TECHNICAL REVIEW Sulfurization of Jojoba Oil
Dr. Jaime Wisniak is professor a t the Depa. ment of Chemical En, neering, Ben Gurion Ui uersity of the Negeu, Eel Sheua, Israel. He reeeiu his E S c . degree from Un ersidad Catolica, Chile, 1957 and a Ph.D. degree pom Purdue University zn 1960. His research interests are in the areas of fats and oils, artificial sweeteners, and thermodynamics.
Introduction This work was undertaken to determine the characteristics of the sulfurization of jojoha oil (Simmondsia oil), a potential substitute of sperm whale oil. In an effort t o protect the world’s dwindling population of sperm whale and seven other whale species, in late 1970 the U.S. Government included them in the list of Endangered Species and banned imports of oil, meat, and other products derived from whales. At that time, sperm oil consumption in the United States was about 40-50 million pounds per year, with half that figure utilized in lubricant applications. The unique liquid wax produced by sperm whales was of importance in various lubricant applications such as automotive transmission fluids, metal working oils, industrial and automotive gear luhricants, and tractor hydraulic fluids (Gisser e t al., 1973). Sulfurized sperm oil was used in many lubricants hecause i t had a comhination of properties not matched by other available additives: solubility in high viscosity paraffinic oils, low tendency to form sludge on oxidation, good antiwear, friction, and extreme pressure (EP) properties, as well as compatibility with other additives such as lead naphthenate. I t was also available in large quantities and a t low cost, a fact that eliminated significant research over the years on synthetic replacements (Peeler and Hartmann, 1972). The ban on the use of sperm oil prompted an intensive search for substitutes. The many approaches tried are summarized: (1) development of a totally synthetic mixture similar to that of sperm oil, and (2) development of a replacement having different chemical composition hut similar cost-performance relationship for sulfurized, sulfurchlorinated, and chlorinated extreme pressure additives. Sperm oil differs in chemical structure from most other fatty oils in that it is largely composed of monoesters derived from C I and ~ CIS alcohols, 60% unsaturated, and CM, C I ~ Cln, , and CZo carboxylic acids, 75% unsaturated. The mechanical properties of the oil are attrihuted to its monoolefinic structure. No single natural or synthetic replacement with the unique qualities of sperm oil has yet been found but several publications have shown that the oil extracted from jojoha (Simmondsia chinensis) may prove to he an excellent substitute if i t becomes available in commercial quantities (Gisser et al., 1973). A recent publication of the Committee on Jojoba Utilization of the National Academy of Sciences (1975) has concluded that among other things jojoha oil can duplicate sperm oil performance and be used as a substitute for the complete range of its uses, without requiring major reformulations. Jojoba possesses several advantageous characteristics over sperm oil: (1) it has no fishy odor; (2) the crude oil Ind. Eng. chem., Prod. Res. Dev., Vol. 14. No. 4, 1975 241
contains no stearins and requires little or no treatment for most industrial purposes; (3) it can take larger amounts of sulfur; (4) it does not darken on sulfurization; and (5) the highly sulfurized oil is liquid, whereas sperm oil when highly sulfurized requires addition of mineral oil in order to remain liquid (Sherbrooke and Haase, 1974). Jojoba is an evergreen shrub of the Buxaceae family that grows in semidesertic areas and yields a nut that contains about 50% of an oil composed of monoesters of the CZOand C22 alcohols and acids, with two double bonds. Details of the chemical and physical constants have been compiled and reviewed (Miwa, 1973; Wisniak and Liberman, 1975) and the report of the Committee of Jojoba Utilization has summarized the total potential of the shrub and its byproducts. Sulfurization processes may be classified from different points of view. The reaction may be carried out with S, sulfur monochloride, and other reagents containing S and chlorine. It can also be carried in the presence of solvents and accelerators. Another possible classification is according to the material being sulfurized: fatty acids, olefins, mono-, di-, triesters, etc., where the common element is the presence of at least one carbon-carbon double bond. The process may also be classified according to the final product. At low sulfur content the material is usually liquid, while at high S contents it is a rubbery solid called factice. S reacts with olefins at 90-160°C in the liquid phase to form several types of polysulfide products. The mechanism of the reaction has not been determined and conflicting theories have been published. Research by Farmer and Shipley (1947) on the reaction of olefins such as cyclohexene and isobutylene at 100-140°C established that the predominant products were alkyl alkenyl polysulfides, that variation in the amount of S used affected the ratio but not the nature of the products obtained, and that no appreciable amounts of thiols or hydrogen sulfide were produced initially, but substantial amounts were formed in secondary reactions if the initial products were heated to 160°C or higher. The results of Farmer and Shipley were reexamined by Bateman et al. (1958), who found that the main product was always a mixture of polysulfides, plus small amounts of thioepoxides. Ross (1961) followed the reaction between S and cyclohexene dilatometrically and concluded that it was autocatalytic. In the early stages the rate was proportional to the square root of S, olefin, and product concentrations. The kinetic and product data were consistent with a polar, rather than a free-radical chain reaction. Bateman and coworkers (1958) and Ross (1961) examined the possibility of an ionic mechanism by addition of substances that characterize this type of reaction. The addition of benzoquinone and iodine did not affect the rate of reaction while acids and bases accelerated it. Ross also examined the influence of solvents on the rate of reaction and found that the rate of reaction of S with cyclohexene increased by 50% when the solvent was changed from cyclohexane to nitromethane (dielectric constant increase from 2 to 39). Glazer and Vidwans (1960) reacted S2C12 with cyclohexene as a model of reaction of SzClZ with an olefin. The rate was found to be first order with respect to each of the two reagents. The effect of temperature on the rate constants was very small; the activation energy varied between 2 and 5 kcal/mol. These low values suggested that the rate-determining step was not a single bimolecular one, and tests were applied for radical or polar mechanisms. The reaction was carried out in solvents and solvents mixtures of increasing polarity; the rate was shown to increase with increasing dielectric constant of the solvent. The effect of HC1 and ionized chlorides (such as tetra-n-butylammonium chloride and cetyl pyridinium chloride) was to in240
Ind. Eng. Chem., Prod. Res. Dev., Vol. 14, No. 4, 1975
