The Dehydrogenation of Cotton Seed Oil Citation House, Harvey Walter (1920) The Dehydrogenation of Cotton Seed Oil. Bachelor's thesis, California Institute of Technology. doi:10.7907/h8am-bw96. https://resolver.caltech.edu/CaltechTHESIS:07232024-204630794 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: Unknown, Unknown Defense Date: 1 January 1920 Record Number: CaltechTHESIS:07232024-204630794 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:07232024-204630794 DOI: 10.7907/h8am-bw96 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 16585 Collection: CaltechTHESIS Deposited By: Benjamin Perez Deposited On: 24 Jul 2024 18:15 Last Modified: 24 Jul 2024 18:15 Thesis Files PDF - Final Version See Usage Policy. 6MB Repository Staff Only: item control page

THESIS THE DEHYDROGENATION OF COTTON SEED OIL BY Harvey W. House Class of 1920 Presented in fulfillment of requirements for the degree of Bachelor of Science in Chemical Engineering INTRODUCTION It is well known that vegetable oils can be hydrogen~ ated by passing hydrogen thru them when they contain a proper catalyst such as finely divided nickle. This thes:ts represents an attempt to reverse the a~ove reaction so that ·hydrogen would be removed instead of added to the oil. The object of dehydrogenation would be to form an oil having greater unsaturation and hence more chemical acttizlity, which might render it useful i-n the preparation of other compounds and mtght also increase its drying properties. Since hydrogenation with finely divtded _nickle as a catalyst proceeds best at a fairly low temperature there was some likelihood that the reaction might reverse by sim~ ply heating the oil at a higher temperature in the presence dS of the same catalyst. The following facts also support this view: 1 1 . Zelinski dehydrogenated cyclohexane by means of heating the latter in contact; with finely divided platinum black. 8 Sabatier 8 showeddthat:lce:r.taitnt br.gant.~r: .compoun.dsnare reducad ·) by;:.passdmg their vapours mixed with hydrogen thru tubes containing finely divided metals, especially Ntckle, at a temperature of about 180 degrees, and that at 250 degrees •the same metal brings about an oxidation, for example, of alcohols to aldehydes or ketones. 1, A, 1911,1,958. 2. A, 1911,1,702. 2 3 The heat of combustion of carbon compounds increases wi.th the unsaturation. According to a general principle, a compqund which ms a high heat of oxidation Bhould be more stable at a hi,gh temperature than one which has a low heat of oxidation for it is difficult to give off heat to surroundings a# a hi-gh temperature. Therefore an unsaturat:ed oil might becmme more unsaturated by heating to a high temperature in the presence of a catalyst which would assist the reaction. Since the iodine number of an oil is an index of its saturation, :tt was decided to determine the iodine numbers of-the va~ious samples worked with. The object of the experimentation then became to increase the iodine number of the oil. The Hanustmethod was used. Precautions given by Smith and Tuttle 2 were observed. Saponification values were also taken by the method given in Leach, p.379. METHOD It was decided to try first to prepare a nickle catalyst by reducing a salt or oxide of nickle in a stream of hydrogen, following a general method used in the preparw ation of nickle for hydrogenation, such as given by Ellis. 3 Samples of oil were then heated without a catalyst for cert ain lengths of t ·i-me to determine the effect of heat alone upon them. Some samples were protected from the air and some were not. Samples were then heated at similar temperatures and for similar lengths of time in the presence of catalysts with or without the passing of hydrogen. In thE , p. 383, 1. Leach, Food Inspection and Analysis, 2. Tech. paper, Bur. Stand. No. 37, "Iodine Number of Linsesi and Petroleum Oils" ' Smith and Tuttle. 3 Ellis, Hydrogenation of oils, p. 54-58. • 3 way the effect of the heat alone would be known, and could be differentiated from the effect of the catalyst. APPARATUS It was first necessary to obtain a source of hydrogen which would contain no trace of arsenic, chlorides, sulphur, or oxygen, phosphorus, ammonia, or oxides of nitrogen, as these substan• ces would poison the ntckle . 