an. I. A. , _ . . . v 2.. .9. e .r _. ... . S H H. ; E N «a S w m m. w mu my N. aw” 3 v 2m m.” «A m 4 1.... Nu». NJ 4 E K h a: u.” t E R .. m 2:. ,u 1;. «4 yr 5 a w any a «3 .- 3:. 3.1. .. {a .5 . H . p, .. .1 . .. 8 f. A ‘ J t .«J .)3 ~— rrséf " is This is to certify that the thesis entitled "The Reaction of Propylene Oxide with Various Alkylmagnesium Chlorides'. presented by Harlan E. Tiefenthal has been accepted towards fulfillment of the requirements for IL degree in Jheznisizry 1 Major professor Date—“1211mm 0-169 ”7.x: C\ =3 - q“ x 5 5' T373813 ABSTRACT 3‘ 5 "’ m3 nmcrxow or momma OXIDE mm vmuons KW AIXYLTmGHESIm! CHLORI 953 By Harlan '25. Tiefenthal Thie investigation hee extended the etudiee that here been eon- ducted on the reeotione of rerioue elkylnngneeium helidee with ethylene oxide, propylene oxide end ieobutylene oxide. It involved the preparation of eeverel Grignerd reegente end oorreeponding di- elkylmegneeium oompounde, together with their eubeequent reeotione with propylene oxide ueing verioue proportione end conditione. Grignerd reegente were prepared ueing ethy1-. n-propyl-. eeo- propyl-, n-butyl-, eeo-buty1-. ieo-butyl, end tert-butyl ohloridee. Yielde were determined by em on)” [Gilmm H... Wilkineon, P. 9.. Fishel, W. P. d.- Heyere, C. ’23., J. Am. Chem. Soc... 32, 150 (1923) ] titretion. Eech Grignerd reagent nee treated in four .YII 1. One mole of Grignerd mgent nee reacted with one mole of propylene «ride without heetingo 2. One mole of Grignerd reegent wee reeoted with one mole of propylene wide end then refluxed with beneene. 3. One mole of Grignerd reegent nee reeoted with two melee of propylene oxide without heating. 4. The dielkylmgneeium compound nee made from the Griperd reegent end nee reacted with two melee of propylene oxide Without heating. Ii Her-lee E. Tiei‘enthel Reeetiene 1 end I Vere elleeed to etend ever night before hw- drolyeie. leeetim 8 end 4 were elleeed te etend until they geve e “El/neptive Miehler'e ketele teet [Guam H. end Sohelee. F” J. An. Chen. 800.. a. 8008 (1925)] . The ether eelutiene er the prodeote of rotation 1, I, end 8 were refluxed with eedinn hydrexide to de- oelpeee en: 1-eh1oro-2-propenol preeert. Anelyeie for ehleride iee eeneentretioe nee nude by the Volherd nethed. The ether leyere fret e11 reeetiene were dried ever peteeeiun eerhenete end freetioneted through e helices peeked. Feneke type eolm. The phyeieel eonetente e! the eleehole were deterlieed end the 8.5-dinitroheneoete derivetivee propered. ,I- PERM “ELIE W ALCOHOLB AND l-CKLORO-I-PROPAIOL I fl 4 mu m. m. CH CH vs, 20!! V3. 8 fig}! Griperd reegeut stag/CH. 8 p 3 8 without with vithont without prepred heet heet heet heet tre- ! F. H "' 2g .. é; .. . 91 .t .. ° E .3 :3 s 32;; 2‘3 54;; B M}; 3‘! A H «.2 e'. '4 I: H on I! q a '4 0' “w 4 : l-ehioreethene 85 87 40 d2 73 £8 60 l-ehleropropene 50 28 85 81 M G! 19 t-ehlerepropene 85 83 58 80 81 84 loehlerobntene 52 41 88 58 77 59 26 z-ehlorohutene 5d 10 I4 16 a 80 9 loohloro-l-nethvl- prepene 58 19 19 40 73 62 88 homered-methyl- prepene 68 O 8 25 61 16 0 Herien 3. Tiei‘enthel A compel-icon of the reeulte of thie inveetigetion with thoee ebteined from the reeotion of propylene oxide with elkylnngneeiun hronidee [Hueton, R. C. end Boetwiok. C. 0.. J. Org. Chem" 32, 331 (1948)] ehowed theta 1. In the 1:1 mole ratio reeotiones e. when but wee not need. the chlorine eotiveted the elkyldnegneeitm bond to give higher yielde of eloohol end, b. when e benzene reflux wee need. the chloridee geve improved yielde of the eeoondery eloohole. 2. In the 1:2 mole ratio reectieml e. the yielde or the eloohole end l-helo-z-propenole were inoreeeed in both oeeee end, 1). these yielde were eeeent ielly the eeme for the cor ree pondi n5 elkyl gran p. LIBRARY - Michigan State University PLACE iN RETURN BOX to remove this checkout from your record, TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 6/01 c,/ClRC/DateDue p65-p 15 e E RflACTION OF PROPYLENE OXIDE WITH VARIOUS ALKYLVAGNESIUV'CHLORIDES By Harlan E. Tiefenthal A THESIS Submitted to the School of Graduate Studies of Michigan State College of Agriculture and Applied Science ‘ in partial fulfillment of the requirements for the degree of DOCTOR.OF PHILOSOPHY Department of Chemistry 1950 ACKNOWLEDG‘TET‘TT With sincere appreciation and gratitude acknowledgment is made to Dr. Ralph Chase Huston for his understanding guidance throughout the course of this study. --Har1an E. Tiefenthal 3 O (U «I A. h V! C33 0 “N 9-1 LA TABLE OF CONTENTS TABLE OF CONTENTS Page INTRODUCTIONOOOOOOOOOOOOOOOOOOOO...OOOOOOOOOOOOOOOOOOOOOOOOOIOOOO 1 HISTORICAIJQOOOQOOOOOOO0.00IOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO. 2 EXERIIWTALOOIOO...OOOOOOOCOOOOOOOOO0.......OOIOOIOOOOOOOOOOOCCO 7 I. mteriaIlBOQCCOOOO..0.COOOOOOOOOOOCCOOCOOCOOOOOOIOO...0.. 7 II. Preparation of Grignard Reagents........................ 7 A. Preparation of Ethylmagnesinm Chloride............ 7 B. Preparation of Other Grignard Reagents............ 8 C. Analysis of the Grignard Reagent.................. 9 III. Reaction of Propylene Oxide With the Grignard Reageflts..o.o.....................o..................... 10 A. Reaction of one mole of Propylene.Oxide with one mole of the Grignard Reagent Idthout heatingOOOOOQOOOOOOOOOOOOOOOOOCCOOOCOOOOO0.0.0.... 