crease the rate. No retardation of the rate occurred when benzoquinone was added. The proposed reaction mechanism was a two-step polar one, with the first stage controlling the rate. The intermediate compound suggested was more active than S2C12 because its electronic charge distribution was more asymmetric. Kaufman et al. (1937a, 193713) studied the action of SzCl2 vapors on thin layers of fats and concluded that the predominating addition reaction was accompanied in some cases by either a relatively small amount of substitution or by further reaction of the S2C12 adduct with formation of HCl in both cases. They also suggested that polymerization reactions accompanying the absorption of SzC12 rendered some of the double bonds more reactive to S2C12. Schiemann et al. (1963) reacted various vegetable oils with sufficient SzC12 to induce factice formation. Their measurements showed that for each S atom a chlorine atom also entered the molecule. In every case the refractive index of the sulfurized product varied linearly with S content and temperature. The viscosity of all the materials tested increased with increasing S and chlorine contents, the largest increase occurring shortly before factice formation. The first published works on sulfurization of jojoba oil and its lubricant uses correspond to the patents issued to Ellis (1936), Flaxman (1940)) and Wells (1948). They were the first to point out the correspondence to sperm whale oil properties and the advantages of the product based on jojoba oil. Ellis' work was intended primarily to the manufacture of a factice which was readily soluble in various aromatic and aliphatic solvents and could be incorporated in rubber, linoleum, paints and varnishes, plastics, and the like. This process was based on dissolution of jojoba oil in an equal volume of benzene and addition of 9-10'36 volume (oil) of S2C12. The reactants were mixed and allowed to stand until reaction was completed. Evaporation of the solvent yielded a rubber-like, sticky, light amber colored, tasteless mass, substantially free of acidity and soluble in various hydrocarbon solvents. The reaction could also be conducted in the absence of a solvent and the HC1 formed was eliminated by hot water washes. Flaxman (1940) and Wells (1948) reported the sulfurization of jojoba oil for use as lubricants and extreme pressure additives. Wells sulfurized the oil by heating it with stirring to 250°C and adding a quarter of the weight of S. When the temperature reached 300'F a second quarter was added; at 350'F a third; and a t 365°F the fourth quarter of S was introduced. The temperature was rapidly taken to 380'F and heating discontinued. Appropriate thermal treatment of the mixture assured that it remained liquid in spite of the high S content (+30%). Gisser et al. (1973) have recently conducted a thorough comparison of the mechanical properties of jojoba oil and sperm oil sulfurized by standard procedures. Both sulfurized oils were diluted to different concentrations in a 100 SUS a t 37.8OC naphthenic oil and in a 150 SUS at 99°C midcontinent bright stock oil. The diluted oils were evaluated on the Four-Ball E P and Falex Testers, and Four-Ball Wear Testers. Results of the experimental work showed that sulfurized jojoba and sulfurized sperm oils were essentially equivalent in improving the load-carrying capacity under extreme-pressure conditions of both naphthenic and bright stock base oils. Both undiluted sulfurized oils exhibited approximately equivalent E.P. properties. Small amounts of sulfurized jojoba and sulfurized sperm oil were also equally effective antiwear additives to the naphthenic and bright stock oils. Shop drilling and tapping operations confirmed that sulfurized jojoba could replace sperm oil in practical operations.
Jojoba oil (Simmundsia)has been sulfurized with sulfur monochloride and S in order to study the kinetics and parameters of the process and the characteristics of the final product. Reaction runs were conducted in the range of sulfur content 0-8%, addition time to 140 min, agitation speeds 100-500 rpm, temperature 20100°C, and solvent dielectric constant range 2-35. Analysis of the results was helped by tracking the reaction with NMR spectra. Reaction between the oil and S2C12 occurs by addition to the double bond, and cross linkage takes place with a corresponding increase in viscosity and molecular weight. Reaction in solvents follows a first-order rate with respect to the percentage of double bonds, and is highly influenced by the dielectric constant of the solvent and the addition of accelerators.
Experimental Section Equipment. The general view and arrangement of the experimental setup appears in Figure 1. The main piece of equipment was the reactor (AA) built from a 1-1. capacity, double-wall flask. The vessel could be emptied through a valve (B) connected to its bottom; this same valve was used for sampling purposes. The cover of the reactor carried five necks that permitted connection of the auxiliary equipment. A dropping-funnel (3), (A) was used to introduce the liquid reagents into the vessel. The reactor contents could be maintained under an inert atmosphere by means of a stream of nitrogen (12) provided from a cylinder (10) connected to the system through a pressure regulator (C, D). An Erlenmeyer flask (14) half-full with water provided visual inspection of the outgoing flow. Mixing of the reactor contents was provided by an anchor-type mixer (1) built of 10-mm glass rod and guided by a Teflon bearing. The mixer was driven by a Servodyne Drive unit (2, 4) manufactured by Cole-Parmer, Chicago, capable of maintaining of constant speed under variable torque. Jacket temperature control was provided by a Haake (Karlsruhe, Germany) cooling unit (8), Model K11, capable of cooling the circulating fluid down to -15OC. Temperature recovery to the desired level was permitted by a thermostatic bath (7) connected in series with the cooling unit. The connecting hoses were arranged in such a way that the circulating fluid flowed in series through the reactor jacket, the refractometer (6) prisms, the cooling unit, and the thermostatic bath. Reactor temperature was measured by a thermocouple immersed in a thermocouple well (9) and connected to a two-pen recorder (5). The recorder also registered the torque required by the mixer. Indices of refraction were measured with an Abbe-3L refractometer (6) manufactured by Zeiss, Germany, that gave direct readings to five significant figures. The temperature control arrangement described before permitted working at prism temperatures up to 70°C. The instrument was carefully calibrated with the test piece supplied by the manufacturer and checked with 1-bromonaphthalene. Viscosity measurements were made with a rotational viscometer manufactured by Contraves, Zurich (Model Rheomat 15), with a range of 1-104 P. For comparative purposes some measurements were also performed with a Haake Falling Ball viscometer, Model B. Molecular weight was determined by cryoscopy. The apparatus consisted of a jacketed beaker of 50-ml capacity provided with a variable speed mixer. Cooling was effected by applying vacuum to the jacket and by introducing the beaker into a Thermos bottle containing liquid nitrogen. Temperature readings were taken with a Hewlett-Packard quartz thermometer. Small-scale runs were done for kinetic purposes using an NMR probe as the reactor and following the reaction by recording its spectrum in a Varian Model XL-100 apparatus. Reagents. The jojoba seeds used in this work were obtained from shrubs grown at the Negev Institute for Arid Zone Research, now a part of the Ben Gurion University of the Negev. Cold-pressing produced a medium color oil that
I
Pi&$I I
I
-I
Q
m c I
I
I
I
AT-
Figure 1. Reactor and auxiliary equipment.