1 To this end a_n electrolytic hydrogen generator was constructed consisting essentially of three U-tubes joined together in parallel, and i-n each side of which was an electrode of sheet iron made from tin-can material. Each electrode was smldered to an iron wire which in turn was soldered to a platinum w&re. The latter was sealed into the side of the tJ-tube and connected to the outside source of current. Ordinary solder gave a mirror test for arsenic, so could not be used. Silver solder, however; was used without analysis. The cathode arms of the tJ-tubes were joined by a manifold so that all the hydrogen generated was brought together &nto one stream. A vertical glass tube about a meter long was sealed to the anode arm of each U-tube so that the electrolyte could be forced up until its hydrostatic pressure would force the hydrogen thru the rest of the apparatus. In In series with the hydrogen manifold was a stop-cock tube, two wash tubes containing alkaline pyrogallol solution, a tube containing soda1:til:.im:e; cl ar;i.d a bulb containing cotton. The stop-cock was used to form a constant resistance to the flow of hydrogen so that the variable resistance to the flow 1 f. ,:· Cata;t;ys'.:i:S 'Jinh Tntlustr1a;ts Cnemhitry. ,·Henderson, ·• p ..;1.03:::·l t '.) f 4 occuring thru the apparatus would be only a small percent of the t,otal resistance, which would make the flow of hydrogen more steady thruout the apparatus. The alkaline pyrogal ::. o::. was placed in the path of the hydrogen in order to absor~ the oxygen whiqh might diffuse over from the oxygen electrode tnto the hydrogen. The sd>d&n1.i me and cotton were to remove moisture and suspended impur1ities. The various parts of the '. apparatus were sealed together so. that .all joints were of glass and all the parts were made of glass. This was done to prevent traces of impurities from the apparatus itself contaminattng the hydrogen . . An approximately 25% solution of sodium hydroxide was used for the electrolyte. The generating current was taken from a s~orage battery having an e.m.f. of 2welve volts. The electric furnace used for maintaining the necessary· II temperatu~es consisted of a silica tube of 2 1/4 diameter and 15 II inside long surrounded with resistance wire and magnesia and asbestos insulation. This tube was placed in a vertical position and arranged SJ) that it could be placed at any de s ired height. Its lower end w~s closed with asbestos. Heating current was furnished at 11b volts. A. thermostat was constructed and used during most of the experimentation which kept the temnerature at any point desired within a variation of about 10 degrees. 5 EXPERI ME NTAL Samples of oil were first heated in the furnace to determine the effect of heat alane upon 'them. Catalysts were then prepared and attempts were made to hydrogenate and dehydrogenate samples of oil. A semi-tabular statement of the experiments carried on and the results obtained follows on the next page. This will in turn be followed by a more detailed discussion of the work done, and by a summary of results. 6 Remarks about catalyst Temp. Time Remarks about oil C hrs. o No datalyst Temp. Time I no. S~. C hrs. K Untreated oil ... . . . :; 108. 3 195. 4 106.4 195 193 1 No catalyst Heated, covered with glass wool 320 4.5 71.8 74,4 2 No catalyst Heated, covered with cork, held by small springs 360 5 94,5 96,7 3 No catalyst Same as above 290 5 104,9 105.9 4 No catalyst Same as above 250 5 106.7 106.9 5 No catalyst Same as above 350 5 104,9 105.0 6 Reduced Ni(OH)2 made from NaOH and Ni(N03)2 7 Reduced.Ni(OH) 2 on pumm1ce 320 5 Tt. half full. Passed H2 200 5 93.3 93,9 240 6 Passed H2 170180 3 106.6 106.6 Not 8. nal. 8 Ni(No) 9 changed to basic carbonate with Na2C03 and .reduced 300320 9 Oil heated with cat. Covered w glass wool. Consistency vaselene. 9 Basic carbonate pptd. from very 'dilute sol' n 310- 5 Covered with slit cork. 220240 2,5 10 Cu soap .prep'd in dil. sol'n,cole. Light blue ppt. Most of s ample spattered out, due to H2B 320 109.1 min, 103.2 11 Cu soap prep'd as above 350Small amt. soap dried and dis. in 360 oil. Dec. on heat., leaving fine suspension. 