10 B. Reaction of one mole of Propylene Oxide with one mole of the Grignard Reagent followed by refluxing With Bemene....................on..." 11 C. Reaction of two moles of Propylene Oxide'with one mole Of Grlgnal‘d Reagenteeeeeeeeeeeeeeeeeeeeee 12 D. Preparation of Dialkylmagnesium.reagents and their reaction.with two moles of Propylene OXideOOOOOOOOOOOOOOO00....0.00.0000...0.0.0.000... 12 IV. Reaction of one mole of Propylene Oxide with one mole of Ethylmagnesium Chloride followed by violent de- composition........................................-...o 14 V. Preparation of the 3,5-Dinitrobenzoates of the alcohols................................................ 15 'VI. Calculation of percentage yields........................ 16 IABLES........................................................... 17 THEORETICAL...................................................... 21 DISCUSSION....................................................... 27 MQMYCOOOOOOOOOOOOOO0.0......I0.0.0.0000...OOOCOOOOOOOOOOOOOOCC 30 MEENCESOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO0.0...OOOOOOOOOOOOOOO 31 INT RODUC TI CN INTRODUCTION As with many other chemical reactions, chemists have made many generalizations concerning the Grignard reagent and its reactions. While it is true that its reaction with propylene oxide gives secondary alcohols, these alcohols are not always the main product of the reaction. This study was initiated to investigate the reactions of vari- oue alkylmagnesium chlorides with propylene oxide and to compare these reactions with those of the corresponding alkylmagnesium bromides with propylene oxide. -1- HISTORICAL HISTORICAL Louis Henry (1) was the first to study the reaction of propylene oxide with alkylmagnesium halides. He Postulated that it is theo- retically possible to obtain several different alcohols when ethyl- magnesium bromide is allowed to react with propylene oxide. The product obtained, according to Henry, depends upon how the oxygen- oarbon bond splits and upon whether or not rearrangement occurs before the propylene oxide reacts with the ethylmagnesium.bromide. If the carbon-oxygen bond splits and addition occurs before isomer- ization takes place, the product would be CH3CHOHCHZCH2CH3 or CH30H2CH(CH3)CH20H. If addition of the ethylmagnesium bromide takes place after isomerization occurs, the product would be (CH3)2COHCH2CH3 or CHscHzCHOHCHZCHs. The latter depends on whether isomerization gives the ketone CH3COCH:5 or the aldehyde CHSCHZCHO. In carrying out the reaction of propylene oxide with ethylmagnesium bromnde, Henry obtained a 60% yield of 2-pentanol. Hess (2) found similar results using ‘( -pyrridyl£magnesium bromide and propylene oxide. Levene and‘Walti (5) condensed optical- ly active propylene oxide with methyl-, ethyl-, and propylmagnesium bromides. They found that similar alcohols were produced and that walden inversion did not occur. Further evidence for this type of addition is given by Newman (4) and Sharefkin and Ritter (5). Newman prepared 1-pheny1-2-propanol by using phenylmagnesium bromide and propylene oxide. Sharefkin and Ritter prepared 4-(4-methyl- phenyl)-2-butanol usingxylylmagnesium chloride with propylene oxide. However, when Norton and Bass (6), carried out the reaction by adding an equinmlecular amount of propylene oxide to ethyl- magnesium.bromide and then heating, they obtained an 11% yield of 3-pentanol. This product indicates an addition which is the re- 'verse of that reported by Henry. It appears that the products formed during the reaction of Grignard reagents with epoxides have been formed by carbanion at- tack at the least substituted carbon atom, OH H ‘ * - éHCH R. aigx + Cr13C\H;H2 ——> CH3 0 2 The reaction is not a simple one, however, as Huston and Agett (7) have shown that dialkylmagnesium is probably an essential reactant. They found'that the addition of one mole of ethylmagnesium.bromide to one mole of ethylene oxide gave an addition complex which was the same as that obtained by passing ethylene oxide into an ethereal solution of magnesium.bromide. The first mole of ethylene oxide ap- peared to react almost exclusively with the magnesium.bromide, in equilibrium with the Grignard reagent and diethylmagnesium, and formed an addition complex. Hydrolysis at this point gave 60% 1-bromo-2-ethanol and very little l-butanol. 2 CHSCHZMgBr= (CHsCH2)2Mg + "gBrz 2 CHZICHZ 2 H o , 2 HOCHZCH28r$ a ..ig(OCHZCH27—3r)2 Grignard (8) proposed oxonium type intermediates for this re- action. While his views were supported by ”eisenheimer and Casper (9) these oxonium salts prdbably do not exist for any great length of time. In order to obtain l-butanol, the mixture must be heated or a second mole of ethylene oxide added: (CHSCH2)2?