had the following characteristics: refractive index (2OOC) 1.4652; iodine value (Wijs), 83.2; melting point, 13.0°; acid number, 6.7; saponification number, 107. The gas chromatographic analysis was performed according to the method suggested by Miwa (1971), namely 3% OV-1 on Gas Chrom Q, 100-200 mesh, stainless steel column 100 cm long and 0.2 cm i.d., temperature programming of 3"/min between 275 and 300°, injection port 350°, flame ionization oven 380°,Packard-Becker Model 417 chromatograph. The following results were obtained: C34 0.1%; C36 1.6%; C38 7%; C40 32%; c42 49%; C44 9%; C46 0.9%; C48 0.1%. The oil was bleached by adding 2% of active earth, heating to 8OoC for 3 hr, and filtering. Sulfur monochloride was purchased from BDH Chemicals, Ltd., England (Cat. No. 30320). Operating Techniques. About 400 ml of jojoba oil, with or without solvent, was introduced into the reactor through port (10) and the mixing system (2) started. The desired agitation speed was attained with the help of control (4) on the Servodyne unit. The contents were heated to the operating temperature by the heating fluid while the temperature was being recorded. When the desired temperature was achieved the system was purged with nitrogen by opening valves (C) and (D) and noting that the gas was slowly bubbling through the water in beaker (14). At this stage the desired amount of S2C12 was introduced into dropping-funne1 (3) and its addition started by opening valve (A). This was considered the start of the run. The operating technique for direct addition of SnClz was similar, except that the reagent was added in one portion through port (10). Several samples were taken during the course of the reacInd. Eng. Chem., Prod. Res. Dev., Vol. 14, No. 4. 1975
249
I
1
I
I
I
1
6
5
I
S o l v e n t : Benzene
Assignment
a. b.
0.8 1.1
32
c.
2.1
Temperature: Probe
d.
2.15
Freq.
Response : 5
Time:
150
Spec. A m p . :
Sweep
i 3
e. 4 f . 5.3
I
I
I
2
1
0
(a)(b)
(e)
(f)
J (d)
(e)
f
CHj(CH&CH2 CH m C H C H 2 (CH2)n CHzCOCH2 (CH2)mn
-
1.9
m = 8.10
Figure 2. NMR spectrum of jojoba oil. tion of introducing a glass rod into the reactor; the attached fluid was examined in the refractometer. For adiabatic operation the cooling fluid entrance was closed and vacuum applied to the jacket. The desired amount of S2Clz was added in one portion as described. The run was continued until a new, higher level temperature, was attained. Most of the runs with solvents were monitored with the NMR apparatus. Solutions of appropriate concentration were prepared by mixing the desired volumes of solvent and oil with a 2-ml syringe. Adequate amounts of tetramethysilane (TMS) and solutions were added to the probe and the spectrum of the mixture recorded. The area of the double bond peak was obtained from the integrator reading. The probe was pulled out and 0.1 ml of S2C12 was added with the aid of a syringe. The probe was rapidly replaced in the apparatus and the time recorded with a stopwatch. At predetermined time intervals the height and area of the double bond were recorded. Runs a t temperatures different from room temperature were performed in a similar manner; the temperature level was determined by observing the shift of methylene glycol. Analytical Techniques. S and chlorine content was determined according to A.S.T.M. method E-443 (1973) with the following minor modifications: sample size was between 20 and 50 mg; the combustion flask contained 40 ml of 0.05 N KOH, with some drops of HzO2. After SO2 absorption the solution was titrated with barium acetate 0.1 N in propanol. The indicator was a 0.1% aqueous solution of 50 ml of carboxyarsenzo, 45 ml of bromophenol blue, and 1 ml of methylene blue, and the end point was a change in color from light blue to deep blue. Iodine values were determined by A.S.T.M. method D1959-69 and by NMR spectroscopy. Results General Description. Over 100 experimental runs were made in the reaction equipment and in the NMR probe. The first runs were made in the large system in order to gain experimental experience and collect enough material to roughly delineate the range of variables. Samples were obtained with 0.54% S, at agitation speeds of 500 rpm, addition times of 6-60 min, and temperature 20-7OOC. A second set of runs was performed using the fixed volumetric ratio of 35 ml of S2C12 and 400 ml of jojoba oil, agi250
Ind. Eng. Chem., Prod. Res. Dev., Vol. 14, No. 4, 1975
