THE ACTIGN BF ALUMiNUM CHLORIDE 0N DffliENN’LETE-WL CAREEML THESIS FOR THE DEGREE OF M. 8 Stuart Vt Wilsey 1933 Wm mum I05! ”9“" uusmo “'5‘ ,b' i E y . W‘ H, - y > ..r|..|'.n I ll 1 titakqlls‘l |1 I ‘Illlf ' . . ,1 y A. l|PI I L\ 7.. I y] '|..Ir.|. V II'Iu’ bp|. DEE [Fillirllplil‘llg v >1 TfUZ ACTION OF ALE-73v II CHLORIDE ON DI PH '- ‘IEYL: .‘THYL CARRIED Tia: l-xc'r'mn 0F ALMS rm: Cl'lgig’zliygg UN TWIN,” II"""*_‘-nr"n C H 3...”. U 4¢L.u’1,{.l.. ,‘A‘.;Au L3‘1,L‘..‘_‘I‘ A Thesis Submitted to the Faculty of the Fichigan St to College of agriculture and Applied Science In Partial Fulfillment of the Requirement for the Dacree of Nastur of Science by Stuart V. Wilsoy Department of Chemistry June 1933 I _I - f" . \ «o. \I.‘ ACKNOWLEDGHCNT The writer wishea to express his gratitude for the helpful and inspiring counsel given by Dr. R. G. Huston while carrying out this work. 5 CD «E 331. COI‘ET? II'PI‘S Introduction Historical Preparation of l. l, 1, Triphonyl Propane Erickson'o Work Uses of Aluminum.0hlorido Dehydration of Carbinola EXperimontul Preparation of Diphenylethyl Carbinol Initial Condensations Preparation of Triphenyl Chloromothane Kittie Reaction Preparation of 1, 1, 1, Triphcnyl PrOpane Identification of Product Factors Influencing Dehydration Discussion Smmnnry Pac- ‘v'- \J 00!“ 10 13 14 16 18 19 20 21 27 51 -1- II'I’I‘Z {ODUCTI OH This problem,erose after the attempts of Harry firickson, working in this laboratory, to prepare 1, 1, l, triphonyl prepane by the conden- sation of diphenylethyl carbinol with benzene in the presence of alum- inum.chloride, gave results other than those normally expected, For some years, the condensations of various alcohols and other aromatic compounds in the presence of aluminum.chlorido have been studied. Until this time, no work has ever been reported on the attempted conden- sation of tertiary aromatic alcohols containing both aryl and alkyl groups. If this particular condensation had occurred as normally expect- ed, it would have fallen in line with the previous work of this type. However, as this work shows, the results are not as were expected. flavor- tholess, they are interesting and will certainly influence attempted condensations of this type in the future. -2- -3- Preparation of l, l, l, Triphenyl Propane 1997 Comberg in 1906 (Gomberg and Gone B. 39, BSGIS was the first to obtain 1, l. l, triphenyl prepane in a crystalline form. he succeeded in preparing the compound by treating triphenyl chloremethane with ethyl magnesium.iodide in an absolute other solution. fie assumed that the thick, light yellow, blue fluorescent oil that he obtained which boiled constantly at 248 - 249° at 41 mm. (228 a 253 at 19 mm.) was triphenyl propane. However, not until he had prepared the next hiuher homolog of the series, 1, l, l, triphenyl butane and obtained it in a crystalline form, was he able to succeed in crystallizing the triphenyl propane. The crystals melted sharply at 51°, and analysis showed that the com- pound wae pure triphenyl prepane. Any following preparation could be crystallized very readily by seeding with one of the pure crystals. Previous to the work of Gomberg, E. and O. Fischer (G. out 0. B. 39. 2962) had attempted to prepare 1, 1, l, triphenyl propane by adding zinc ethyl to a solution of triphenylchloremethane in benzene. After decomposition of the reaction product with water the residue was allowed to stand in benzene and from this triphonylmethane crystallized out. From twenty-five grams of triphenylchloroncthane they obtained, in the above manner, sixteen grams of triphonylmethano. Had a quantitative yield of triphenylmethane been formed, they shOuld have obtained twenty- two grams. ~4~ Gomberg duplicated their work and he, too, obtained triphenyl- methane, but the few grams of oil that were left were carefully fraction- ated and found to boil at 235 - 230° at 19 mm. after dissolving-the oil in alcohol and seeding with a crystal of triphenyl propane, it all cry- stallized and was found to be pure, 1, 1, l, triphenyl propane. The yield from twenty-five grams of triphenylchloromethane was only three grams. Harriet and St. Piers (C. and C. B. 59. 