[FROM T H E
KEDZIECHEMICAL LABORATORY, MICHIGAN STATE COLLEGE]
T H E REACTION OF ETHYLENE OXIDE WITH ALKYLMAGNESIUM CHLORIDES R. C. HUSTON
AND
C. C. LANGHAM
Received July 2, 1946
Earlier work by Schlenk and Schlenk (I), Ribas and Tapia (2), and Huston
and Agett (3) indicated that the reactions which take place when one mole of ethylene oxide is added to one mole of ethylmagnesium bromide may be represented by the following equations:
Direct hydrolysis of the addition product, without heating, gave sixty to seventy per cent yield of ethylene bromohydrin and little or no n-butyl alcohol (3). The principal reaction may be represented according to the equation:
(111)
Br2C4HsOzMg
+ 2Hz0
--$
2BrCH2CHz0H
+ Mg(0H)z
Good yields of n-butyl alcohol were obtained either by heating the reaction mixture :
+
+
( C Z H ~Mg Z B ~ Z C ~ E G O+ Z M(C4H90)~Mg ~ Mgh-2 (IV) or by adding a second mole of ethylene oxide:
+
(CzH5)zMg 2(CHz)z0 + (Ci"0)zMg (VI Hydrolysis of the addition product obtained by use of two moles of ethylene oxide gave a mixture of n-butyl alcohol and ethylene bromohydrin. The purpose of the present investigation was to examine the behavior of alkylmagnesium chloride solutions and ethylene oxide under similar conditions. Hydrolysis of the reaction mixture formed by adding ethylene oxide to ethylmagnesium chloride in a 1:1 molar ratio, without heating, gave a 54.6% yield of n-butyl alcohol and 22.0% of ethylene chlorohydrin based upon the titrated Grignard reagent (4). These results represent only the amounts actually recovered from the ether layer. They do not take into account the appreciable quantity of ethylene chlorohydrin destroyed during hydrolysis and that remaining in the water layer after extraction with ether. A precipitate was formed by adding ether to the reaction mixture obtained from equal molecular quantities of ethylmagnesium chloride and ethylene oxide. This precipitate upon hydrolysis gave n-butyl alcohol. The chlorine content of nine precipitates varied between 27.03 and 36.34% and the magnesium content between 16.47 and 18.21y0. In some cases the ethylmagnesium chloride was prepared by use of ethylmagnesium bromide as a catalyst; in other cases no starter was used. The 90
ETHYLENE OXIDE WITH ALKYLMAGNESIUM CHLORIDES
91
presence or absence of a catalyst seemed to make no difference either in the results of analysis or in the yields of alcohol and chlorohydrin. Since butoxymagnesium chloride contains 26.7% chlorine and 18.3% magnesium, the values obtained indicate that the precipitate may have consisted largely of either butoxymagnesium chloride resulting from the reaction :
+
C2HbMgC1 (CHJZO -+ C4HgOMgCl (VI) or (C*H$,O)Z MgeXMgC12 formed by the reactions:
(VW (VIII)
+ 2(CHz)20 ( 0 ) z M g (C4HsO)zMg + XMgCl2 (C4H90)2Mg-XMgClz (CzHs)zMg
+
-+
Hydrolysis of either intermediate would yield n-butyl alcohol. However, when the molar ratio of ethylene oxide to ethylmagnesium chloride was 2:1, there was obtained an 80.3% yield of n-butyl alcohol and 69.4y0 yield of ethylene chlorohydrin. If the intermediate compound leading directly to the formation of a large amount of n-butyl alcohol were butoxymagnesium chloride, insufficient magnesium chloride would be left in solution to give this yield of ethylene chlorohydrin. Even so, butoxymagnesium chloride may still be considered a possible intermediate, if it is assumed that as the solution becomes depleted of magnesium chloride, decomposition occurs according to the equation :
(IW
2C4HsOMgC1-+ (C4HgO)eMg
+ MgClz
Thus it seems evident that the immediate precursor of the n-butyl alcohol is dibutoxymagnesium which could be formed by the reaction of ethylene oxide with either ethylmagnesium chloride or diethylmagnesium. The reaction of the epoxide with diethylmagnesium appears more probable in view of the fact that ethylmagnesium chloride, according to Schlenk ( 5 ) , is approximately eighty-five per cent dissociated into diethylmagnesium and magnesium chloride. Evidence submitted by Noller and Castro (6) substantiates the theory that dibutoxymagnesium combines with varying amounts of magnesium chloride. They observed that the precipitate formed by the oxidation of n-butylmagnesium chloride removed magnesium chloride from solution. Noller and Castro suggested that if butoxymagnesium chloride is the product of oxidation, it should have the property of forming a complex with magnesium chloride, probably by virtue of the unshared electrons on the oxygen atom. Obviously dibutoxymagnesium should behave like butoxymagnesium chloride in this respect. Further evidence for the precipitation of (C4H90)2Mg.XMgClzwas provided by analysis of the precipitate formed by the use of one-half mole of ethylene oxide to one of ethylmagnesium chloride. In this case the chloride content was found t o be 40.8% and the magnesium content was 19.670. These are much higher values than the ones obtained when equimolecular quantities of the reactants were used. This is the result that should be expected under these