tation speeds between 100 and 500 rpm, and addition times 18-144 min. Measurements of the refractive index and S content were done in enough samples to determine the relation between these two parameters and their variation with reaction time. The kinetics of the reaction was studied in the absence of solvents in the large reactor, and in the NMR probe in the presence of solvents. The influence of the concentration of S2C12 on S content was studied at the fixed conditions of 3OoC, 500 rpm, and 400 ml of jojoba oil; the added volume of SpClp varied between 10 and 50 ml. The influence of reactor temperature was investigated in the range 20-70°C using dropping times of 18 and 40 min. The influence of the polarity of the solvent on the rate of disappearance of the double bonds was determined using the NMR techniques described previously. The variable selected was the number of double bond protons, and it was registered against time, at 25OC, using solvents which had dielectric constants in the range 2-35. The probe rotational speed remained constant at 1200 rpm and contained 0.1 ml of jojoba oil, 0.4 ml of solvent, and 0.02 ml of S2Clz. The influence of SzC12 concentration was studied by changing its initial volume between 0.01 and 0.04 ml. The temperature influence was also investigated in the range 30-70°C using the fixed ratio of 0.1 ml of jojoba oil and 0.4 ml of solvent. The influence of the concentration on the temperature rise was studied operating the large reactor under adiabatic conditions, using a fixed volume of 300 ml jojoba oil, and adding 120 ml total of the two other reagents. Of the latter, the volume of S2C12 varied between 5 and 60 ml. The variable studied was the adiabatic temperature rise with reaction time, when the solvent used had dielectric constant between 5.6 and 35. Molecular Description of Jojoba Oil. Jojoba oil is a liquid wax that contains two double bonds in each constituent molecule. The NMR spectrum of the raw material was studied in a 1:l volume solution in benzene and is reported in Figure 2. The hydrogen atoms next to the double bond and the allylic hydrogen appear with a shift of 5.5 ppm and 2.1 ppm, respectively. The hydrogen atoms in position (Y to the carboxylic group appear with shifts 4 and 2.15 ppm. The aliphatic hydrogens and those in the -CHz- groups appear with shifts of 1.1 and 0.8 ppm, respectively. The area of the different peaks shows that that of hydrogen atoms
8
1
'5
6
S o l v e n t : Benzene
F r e q . Roeponsa : 4
a . 0.98
e. 2.8-33
Swoop l i m o : 2 5 0
b.
f . 4.1
Spec.
Amp. 3 2
T e m p e r a t u r e : Probe
1.5
2
3
4
0 ppm
1 (c)
(HI
Assignments
(d)
(PXe)
(1)
(b)
(a)
CI
?! - CH=CHCHz(CHZ) CHn C O C H 2 ( C q m kHCHCH2(CH&
d
c.
2.1
g . 4.3
m = 8.10
d.
2.2
H.5.3
n 17.9
CHS
I
S I
YHCH CI
Figure 3. NMR spectrum of jojoba oil sulfurized with SzC12.
I
I 6
1
Solvent : Banzene
I
I
I
I
5
I
3
2
A ssignments
F r e q . Response : 4
a . 0.98
Sweep Time ; 2 5 0
b.
e, I
0
{CH~),, S p e c . Amp :
20
Temperature : Probe
c. d.
cn2cocn2(cn2),cn2
1.5 2.2
2.11-3.5
m = ai0
n
a
I 0 ppm (d)
(e)
(C)
I I
(b)
(*)
FHCHCH~(cn2hcn3 fx :a -cn C H -
29
Figure 4. NMR spectrum of jojoba oil sulfurized with S. belonging to the double bonds is double that of hydrogen atoms in position a to the oxygen. Also the area of the aliphatic hydrogens is substantially larger than that of the other peaks. Iodine Value and Number of Double Bonds. The influence of S and chlorine content on the iodine value and the number of double bonds was studied with the aid of NMR spectra (Figures 5 and 6). It is seen that there is a linear relation between these parameters and that the intersection at zero iodine value occurs at 8.7% S and 9.5% chlorine, while that at zero double bonds occurs at 8.5% S. Refractive Index. The variation of the refractive index with S and chlorine contents was studied in the temperature range 20-70°C (Figures 7 and 8). It is seen that there is a linear relation between these parameters so that the refractive index can be used as a good indicator of the S and chlorine contents.
For the case of S content the relation between the parameters could be described by the following equation
R.I. = 0.004358 - 0.00035(t - 20)
(1)
where R.I. = refractive index at t°C and S = S percent. Viscosity. The variation of viscosity with S content appears in Figure 9, for S contents between 0 and 8%, for reaction with S2C12, and 0-25%, for reaction with s. In both cases it is seen that sulfurization causes a significant increase in viscosity (about 1000 times) and that for equal S contents the product obtained by reaction with S is less viscous. The influence of temperature on the viscosity was studied in a sample containing 7.42% S and 8.5% in the temperature range 17-100OC. The results obtained appear in Figure 10 and show that the viscosity at 100°C is about 10%of that a t 17OC. Ind. Eng. Chem., Prod. Res. Dev., Vol. 14, No. 4, 1975
251
I
I
c
z Y
c
U
E
I
U
L Y
L Y Y
z 2
Y
i
s
0
U
;
0 OI
L
I
0
a
Y _1
d L
a
a
UI
UI
0
10
30
20
50
40
IODINE
VALUE
IO
10
I O
( WlJS)
Figure 5. Iodine value as a function of S or chlorine content. REFRACTIVE
1
~
~-
2
3
~
SULFUR
_
1
Figure 7. Refractive index as a function of S or chlorine content.
-_
5
INDiX
1
1
1
8
8
PERCENI
Figure 6. Percent double bonds as a function of S content. Molecular Weight. The results obtained by cryoscopy indicated that for S contents between 0 and 8% the average molecular weight of the sulfur-chlorinated oil varied between 600 and 3400 (Figure ll).