“Yg + l-’g(OCH2CHz*-3r)2 —-» "g(OCH2CH2CH2CH3)2 + MgBrz 2320 T J EEOCHZCHZCHZCfiat— or (CHSCH2)2Mg + ZCHZCHZ —-9 ”g(OCH CH CH CH ) 2 2 2 3 2 O ZHZO H i H £—— 2 OCHZCHZCHZC 3 Huston and Langham (10) have shown.that when ethylene oxide is reacted with ethylmagnesium chloride the intermediate compound lead- I ing to 1-butanol (in 55% yield) is (C4H90)2Hg from (CZH 4g or 5)2' C H OMgCl from, C H 49 25 While studying the reaction of n-butylmagnesium bromide with ethylene oxide, Cottle and Hollyday (11) obtained a mixture of l- hexanol and 2-hexanol. They attempted to explain these products by proposing a mechanism involving the nucleophilic attack of the carbanion of the Grignard reagent upon one of the carbon atoms of ethylene oxide H" H - c- c" _ ‘ C49 + c\2/c:n2 -—-’ 4H9112CH20 or on the BrCHZCHZO- particle from the magnesium di-(l-bromo-Z-ethoxide) C4H9 + BrCHZCHZO --) C4H90H2CH20 4 Br . However, neither of these nucleophilic reactions should give re- arrangement and therefore would not account for the 2-hexanol pro- duced. They also found that: l. when.the mixture from the reaction- between ethylene oxide and n-butylmagnesiunibromide was heated in such a manner that a violent reaction took place, more 2-hexanol was formed than in similar experiments when there was no violent reaction; and, 2. an acetaldehyde resin was formed when the ethylene oxide and magnesium bromide reaction product was heated in a bomb. This led them to postulate that the BrCH CHZO- particle may, upon heating, loose 2 a bromine ion and then rearrange to acetaldehyde. Further proof that the addition complex decomposes on heating to form a carbonyl com» pound was found by Huston and Bostwick (12). CH O H ' 3 (Brcnzcno)2 rig—L— CHSCCHS 4 HBr + Mgo + CH CH =CHBr. 3 Huston and Brault (13) have shown that the magnesium di-(l-bromo-Z- methyl-Z-propoxide) rearranges spontaneously to isobutyraldehyde. Actually, Huston and Postwick found acetone among all of their reaction products and Huston and Brault found isobutyraldehyde in all of their reactions. Norton and Hass (6) reported a 23% yield of 2-pentanol when they reacted diethylmagnesium with propylene oxide. Other reported reactions of epoxides with dialkylmagnesium.compounds indicate only simple addition products even when the reaction ndxture is refluxed (7, 14, 15, 16 and 17.) -e- I. II. EXPERITQITAL Materials Propylene Oxide (B.P. 33.4-34.400.). Furnished by the Dow Chemical Company. Used without further purification. Ethyl Chloride. Secured from the Ohio Chemical Company. Alkyl chlorides. Secured from the Columbia Chemical Company. Vagnesium. Secured from the Dow Chemical Company special for Grignard reactions. Dried at 50° C. before use. Anhydrous Diethyl Ether. Dried over sodium wire for at least one week. Benzene. Thiophene free and anhydrous. Dried over sodium wire for at least one week. Dioxane. Purified by refluxing over dilute hydrochloric acid, drying, refluxing over sodium and distilling. Potassium Carbonate. Anhydrous, C. P. Ammonium Chloride. C. P. Silver Nitrate solution, approximately 0.1 N. Prepared from Bakers' Analytical Grade AgNOS. Potassium Thiocyanate solution, approximately 0.1 N. Standard- ized against the silver nitrate solution. Sodium.Hydroxide solution, approximately 0.2 N. Standardized against sulfamic acid. Sulfuric Acid solution, approximately 0.15 N. Standardized against the standard sodium hydroxide solution. Preparation of Grignard Reagents A. Preparation of Ethylmagnesium Chloride Two and one-quarter moles (54.7 g.) of magnesium.turnings were placed in a dry, three-liter, three-necked, round-bottomed flask which was placed in an ice, salt bath. The flask was fitted with a glycerol sealed stirrer, a dry ice condenser (fitted with a calcium chloride:soda-lime tube) and an inlet tube for ethyl chloride. The ethyl chloride inlet tube ex- tended as near the bottom of the flask as the stirrer would allow. Six hundred milliliters of anhydrous ether were added to the flask and the condenser was then filled With a dry ice and acetone bath. Ethyl chloride was bubbled into the flask through a potassium hydroxide pellet tower and then through concentrated sulfuric acid for two hours while the mixture was stirred. The stirrer was then.stopped and the mixture allowed to stand for an hour. During this period the reaction usually began but in those cases in which it did not, a charge of ethylmagnesium bromide was added to start the reaction. Stirring was resumed and ethyl chloride added at a rate sufficient to maintain gentle reflux. After most of the magnesium had reacted, addition of the ethyl chloride was discontinued and stirring maintained for one hour. The stirrer was then stopped and the reaction mixture allowed to come to room temperature by standing overnight. Preparation of Other Grignard Reagents Three and one-quarter moles (79.0 g.) of magnesium turn- ings were placed in a dry, three-liter, three-necked, round- bottomed flask which was placed in a container so that -8- C. external cooling was possible when needed. The flask was fitted with a glycerol sealed stirrer, a bulb condenser (fitted with a calcium chloridezsoda-lime tube), and a Hershberg dropping funnel. Four hundred milliliters of anhy- drous ether and ten to twenty grams of the alkyl chloride 'were added to the flask. The reaction was started by the addition of a charge of the alkylmagnesium.chloride prepared in a test tube. The rest of the three moles of the alkyl chloride was mixed with an equal volume of anhydrous ether and added at a rate sufficient to maintain a gentle reflux. External cooling was used only when necessary. After addi- tion was complete, stirring was continued for an hour and the mixture allowed to stand overnight. Analysis of the Grignard Reagent The Grignard reagent was