An Attempt to Dehydrogenate Cottonseed Oil Citation Gillies, Robert W. (1922) An Attempt to Dehydrogenate Cottonseed Oil. Bachelor's thesis, California Institute of Technology. doi:10.7907/6E4J-V604. https://resolver.caltech.edu/CaltechETD:etd-04232004-143458 Abstract No Abstract. Item Type: Thesis (Bachelor's thesis) Subject Keywords: Chemical Engineering Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemical Engineering Thesis Availability: Public (worldwide access) Research Advisor(s): Unknown, Unknown Thesis Committee: None, None Defense Date: 1 January 1922 Record Number: CaltechETD:etd-04232004-143458 Persistent URL: https://resolver.caltech.edu/CaltechETD:etd-04232004-143458 DOI: 10.7907/6E4J-V604 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 1474 Collection: CaltechTHESIS Deposited By: Imported from ETD-db Deposited On: 27 Apr 2004 Last Modified: 03 Oct 2019 23:08 Thesis Files Preview PDF (Gillies_rw_1922.pdf) - Final Version See Usage Policy. 995kB Repository Staff Only: item control page

AN ATT 2MPT TO DEHYDFOGEN ATS COTTONS 'aD OIL Thesis by ROBART W. GILLIS In partial fulfillment of the requirements for the degree of Bachelor of Science In Chemical ingineering. CALIFORNIA INSTITUT! OF TACHNOLOGY Pasadena, California. 1922. AN ATTEMPT TO DSHYDROGSNATZ COTTON SSED OIL. INTRODUCTION. | It is now a well known fact that certain vegetable and animal oils can be hydrogenated or hardened, by heat- ing them in an atmosphere of hydrogen, with a suitable catalyst, such as nickel or phatinun. It is the purpose of this thesis to carry on an investigation of the re- action and to attempt to change the equilibrium conditions so that hydrogen is removed from the oil instead of being absorbed by it. ‘Removing hydrogen would give a more unsaturated oil, i.e. one containing less hydrogen and having more double bonds. Such a procedure would give an oil with different and interesting properties. For instance it should be more active chemically and should have better drying qialities. Altho the exact mechanism by which the so called "dry- ing oils" dry and become hard is not definitely mown, it undoustedly is an oxidation process in which the double bonds in the carbon chain unite with oxygen to give a hard elas- tic product. As a general rule the saturated oils are non- drying and as the number of double bonds increases, the ability of the oil to absorb oxygen also increases, and therefore the drying properties are improved! 1. Lewkowitsech, Chem. Tech. & Anal. Oils, Fats & ‘axes, (1913) P.466. Re From this it can be expected that if an oil were to be dehydrogenated, it would dry better than the original oil which was less unsaturated and the treated oil could thus be used as a substitute for linseed and other drying ofls. “The hydrogenation of an oil is extremely slow if no catalyst is present but if a suitable catalyst is used the reaction takes place readily. . When nickel is used for the catalytic agent the hydrogenation takes places most rapidly at a temperature of 175 to 200°C. The speed of the reaction increases with rising temperature up to 180°C and then de- creases as the temperature is raised further. This effect is an indication of an equilibrium condition. The speed of the reaction is also greater with increased pressure. There- fore by raising the temperature and reducing the pressure it might be expected that the equilibrium would be changed and that the reaction would reverse. This view is supported by the following facts: (1). Zelinsiit dehydrogenated cyclohexane by heating it in contact with finely divided palladium black. The reaction commences at 170°C and proveeds rapidly at 200 to 300°C. At lower temperatures the reverse change occurs, passage of hydro- 1. C.A. 5, 958, (1911) gen and benzene over the catalyst at 100 to 110°C result- ing in the hydrogenation of the benzene. Equilibrium is established at 200°C and above this temperature dehydro- genation occtrs despite the presence of hydrogen. (1). sabatier™ showed that although certain organic com- pounds are reduced at a temperature of 180°C by passing their vapors mixed with hydrogen thru tubes containing fine- ly divided metals, especially nickel, at a temperature of 250°C the same metal brings about an oxidation of the com- pound, for example of alcohols to aldehydes and ketones. (2). Shaw” found that distilling oleic acid thru fresh cat- alyst under a pressure of 100 mm. reduced the iodine number 5%, and that redistilling as before made no further reduc- tion. Similar results were obtained at 150 and 200 mm. re~- ducing the iodine number 13 and 20% respectively. «This seems to indicate that there is a definite equil- ibrium between stearic acid, oleic acid and hydrogen and that the amount of reduction was constant for any given pressuree Wieland” assumes that metallic hydrides unite as such with the unsaturated compound at the double bond and the labile addition product formed breaks down with the retention le CeAs 5, 702. (1911). 