UIU 1.k111
0
1
I
I
1
1
1 SULFUR
Influence of Operating Variables Agitation. The influence of the agitation rate on the rate of addition of S was studied at 5OoC under the two modes of addition of S2C12: addition time of 138 min and direct addition. The experimental results showed that there was no influence of the agitation rate in the range 100-500 rpm. Addition Time. A series of runs were made a t 40°C and 500 rpm using SzCl2 addition times of 18, 36, 60, and 144 min. Figure 12 shows that there was no difference in the rate of S addition for SzC12 addition times between 18 and 60 min. As an addition time of 144 min there was a marked decrease in the rate. Viscosity-Torque. The experimental setup allowed registering the time change of mixing torque, caused by the in252
Ind. Eng. Chem., Prod. Res: Dev., Vol. 14, No. 4, 1975
1 4
I 5
1 1
I 7
I 1
PERCENI
Figure 8. Refractive index as a function of temperature. crease in viscositv. Two runs were performed a t 3OoC and 500 rpm to test the influence of the-amount of SzC12 added: one with more than the stoichiometric amount for full reaction, and the other with less than the stoichiometric amount. Samples were taken during the course of each run and their S content was determined from the measured refractive index. Both runs showed that the S content rose much more rapidly than the viscosity and the mixing torque. With less than the stoichiometric ratio of S2Clz the torque and the viscosity rose to an almost constant value toward the end of the reaction (Figure 13).
0
IO
IO0
ZOO
I50
l l M E , MIHUlE
Figure 12. Rate as a function of addition time.
/
Dimst
eddition
Of
S)CI>
,
HINUlE
10
1EMPERATURE
%
Figure 10. Viscosity as a function of temperature. 0
IO0
50 llHE
130
Figure 13. Viscosity, S content, and torque change during reaction.
I 0
i
I
ZOO0
-
0
SULFUR
1
1
I
I
6
I
PERCEHl
Figure 11. Molecular weight as a function of S content.
Reaction Kinetics and Mechanism Influence of Solvent Polarity. In order to study the mechanism of the reaction, runs were performed using solvents having varying polarity. These included nitrobenzene, tetrahydrofuran (THF), chlorobenzene, l,&-chloro-
benzene, 1,4-dioxan, benzene, toluene, xylene, cyclohexane, and cc14. Most of the results are reported in Figure 14. I t is seen that there is a clear relation between the rate of disappearance of the double bonds and the polarity of the solvent; that larger dielectric constants are conductive to faster rates of reaction. With nitrobenzene, for example, the double bonds totally disappeared within 1.5 min. The value of the dielectric constant was not the only significant parameter; in Figure 15 it can be seen that there was a significant difference in the rate of reaction in benzene and cyclohexane, in spite of the fact that both solvents have essentially the same dielectric constant. The reaction rates with or without cc14 were similar, except for the fact that CC14 did not dissolve the reaction product. Catalyst and Accelerators. Some qualitative tests were done on the influence of adding catalysts and accelerators to the reaction mixture. For example, when 0.1% pyridine was added to a solution of oil in cyclohexane the reaction rate increased to that of a solution of oil in benzene. Addition of small amounts of dimethylamine and butylamine caused temperature increases much larger than those encountered in adiabatic operation in the presence of benzene. Addition of 5% parabenzoquinone to a solution of oil in T H F did not affect the rate of reaction. Influence of SzClz. The influence of the amount of S&l2 added was studied in the absence of solvents in the large reactor and in the NMR probe in the presence of solvents. The common pattern to all the S2Clz/oil ratios studInd. Eng. Chem., Prod. Res. Dev., Vol. 14, No. 4, 1975
253
IO'
51
c
z
Y
U I/ 0 Y
0
z
2 Y
2
3 I
: t o
,\ IO
0
20
10
40
SO
IO
70
ID
I
I
I
I
1
1
IO
IO
10
40
so
IO
IO
l l M f ,MINUTE
Figure 14. Solvent influence on the disappearance rate of double
bonds.
TIME , M I N U T E
Figure 16. Disappearance rate of double bonds as a function of SzC12 concentration.
SI-
10
-
0.5
I I5
I MOLAR
RATIO
e
I I
Figure 17. Temperature rise difference as a function of molar ratio of SzClz and jojoba oil.
Figure 15. Disappearance rate of double bonds as a function of the dielectric constant.
ied was that given in Figure 12, the only difference being in the final sulfur content. The presence of solvents caused a significant change in the rate of disappearance of the double bonds (Figure 16), under similar operating conditions. Doubling the concentration of S2C12 increased the rate in 1,2-dichlorobenzene by about 300% and between 300 and 500% when the solvent was tetrahydrofuran. The influence of the concentration of S2Cl2 on the adiabatic temperature rise was also studied in different solvents. The maximum temperatures were achieved with nitrobenzene when jojoba oil and S2Cl2 where reacted in the volumetric ratio 5:l. A plot of the adiabatic temperature rise against the molar ratio S2Cl2/oil showed that the high254
Ind. Eng. Chem., Prod. Res. Dev., Vol. 14, No. 4, 1975
est temperature occurred when the molar ratio was about 1.5 (Figure 17). Influence of Temperature. The influence of reaction temperature was studied in the large reactor and in the NMR probe, with and without solvents. It was found that with an addition time of 18 min the change in rate was very small between 20 and 50°C. When the addition time was increased to 140 min there was no temperature influence on the rate between 20 and 70'. Striking differences were present when the reaction was conducted in the presence of polar solvents. Figure 18 shows that the rate decreased when the temperature was increased; this behavior was opposite to that predicted by a simple Arrhenius equation.