92
R. C. HUSTON AND
C. C. LANGHAM
conditions since more magnesium chloride would be available to precipitate with dibutoxymagnesium. Theoretical values for (C4HsO)zMg.2MgC12 are 39.3% chlorine and 20.2% magnesium. There are two possible intermediates for ethylene chlorohydrin :chloroethoxymagnesium chloride formed by the reaction :
(X)
MgClz
+ (CH2)ZO + ClCzH40MgCl
or dichlorodiethoxymagnesium formed by the reaction :
(XI)
MgCh
+ 2(CHz)z0
+
(ClCdL0)zMg
According to equation (X) the theoretical yield of ethylene chlorohydrin per mole of Grignard reagent is one-half mole. However, if the percentage yield is calculated on this basis, when ethylene oxide and ethylmagnesium chloride come into reaction in a 2:l molar ratio, the amount of halohydrin actually TABLE I PERCENTAGE YIELDSOF ALCOHOLS AND ETHYLENE CHLOROHYDRIN BASEDUPON TITRATED GRIGKARD REAGENTS
THE
RMgCl -t 2CzH40 NOT HEATED
Alcohol
1
Alcohol
Chlorohydrin
45.9 49.9 46.5 45.8 35.1 29.8
16.4 32.0 19.5 36.8 23.9 25.0 23.3
-___-
Ethyl ............................... n-Propyl ............................ Isopropyl ........................... n-Butyl ............................. &Butyl.. ........................... Isobutyl.. .......................... &Butyl. ............................
54.6 39.5 34.5 39.0 25.6 23.6 0.0
22.0 31.5 35.1 25.5 30.9 34.6 21.3
0.0
*lcohol
80.3 66.4 54.8 69.6 40.9 35.2
I
Chlorohydrin
69.4 54.0 58.0 60.1 61.5 60.4
obtained leads to far more than a hundred per cent. Hence, we feel justified in the assumption that the intermediate leading to ethylene chlorohydrin is dichlorodiethoxymagnesium . A survey was made of the reaction of ethylene oxide with the Grignard reagents of several alkyl chlorides. The yield of the alcohol and of ethylene chlorohydrin was determined in each case. The Grignard reagents were prepared from the alkyl halide and pure magnesium turnings. All reactions and transfers were carried out in an atmosphere of nitrogen. Each Grignard reagent was treated in three different ways: 1. Treated with one mole of ethylene oxide and hydrolyzed without heating. 2. Treated with one mole of ethylene oxide and then refluxed with benzene before hydrolysis. 3. Treated with two moles of ethylene oxide and hydrolyzed without the addition of benzene or heating. The results of the three methods of procedure are compared in Table I. The table shows that three of the Grignard reagents reacted with ethylene oxide in a mole-per-mole ratio without heating to give the alcohol as the principal
ETHYLENE OXIDE WITH ALKYLMAGNESIUM CHLORIDES
93
product. s-Butylmagnesium chloride and isobutylmagnesium chloride gave somewhat higher yields of halohydrin than of the alcohols. In agreement with results reported by Whitmore (7), t-butylmagnesium chloride, when treated with ethylene oxide, yielded only ethylene chlorohydrin. The data also reveal that in respect to the two secondary Grignard reagents and isobu tylmagnesium chloride, heating the reaction product before hydrolysis appears to enhance somewhat the yield of the alcohol a t the expense of the halohydrin. In other cases the result of heating was to increase the yields of both the alcohol and of the ethylene chlorohydrin. The reaction product from ethymagnesium chloride and ethylene oxide could not be heated successfully because the mixture solidified upon attempting to distil off the ether prior t o heating with benzene. This resulted in considerable decomposition of the intermediate product due to excessive heating near the side-walls of the flask. In most cases addition of two moles of ethylene oxide to one mole of the Grignard reagent greatly increased the yield of both the alcohol and the ethylene chlorohydrin. When the boiling points of the alcohol and halohydrin were too close together t o allow a good separation by distillation, the two compounds were collected together. The chlorohydrin was determined by hydrolysis with sodium hydroxide and titration of the chloride ion by the Volhard method. The alcohol was determined by difference. Attempts to prepare dichlorodiethoxymagnesium in the pure state by treating a stirred suspension of dry magnesium chloride with ethylene oxide were not successful. This was accomplished by the treatment of diethylmagnesium in ether solution with the molecular equivalent of ethylene chlorohydrin. The solid compound dissolved readily in cold ethylmagnesium chloride. Hydrolysis of the precipitate, either directly or after dissolving in Grignard reagent (C2HJ\lgC‘1), gave ethylene chlorohydrin. EXPERIMEXTtlL