2963) state that uncrystall- inable products were formed when ethyl iodide reacted with triphenylmethyl potassium. Gomberg duplicated this work and found that the product In! " chiefly l, 1, 1, triphonyl propane. From twenty-five grams of triphenyl- methane and four grams of potassium he obtained eeVen grams of oil distill- ing at 228 - 233° at 19 mm. which was pure triphcnyl propane. It crystall- ized after seeding. ‘he trinitro derivatiVe of l, l, 1, triphenyl propane was prepared by Gonberg (B. 59, 2965) by treating one part of the hydrocarbon with fifteen parts of fuming nitric acid at about «10°C. It was interesting to note that nitration occurred on only one phenyl group at a time. Thus, in order to obtain the trinitro derivative, the compound had to be nitrated three times. The trinitro 1, 1, 1, triphenyl propane melted sharply at 194 - 195°. ~5- Hrickson‘s Work Barry Erickson, working in this laboratory, attempted to prepare 1, l, 1, triphenyl propane by the condensation of diphcnylethyl carbinol with benzene in the presence of aluminum chloride. However, instead of obtaining the desired compound he states that alwminum chloride had a degratory effect upon the carbinol, causing it to break down into benze- phenone and ethane. he believed that diphenylethyl carbinol on standing, would after two or three months be completely broken down in the above manner. Furthermore, that in the condensation.mixture aluminum chloride merely hastened this break down. -5- Uses of aluminum Chloride During the past few years, aluminum chloride has become a very important agent in the preparation of organic compounds. Probably its outstanding use has been in the Friedol and Drafts reaction, by means of which such substances as hydrocarbons, aldehydes, carbonylic acids, and ketonos have been synthesized. It has been used in dehydrogenation reactions, addition reactions, condensation between molecules by the elimination of water, introduction of halogen and nitro groups, and other processes that have theoretical importance as Well as practical value. he long ago as 1880, aluminum chloride was used as a dehydrating agent in organic chemistry. Here and Wcith in 1881 (B. 14, 189) succeed- L//// ed in preparing diphenyl other by refluxing phenol over anhydrous alumin- um.chloride in the cold or at ordinary temperature. Woes (B. 15, 1128) prepared diphenyl and triphenyl ethane by treating benzene with dichloroethyl oxide in the presence of aluminum chloride. Graebe in 1901 (B. 34 1778) was successful in preparing aniline by the action of hydroxylamine on benzene in the presence of aluminum chloride. nlthourh the yields were poor, Graebe was successful in using this method for the introduction of amino groups in several different mole- Oulea. Frankfurter (J. A. C. S. 37, 385) in the introduction to his work on aluminum chloride as a dehydrating agent, states that he believed that aluminum.chloride would have as great a use as a dehydrating agent, as did -7- same or the more common agents employed for the elimination of water beu tween molecules. He showed that aluminum chloride would bring about the same condensation between benzene, toluene, xylene, and chloral as would the other condensation agents used in the past. Until use was made of aluminum chloride as a dehydrating agent to bring about the condensation of aromatic alcohols with aromatic compounds, such substances as sulphuric acid, hydrochloric acid, acetic acid, abso- lute alcohol, phosphorus pentaehloride, etunnic cthrido, and others were used. although aluminum chloride has been used for dehydration by a few workers other than those mentioned above, the work of R. C. Boston and his co-workers has been the outstanding contribution to our knowledge re- garding the behavior of aromatic alcohols in the presence of other aromatic groups. huston and Friedman (J. n. C. J. 58, 3537) started the series of condoneations that were to follow by preparing diphenyl methane by the reaction between benzyl alcohol and benzene in the presence of aluminum chloride. Later the same workers (h. and F. J. a. G. 3. 