Discussion of Results Product Characterization. The reaction between jojoba oil and S2C12 caused addition of chlorine and S in the
I
1 E M f E IA 1 U IE
.!.!3e(1
Figure 18. Disappearance rate of double bonds as a function of temperature. same ratio in which they were present in S2C12, that is, 1:l. The results plotted in Figures 5 and 6 pointed out that when the double bonds were completely saturated the content of S was 8.7% and that of chlorine was 9.5%. The linear relation found between the number of double bonds and the chlorine and S contents (Figures 6 and 7) indicated that the basic reaction was one of addition to the double bond, with little or no substitution. The latter phenomenon would have increased the S and chlorine content without changing the iodine value. The molecular weight and viscosity of the sulfurized oil showed a parallel behavior with S content (Figures 9 and 11); a large increase in both parameters was observed in the small range of 7 to 8.5%S. These two facts indicated that a polymer was being formed, an assumption that was further supported by the proposed reaction mechanism. On the basis of these facts emerged the following picture of the attack of the double bond by S2C12
IP
-c-cI s I
-c-c-
I c1I
An S bridge was normally formed between two different oil molecules, This structure is similar to that suggested by Flint (1955) and Glazer and Vidwans (1960). Each molecule of jojoba oil contains two double bonds so that theoretically it is possible to achieve a large degree of intermolecular polymerization. I t has already been pointed out that there was a significant increase in the molecular weight and viscosity when the oil was almost saturated. At this stage apparently there was a better chance that two or more molecules would bridge together. Flint [1955] showed that in the sulfurization of triglycerides there was a possibility of intermolecular bridging. In the case of jojoba oil this is the only possibility (monoester), as can be shown graphically as follows. Assume that structure I represents the molecule of jojoba oil, a black dot represents a double bond, an empty dot represents a saturated bond, and a double line represents an S bridge.
I1
I11
One possibility of sulfurization is the one represented by structure 11. Here the iodine number is zero and the bridging is internal. Case I11 corresponds to a dimer with iodine number zero. A dimer may also be formed with half the original iodine value (structure IV). Larger degrees of polymerization are illustrated by structure V. All these structures are possible, but V seems to be the preferred one. Inspection of the NMR spectra of jojoba oil sulfurized with SzClp (Figure 3) and S (Figure 4) showed the similar nature of the S peaks and the absence of the chlorine peak a t 4.3 ppm. The NMR spectrum of jojoba oil sulfurized with S did not allow us to determine clearly the degree of bridging, but a reasonable guess was that the bridge would be built by two S molecules so that the amount of S needed to open the double bond would be double of that needed for sulfurization with S2C12. The height of the chlorine peak a t a shift of 4.3 ppm was equal to that of the S peak, corroborating the assumption that the opening of every double bond involved the addition of one chlorine atom and one S atom. This hypothesis was strengthened by the findings of Hotten (1972) on the sulfurization of 1-decene. The product distribution corresponded basically to thioether and dithioether. The difference between sulfurization with S and sulfurization with S&lz is then the formation of two bridges between two S atoms in the first case, and only one bridge in the second case. This difference in structure permits to advance the idea that the mechanical properties of jojoba oil sulfurized with SpClz will be substantially different from those of jojoba oil sulfurized with S. Schiemann and coworkers (1963) studied the variation with S content of refractive index, iodine value, viscosity, and molecular weight of a number of fatty materials. Some of their findings are reported in Figure 19, together with the ones obtained in this work. The similar slopes of the curve index of refraction/S content of sulfurized jojoba oil, methyl linoleate and linseed oil illustrate the similarity in the way S2C12 reacts. The relation between iodine value and S content are straight lines of different intercept due to the difference in the initial iodine value; nevertheless linearity corresponds to an addition reaction in every case. I t can be seen that the overall behavior of sulfurized products is similar and independent of the fact that the starting material is a triglyceride, a diester, or a monoester. The viscosity of the sulfurized oil increased from 20 to 600 P when the S content increased from 7 to 8% (Figure 19); in the same S content range the molecular weight grew from 1800 to 3400. Here jojoba oil is distinctly different from other fatty materials. In the case of triglycerides, about 50% of the double bonds are saturated by intramolecular bridges (Harrison, 1953; Flint, 1955), but the intermolecular bridges are the most important factor for an increase in viscosity and molecular weight. The nature of jojoba oil, a monoester with two noninteracting double bonds, gives place to a number of intermolecular bridges larger than those that are possible when sulfurizing methyl linoleate, also a monoester with two double bonds, but capable of conjugation. This conclusion is reinforced by the behavior of ethyl oleate: the increase in molecular weight is linear with the S content. Only one double Ind. Eng. Chem., Prod. Res. Dev.. Vol. 14, No. 4, 1975
255
I I
id
c Y z
d
U
. i
P Y
0
z
*0 d Y
*a 0
,.iL JI
SULFUR
nI
I
1
S I
11
I SI
I II
T l Y t i YlNUlE
PERCENT
Figure 19. Viscosity as a function of S content.
Figure 20. Solvent effect on mass transfer resistances.
bond is present so that the largest chemical species possible is the dimer. Since the two double bonds cannot interact, they can be looked upon as two functional groups capable of independent polymerization. This situation can be analyzed by the methods described by Billmeyer (1962) for calculating the molecular weight of a polymer, under the assumption that the rate of reaction is independent of the molecule size. The corresponding equation is
constant value and the following increases were very slow because of the resistance to SzCl2 diffusion caused by the higher viscosity. The phenomenon described above of significant mass transfer resistances did not present itself when the reaction was carried out in the presence of solvents. The dilution effect of the solvent was instrumental in maintaining a low relative viscosity during the course of the reaction. This fact was illustrated by plotting in semilog graph the variation of double bond contents or S contents with time, when SzC12 was added in one portion, with or without solvents (Figure 20). A straight line was always obtained when the reaction occurred in the presence of a solvent. Examining again the results plotted in Figure 14 shows that addition of solvents like cyclohexane, toluene, and xylene caused a modest decrease in the number of double bonds present. All these solvents had a relatively low dielectric constant and the overall effect could be considered as one of dilution alone. Addition of a solvent of this class decreased the viscosity and increased the dilution and the probability that two reacting molecules would be within a certain distance, a t a given time. For higher dielectric constants a new factor made an important contribution to the overall phenomena (Frost and Pearson, 1961; Caldin and Peacock, 1965). Higher dielectric constants were seen to give place to faster rates of double bond disappearance. A polar solvent will attract intermediate compounds that are in a polar form and will enlarge their average life by enough time to enable them to react before they disappear. Frost and Pearson (1961) pointed out that the activation energy of a reaction did not depend on the presence of a solvent so that the solvent should be particularly useful when the control of the reaction is by physical means. I t has been mentioned that in the early stages of the process the reaction is almost instantaneous so that addition of a solvent is of little benefit; the following increase in viscosity caused the transfer of the controlling stage to the physical step and here addition of a solvent should certainly enhance the rate of reaction.