poured, under a stream of nitrogen, into a graduated cylinder and the total volume measured. Two milliliter portions were then taken and analyzed by the Gilman procedure (18) using methyl red as the indicator. The solution of the Grignard reagent was then divided into four approximately equal parts and each was treated in a different way. A typical run for each of the four procedures will be described. III. Reaction of Propylene Oxide with the Crignard Reagents A. Reaction of one mole of Propylene Oxide with one mole of the Grignard Reagent without heating Approximately one quarter of the total Grignard solu- tion was placed in a one-liter, three-necked, round-bottomed flask. The flask was fitted with a bulb condenser, a glycerol sealed stirrer and a dropping funnel. The condenser and the funnel were fitted with calcium chloridezsoda-lime tubes to protect the solutions from atmospheric moisture and C02. The required amount of propylene oxide was weighed out and placed in the dropping funnel along with an equal volume of anhydrous ether. This solution was added dropwise with stirring as fast as possible and still maintain a gentle re- flux. After the addition had been completed, the mixture was stirred for an hour and allowed to stand overnight. The mixture was then hydrolyzed, without external cool- ing, by adding a saturated solution of ammonium chloride dropwise, with stirring, at a rate sufficient to maintain gentle reflux (19). About one hundred milliliters of this solution were required to reach a point where a clear separation occurred. The ether solution was decanted from the precipitated magnesium salts and the dense precipitate was washed with one or two portions of ether. The combined ether solution was then placed over 40 grams of sodium.hydroxide pellets and 150 milliliters of -10.. B. water were added slowdy, with stirring. This mixture was then refluxed, with stirring, for seven hours to convert the l-chloro-Z-propanol to propylene oxide and sodium chloride. It was then allowed to stand overnight and come to room temperature. The layers were separated and the ’ water layer extracted with ether. The combined ether solu- tions were placed over anhydrous potassium carbonate to dry , and the water layer, containing the sodium chloride, was diluted to one liter for analysis. Two milliliter portions were titrated for chloride ion by the Volhard method to give an accurate measure of the yield of l-chloro-Z—propanol. The dried ether*solution of the products was distilled at atmospheric pressure through either a 9.0 inch or a 16.5 inch Fenske-type column packed with 5/52 inch glass helices and fitted with a total reflux, partial take—off type head. Heat was supplied by a Glas-col mantle. All like runs were Combined after one fractionation and a refractionation made before physical constants were determined and derivatives made. Reaction of one mole of Propylene Oxide with one mole of the Grignard Reagent followed by refluxing with Benzene Approximately one quarter of the total Grignard solution was placed in a one-liter, three-necked, round-bottomed flask fitted as described in III A. The addition of the propylene oxide was carried out as described in III A. and allowed to -11.. C. D. stand overnight. The flask was then placed in a Glas-col mantle, the condenser set for distillation and a thermometer fitted so that the bulb was immersed in the liquid. Approxi- mately one-half of the ether was removed by distillation and than twice this amount of anhydrous benzene was added to the distillation flask. This solution was then distilled until the temperature of the liquid reached 75° C. The condenser was then.reset for reflux and the mixture refluxed with stir- ring for seven hours and allowed to stand overnight. The mixture was hydrolyzed, extracted, dried and fractionated as described in III A. The analysis for chloride ion concentra- tion was also carried out as in III A. Reaction of two moles of Propylene Oxide with one mole of Grignard Reagent The other solution of propylene oxide was added to the Grignard reagent as in III A. The reaction mixture was stirred for an hour after addition.was completed and then allowed to stand until the test with Michler's ketone was faint or negative (20). By this time the reaction mixture had set-up like a gel but was easily hydrolyzed, extracted, dried and fractionated as in III A. The analysis for chloride ion concentration was also carried out as in III A. Preparation of Dialkylmagnesium reagents and their reaction with two moles of Propylene Oxide Approximately one quarter of the total Grignard solu- tion was placed in a one-liter, three-necked, round-bottomed flask fitted as described in III A. Slightly more than an equal molar amount of dioxane was mixed with two hundred milliliters of anhydrous ether and added dropwise, with stirring, so that a gentle reflux was maintained. When addi- tion was completed, the stirring was continued for an hour and the mixture allowed to stand overnight. The solution and precipitate were then poured into centrifuge bottles under a stream.of nitrogen and stoppered tightly. They were then centrifuged at 1500 r.p.m. for 15 minutes, the ether solution decanted into a graduated cylinder under a stream of nitrogen.and the total