40, 785) instead of using a primary alcohol as they had in their first work, showed that secondary alcohols would condense equally as well. Fran phenyl methyl carbinol and benzene they obtained diphenylethanc; from phenylethyl car- binol they obtained triphenyl prepane; and from diphenyl carbincl (benz- hydrcl) and benzene they obtained triphenyl methane. Having shown that both primary and secondary aromatic alcohols co ld be condensed with benzene, the field was broadened and derivatives of the above groups were employed. Huston (J. A. C. 3. 46, 2775) working alone, —8- showed that benzyl alcohol would condense with phenol in the presence of aluminum chloride to give p-hydroxy diphonyl methane . This showed that the hydrogen in the position para to the hydroxyl nroup in the phenol was removed. huston, Lewis, Urotomut, continuing the work on the condensation of secondury alcohols (J. A. J. J. 49, 1365), condensed phonylmcthyl carbinol and phenylothyl carbinol with phenol. Here again, the para hydrogen in the phenol molecule was removed and in the first case they obtained p- hydroxy diphenyl ethane and in the second p-hydroxy diphcnyl propane. Using a still more complicated group, huston, dwarthout and Ward- well (J. A. C. 3. 52, 4464) prepared the p-benzyl derivative of o-cresol by treating benzyl alcohol with o-croeol in the presence of aluminum.chlor- ids. Hueton and Lewis (J. A. C. 3. 55, 23 9) prepared the ortho benzyl derivative of p-cresol in the same manner. Inter, to complete the series on the benzylation of the cresols, huston and houk (J. a. C. S. 54, 1306) prepared benzyl derivatives of mpcresol and obtained three mono and two di derivatives in the same manner as described above. Saturated aliphatic alcohols have never been shown to condense with benzene, but the unsaturated aliphatic alcohol, allyl alcohol , was made to condense with benzene in the presence of aluminum chloride to give allyl benzene. (Huston and Sager J. h. C. S. 48, 1955). Other work carried out in this laboratory, and as yet unpublished, has shown that phenol will condense with phonylbutyl carbinol or with phonylpropyl carbinol under the influence of aluminum chloride, to give p-hydroxy diphcnylbutyl methane and p-hydroxy diphcnyl butane respective- ly . Benzaldehyde has been shown to condense with two molecules of benzene to give triphenylmethane. (Schaarschmidt, homann, Uzemo B. 58, 1914). All of the above described dehydration reactions attributed to the influence of aluminum chloride, have been intermolecular reactions and involve the removal of the hyurogen from one molecule and the hydroxyl from another. In the literature, I have found only one recorded example of an intramolecular dehydration due to the action of aluminum chloride. This reaction consists in the dehydration of 1 nitro 2 methyl anthra- quinone to give anthraquinone 1, 2, isoxazolc. (I. G. Farbenindustrio A. - u. Tu., D. R. r. 479 350). A brief summary of the recorded uses of aluminwm chloride as an agent for dehydration, shows that primary and secondary aromatic alco- hols or certain unsaturated aliphatic alcohols, can be condensed with other aromatic rroups. The temperature maintained, and the solvents used, greatly influence the reaction. She nature of tie molecule to be condensed also materially influences the OUtCOflO, chiefly in respect to the speed with which the reaction is completed and the amounts of the products to nod, rather than the directing of the reaction in an entire- ly different course. ~10- Dehydration of carbinole a short study of Various alcohols and their prepcrtiee in r,gards to the cane in which they may be dehydrated internally, shows that primary aromatic alcohols can not be dehydrated in tuie manner. A true secondary aromatic alcohol, benz ydrol, likewise can not be made to lose water Erma within the molecule. A mixed secondary alcohol; namely, phenylbenzyl car- binol, will lose a molecule of water by distilling the compound at atmos- pheric pressure. (findboronch doc. 67, 604). The difference between this carbinol and hcnzhydrol is that the latter does not have the second phenyl group attached directly to the carbon to which is connected the