2. Seifer Stg. 1912, 713. 3. Ber. 45, 484. (1912). 4. of nydrogen and the elimination of the metal, the later then being in a condition to take up more hydrogen and again form the hydride. The following equation shows the assumed equilibrium using palladium as Catalyst, but other metal would act similarly. Sc Co PdH, = 26 -G = | H. bday \¢ mc ' From a thermodynamic standpoint the equation appears to be a reversible one. Sabatier? also believes that the activity of the cat~ alyst is due to the formation of hydrides and that varia- tions in the activity of the catalyst are due to different dydrides formed, Thus Nilo is more active than Ni -.it H 4H If this assumption is correct it would appear as a consquence that nickel and other metals should not only effect a union of hydrogen but also that hydrogenating com~ pounds should suffer a removal of their combined hydrogen, the metals acting as dehydrogenators. This actually proves to be the case. Between 200 and 300°C finely divided copper readily acts as a dehydrogenating catalyst, converting al- cohols into ketones and aldehydes. 1. Die Hydrierung durch Katalyse, Leipsic (1915), 17. It is also known that the heat of combustion of car- ‘pon compounds increases with the unsaturation. According to a general principle compounds with high heats of com- pustion should be more stable at high temperatures, than ones with lower heats of combustion. Therefore an unsat- urated oil might become more unsaturated by heating to a high temperature in the presence of a suitable catalyst. On the other hand, according to Padoa and Fabist under atmospheric pressure hydrogenation and dehydrogenation are distinct processes. In most cases nickel ‘ean effect either reaction, but on certain compounds the nickel acts in only one way. Under increased pressure the two limits of tem- perature are close together and under certain conditions the two reactions may proceed simutaneously, until equili- prium is attained”. METHOD, The effect of heat alone upon the cotton seed oil was first studied and the effect of heat upon the oil mixed ¢ with various catalyzers. Samples of the oil were therefore heated over a range of temperatures varying from 175 to 340°C 1. 3.8.0.1. (1908), 1083. 2. This may account for the failure to obtain any dehydro-~ genation.e , Ge for various periods of time, exposed to air and also in ‘elosed tubes, in order to determine if the oil was seri- ously darkened or decomposed. It was found that the oil was effected somewhat but not enough to interfere with the experiments to be performed later. Several nickel catalyzers were then prepared, using the methods outlined by Mr. Trowbridge and Mr. zoul” re~ spectively. In the preparation and use of the catalysts great care was taken to protect them from impurities that might poison them and reduce their activity®. These catalysts together with a sample of catalyst ob- tained from the Globe O11 Mills, Los Angeles, were used in the experimental work. They were first tested out to make ‘sure that they were active, by using them to hydrogenates samples of cotton seed o11. If they were found to be suffi- ciently active they were used for further experimentation in attempts to dehydrogenate the oil. The hydrogenation tests were made at temperatures be- tween 175 - 200°C, by passing hydrogen gas thru the heated oil and catalyst. The dehydrogenation was attempted at tem= peratures between 200 - 340°C by heating the o11 and cataly- 1. Method used by the Globe Oil Milis, Los Angeles, to pre- pare their catalys for hydrogenating oils. 2. Of the Celite Products Co. previously with Proctor and Gamble. 5. There are many things that poison nickel, namely, Pb, “Zn, He, S, Si, P, Cu(OH), As,0, , HC1, H,SO,, KOH, NaOH, and soaps of K,Na, Li, Mg, Ba, Fe, ce, In, Cd, Pb, He, Zn, Sn, and Bi. C. A. Vol. 13, 385-4. (1913). Fe ger in special heating tubes. In some cases the outside air was excluded by closing the tube except for a piece of capillary tubing, in others the air pressure was reduc- ed to about 0.5 atmosphere and in the remainder, carbon dioxide gas was bubbled thru the oil so as to displace the MaTHODS OF ANALYSIS. In determining the effect of the treatment, three tests were used, namely the iodine number, acid value, and refrac- tive index. The physical appearance of the oil such as col- or, and viscosity was also useful in some casese The Iodine mimber is a measure of the unsaturation