w,= x ( 1 -
pp.px-1
(2)
where p = probability that a functional group will react at time t , x = number of repeating units, and W, = weight percentage. It can be shown that Wx has a maximum value at a molecular weight equal to the number molecular weight. Calculations were made assuming that p is proportional to the percentage of S, with p = 1.0 for %S = 8.7. A good correlation was found for S contents up to 6%. Above this figure the basic equation involved in eq 2 apparently was no longer valid. Mass Transfer Resistances. In the course of the sulfurization reaction the oil became more viscous, particularly toward the end of the saturation process. Plots of the change of S content with time indicated that in the early stages of a dropwise addition of SzC12 the rate was very fast in relation to the overall time, and that an increase in agitation did not improve the reaction rate. The larger interfacial area created by a higher speed of mixing did not influence the overall rate of reaction; the only important variable was the rate of addition of SzClz. On the other hand, when SZC12 was added in one portion, both reagents were present in high concentration, the increase in S content and viscosity was fast so that the unreacted material faced a larger resistance to mass transfer. Kaufman et al. (1937a, 1937b) reacted different liquid fats with SzC12 vapors in a desiccator. The analysis of the S content showed that within an hour it achieved an almost 256
Ind. Eng. Chem., Prod. Res. Dev., Vol. 14, No. 4, 1975
It has been stated before (Figure 15) that the value of the dielectric constant is not enough to describe the overall picture. Solvents with the same dielectric constant did not necessarily produce the same rate of reaction; the difference was due to the difference in their molecular structure. With solvents of the same family such as benzene, monochlorobenzene, and 1,2-dichlorobenzene, the rate of reaction varied almost linearly with the dielectric constant. On the other hand, solvents such as THF, with an ether linkage, or benzene and xylene with 7r electrons, deviated substantially from the linear behavior. These results ratified the suggestion of Caldin and Peacock (1955) that for the purposes explained here, solvents should be classified in categories such as aromatic, aliphatic, hydroxylic, etc. Reaction Mechanism. The reaction between S and olefins has been explained by polar and radical mechanisms (Bloomfield, 1947; Bateman et al., 1958; Ross, 1958; Caldin and Peacock, 1955) with the possibility that the different stages of the reaction are associated with different mechanisms. According to Bateman et al. (1958), the opening of the bond between S atoms takes place by a radical or ionic step, but the addition to the double bond was strictly ionic, since its rate was increased by the addition of polar solvents and decreased by the addition of radical inhibitors like benzoquinone. In order to test these hypotheses, Ross (1958) performed his experiments in the presence of different solvents with dielectric constants between 2 and 35 and addition of radical traps. The rate of reaction in the presence of benzene and nitrobenzene grew by a factor of 60 although the dielectric constant grew only 15 times. On the other hand, reaction with cyclohexene in the presence of cyclohexane and nitromethane produced a rate increase of 50% in spite of the fact that the dielectric constant was 26 times higher. Addition of 5% of benzoquinone to a solution of oil in nitrobenzene and T H F did not affect the rate of reaction. From these two facts it could be inferred that the mechanism of the reaction of SzC12 with jojoba oil was a polar one and no free radicals were involved. Glazer and Vidwans (1960) reacted S2Cl2 with cyclohexene as a model of the reaction between S&12 and an olefin. The two double bonds in jojoba oil bear no relation between them so that its sulfurization should parallel that of an olefin, similarly to the sulfurization of sperm oil (Hotten, 1972). Bateman and Moore (1961) described a possible mechanism for the sulfurization of olefins. The mechanism was a polar one, composed of two steps in series; the first one controlled the rate. The intermediate product was more reactive than SzC12 because its electronic charge distribution was asymmetric, in contrast to S&12 that presented a symmetric configuration. By comparison, the following mechanism can be proposed for the sulfurization of jojoba oil
-C
I -C
I
I I( + s*c12
I
K
-C-S-S-CI
I
-c-CI
I
Stage I is the one that controls the rate of reaction because the intermediate is more reactive than SzC12. Aryan (1961) pointed out that two polar structures may be proposed for
S2C12 but the preferred one should be that involving the heterolytic decomposition of the molecule. The two-stage reaction mechanism described above may be analyzed from the viewpoint of electronic charges and the influence of polar solvents on the rate. In the first stage the excess electron charge on the double bond and the polar solvent gives place to formation of the following intermediate
This intermediate changes immediately to
I I11 I -c-c1 I
-C--S-S(+)
Intermediate I11 is more active than SzC12 since its electron distribution is not symmetric. The role of the polar solvent is a double one: (1) displacement of the charge in S2C12 and its subsequent heterolytic decomposition, and (2) stabilize the intermediate I11 so that it may react with the double bond in the following way