hydroxyl, but is attached to an alpha arbon atom, upon which there are two hydro- gene . It is not necessary that the secondary alcohols contain two phenyl groups in order to underpo this loan of water, for at high tenperaturce and under certain conditions those with one, behave in a similar manner. Hell and Bauer (B. 36, 205) showed that by the repeated distill tion of ethylphonyl carbinol at ordinary pressures that it was converted into 1 phenyl l propane. The same unsaturated compound ia fOHde eVen more readily by the conversion of the carbinol into the chloride by treatment with phosphorus pentachloride and teen heating. (Wagner Jour. Russian Phys. Chem. Soc. 16, 324). Triphenyl carbinol, which we may define as a true tertiary aro- matic alcohol, may be repeatedly distilled at ordinary pressures without -11-- undergoing any change. It also dietille over soda lime unchanged, and in no manner.mey be made to undergo internal dehydration. (Hemilien B. 7, 1206). Kixed tertiary carbinole, however, such as methyldiphenyl curbinol, ethyldiphenyl curbinol, and prepyldiphenyl carbinel are very easily do. hydrated by heating with such agents as pyridin or acetic anhydride; or by merely distilling at ordinary pressures. It is not necessary for the tertiary carbinole to have two phenyl groups to behave in that manner, for such an alcohol an butylphenyl cerbinol, may be mode to lose a mole- cule of water. Although it is not an easily accomplished on in these compounds containing two phenyl groups, heating then with crystalline oxalic acid will result in its elimination and the formation of file re- lated unsaturated compound. (Tiffenenu C. 4. 143, 649). This brief survey shows that when cerbinole underwent internal dehydration, that the hydroxyl group combined with a hydrogen on an alpha carbon atom to form the molecule of water that wee eliminated. The pro- 0983 took place more readily with tertiary than with secondary alcohols. Also, the preselce of two phenyl groups greatly increased the once with which dehydration occurred. Diphonylethyl carbirnl hoe been prepared in eevorel different ways. By treating benzephenonigtith ethyl newneeium bromide in other. hell and Bauer B. 37, 83 ). By the reaction of sodium end.mercuric dicthyl on benzophenono in ether in an atmosphere of hydrogen or nitro- gen. (Schorigin B. 41, 2719). By the reaction of sodium on a mixture of benzephenone and ethyl iodide in benzene. (Schorigin B. 41, 2715). ~13- Alec prepared by Henson (0. r. 135, 553) by treating an ester of propionic acid with phony]. magnesium bromide. L" It wee a cmrstullino alcohol having; :1 melting; point of 94°. It be- haved as some described above and was easily dehydrated. ”:‘fhen distilled at ordinary pressures or when heeted with acetic enhydride it for-med l, l, diphenyl 1 prepene. This unsaturated hydrocarbon when treated with bromine in chloroform or other eelutim formed a mono-brom derivative having a . ‘ o u v- i ' f 0 Y“ meltihv point of 4b - 4-9 . (hell and finder i3. .57, do ). ~13- ‘1w*"'.) 3‘?' “"f’.1 :1. 5; I. {.[i M.’ 11-14 -14- Preparation of Uiphenylothyl darbinol Diph yict’. yl eta-bins]. was prepared by means of the C—rigmard reaction. In a tJo liter round bottom flask, 24 rr no of clean, dry, magnesium ribbon W:'~s:ms3'cx;;1lod in 500 see. of .1' my 1011:; ethyl ether. Lho flask one connected with a water cooled reflux condenser and 110 Crane of ethyl bromide was added. In OPGOL‘ to start the reaction, a small crystal of iodine was added, and then a cvlcium chloride tube was attached to the condenser. The reaction, as rule, proceeded quietly, bit sometimes it use necessary to cool the flask to keep it under con- trol. Durirn Le course of the reaction, the flask was shaken frequent- ly. After the spenteneoue re ctiln h l ceased, the flue: w s warned on a water bath for about an hour. After a short tine, the ”lost as surrounded with Ln ice water bath. To it Was slowly added t rough tie Condenser, n sr>lutior of 180 grams of Lennopkonone inc on; :Wt;r of dry other. The flzsi was shaken ‘fio?on and after the rerct'uh had “3'rently ceased, the hixture was nrain us11xc ed on the water