of the oil and was used to determine the amount of hydrogena- tion or dehydrogenation of the oil. However the oil can become saturated by the addition of oxygen aswell as hydro- gen, and the Iodine number does not differentiate between tke two additions. Consequently another constant was needed, and was found in the refractive index. This valve was found to increase if there was oxidation and decrease if there was reduction. A decrease in both the Iodine number and re- fractive index indicates hydrogenation, while a decrease in the iddine number accompanied by an increase in the refractive index indicates oxidation. Be The acid value gives the amount of free fatty acid ‘in the oil. This is normally very small but increases if there is decomposition, and so this test was used to deter- mine if the oil was broken up into glycerine and free fatty acids by the treatment. | the Todine number was determined by the Hanus methoat taking the precautions given by Tuttle and Smith”. The Acid value was determined by the method advanced by Steele and sward? This is much faster than the older analyses and gives more reliable results. The refractive indices were obtained by using an Abbe Hefractometer. APPARATUS AND MATMRTALS. All of the heating was done in an electric heating unit consisting of a porcelain tube wound with resistance wire, and insulated with magnesia and asbestos. The heater was placed in a vertical position and the ends closed with asbestos plugs. The temperature was regulated with a mercury thermestatic regulator that kept the furnace within 10°C of any desired temperature. 1. Leach Food Inspection and jinalysis, p 5835. 2. Tech. Paper, Bur. Stand. No. 37, Iodine liumber of Linseed Oil. 3. Jour. Ind. "ng. Chem., 14, 57 (1922). De The oil was contained in a special heating tube of pyrex glass made as shown in the drawing. This tube could be connected to the vacun “line and the pressure on the oil reduced, or a gas could be bubbled thru the oil, or the tube closed with stoppers and the oute side air excluded. The nickel catelyst was reduced in the tube as shown below. The tube was of pyrex glass 25 mm. by 300 mm. closed at each end with cork stoppers, protected by glass wool. The heating was done with two fish tail burners. Hydrogen gas was led from a cylinder thru a flask contain- ing oi], to act as a bubble counter, and then thru a caleium chioride drying tube, and to the pyrex tube. The used gas was passed thru another flask partially filled with oil to act as a safety device. The temperature was measured by means of a thermo-couple.placed inside of the pyrex heating tubee Cottcnseed oil was used in this work because it is a - gemi-drying oil, it is fairly inexpensive, and a pure supply could be easily be obtained from the Globe 041 Millis, Los angeles. ‘iinter oil, bleached, but with the stearin not removed was used, for the first experiments. Later, this was removed by cooling in ice water, and filtering off the precipitated stearin. This oil is used by the Globe Oil Mills to hydrogenate conmercially and so it did not con- tain anything liable to poison the catalyst. The hydrogen gas used was obtained in cytinders from the California Compressed Gas Co., Los .ingeles, ana is manufactured by clectrolytic processe It contained about 0.5% oxygen. The nickel was obtained from the Globe oil Wills and is the same as is used by them. It is electrolytic and is claimed to be practically 100% pure. The kieselghur was obtained from the Celite Products Co. and was their best grade. The carbon dioxide gas was obtained in cylinders from the Liviti Distilled ‘iater Co., Los Angeles, and con- tains about 3% air. RESULTS . It was found that heating samples of cottonseed oil decreased the Iodine number and increased the acid value ll. and refractive index. The oil was also darkened and be- -game more viscous. <A precipitate was formed when the oil was heated above 275°C. . | The magnitude of these effects depended upon the temper- ature, the time of heating and the amount of air in contact with the oil. Nickel oxide and the reduced nickel oxide catalyzers were found to increase this effect, the former to a greater extent than the latter. The nickel catalyzers prepared and used in the experi- ments were tested and found to be as active as a sample of catalyzer obtained from the Globe Oil Mills, Los Angeles, which was known to be catalytically active. Cottonseed oil was successfully hydrogenated to a hard oil. Heating under diminished pressure of about atmosphere the oil darkened considerably and the viscosity increased more than when the oil heated under atmospheric pressure. The effect was greater when nickel was present. From the oils