5 90.7 90.3 Ni soap in oil decomposes, fine suspension 6 88,8 88.3 12 Prepared Ni soap as abovB 320 350360 107.3 106.4 5 7 All the heating af catalysts and oil samples was done in five inch test tubes suspended vertically in the furnace. The various precipitates which were reduced were placed in the test tubes and hydrogen passed over them while in the furnace at the temperatures indicated. The test tiube used with samples 6 and 7 had the hydro~en tube sealed to its lower end from whence the latter was bent around and upward, passing out of the top of the furnace. The hydrogen therefore bubbled up from the bottom. The o i. l was added to the same tube that the catalyst was prepared in. Between samples the hydrogen tube was broken neaE the furnace and the oil-tube removed from the furnace f'or clean ing, after which it was sealed back onto the aparatus. Because of the inconvenience of this method, the rest of the experiments were carried on by placing the catalyst or oil into an ordinary test tube and passing hydrogen by the use of a straight glass tube leading down into the test tube and nearly touching the bottom. I Samples 10, 11, and 12 did not re quire the passing of any hydrogen and were kept from exposure to the air by being covered with round disks of grounq glass. The top of each test tube was also ground to a plane surface so that a tight joint was made simply by the glass disk resting upon the top of the test tube. This joint was made especially air-tight by the oil itself. The catalyst of sample 6 was made by reducing nicklehydroxide formed by adding normal sodium hydroxide to a solution of ~O cc. nearly saturated nickle nitrate solution 8 in 100 cc. water. The precipitate was filtered off, placed iri. test tube and reduced in electric furnace by a stream of hydrogen from the generator. The temperature was maintained at 200 degrees and hydrogen passed 5 hours. The test tube was stoppered with glass wool. A fair suspension of nicklm was obtained but the particle,s were large enough to be easily seen with the naked eye. There were undoubtedly some finer • particles present, however, The oil used in this experiment was partially solidified the day after hydrogenation, It was warmed and centrifuged before its iodine number was obta ined. In sample 7, 7 grams of powdered pummic,e stone were added t ,o 600 cc. water, and the suspension so formed was boiled and decanted. 30 cc. nickle nitrate solution and 600 cc. more of water were added to this decanted pu.mmice stone. The nickle was then precipitated with 1/5 normal sodh1m hydroxide, which gave an intimate mixture of nickle hydroxide and pu.mmice. This was filtered with suction, dried in an evaporating dish, and a considerable amount poked into the test tube. Silver nitr a te was then used to test the sodium hydroxide and nickle nitrate for chlorides. The sodium hydroxide anntained enough chloride to make a light milky silver chloride precipitate. Some sodium bicarbonate and sodium carbonate was then tested. Of these the sodium carbonate gave the faintest test, which was mu.ch fainter than the test given by the sodium hydroxide formerly used. This sodium carbonate was thereafter used in precipitating the nickle. 9 After finishing the work on sample 7, a basic nicklecarbonate precipitate was prepared in the following manner, using dilute solutions to prevent the occludLng of impurities, especially chlorides, as much as possible. Dissolved 5 grams sodium carbonate in 300 cc. water, added to a solution of 20 cc. nickle nitrate in 300cc. water, slowly with stt.rring. A very fine precipitate sepe,'rated which was filtered with suction. Dried the above over sulfuric acid in a dessicator for six days, placed in tube and in furnace, and passed hydrogen. Later, during the heating of the oil, the thermostat broke which allowed the temperature of the furnace to go high enough· to boil off and char all the oil and distil the mercury of the regulato~ to the outside of the furnace. After this the temperature was adjusted by regulating a lamp-bank resistance by hand, which @ethod gave a control of temperature which was sufficiently accurate. The catalyst used in sample 8 was prepared in the same manner as that used in sample 7, except the.t the precipitate stood in the dessicatCT' onJ_,Y one day. The catalyst in its final form was black, in small chunks in the bottom of the test-tube. An attempt was made to