volume of the solution.measured. Two milliliter portions: were taken and analyzed by the Gilman procedure using methyl red as the indicator. This solution was then poured under a stream of nitro- gen into a one-liter, three-necked, round-bottomed flask fitted as in III a. The calculated amount of propylene oxide mixed with an equal volume of anhydrous ether was added drop- wdse, with stirring. The stirring was continued for an hour and the mixture allowed to stand until the test with Michler's ketone was faint or negative. Hydrolysis was carried out as described in III A. In some of these reactions there was a quantity of solid material in.the bottom of the flask which did not hydrolyze. It was found to contain magnesium and chlorine. The chlorine might have come from the ammonium chloride used for hydrolysis. Analysis showed that this solid contained 20.0 percent magnesium. -13- The ether solution of the products was placed over anhydrous potassium carbonate to dry. The dried solution was then fractionated as described in III A. IV. Reaction of one mole of Propylene Oxide with one mole of Ethyl- magnesium Chloride followed by violent decomposition The ethylmagnesium chloride was prepared and the propi- ene oxide added as in the regular 1:1 reaction. After the mixture had stood overnight the apparatus was arranged for distillation and thermometers were placed in each of the three necks of the flask (one was immersed in the liquid, one was near the goose-neck leading to the condenser and the third one registered the temperature of the vapors in the flask.) The distillation was started and'the ether began to come off when the temperature of the solution had reached 37° 0. Meet of the ether had been distilled by the time the temperature of the liquid had reached 600 C. but it was necessary to increase the temperature of the liquid 60 84° C. before the following violent decomposition occurred. At this point the black, tarry looking mass swelled rapidly (almost violently) to a puffy white solid with the evolution of a dense white gas that carried over into the receiver. The hottest temperature recorded was near the bottom.of the flask where it reached 2800 C. At the same time the temperature near the goose-neck reached 1200 C. The reaction flask was then set up for refluxing and a volume of dry benzene equal -14.. to the amount of other which had been removed, was added. The resulting mixture was refluxed for’seven hours with stirring and allowed to stand overnight. The benzene solu- tion of products was hydrolyzed, extracted, dried and frac- tionated as in III A. Analysis for chloride ion concentra- tion was made by the Volhard method. The results of magnesium analyses on the white puffy solid, and the yields of l-chloro-2-propanol and 2-pentanol may be seen in Table I. V. Preparation of the 3,5—Dinitrobenzoates of the alcohols (21) Two grams of the alcohol were placed in a large test tube containing five grams of 3,5-dinitrobenzoyl chloride. A large excess of anhydrous pyridine was added. The test tube was fitted with a calcium.chloride tube to protect the contents from moisture. The mixture was then placed on the steam bath and allowed to heat for an hour with occasional shaking, following which, it was poured into water to hydrolyze any remaining 3,5-dinitrobenzoyl chloride. The water mixture was extracted with ether and the water layer discarded. The other layer was then washed successively with 5 percent sul- furic acid, 10 percent sodium carbonate and finally, water. The ether was evaporated on the steam bath and the residue taken up in hexane. This solution was boiled with Norite to decolorize the solution which was then filtered. Where possible, the 3,5-dinitrobenzoate was allowed to crystallize -15- at room temperature. Otherwise, the solution was placed in the refrigerator. The esters were recrystallized from hexane in a similar manner. As each alkylmagnesium chloride was treated in four dif- ferent ways, a total of four esters of the same alcohol were prepared in most cases. These were mde using the alcohol fraction obtained from the combined fractionations of all like runs. The melting points of these derivatives were compared with those given in the literature. Further proof of the identity of these alcohols was obtained by taking mixed melt- ing points of their derivatives with those of corresponding authentic compounds . VI. Calculation of percentage yields The percentage yields of the alcohols were based on the titrated Grignard reagents. The percentage yields of l-chloro-Z—propanol were based on the titrated Grignard reagents. _15- TABLES TABLE I ANALYSIS OF SOLID RESIDUE FROH'VIOLENT DECOVPOSITION Run Percent Percent Percent 1-Chloro-2-Propanol Alcohol Magnesium. l 6.4 37.0 2 10.9 42.9 3 0.0 45.5 4 18.7 41.0 19.5 Analyses for Vagnesium 18.9 20.5 Content 20.1 20.4 18.9 Average 19.5 Analyses of Solid Residue 18.9 16.9 For Hagnesium Content after 20.9 23.9 Extraction with Ether 21.9 21.8 Average 20.7 Theory for CHscHZCHZCHOMgCl + MgCl2 is 20.1 CH:5 There is definite evidence, however, of the presence of a 1-chloro-2-propanol intermediate. -17- TABLE II PSRCENT YIELUS CF ALCOHOLS AND 1-CHLORO-2-PROPANOL BASED UPON THE TITRATED GRIGNARD REAGENTS i RWgCl + RHgCl + RNgCl + RZHg'+ CflsqggHz