bath for about half an hour. After cooling the mslntents we: soured slowl into a beaker con- , . taining cracked ice and hydrochloric acid. After two well defined layers forned, the other layer was sexarated and dried over anhydrous sodium sulphate. The other solution was then decanted off and allowed to evapor- ate in a large evaporating dish. The oily crystalline material that re- 4 ~15- mained was placed on filter paper and the oil proceed out. The solid ma- terial Wre recrystallized from ethyl alcohol. Average yield of diphenyl- _ 1 o ethyl curbinol was 53 grams .. r. 94 . Initial Condonsations Six condensations were attempted using diphenylotnyl carbinol and benzene. The solVents and the amounts of aluminum chloride differed in each ones. The carbinel tonotncr with the benzene was suspended in a small condensation Jar fitted with a mechanical stirrer. The aluminum chloride was added gradually so that the temperature could be maintained at the desired point. Wtirrinr woe Continued throughout the reaction and for about two hours after it had apparently ceased. The reaction mix- ture turned a reddish brown color and large amounts of hydrochloric acid gas were evolved. The reaction product was allowed to stand OVCr nirht and then do- composed by pouring into ice water containing hydrochloric acid. nfter decomposition was conploto, the solvent layer was removed, dried over calcium chloride and allowed to evaporate SletHflOOUSlyc icon evapor— ation of the solvent, on oilv crystalline residue rereined. After press- ing out the oil and recrystallizing the solid motericl frofi alcohol, shiny, white plates formed which melted at 49 - 50°. In solution this product gave a blue fluorescence. -17- Condeneations at 20° Carbinol Benzene A1 013 Solvent Product Yield 7.063. 2.63. B g.(.5 mol) 1 Pctmother 3.23. ’ 3 7.065;. 2.63. 4 g.(1 mol) " 5.4g. 68;; 7.063. 2.63. 3 g.(.5 moi) Carbon Dieulfide 2.83. 44% 7.063. 1 2.6g. , 4 3.(1 mol) " " 3.0g. 47% 7.063. 2 g.(.5 mol) Benzene 5.03. 4753 7.063. 4 g.(1 ml) .. 3.4g. 52.3 It was then assumed that the condensation had occurred an expect- ed, for the melting point of the compound obtained, checked within a degree of the melting point or 1, 1, l, triphcnyl propane as recorded in the literature. It appeared that an easy way to check the identity of the compound, would be to prepare triphenyl propane by means of the Fittig reaction, us- ing triphenyl chloromethene and ethyl bromide. -18- Preporetion of Trlphenyl Chloromethene A.mlxture of 80 grams of pure dry carbon tetrachloride and 200 creme of benzene were placed in e one-liter round bottom flask. The flask was connected with a water cooled reflux condenser to which won attached a cal- cium chloride drying tube. To this mixture nae added 60 grams of aluminum chloride, slowly enough to keep the reaction from becoming too violent. after all the aluminum chloride had been added and the main reaction was over, the flask was heated on a water bath for about an hour. The reaction pro- duct was then decomposed by pouring into a beaker containing about 150 grams of lee and 200 cos. of concentrated hydrochloric acid. After two well de- fined layers had formed, the benzene leyer~wne separated and dried over eel- clum chloride. As much benzene as could be easily evaporated on the steam bath was removed, and after cooling, the residue was digested with an equal Volume of ether and cooled for at least two houre in ice. The crystalline sludge was filtered through e large auction filter and washed several timee with ice cold ether. Yield 100 grams, melting point 112°. 