heated to temperatures above 275°C there was obtained flocculent precipitats which came down after the oil had stooc for a day or two. This appeared to be stearin, but it was not definitely proven to be so. If it were stearin, it may have been already present as such, and le. precipitated by the heat treatment, or it may have been freshly formed by two molecules ot olein reacting with each other in such a way that one molecule gave up part of its hydrogen to the other, forming as a result one molecule of stearin, and one of linolic tri-slyceride. No denydrogenation of the cottenseed oil was obtained in any of the experiments performed. Atabular statement of the procedures and the results is given and will be Pollowed by a more detailed account of each, and a discussion of the results. Refract Time Condition Cet. 1 Ho Tnésx Aeid No, flemarks, - Ee NaS om ee = oe vo nae OEE NOlnNCE nan SvO STS RE CCP CETD nara renner ERED ern od a a a * Sane 7 ~ Ny eS on Ny NO Fe. & Gi cS oOomn on a yh Cs N OF C4 o nm ~ Faw O12 105.0 1.4686 ~1G55 Ol ear, Licht yellow a * stearine 10 1141 Wo change renoved 1 Yeated exvosad to 104.5 1.4680 ~LE64 oO. change air on ct tx Pa a +: rap) oO © o Oo a) le " 102.7 1.4686 .1281 So change 1 ) fe 2 u 102.4 1.4679 0.2285 Q change 22. " 64,0 1.4706 “uen darker, more viscous hs 76.1 11,4755 About the sams eas Wo. 4. oi toad he . iPS Oo ren) oy C2 aa £4) 3) 8 ‘we a | ed uyQ ae cr “g o Q Q " 1.4592 Slightly darker than Wo. Or | ed mn ed nie b. * Ci Py " 1.4695 - sliehtly darker than No. 7 i av) i if ~J oO af slightly derier than lo. 8 it Tales i and t 3 e Bs ~J ond Oe) 2 4725 slightly darker than Mo. 10, small not. i n 1 an mH a) 3 re) tae? Rel 4) 2 FPeated in test-tuba 98.5 11,4691 elosed with eork. oO 29 O i a o ~ 24 Heated in Feet tube closed with capillary. 102.9 1.4680 " a Refract Temp. Time Condition Cat. 1 No {Index Acid No. Remarks BY5~3 2: Heated 5 101.1 Darkened, floculent ppt. - sad with more viscous. it Ne wr * oy 1.4578 Bo nar ns ~ oF a I. i or MS to ts a 2; ook nN cr 3 an ior °. Ke a) . ,, nN . 0.274 yok Sad ~<a 3 od al nO = e, i) oJ °C + 3 os “ i : et W i ti 175-200 3 Yeaated in flask 5 O6.5 Yo A& KM. . ra tot 2 OK LEE No chanra 220-50 es COp passed thru % 105.9 1.4660 No chanrea. 2 a tx a co) nN . J _ 3 O % r?) ee ble ct ah 2 t 3, a 2) Ry “ neces cous, cous, 2O opt. N Senet oO ol * C3 aa e Fabs 7 Od ce x 195-215 * H Ww f . f 1 fe su 4 Olen n' xe we ae oS oe B15-35 86 ' i " B 96.0 1.4695 still serker, more vis ecus, small pot. ' " 3 9 Slightly darker enc more O71 eG S n " a O71 Owe ETI tN ELEY ON Oe ae ae 7Q 85 ie] an - viscous. 290-310 2: & Atm, Dress. 5 | _ — . fused) 76.0 1.4716 9.56 Very der, viscous, oil epéonec on standing. 3l5-35 22 = " 4 92.6 1.4692 some darker, viscous, td. not SES os smell pot. Yory Carl, and viscous, large ppt, ea oi i NO ah no cole oie ——, cr va 3 Oo sav) * ee . et {4 * oa madd * 1 =i Do { th The first experiments made were to heat the raw oil in 7 inch test-tubes exposed to the air, to various tenper- atures in order to ascertain huw heating alone effected the “oil No. 1, 2& 3 were heated to a temperature of 175°C and p showed that the oil was not effected very much at this temper- aturee There was however a small decrease in the Iodine number showing that there was some oxidation, and a small in- erease in the acid value showing thet there was a little de- composition. These effects were too small to be shown by the refractive index. There was no change in the physical appearance of the cil. In No. 4 and 5, the oil was heated to a higher tempera- ture, in the first case without, and in the second case with a catalyzer. Both were heated exposed to air, for the sane time and at the same temperature. The Iodine mumber was lowered and the refractive index raised, but the effect was greater in the one containing ths catalyzer then in the other one. ‘the only difference in the conditions of the two was the catalyzer and so it appears that it helped in the oxid- ation of the oi1t 1. Tne oxidizing effect of the catalyst may be due to the oxygen in the catalyst. If 2% catalyst is used, there would be 0.4 gm. of nicke, which if entirely oxidized, would furnish 0.109 gm. of oxygen. This would he suf ficient to reduce the Iodine number to 55.6. ; Assuming that the reduced catalyst contained 20% ox- yeen, there would be sufficient to decrease the fodine nunber to 94 .5. Experiments 5, 7, 5, 9, 10 and 11 were made by heat- ing the of] in open test-tubes et a temperature of S15 _o a to S385 C end takine samples every half hour. Toe pefractiv: index of these sarnles showed a gradual increase snowing trat the amount of oxidetion denencs wnoen tre lLengeth of time that the ofl is hastad, The