CquggHz 20H5§§PHZ 20H30%CH2 Without ‘With Without Without heat heat heat heat Ethylmagnesium chloride » 2-Pentanol 27 42 56 60 l-Chloro-Z-propanol 55 40 75 : n-Propylmagnesium chloride : 2-Hexanol 28 51 62 19 1-Chloro-2—propanol 50 55 64 : sec-Propylmagnesium ch oride : 4-methy1-2-pentanol 25 50 46 54 l-Chloro-Z-propanol 55 55 81 ‘ ' n-Butylmagnesium chloride : 24Heptanol 41 58 59 26 l-Chloro-Z-propanol 52 28 77 : sec-Butylmagnesium chloride i 4-"ethyl—2-hexanol 10 16 50 9 l-Chloro—2-propanol 54 24 69 : iso-Butylmagnesium.chloride f S-Nethyl-Z-hexanol 19 40 62 55 l-Chloro-Z—propanol 58 19 75 : tert-ButyImagnesium chIoride 4,4-Dimethyl-2-pentanol 0 25 15 0 l-Chloro-Z—propanol 65 5 61 -18- PHYSICAL CONSTAHTS OF THE ALCOHOLS TABLE III 20 ALCOHOL nD B.P. 00. Ref. 2—Pentanol 1.4068 120.1/743 mm. 12, 22, 23 Z-Hexanol 1.4140 139.2/743 mm. 12, 23, 24 4-nethy1-2-pentenei 1.4132 131.4/745 mm. 12, 24, 25 ZéHeptanol 1.4218 156.8/745 mm. 12, 23 4-Methy1-2-hexanol 1.4236 152.2/744 mm. 12, 26, 27 5-“ethyl-2-hexanol 1.4194 150.4/744 mm. 12, 28, 29 4,4-Dimethyl-2-pentanol 1.4241 135.4/742 mm. 12, 30 -19- TABLE IV 3,5-DINITROBENZOATES 0F THE.ALCOHOL8 ALCOHOL n.5,. °c. Ref. 2-Pentanol 60-61 12, 51 ZéHexanol 56-57.5 12, 24 4-Vethyl-2-pentanol 60.5-61.5 12, 24 2-Heptanol 47.5-48.5 12, 31 4—"ethy1-2-hexanol 50-51 12, 26 5-Hethyl-2-hexanol 54.5-55.5 12, 29 4,4-Dimethyl-2-pentanol 95-94 12, 50 -20.. THEORETI CAL THEORETICAL Previous work has shown that the reactions of ethylene oxide with alkylmagnesium bromides (7) and ethylene oxide with the corres- ponding a1kylmagnesium.chlorides (10) are fundamentally different. It is also true that there are fundamental differences between the reactions of propylene oxide with alkylmagnesium bromides (12) and propylene oxide with the corresponding alkylmagnesium.chlorides. When ethylmagnesium.bromide and ethylene oxide are reacted in a mole per mole ratio without heating, the only product reported was l-bromo—Z-ethanol (52). Blaise proposed that the Grignard reagent was broken between the magnesium and bromine bond according to the following scheme: ‘iH2\ O OH2/ + CZHSMgBr -—-—) BrCHZCHZOCZH5 This product would give l-bromo-Z-ethanol on hydrolysis. Grignard (8) proposed that an oxonium salt was formed. This salt would hydrolyze as follows: CH~\‘ c H CH 2 0” 2 5 + ZHOH -+ 2 | 2\O + 20 H 9 "gBr . ng(OH) cng” “ug3r cné” 2 6 2 2 A'Wurtz type reaction would then follow, giving l-bromo-Z-ethanol. 2 €32\\ cné” 0 + 2303 + ”gBrg -+ 2 BrCH CHZOH + “2(OH)2 2 -2 1- Grignard found that heat produced a "second phase" reaction giving a considerable quantity of the expected l-butanol. He proposed that the oxonium salt was ruptured as follows: I Hé\‘ O/CZH5 -—51-e' C H CH CH ouger CH./’ O\\,gBr 2 5 2 2 The product would give l-butanol on hydrolysis. Heisenheimer and Casper (9) presented analytical data supporting Grignard's views. Ribas and Tapia (55) proposed that the addition product consisted of a mixture of (BrCHZCHZO)2??g and Br HzCHZOTigBr. However, their analytical data supports the.formu1a BrCHZCHZOVgBr. Huston and Agett (7) showed that the intermediate compound is (BrCHZCH20)2Hg and is formed by the following scheme: ZCszligBr.= (C2H5)2‘ig 4- ligBr2 zcnzcnz \O/ P ZHOH 7 I V 28rCH20H20H2£————-— (arci tho)2 g When ethylmagnesium chloride and ethylene oxide were reacted in a mole per mole ratio without heating, a 55% yield of l—butanol was obtained along with 22% 1-chloro-2-ethanol (10). Huston and Langham have shown that the intermediate compound leading to 1-butanol may have consisted largely of either H 0"gC1 produced from C4 9 CH CH2 Rearrangement (0439:)2-5 + "F012""J -22- OP (C4H90)2"g produced from cs2\ (023552 + 2 L/O ——9 (C4EIQO)2‘Tg 1.2 Hydrolysis of either intermediate would give l-butanol. Huston and Langham further showed that (ClCH CH20)2Hg was the intermediate 2 precursor leading to l-chloro-Z—ethanol. When ethylmagnesium bromide was reacted with propylene oxide in a mole per mole ratio without heat, the intermediate reaction products 'were reported,by Huston and Hostwick (12), to be predominately of the magnesium di-(l-hromo-Z-propoxide) type, (BrCHzggg)2-Hg. The mechan- ism proposed for this reaction was: CH5 (auger =2 ng'rg - "gBr2) + CHSCISCHZ -—+ (BrCHzCHO)2?'.g (1) CH3 (RCHZ CHO)2!’Ig (II) CH3 CH :2 r6112 02011500213112 R ( I II) CH BrCELZ 0118*ng (IV) OHS RCHZ CHOTTgR (v) R”gBr (VI) azi-rg (VI I ) tigerz (VIII) When the ratio of reactants was 1:1, the first, fourth and seventh intermediates were present in large amounts. -23- When ethylmagnesium chloride and propylene oxide were reacted in a mole per mole ratio without heating, a 57H yield of the 2-pentanol was obtained along with a 55% yield of l-chloro-Z-propanol. The re- sults of the reaction of alkylmagnesium chlorides with propylene oxide in a mole per mole ratio without heat, indicate that the alkyl-magnes- ium.bond is ruptured more easily when the a1kylmagnesium.chlorides are used than when the a1kylmagnesium bromides are used. If the mole per mole reaction mixture of ethylmagnesium bromide and ethylene oxide was refluxed with benzene, the yield of l-butanol was greatly increased at the expense of the l~bromo-2-ethanol. ‘ I ,, heat ,, 32H" (BrCLzCHZO)21.g -—-+ (RCHZCH20)2Hg 4 yigerz (RCHZCHZOM-“g + ZHZO —-—-) ZRCHZCHZOH + 2.