic acid was not detected in the analysis
lysts toward the formation of nonanoic and azelaic acids were of the reaction products with 99% oleic acid. Figure 7 shows
less compared to tungstic oxide catalysts (see Figs. 1 and 5). the profile of products formed from oxidation of pure and
The reactions with the unsupported tantalum oxide catalyst technical grade oleic acid with tungstic oxide catalyst sup-
appear to generate a greater quantity of the chain degradation ported on silica. This figure reveals that the impurities in tech-
products such as octanoic and suberic acids compared to the nical-grade oleic acid did not interfere with the reaction.
unsupported tungstic oxide catalysts. In the case of unsup-
Silica-supported tungsten oxide. One of the objectives of
ported tantalum oxide catalyst, the average concentration of this study was to develop a feasible process scale-up scheme.
octanoic and suberic acids formed was about 11 and 13 A factor of particular interest in the process scale-up is the re-
mol%, respectively (Fig. 5), as compared to about 10 and 4 action residence time. It was emphasized earlier that reactions
mol% for the unsupported tungstic oxide (Fig. 1). The rela- with tungstic oxide supported on silica resulted in the highest
tively high concentration of the lower carbon mono- and conversion after about 1 h. Examination of Figure 2 also re-
veals that the maximal amount of azelaic acid, the di-acid of
interest, was about 32–33 mol% only after 1 h of reaction.
Therefore, it appears that after the first hour of reaction, the
rate of oxidation of oleic acid is almost equal to the rate of
FIG. 7. Reaction results for the oxidation of oleic acid, subject to sil-
FIG. 6. Reaction results for the oxidation of oleic acid, subject to un-
ica-supported tungstic oxide catalyst: (◆), nonanoic acid; (■), azelaic
supported metallic tantalum oxide catalyst: (◆), nonanoic acid; (■), aze- acid; and (●), oleic acid. Solid and open symbols indicate 99 and 90%
laic acid; (▲), octanoic acid; (●), oleic acid; and (❋), suberic acid.
oleic acid, respectively.
JAOCS, Vol. 76, no. 3 (1999)

CATALYTIC OXIDATION OF UNSATURATED FATTY ACIDS
311
chain degradation. This explains the somewhat steady con-
centration of nonanoic and azelaic acids in most cases after
the second hour of the reaction (see Figs. 1–7). Therefore,
from a scale-up and design perspective, it appears most logi-
cal not to continue the reaction longer than 1 h.
The oxidation reaction was investigated further for silica-
supported tungstic oxide during the first hour of the reaction.
The rate of reaction was compared for silica-supported
tungstic oxide and unsupported tungstic oxide. The reaction
results for the formation of nonanoic and azelaic acids during
the first hour of the reaction are presented in Figures 8 and 9,
respectively. Data obtained for the uncatalyzed reaction are
also included in these figures. Initially, the unsupported cata-
lyst showed the highest activity. However, while the activity
of the unsupported catalyst stayed constant throughout the
whole 60 min of the reaction, the supported catalyst resulted
in a significant oxidation after the first 20 min of the reaction.
The low activity during the initial stage of the reaction with
the supported catalyst may be due to high pore diffusion re-
sistance that is caused by the support. Once the initial resis-
tance is overcome, the rate of reaction turns out to be much
faster than the case of unsupported catalyst. After 1 h of reac-
tion, the concentration of nonanoic and azelaic acids was 36
FIG. 9. Formation of azelaic acid in the oxidation of oleic acid: (■),
supported tungstic oxide on silica; (◆), unsupported tungstic oxide; and
(▲), without catalyst.
and 32 mole percentage, respectively, as compared to 29 and REFERENCES
3
0 mole percentage in the case of unsupported catalyst. In the
1
. Johnson, R.W., Dibasic Fatty Acids in Fatty Acids, in Industry,
edited by R.W. Johnson and E. Fritz, Marcel Dekker Inc., New
York, 1989, pp. 327–351.
. Budde, W.W., Jr., General Physical and Chemical Properties of
Fatty Acids, in Fatty Acids and Their Industrial Applications,
edited by S.E. Pattison, Marcel Dekker Inc., New York, 1968,
pp. 47–75.
case of uncatalyzed reaction, the concentration of nonanoic
and azelaic acids was 6 and 7 mol%, respectively, after 1 h of
the reaction.
2
ACKNOWLEDGMENTS
3
4
5
. Kadesch, R.G., Dibasic Acids, J. Am. Oil Chem. Soc. 31:568–573
1954).
. Kadesch, R.G., Fat-Based Dibasic Acids, Ibid. 56:845A–849A
1979).
. Nieschlag, H.J., W.H. Tallent, I.A. Wolff, W.E. Palm, and L.P.
Witnauer, Ester Plasticizers from Mixed Crambe Dibasic Acids,
Ind. Eng. Chem., Prod. Res. Dev. 6:201–204 (1967).