previous tests were all made with the ofl exposed to the air but as the attempts to dehydcroranate would be mn made in a closed tubs a series of experiments were next mace bee it “ by heating the oil in tubes with only a small amount of al t i 1 present. Experiment 12 was mace by closing a test-tube with a cork to prevent oxidetion from exposure to the air. The high temperature charred the cork so much that another method of excluding the air was necessary. Wrapping the cork in as- bastos oaner, end coverings the surface with a coating of gpound asbastos in sodium: silicate, nelpecd somewhat, but was not entirely satisfactory, To overcome this difficulty a specisl heating tube was 2 4 made as previously deseribed, and used for the remaining procedures excent where otherwise stated. 1. ‘yhen the tube was filled there was about 20 cc. of air left in it. Por an oil with an Todine number of 105 it would take 4435 ee. of air to completely oxidize the oil. Thua 20 ce. of air would lower the Todine number about 2. ©. points. Heating in closed tubes with out any catalyst gave a reduction of the Iodine number of about this magnitude. 15. In experiment 15, the tube was closed by fasten- ing ea piece of capillary tubeing having a bore of inm. : and about 20 cm long, to the ends of the heating tube. This allowed the air to expand upon heating but prevent- ed the diffusion of the outside air into the tube to take the piace of that which might be absorbed by the oil. as shown by thse Iodine nunber and refractive index, the capill- ary tube prevented oxidation more than the cork did. The experiments so far showed that heating in the air as high as 330°C caused some oxidation, and a little de- composition with liberation of free fatty acids, but that if the air was excluded the oxidation was small. The next step was to prepare a catalyzer. The first catalyzer made was prepared from directions given by lir. Trowbridge. The formula used was; 1 part nickel (Slectrolytic, aporoximately 100%) 4 parts nitric acid (C. FP Concentrated) 4 parts distilled water (Free from chloride) 3 parts kieselghur (Best grade and washed with acid) The nickel was dissolved in the nitric acid, dilutéd with the 4 parts of water, and the kieselghur {previously washed for two days in 2% nitric acid) and 4 liters of water added in a 6 liter flask. C. P. ammonia hydroxide was then added until the nickel dissolved and the solution turned a deep blue color. <i ad -s 2 oS t hart The flask was then heated at a temperature of SO°C and sir blown thru until the nickel was precipitated on the -kieselghur. The precipitate was then washed thoroly with my a liped wnttl it turned dark gray. It was an) hot weter unk then ground to a fine powder with a mortar and pestle, and reduced in the reduction tube (previously described) at a temperature of 500°C in a stream of hydrogen gas. The re- @uced nickel was allowed to cool in an atmosphere of hydro- gen was then poured directly into cottonssed o11 so that 10 em. of the mizture contained 1 gm. of the nickel. A solution of nickel nitrate was made up, and this was used for making various batches of catalyst as described above. In the case of a later bateh, the nickel was pre- cipitated by sodium carbonate instead of by anmoniun hydro- xice. The precipitated nickel carbonate was treated as be- fore. Catalysts Wo. 1 and 2 were prepared from the same nick el hydroxide but were reduced at different times. No 5 was prepared from the nitrate solution. No. 4 was obtained forn the Globe O11 Mills No. 5 was prepared from the nickel car- bonate precipitate. No. 6 was a sample of nickel oxide pre- pared from the carbonate by heating. It was not reduced. ixperiments 16 to 20 were made to test the catalyzers for activity. The tests were made in the heating tube and in a special fiask. 7. As Was shown by the Iodine number catalysts No. 1 -and 2 were only very sligntly active. No 3 and 5 were as active as No. 4 which was ontained from the Globe Oil Millis ~ ana was known to be active. In experimsnts 19 and 20 cate alyst No. 3 reduced the o1i1 to a hard fat. The flask used in experinents 18 and 19 for testing the catalysts was a conical 300 cc. flask fitted out with a 3- hold stopper carring a long tube for bubbling gas thru the oil, a thermometer and a ahorter tube for an outlet with a ballon attached to the end. - To A, Tank The flask containing the oil and catalyst to be tested was put on an elect ric hot-platse and heated to a temper- ature of 180°C (temperature of maximum hydrogenation). Hydro- gen was then bubbled thru the oil from the cylinder until the ballon was partially filled and assumed the position shown by the dotted lines. The flask was then shaken