‘.'.'g(0H)2 There was but slight reaction between RZHg and (CICH CH20)2¥g when 2 ethylmagnesium chloride and ethylene oxide were refluxed with benzene. When propylene oxide and ethylmagnesium bromide in a 1:1 ratio, were refluxed with benzene, there was only a slight increase in the yield of 2-pentanol. When propylene oxide and ethylmagnesium chlorides in a mole per mole ratio, were refluxed with benzene, the yield of 2-pentanol was increased and the yield of 1-chloro-2-propanol was decreased. (Table II). This increase in alcohol yield was probably due to a slight interaction of the magnesium.di-(1-chloro-2-propoxide) with the di- ethylmagnesium. The rest of the l-chloro-Z-propanol intermediate apparently decomposed to propylene oxide. -24- Excessive heating of the magnesium di-(l-bromo-Z-propoxide) from the propylene oxide and ethylmagnesium.bromide reaction mixture caused the formation of acetone along with a complete loss of l-bromo- 2-propanol and only a slight increase in the yield of 2-pentanol. Excessive heating of the magnesium di-(1-chloro—2-propoxide) from the propylene oxide and ethylmagnesium.chloride reaction mixture gave no evidence of acetone formation but there was a great loss in the yield of l-chloro-Z-propanol and only a slight increase in the percent of 2-pentanol formed (Table I). The strong heat generated during this reaction decomposed the l-chloro-2-propanol intermediate. The product of this decomposition was evidently propylene oxide since no evidence of either acetone or propionaldehyde was found among the products. The slight increase in alcohol yield was probably due to a slight interaction of the magnesium.di-(l-chloro-2-propoxide) with the diethylmagnesium. This was also found in the benzene refluxed reactions. One mole of alkylmagnesium bromide reacted with two moles of ethylene oxide to give increased yields of both l-bromo-2-ethanol and the l-butanol: 22133; + l'gBrz + 4(CHZ)20 -—-> (RrCHZCHZO)2Hg + ( RCH2 CHzO)2Ng (BrCHZCH20)2‘g 4 (RCHZCHZO)2‘Tg 4 4320—) 23rCHZCH20H + ZRCHZCHZOH + 2"g(OH)2 -25- The same scheme may be used to represent the reaction of one mole of alkylmagnesium chloride with two moles of ethylene oxide if it is assumed that dialkylmagnesium leads to the formation of the alcohol. When two moles of propylene oxide were reacted with one mole of a1kylmagnesium.bromide, the first three intermediates, in the mechanism proposed by Huston and Bostwick, were present in large amounts. This gave increased yields of both l-bromo-2-propanol and the alcohol. The reaction of one mole of alkylmagnesium chloride with two moles of propylene oxide also gave greatly increased yields of both the 1-chloro-2-propanol and the alcohol (Table II). When two moles of ethylene oxide were reacted with one mole of dialkylmagnesium, the expected alcohol was obtained in good yield except when tertiary alkyl groups were involved. When two moles of propylene oxide were reacted with one mole of dialkylmagnesium, the expected alcohol was obtained, except when tertiary alkyl groups were involved. The yields, however, are much less than when ethylene oxide was used. DISCUSSION DISCHSSION When propylene oxide was reacted with various alkylmagnesium chlorides in a mole per mole ratio, without heating, the principal product in most cases was l-chloro-2-propanol, (Table II). A compari- son of similar reactions of propylene oxide with various alkylmagnesium bromides (12) shows that the yield of l-chloro-Z-propanol was only slightly less than the yield of l-bromo-Z-propanol. It also shows that when the alkylmagnesium chlorides were used the yields of alcohol were considerably greater than when the alkylmagnesium bromides were used. These facts point out that the alkyl-magnesium.bond is much more easily broken when the alkylmagnesium.chlorides are used than when the a1kylmagnesium.bromides are used. These facts also show that the ease of rupture of the magnesiumpchloride bond is not greatly diminished when the alkylmagnesium.chlorides are used. When the mole per mole reaction mixtures of propylene oxide and a1kylmagnesium chlorides were refluxed with benzene, the yields of alcohol were increased in all cases. The increase in alcohol yield was most evident in the iso-butylmagnesium chloride and tertiary- butylmagnesium.chloride reactions. In most cases, with the exception of those in which ethylmagnesium chloride and secondary-propyl- magnesium chloride were used, the yield of l-chloro-Z-propanol was greatly reduced. The greatest reduction in the yield of l-chloro- 2-propanol occurred when.iso-butylmagnesium chloride and tertiary- butylmagnesium chloride were used. -2 7- When the ether was completely distilled from the reaction mix— ture of propylene oxide and ethylmagnesium chloride in a mole per mole ratio and the temperature in the reaction mixture had reached 84° C., a violent reaction took place. The resulting white solid was then refluxed with benzene. This reaction caused almost a com- plete loss of l-chloro-2—propanol and a yield of alcohol very similar to that obtained in the benzene refluxed reaction (Table I). No evidences of acetone or propionaldehyde'were found in the final re- action mixture or in the material that distilled over during the violent reaction. Therefore, as