1-19- Fittig Reaction with Ethyl Bromide and Triphenyl Chloromethune To a half—liter round bottom flask containing 100 cos. of dry, ul- cohol free, other was added 10 grams or triphenylchloromethnne, 3.9 grams of ethyl bromide. and finally 1.6 grams of clean, freshly cut, finely dia vided sodium. The flask was connected with a reflux condenser and a dry~ lug tube and allowed to etend for 48 hours. The reaction.mixture was then filtered, to free the materiel from sodium and any inorganic cult formed. The other one evaporated and a solid mace remained. Recrystallization from methyl alcohol gave shiny, white needles which melted at 81.5". ”he same experiment was duplicated, except that petroleum other was used instead of ethyl ether. The results were the name as in the first case. For some tune, the identity of this compound was not known. It was finally diecOVerod that no reaction had occurred and that tne solid mess that was obtained after the eVuporetion of the other, was unchanged triphenyl~ chloromethene. Furthermore, that in dissolving this material in.methyl e1» cohol. the heat necessary to bring about solution was sufficient to cause the elimination of hydrochloric acid and resulted in the formation or the 0 methyl ether of triphenylmetnnne. fieltinc point 81.5 o 0-20— Preperation of l, l, l, Triphenyl Prepane by lfethod of Gomberg A mixture of 4.8 grams or magnesium and {53 grams of ethyl iodide was placed in a one liter round bottom flask and suspended in absolute other. after the mmesium had all been converted into the Grimard reagent, the flask wee placed in a freezing; mixture. A solution of 2'? {fir-7.1113 or trip‘nenyl chlorcmot‘arene in ether was added slowly from a dropp- ing: fumzel. Finally, the reaction liquid wns heated to boiling; for a short time. .from the reaction liquid, 4 game of triphenyl methunc was isolat- ed by cnretallizetion from benzene. The oil that renmined was than care:— fully distilled under reduced pressure. Almost the entire product die- tilled at 220 ~ ziafiounder 13 mm. pressure. The oil that was obtained in this munner, had a boiling point which checked with the one recorded by Gomberg end was prepared in the ewe mnner. It was thus aesmned to be 1, 1, l, triphenyl prepene. It was a. thick, light yellow, blue fluorescent oil, but all ettmmts to crystall- ize it from different solvents and under different conditions were futile. «.21- Identification of Product It seemed peculiar, that in the first condensations, the desired product crystallized so readily. It, of course, seemed logical to believe that the desired compound had been obtained at the outset, because its melting*point checked with that of the desired compound. The oily triphenyl propane prepared by Gomberg's method was seed- ed with one of the crystal. obtained by the condensation method. The oil itselfg as well as the oil in several different solvents was seeded, but in no way could crystallization be induceda This, of course, indicated that one of the two preparations had given rise to something other than the expected thing. All evidence pointed to the product of the condens:tion as the uneXpectod ones The condensation product had never been distilled because the greater portion of it always crystallized upon removal of the solvent. It was then distilled and found to have a boiling; point of 280° at atmos- pheric pressured This could not be triphenyl prepane because that had a boiling point of 220 h 225° at 12 mm. or pressure. Nitration of the compound, failed to give the trinitro derivative of triphenyl propane as recorded by Gomberg. .Analysis showed that it contained no hologen. A study of the literature for some canpounds that might have been formed due to a degratory action of aluminum chloride on the carbinol, showed that there were a few possibilities. BonZOphsnone, having a melt- ing point of 48.5o loomed as a possibility. however, the unknown com- pound would not fermion oxime nor a hydrazone; it gave no color when heated -22.. with metallic sodium; and had a boiling point of twentyasix degrees below that of benzephenonc. These tests showed conclusively that it was not benchhenone. Another possibility which seemed to meet the required physical pro- parties cVen more closely was that of dibenzyl. This compound had been prc~ pared by the action of aluminum chloride on ethylene dichloride and benzene in the past. It was possible to conceive of aluminum.chloride having had some kind of a peculiar effect on the Carbinol, such that when benzene was present that it gave rise to the formttior of’dibenzyl. In order to see if this assumption was true, dibcnzyl was prepared by refluxing benzyl chloride over sodium.and than distilling off the