grind up these chunks with the end of the hydrogen tube. This turne,d the niOkle into a faiily fine powder, but not fine enough to form the kind of suspension desired. Oil was added to within one inch of the top of the test tube, and hydrogen was passed during the additron. The hydrogen tube was then removed and the test tube stoppered with glass wool which was saturated with the oil so as to exclude the air better. After heating this sample for 14 hours at 310-320 degrees, it was dark brown, nearly opaque, and when 10 cool of about the same consistency as vaselene. From this it appeared that in spite of precautions t aken, the oxygen of the air had oxidized the oil, for sample 2 which had been heated even up to 360 degrees for a period of five hours ·, when stoppered with a cork, showed almost no change in physical appearance. The iodine number of this sample, therefore, was not taken as it was bound to be very low duEeto oxidation, just as in the case of sample 1. The possibility that the great increase in the viscosity of the oil could have been due to the action of the catalyst and not to the air is eliminated by the failuDe of other samples to thicken when heated with a catalyst and well protected fro~l the air. The catalyst used in sample 9 was prepared- with much more dimmte solutions than formely used in order to get a purer precipitate. Dissolved 5 grams sodium carbonate in 2 L water- j:andl i. aft!sd 20 cc. nickle nitrate in 4 L water. Added the sodium carbonate solution to the nickle nitrate solution slowly with stirring. Decanted, added 2 L water, stirred, _dedanted again, filtered and stood over night on filter. Filled test tube about· half full of the voluminous precipitate so obtained, covered with a slit cork thru which the hydrogen tube protruded, and left in furnace for a week. Then passed hydrogen, 310-320 degrees, 5 hours. Black solid in small lumps settkd to bottom, with smme fine dust mixed mn. Added a half test tube of oil, passed hydrogen at 220-240 degrees for 2 1/2 hours. Centrifuged before obtaining iodine number. The iodin~ number showed the oil to be saturated to practically the same 11 extent as the untreated oil, so the catalyst evidently had no effedt in increasing the unsaturation. It was then thought that a much more finely divided form of nickl e mi ght give the result desired so it was detUded to try to decompose a nickle soap in solution in the oil. This e_x periment was first carried on 'flith the copper. A potassium soap was first made by saponifying 30 grams of cottonseed oil, (the same oil as that used thruout the ex periment), with 7 grams of potassium hydroxide dissolved in water. The mixture was stirred and heated until a nearly solid mass wa s formed. There was considerable unsaponified oil left, altho the amount used wa s calculated from the saponificati on constant of the oil. This uns aponified oil was drained off and the soap dissolved in 300 cc. water and heated to boiling. Dissolved 24 grams copper nitrate , Cu(N03)2.3H28, which ws equivalent to 6 grams potassium hydroxide, in 200 cc. water, added slowly with stirrirg to the soap solution which was still hot. A black, tarry precipitate separated out and floated on top. Poured out the water and placed 20 drops of the tarry matter in a test tube half full of oil and placed in furnace, in which the temperature was about 400 degrees. In about half a minute there occurred vimlent boiling and slight gas explosions inside the furnace. It was thought at -first that the.se explosions mi ght be due to gaseous decomposition products of the oil and possibly even to hydrogen, but subsequent experimenting whowed that '. the phenomena was probably due to the vaporizati on and burning of the oil as it spattered onto the hot furnace. :12 A similar soap sol_ution to the above was again prepared, but dissolyed in 800 cc. water instead of 200 cc. 50 cc of this solution was then diluted to 500 cc. with cold water. To this was added slowly, with stirring, a a solution of 2 grams coppernitrate in 300 cc. water. This time a beautif~l light bl1-e2 precipitate was obtained, which coagulated well. After decanting and filtering, about 1/2 cc. of this precipitate, undried, was smeared around the inside of a test tube, oil was added until the tube