the yield of 1-chloro-2-nropanol was small, it is assumed that the intermediate leading to its formation was decomposed to give propylene oxide. When one mole of alkylmagnesium chloride was reacted with two moles of propylene oxide without heat, the yield of l-chloro-Z-propanol was at least 60 percent in all cases and the yields of the alcohols ‘were greatly improved. A comparison of these results (Table II) wdth those obtained when propylene oxide and alkylmagnesium bromides were reacted under similar conditions and in like proportions (6), shows, with the exception of the iso-butylmagnesium halide reactions, that the yields of alcohol and l-halo-2-propanol for corresponding alkyl- magnesium halides are, within experimental error, identical. These facts, when compared with the results of the mole per mole reactions, show that when alkylmagnesium chlorides are used the alkyl-magnesium bond is ruptured much more easily than when the alkylmagnesium bromides are used. -28- No precipitate was obtained in any of the 1:1 ratio reactions but all of the 1:2 ratio reactions formed a gel-like solid before becoming negative to the Michler‘s ketone test. In the reaction of one mole of dialkylmagnesium with two moles of propylene oxide the expected alcohol was obtained in all cases except when di-tertiary-butylmagnesium was used. The yields of the alcohols were considerably less than those obtained when ethylene oxide was used (7). Huston and Agett allowed the reaction mixture to stand only three hours before hydrolysis. The reactions of the dialkylmagnesium compounds with propylene oxide were allowed to stand until a negative Michler's ketone test was obtained. Therefore, a comparison of these results is not entirely valid. -29... SU.’ VARY SUMHQE: 1. No evidence of the formation of acetone was discovered in any of the reactions of propylene oxide with the various a1kylmagnes- ium chlorides. No evidence of the alcohol that would have resulted from.the reaction of acetone with any of the various alkylmagnesium chlorides was found. 2. Data is presented which shows that the alkyl-magnesium bond of the alkylmagnesium chlorides was ruptured with greater ease than the alkyl-magnesium bond of the alkylmagnesium bromides. 3. The results obtained by refluxing equal molecular quantities of propylene oxide and alkylmagnesium chloride show that increased amounts of the alcohols are obtained by forcing the reactions. 4. The reaction of two moles of propylene oxide with one mole of the various a1kylmagnesium chlorides greatly increases the yield of l-chloro-Z-propanol as well as the yield of the alcohol. These results are very similar to those obtained when the corresponding a1kylmagnesium.bromides are used. -30.. S v-nwww-nr v1 .3: .5 tunnCJJ R REFERENCES Henry, L., Compt. rend., 145, 453 (1907). Hess, K., Ber., fig, 3113 (1913). Levene, P. A. and Halti, A.,J. Biol. Chem., 92, 81 (1931); 94, 367 (1931). "' Newman, V. 8., J. Am. Chem. Soc., 62, 2295 (1940). Sharefkin, J. G. and Titter, J. J., J. Am. Chem. Soc., 63, 1478 (1941). Norton, F. H. and Bass, H. 8., J. Am. Chem. Soc.,‘§§, 2147 (1936). Huston, R. C. and Appett, A. 3., J. Org. Chem., g, 123 (1941). Grignard, V.,3u11. soc. chim.,.29, 944 (1903). Veisenheimer, J. and Casper, J.,Ber., E4, 1555 (1921). Huston, 3. C. and Langham, C. C.,J. Org. Chem., 133 90 (1947). Cottle, D. L. and Hollyday, H. C., Jr.,J. Org. Chem., 12, 510 (1947). Huston, R. C. and Fostwick, C. 0. J..Crg. Chem., lg, 331 (1948). Huston, Q. C. and Brault, R. G.,Ph. D. Thesis, Nichigan State College, 1948. Columbic, C. and Cottle, 9. 1e,J. Am. Chem. Soc., El, 996 (1939). Kharasch, U. s. and Clapp, H. C., J. Org. Chem., 3, 355 (1939). Tiffeneau, V. and Fourneau, E.,Compt. rend., léé' 697 (1909). 9artlett, P. D. and Berry, C. "., J. Am. Chem. Soc., §§, 2683 (1934). Gilman, H., Wilkinson, P. D., Fishe1,'fi. P. and ”eyers, C. H., J. Am. Chem. Soc., 45, 150 (1923). Fieser, L. F., "Experiments in Organic Chemistry," 2nd Ed., D. C. Heath and Co., Poston, "ass., 1941, p. 410. Gilman, H. and Schulze, 7., J. Am. Chem. Soc., 31, 2002 (1925). -31- 22. 23. 24. 25. 26. 31. 32. 33. Shriner, R. L. and Fuson, R. C.,"Identification of Crganic Com- pounds," 3rd Ed., John Kiley and Sons, New York, N. Y., 1948. Sherrill, h. L., Valdwin, C. and Hass, 0., J. Am. Chem. Soc., 51, 3036 (1929). Ellis, L. ‘1'., Jr. and Reid, 73. 3., J. Am. Chem. Soc., 54, 1678 (1932). Sutter, P.,Helv. Chin. acta, 31, 1269 (1932). Brunel, R. 9.. J. Am. Chem. Soc., 33, 1337 (1923). Cymerman, J., Heilbron, I. ”. and Jones, E. R. 3.. J. Chem. Soc., 90 (1945). Levene, P. A. and Narker, R. 3.: J. 3i01. Chem., 21, 76 (1931). Rohen, 3;, Ann., 190, 309 (1878). Tuot, r., Compt. rend., 202, 1339 (1936). Whitmore, F. C. and Homeyer, A. H., J. Am. Chem. Soc., 55, 4194 (1933). “alone, G. 9. and Reid, 3. 3.,3. Am. Chem. 300., 31, 3426 (1929). Blaise, E. E..Compt. rend., 134, 552 (1902). '1 Ribas, I. and Tapia, 3., Anales soc. espan. fis. quim., 30, 944 (1932). JUL 6 - 153 ~ .3 "“,"'5 n 1‘) ~ I L 1 1 d2. " ‘r _ . ”-3“! 3.. V... . 'd ’3) 1'13,”— 1‘: ‘(3>°"‘31‘-j..; " ' . CHMSTRY LIBRARY " 273695 3-7 3 361. Tiefenthal rari ( CHEMISTRY ILLbMuJ T547.3 273 T561 Tiefenthal The reaction of propy- lene oxide with various a1kylmagnesium chlorides. P.) 9 U] w1191111111