pro- duct. The Command time o'ct‘jzined boil-ed at 2534.0 and upon crystallization from alcohol nge shiny, snail plate, crystals which melted sharply at 50.50. The mixing.nelting point test was employed and after mixing a little dibenzyl with a little of the unknown, the mixture was found to melt at 22° which showed that the compound in question was not dibenzyl. another possibility remained; namely, that the unidentified comp pound might have been 1, l. diphenyl l propane. This was prepared by reu fluxing ton grams of diphenylethyl carbinol with ten cos. of acetic an- hydride. After recrystallization of this product from.ethyl alcohol, 5 grams of shiny. plate—like, white, crystals were obtained which had a . . o melting point of 49.5 . Some of this material was mixed with the unknown condensation product and the melting point of the mixture Was found to be ~23- ctill 49.50. From all appearances, the physical prepertios of the two comp pounds seemed identical. In order to prove more definitely the identity of the compound, the bromine dcriVetive was prepared. A solution of 10 grams of the material in 25 cos. of chloroform Woo cooled in an ice bath. With constant stirring, 2. 55 cos. of liguid bromine was added drop by drop. The bromine was irmmflietely Completely docolorizcd after the addition of each drop. There was $130, no QVQlutiun of hydrobrcmic acid gee. The solution use then rcmoved from the ice bath, and the chloroform was allowed to evaporetc slowly. he tto solvent gradually evaporrtod, E Br began to be gradually liberated, and by the time htt the solvent was nzarly {one the F. Pr Cure of? in clouds. Tho residue wxo then placed on the steam both until the oVoTuiinn of 3 Br won complete. The remaining oil was corefully fractionated under reduced pressure and almost the entiro product distilled constantly at 178 - l?d° at 14 mm. proeeure. This boiling point checked with the one recorded in the literature for this substance, but it could not be made to crystallize as had been re- pcrted. It was analyzed for bromine contort aid gave the following results: Calculated 89.55 Found 29.1% This definitely proved that the compound obtained from each attempt- ed condensation was not the product of an intermolecular condeneotion bow tween the diphanylethyl curbinol and benzene, but that it was 1, 1, diphenyl 1 propane resulting from.an intramoleculer condensation or dehydration. .34.. This fact was further substtcntiatm‘x by showing; that when the car- binol was suspended in petroleum ether and treated with ulnmimm chloride, as daacribed previously, that the same product was obtained when benzene Was omitted from the reaction mixture. ~25- ltctoro nfluoncing benydretion A study Was then launched to determine the effect of varying the temperature, the nolvent and the amounts of aluminum.chloride upon the ,utcoma of the éohydration. The results of these dehydre.iona, which were carried out in exactly the game manner as dcscribeé for the initial candoneatinne, ere tabuletoa below. {any or the ring, as tabulatefl, were duplicated Beverol timee. In any onee where the product was distilled rather then allowed to cryeu tullize after deoomgonitLOu with water, almost the entire product was 0 always found to fiietill in the 'enue of 2V5 - 230 at atmosyheric pressure. “all of the oil Would not crystallize even thoagh it all distilled over the same range. Thie probably was due to the formation of a small amount of an isomeric fonn of the compound. On one run, the torperuture of the reaction mixture during the dehydration was kept at 40°. The product ieoleted from this reaction had a melting point of 81°. .Altnough the identity of this omgpound has not as yet been proven, it in probable that it was 1, B, diphenyl l propane which melts at this temperature and 13 on isomeric form of l, l, diphenyl 1 propene. A subsequent dehydration reaction in which the temperature was 0 again maintained at 40 resulted in a ext yield of the same product as had been obtained at the lower temperatures. .86- .I II... S .. :8. 1.36m. .mooJ .I. all... MS .. :2. dream. .e84 en in; con .. Son 135.“ .eooJ 3 .3.