was 2/3 full, and the precipitate was caused to dissolve by being stirred into the oil with a glass rod. Heated this sample, which was numb:er 10, five minut:es at 320 degrees, during which time most all- the oil spattered out because of the water which was introduced with the predipitate. The_ oil which was left was dark greenish brown and opaque. For sample 11 a perfectly dry test tube was used. About 1/10 cc. of the samf3 preci,pi tate was smeared around the inside O'f the t-est tube and sliglitlf heated · to drive o'ff some of the mmisture. A few cc. of oil was thenadded, the precipitate dissolved, and the solution heated over a bunsen flame until all the popping due to evaporation of water had ceased. The test tiube was then nearly filled with oil and covered with a disc of ground glass as previously described, making an air tight joint. Heated to 350-360 degrees for five hours, after which the oil was observed to contain a very fine copper suspension. The suspension had the same color as copper itself, the particles could not be se en with the naked eye,. and the suspension was nearly opaque. The iodine number of this oil 13 showed that there was a decrease in the unsaturation rather than an increase. Several attempts were then made to prepare a nickl~ soap using the nll:ckle carbonate which had been precipitated for use with_ sample 9, as it was thought that this precipitate was especially pure. The nickle solutions thus prepared gave emulsions when added to the soap solutions. This was probably due mostly to acidity of the nickle solutions so prepared, which liberated the fatty acid from the soap. A nickle soap was finally prepared from a very dilute solution of nicklenitrate which coagulated in the same manner as the copper soap previously prepared. Similar amounts of solution were used as when the copper precipitate was prepared. The nickle preci_p~tate wa s decanted and filtered. About 1/10 cc. of i t was smeared around the inside of a perfectly dry test tube, heated, dissolved in a small amount of oil and heated over a bunsen burner until all water was driven off, in the same way as the last copper precipitate was treated. The test tube was then nearly filled with oil, covered with a glass disk, and heated at 350-360 degrees for six hours. The nickle soap did not decompose as rapidly as did the copper soap, but: finally gave nearly as good a suspension, which was black and nearly opaque. There waBe some larger particles than in the case o:f the copper and these settled to the bottom of the test tube, but most of ta the nickle remained in a fine suspension. The iodine number of this sample, however, was even less than that obtained from sample 11, sh owing that no dehydrogenation had taken place, or else that whatever dehydrogenation that may have taken place 14 was more than counteracted by other actions of an oppostte nature. In connection with these last two samples containing the copper and nickle suspensions, wh-ich ·were heated to 350-360 degrees for 5 and 6 hours respectively, it is interesting to note that their indine numbers were reduced to about 90 and 88 respectively by the treatment, whereas sample 2 which contained no suspension had its iodine number lowered only to 95 and sample 5 which also contained no suspension and which was heated to 350-360 degrees for 5 hours did not have its iodine number changed appreciably at all .. It seems therefore p : probable that the presence of the copper and nickle suspensions in the oil must have assisted chemical changes within the oil, resulting in lowering the ct6din~ number to a greater extent than when no suspension was used. ' An interesting experiment wa s done with the cottonseed oil which d~d not have any particular bearing upon the problem of this thesis but which may be worth reporting. 