. Kenneth, D.C., V.E. Sohns, R.B. Perkins Jr., and E.L. Huffman,
Brassylic Acid: Chemical Intermediate from High-Erucic Oils,
Ibid. 16:95–100 (1977).
The authors express their gratitude to U.S. Department of Agricul-
ture Cooperative State Research Service and Nebraska Soybean
Board for their support of this work.
(
(
6
7. Nieschlag, H.J., J.A. Rothfus, V.E. Sohns, and R.B. Perkins Jr.,
Nylon-1313 from Brassylic Acid, Ibid. 16:101–107 (1977).
8. Goebel, C.G., A.C. Brown, H.F. Oehischlaeger, and R.P. Rolfer,
Method for Making Azelaic Acid, U.S. Patent 2,813,113 (1957).
9. Nieschlag, H.J., I.A. Wolff, T.C. Manley, and R.J. Holland,
Brassylic Acid from Ozonolysis of Erucic Acid, Ind. Eng. Prod.
Res. Dev. 6:120–123 (1967).
1
0. Garti, N., and E. Avni, Permanganate Oxidation of Oleic Acid
Using Emulsion Technology, J. Am. Oil Chem. Soc. 58:840–841
(
1981).
1
1
1
1. Zaldman, B., S.Y. Kesllev, and N. Garti, Double Bond Oxida-
tion of Unsaturated Fatty Acids, Ibid. 65:611–615 (1988).
2. Hill, J.W., and W.L. McEwen, Azelaic Acid, Org. Syn., Coll.
2
:5–55 (1943).
3. Ayorinde, F.O., G. Osman, R.L. Shephard, and F.T. Powers,
Synthesis of Azelaic Acid and Suberic Acid from Vernonia
galamensis Oil, J. Am. Oil. Chem. Soc. 65:1774–1777 (1988).
4. Passi, S., M. Picardo, C. De Luca, M. Nazzaro-Porro, L. Rossi,
and G. Rotillo, Saturated Dicarboxylic Acids as Products of Un-
saturated Fatty Acid Oxidation, Biochim. Biophys. Acta
1168:190–198 (1993).
1
FIG. 8. Formation of nonanoic acid in the oxidation of oleic acid: (■),
supported tungstic oxide on silica; (◆), unsupported tungstic oxide; and
(▲), without catalyst.
JAOCS, Vol. 76, no. 3 (1999)

312
H. NOUREDDINI AND M. KANABUR
1
5. Shen, H.C., and H.S. Weng, Liquid-Phase Oxidation of Cyclo- 22. Metcalfe, L.D., and A.A. Schmitz, The Rapid Preparation of
hexanone to Dibasic Acids with Immobilized Cobalt Catalyst,
Fatty Acid Esters for Gas Chromatographic Analysis, Anal.
Chem. 33:363–364 (1961).
Ind. Eng. Chem. Res. 27:2254–2260 (1988).
1
1
1
6. Shen, H.C., and H.S. Weng, Liquid-Phase Oxidation of Cyclo- 23. Metcalfe, L.D., A.A. Schmitz, and J.R. Pelka, The Rapid Prepa-
hexanone over Cerium Oxide Catalyst, Ibid. 29:713–719 (1990).
7. Fujitani, K., and M. Nakazawa, A Process for Preparation of
Carboxylic Acid, Japanese Patent 240,129 (1986).
ration of Fatty Acid Esters for Gas Chromatographic Analysis,
Anal. Chem. 38:514–515 (1966).
24. Bannon, C.D., J.D. Craske, N.T. Hai, N.L. Harper, and K.L.
O’Rourke, Analysis of Fatty Acid-Methyl Esters with High Ac-
curacy and Reliability, II. Methylation of Fats and Oils with
Boron Trifluoride-Methanol, J. Chromatogr. 247:63–69 (1982).
8. Nakano, Y., and T.A. Foglia, Oxidation of Unsaturated and Hy-
droxy Fatty Acids by Ruthenium Tetroxide and Ruthenium
Oxyanions, J. Am. Oil Chem. Soc. 59:163–166 (1982).
1
2
2
9. Rushgen, M., P. Klaas, and S. Warwel, Transition-Metal Cat- 25. Chang, I., and W. Sandera, Fast GC Analysis of Free Volatile
alyzed Cleavage of Unsaturated Fatty Acids, Fat Sci. Technol.
7:359–367 (1995).
0. Willms, R.S., D.D. Reible, D.M. Weitzel, and D.P. Harrison,
Aqueous-Phase Oxidation: Rate Enhancement Studies, Ind.
Eng. Chem. Res. 26:606-612 (1987).
Organic Acids, Fatty Acids, and Their Isomers Using HP IN-
NOWax Columns and EPC, Hewlett-Packard Company publi-
cation, Copyright 1993.
9
1. Pendyala, K.S., Catalytic Oxidation of Unsaturated Fatty Acids,
M.S. Thesis, University of Nebraska-Lincoln, December 1995.
[Received May 26, 1998; accepted November 20, 1998]
JAOCS, Vol. 76, no. 3 (1999)