and the ballon would collapse, due to the hydrogen being absorbed by the oil, if the catalyst were active. kixperiments 21 - 26 were heated at the same temperature bubbling carbon dioxide thru the oil. This served to stir 18. up the catalyst and prevent it from settling out at the pottom. It also displaced the air and prevents oxidation and the bubbles going up thru the oil might tend to drag hydrogen along with it if any were liberated from the oil. In all cases the Iodine number was increased to only a slight extent and the refractive index was only slightly reduced. There did not appear to be any dehydrogenation of the oil, or if there was it was obscured by the oxidation. iixperiments 16 and 17 were each carried out by heating the oil at the same temperature, for the same time, but 16 had a catalyst, and 17% did not. In this case again the catalyst seemed to have assisted in the oxidation of the oil as shown by both the Iodine number and the refractive index. axperiment 25 was carried out by heating a sample of the oil to a temperature of 270 - 285°C for 6 hours, bubbi- ing carbon dioxide gas thru the oil while heating. lven by heating for this length of time there was no appreciable dehydrogenation, altho the oxidation was not very large. =axperiments 26 - 28 were heated under # atmosphere pressure. In all of thes e€ a precipitate was obtained that appeared to be stearin but it was not definitely proven to be soe In experiment 20 the Iodine number was reduced to 76.6 by heating to a lower temperature than in experiments 27 and 28. This was because a catalyst was used that had already been used. in hydrogenating and some of the hardened 19. ‘oil still remained in the nickel. This sample harden- ed upon standing. . These last three oils all became very viscous and it is thought that this was due to oxygen diffusing back thru the vacuum line. fo test this out a capillary tube of lm. bore and 20 mm. long was placed in the vacuum line to prevent this. However practically the same changes occured as when this was not used, thus disproving the assumption. At temperatures above 300°C the ofl fumed considerably in air, SUMMARY. Le Samples of cottonseed o11 were heated with and without a catalyst and it was found that the Todine number decrcas= ed, the acid value and the refractive index increased, and that the oil darkened and became more viscous. This show- ed that there was oxidation and some decomposition, with a liberation of free fatty acids. Qe At temperatures below 260°C the effect was very small on e1ill of these properties. Se As the temperature increased the oil darkened and the Todine number decreased, and the acid value and refractive index increased. These effects inersased more rapidly above 260°C. be As time of heating increased the above effects increas- ed. (20. Be In closed tubes the change in Todine number and refractive index was less than in open tubes. The other -ghanges (increased) regardless. Ge The addition of reduced nickel or nickel oxide appear- ed to increase the oxidation of the oil. The later had more effect than the former. Te Catalysts were prepared and found to be catalytically active as shown by their ability to harden the cottenseed oil. 8. Attempts were made to dehyd rogenate the oil by heat- ing to a high temperature and passing carbon dioxide thru the tube. They were unsuecessful however and oxidation took place, probably caused by the oxygen (3%) in the carbon dioxide gas. 9. The oil] heated in the carbon dioxide atmosphere was not oxidized as much as if heated in air or in a closed tube. This was because the carbon dioxide gas displaced the air. 10. Samples were heated under a pressure of abput % atmos- phere and it was found that the Iodine mumber was reduced, and the acid value and refractive index increased to a great- er extent than in the other experiments. 11. Inserting a capillary tube liom. in diamiter and 20 em. long into the vacuum line did not oive different results than when it was not used. 12. At temperatures above 200°C the oil fumed and it is probably that under reduced pressure part of the more vol- atile portions of the oil distilled off. This would account ele for the large changes in the fodine number and the very large inersase in the acid value as well as the increase in viscosity. 13. Samples heated to high temperatures were found to’ grad- ually form a precipitate upon standing. ‘his appeared to be stearin but this was not definitely pyoven to be so. 44, it was found that the oil was oxidized and decomposed ‘before a high enough temperature could be reached to bring about a dehydrogenation of the oil or else that any hydro- ‘genation obtained was obscured by the oxidation.