100 cc. oil ·was placed in a 250 cc. distilling flask which was provided ~·'" with an a ir-cooled Leibig condenser. Compressed air was forced thru the co ndenser from the compressed air line. The oil was heated with a bunsen burner and three fractions were distilled off. During the distillation large quantities of acrolein were evolved due to the decomposition of glycerine in the oil. The oil began to give off "steamy" vapors at 303pdegrees, bubbled slightly at 350 degrees, and started to boil at about 360 degrees.The thermometer bulb in the above instances 15 - . was immersed in t:tie oil, ·and th:e, Eeadings are not corrected. ThB following readings were taken with the bulb just below the outlet of the ctistilling flask~ The temperature remainea at 350 degrees during practie.ally the whole ot the ct.istillati9n. Three portions of distillate were collected, as follows: Frae- voltion ume Color Physical state Saponification K I 12cc cream like solid cooking 1·at 166.6 - l66.1 II 25cc straw ii 170.6 - 176,2 III 20cc amber " 171,3 - 172.8 Resi- 25cc ,.due greenblack tarry 83.4 ... 89.6 It was very difficult to titrate the residue in getting lts saponirication value, due to the very dark color of the solution. Tne most n interesting result of the above experiment was that each distillate crystallized on coibling into a solid fat. 'I'his may have been due to an increase in the percentage of free fatty acid, due to ctecomposi tion of the glycerides, which W'.:mld raise the mel tin~ point, and also to polymeriz~tion. Free tatty acids would raise the saponificati on va lue, however, but since the latter value was c ons i derably less in the distillates than in the original oil, there is an incticati on that most 01· the change was due to polymerization, smnce only the latter explanation will accouht tor the decrease in the saponification value. This also would explain the very low saponu•ication value 01' the residue, tor pmlymerization products occurring in the 16 distilling portion woula have a higher boiling point than the rest of the oil and hence remain over in the residue. Unsaponiria- ble decomposition proctucts may also have · been 1·ormed. SUMMARY 1 Samples o:r cotton seed oil were heated in test tubes protected trom air, each tor a period of five hour--s, a:A.t temperatures of 250, 290, and 350 degrees .tbe;·-ioctine '.!!urnber was but very slightly lowered. At 360 degrees it was lowered about b per-cent. It was also shown that at 360 degrees distillation and decomposition or oil begins; but at temperatures below 3b0 degrees heat alonw has but little effect. (Sa~ples 2,3,4 and 5.) 2 Sample heated 4 1/2 hours, 320 degrees, not protected from air (stoppered with glass wool) had iodine numeer reduced to about 73. (Originally 107). Since the introduction of air was the only variation between this and the samples heated while protected from the air, this reduction of the iodine number must have been due to _oxids.tion. ( Sample 1). 3 No satisfactory hydrogenation and no dehydrogenation was obtained by heating the oil in the first case with and in the sec ond case without the passtge of hydrogen, using nickle catalyst prepared by reduction of basic nicklJ carbonate in stream of hydrogan. (Samples 6. 7. 8. 9,) 17 4 Very fine suspens i ons of copper and nickle were obtained by making cotton seed oil soaps of copper and nickle respectively, dissolving these in the oil, and decomposin~ by heating. These s amples were well protected from a ir and heated to 350-360 degrees for 5 and 6 hours respectively .. The iodine numbers of the oils were lowered to about 90 _and 88. Since oil con taining no suspension and heated to 360 degrees continuously for 5 hours (s ample 2) had iodine number / lowered only to about 95 and the iodine number of oil heated 3 01) ~t to 350 degrees , 5 hours (sample 5) was not affected, the presence of the copper , or nickle suspension in the 011 must have assisted chemical changes within the oil, resulting in lowering the iodine number to a grea ter extent them when no suspension wa s used. 5 Yvt~ G-e. y..d,~ 1; 7¼ .~tt.t,t/,y ()-1.,~ , ~...U. V'i""~,,.J..& ~ ~I ,.. , bf._. µ-1~ ~~• v>17ALJ{ k,, -€,r,..~ ~ A sample of cottonnseed oil distilled with decomposition at 350 degrees, giving off large quantities of acrolein. The saponification constants of the fr a ctions distilled inc~eased slightly as the distillation progressed, but the s a ponification constant of the residue was very low. This result may have been due to decomposition of the glycerine, giving rise to a higher proportion of free f a tty acids in the distillate and a lower proportion in the residue, to polymerization, and to formati on of uns a ponifiable decomposition products in the residue.