“ no." .333 :3 913$. .wooé Ev .mo.» 08 .. c :08 rtflquoo .moo.» an .3.» o3 .. .. Zoe 5.8.4 .30; 5 .eo.» o3 .. .. :8 o...eo.m .39.“. on .3,» Goa Henna Jam :8. nJowoé .30.“. no .3.» o8 .. :3 $.84 $84. 5 .3.» com 3338 :9. a. 73.x .mooé : ea.» o8 .. . :8 3.3.. .33 3 _ .mm.» 98 83.358 gonna :8 n45?» 58:. N0 omwon DON .. : Ada Hvowoov owoOob on .3.» can .53 Joe :8 €73.» .weo.» Mama 3930 «duo Mom i 6H0“ «A809 «abdom «g «g hitti— Discussion A glance at the data shows that the solvent was not an important factor in the dehydration. Likewise at 20° and below, the temperature was not a factor that greatly influenced the amount or type 01‘ product formed. As yet, no statement can'be.msde as to the affect of higher temperatures upon the reaction product. In regards to the amount of aluminum.chloridc present, the data shows that within the ranges of one half to one and one half mole the amounts acre insignificant. nrter more than one and one half mole was used the yield was not increased, but more terry residue was form» ed. If less then one third of a mole of aluminum chloride wee need, the reaction did not eccur, and the carbinol was recovered unchanged. The action of the eluminum.chloride was, therefore. more than catalytic. This work has shown that diphenylethyl curbinol, a tertiary alcohol containing two phonyl and one alkyl groups, did not condense with benzene in the Deviance of aluminum chloride. Basing ene'l predictions on the fact that primary and secondary alcohols could be condensed with benzene in this manner, it seemed logical to have believed that the reaction would have occurred in this manner. ' + u (3:: \. ,7... I H. < > C-C~ , .1 H on I However, this did not occur. Instead u reaction of this type tmk lllétc 8 g + “1.0 o: - w ”m (-7 - (I; (ti-EH C: H - (Lu OH “ By use of the electronic conception of Valence, it is possible to explain why dehydration rather than condensation occurred. If it is assumed that in e.molecule of diphenylethyl carbinol, that the two phenyl groups are strongly positive, then the electron puirs between the phenyl groups and the Llphu carbon atom will be closer to the phenyl groups than to the carbon. This will make that carbon relatively more positive. Then, in the same way, if the alpha carbon atom is relatively more positive, the lectron pair between the alpha and beta carbons will be closer to the alpha one, which will in turn make the beta carbon relatively.more posi- tive. The positive hydrogen attached to this carbon will then be less firmly attached and more easily subject to removal. While assuming that the phenyl group itself is positive, we must remember that the field of force around it as a whole, has a strong elec- tron repelling power. This would repel any strongly negative group which was close to it. The hydroxyl group is then repelled or pushed away by the influence of the phenyl groups, and thus it, too, is easily subject to removal. Therefore, under the dehydrating influence of the aluminum chloride, the tendency would be to remove the.moet easily formed molecule of water, and so the carbinol which is in a condition of strain, readily loses a molecule of water and is converted into its more stable deri- vative 1, l, diphenyl l prepeneo A slight modificetlon of exactly the same eXplenetion would account for the fact that the bramiuation of the unsaturated cqmpound re- sults in the isolation of a monobrom rather than a dibram derivative. The chief difference being that in the latter case the condition or strain 18 so great that merely raising the temperature a very little results in the loss or halogen acid. This work bps chown that at temperatures of 20° or below, di- phorglethyl carbinol con not he made to condense with bensexo, by use or aluminum chloride as c dehydrating agent. Ihotond, dehydration occurs within the curbinul molecule, rceoltihg in the formation of l, l, diphenyl l propane. Although one dehydration carried out at 40° resulted in the formation or some unidentified compound, duplication of the procedure at the same temperature yielded the unentureted product as expected. It is predicted that similar dehydration will occur at all temperatures.