The suggested mechanism is corroborated by the findings on the influence of temperature on the rate of reaction (Figure 18). If the mechanism was one involving two molecules there should have been a significant influence of temperature, contrary to the experimental findings. An increase in temperature produced a decrease in the rate of reaction (Figure 18) because the dielectric constant of most solvents decreased substantially with an increase in temperature. The reaction between S2C12 and cyclohexene is second order, one with respect to each.reagent (Glazer and Vidwans, 1960). Since jojoba oil contains two double bonds we may assume that the order will be one with respect to the oil and two with respect to S2C12. This hypothesis was verified using the experimental technique suggested by Williams (1974) where the reaction is conducted adiabatically and the maximum temperature rise is plotted against the reagents ratio. This is shown in Figure 17. The maximum of the curve occurred at a molar ratio S2Clz/jojoba between 1 and 2, pointing out that the reaction probably took place between 1 mole of oil and 2 of S2C12. In addition, the overall reaction could be considered to be pseudo first order with respect to the concentration of the double bonds (Figures 14 and 16). Kaufman et al. (1937a, 1937b) also suggested that the double bonds could be used as a measure of the rate of reaction. Kinetics of t h e Reaction. A semilog plot of the change in double bonds concentration against time gave a straight line. so that
Integrating In [(I.V.)/(I.V.)o] = Kt Ind. Eng. Chem., Prod. Res. Dev., Vol. 14, No. 4, 1975
(4) 257
where (I.V.)o = initial iodine value, (I.V.) = iodine value a t time t , and K = reaction constant. The reaction in the presence of solvents took place according to eq 3 where K would change according to the solvent, SzC12 concentration and the temperature (Figures 15, 16, and 18). A change in the concentration of S2C12 when the solvent was T H F did not give a clear picture regarding the possible order of the reactor with respect to SZC12, possibly because of the activity of S2C12 with THF. On the other hand, when the solvent was 1,2-dichlorobenzene the rate grew by about 300% when the concentration of SpClp was doubled from 0.1 to 0.2 and to 0.4. From this fact it can be inferred that the order with respect to SZC12 is about 1.5, similar to that obtained from Figure 17. This kind of information was not available for reactions without solvent because of the strong influence of viscosity. According to Ross (1958) the reaction between S and olefin is autocatalytic and in three steps: (1)initial rate, and fast accelerating, (2) constant rate, and (3) diminishing rate. For jojoba oil the rate can also be assumed as autocatalytic if we assume that structure I1 is the one that enhances the rate. Harvey and Schuette (1930) pointed out that fish oils react very fast with S2C12 because of the hydroxy acids and the amines that are naturally present in the raw material. The active role of these materials has also been shown in the present work. Addition of pyridine to a solution of oil in xylene increased the rate by about 500%. When the solvent was cyclohexane (slow rate) addition of 5% pyridine increased the rate to about that in the presence of benzene. In order to further investigate the influence of catalysis, runs were also made in the large reactor operating adiabatically, the index here was the rise in temperature. The results indicated that butylamine and dimethylamine enhanced the rate of reaction.
1975. Aryan, Z. S., Wiles, L. A., J. Chem. SOC..4510 (1961). A.S.T.M., Book of Standards, American Society for Testing and Materials. Philadelphia, Pa., 1973. Bateman, L., Moore, C. G., "Organic Sulfur Compounds", Vol. I, p 97, Pergamon Press, London, 1961. Bateman. L., Moore, C. G.. Porter, M.. J. Chern. Soc.. 2866 (1958). Biilmeyer, F. W., "Textbook of Polymer Science", 2nd ed, p 268, Wiley, New York, N.Y., 1962. Bloomfield, G. F., Naylor. R. F., "Proceedings of the XI International Congress of Pure and Applied Chemistry", Vol. 2, p 7, London, 1947. Caldln, E. F., Peacock, J.. Trans. Faraday Soc., 51, 1217 (1955). Ellis, C., U S . Patent 2,054,238 (1936). Farmer, E. H., Shipley, F. W.. J. Chem. SOC.,1519 (1947). Flaxman, M. T., U.S. Patent 2,212,899 (1940). Flint, C. F., Roc. Inst. Rubber Ind, 2, 151 (1955). Frost, A. A.. Pearson, R. G., "Kinetics and Mechanisms", p 123, Whey, New York, N.Y., 1961. Gisser, H.,Messina, J., Hassan, D., "Jojoba Oil as a Sperm Oil Substitute", Publication U S . Army, Frankford Arsenal, Pittman-Dunn Laboratory, Philadelphia, Pa., 1973. Glazer, J., Vidwans, D. B., Rubber J. Inter. Plastics, 312 (1960). Harrison, J. B., Trans. Inst. RubberInd., 28, 117 (1953). Harvey, E. H.,Schuette, H. A,, Ind. Eng. Chem.. Anal. Ed., 2, 42 (1930). Hotten, B. W., NLGI Spokesman, 174 (Aug 1973). Kaufman, H. P., Baltes, J., Marden, P., FeffeSeifen, 44, 337 (1937a). Kaufman, H. P., Baltes, J., Marden, P., Fette Seifen, 44, 1390 (1937b). Miwa, T. K., J. Am. OiiChem. Soc., 42, 259 (1971). Miwa, T. K., Cosmet. Perfum.,88, 39 (1973). Peeler, R. L.. Hartmann, L. M.. "Evaluation of Sulfurized Sperm Oil Replacements", presented at the 40th Annual Meeting of the National Lubricating Grease Institute, Denver, Colo.. Oct 1972. Ross, G. W., J. Chem. Soc., 2856 (1958). Schiemann, G., Duering, H., Koerner, H.. Deut. Farber Z., 17, 408 (1963). Sherbrooke, W. C., Haase, E. F., "Jojoba: A Wax-Producing Shrub of the Sonoran Desert. Literature Review and Annotated Bibliography", Office of Arid Lands Studies, University of Arizona, Tucson, Ariz., 1974. Wells, F. W., U.S. Patent 2,450,403 (1948). Williams, R. D., Chem. Eng. Ed., 8, 28 (1974). Wisniak, J., Liberman, D.. J. Am. OiIChem. Society, 52, 259 (1975).
Receiued fo; reuiew July 15, 1975 Accepted September 4,1975
Literature Cited "Products from Jojoba: A Promising New Crop for Arid Lands", Committee on Jojoba Utilization, National Academy of Sciences, Washington, D.C.,
258
Ind. Eng. Chem., Prod. Res. Dev., Vol. 14, No. 4, 1975
This work was partially financed by a grant from the United States-Israel Binational Science Foundation.