Preparation of ethylmagnesium chloride. Two and one-quarter moles (51.7 g.) of magnesium turnings was placed in a dry tor-o-liter, three-necked, round-bottomed flask. The flask was fitted with a mercury-sealed stirrer which was equipped with an inlet tube for nitrogen, a “dry-ice” condenser, and a n inlet tube for ethyl chloride. The ethyl chloride tube extended nearly to the bottom of the flask. After flushing out the system with dry nitrogen and ethyl chloride, one liter of anhydrous ether was added to the flask and the reaction catalyzed by the addition of a small amount of ethylmagnesium bromide. “Dryice” was then placed in the condenser. Ethyl chloride mas bubbled through concentrated sulfuric acid1and added at such a rate as to maintain a moderate rate of reaction. hfter most of the magnesium had reacted, addition of the ethyl chloride was discontinued and stirring maintained until the reaction mixture reached room temperature. This latter stage of the reaction was carried out in a very slow stream of nitrogen. The flask was sealed off from air and allowed t o stand overnight. Preparation of other Grignard reagents. Two and one-quarter moles (54.7 9.) of magnesium turnings was placed in a two-liter fiask fitted with a mercury-sealed stirrer which was equipped with an inlet tube for nitrogen, a reflux condenser, and a dropping-funnel. After flushing out 1 he system with dry nitrogen, four or five grams of the alkyl halide mixed with one hundred ml. of anhydrous ether was added t o the flask. The reaction was started by the addition of a small amount of ethylmagnesium bromide and a n additional hundred ml.
94
R. C. HGSTON AND C. C . LANGHAM
of ether. The balance of the alkyl halide to make a total of two moles was mixed with eight hundred ml. of ether and added a t such a rate as t o maintain moderate refluxing. After addition was complete, a slow stream of nitrogen was introduced and stirring continued for two hours. The product was allowed t o stand overnight. The preparation of t-butylmagnesium chloride was carried out very slowly. Reaction of ethylene oxide and Grignard reagent in a 1:l molar ratio. A standard acid-base titration (4)was used to determine the concentration of the Grignard reagent. The total volume of the solution was measured, and i t was transferred (in an atmosphere of nitrogen) t o another three-necked, round-bottomed flask. This flask was also fitted with a mechanical stirrer, reflux condenser, and a dropping-funnel equipped with an outer metal jacket to permit the use of “Dry-ice” as a cooling agent. The Grignard reagent was cooled in an ice-salt bath. One hundred ml. of anhydrous ether was placed in the dropping-funnel and surrounded with “Dry-ice.” The calculated amount of ethylene oxide which previously had been dried over soda-lime was weighed out and placed in the dropping-funnel. The ethylene oxide-ether mixture was cautiously added to the Grignard reagent. As soon as addition was complete, the ice-salt bath was removed and the mixture was allowed t o stir for an hour. The product was hydrolyzed by pouring i t on chipped ice. After hydrolysis the ether layer was decanted. The pasty magnesium hydroxide was extracted three times with small portions of ether, neutralized with dilute hydrochloric acid (iced), and again extracted three times with ether. The ether extracts were combined, dried over sodium sulfate and fractionated through a twelve-inch, Fenske-type column packed with & inch glass helices. Reaction of ethylene oxide and Grignard reagent in a 1:i molar ratio followed b y heating. After adding the ethylene oxide to the Grignard reagent as described above, the condenser was set for distillation. The reaction-flask was placed in a Glas-Col heating mantle and about seven hundred ml. of ether removed. The temperature of the reaction mixture was not allowed to exceed 45”. Five hundred ml. of anhydrous benzene was added and gentle refluxing continued for six hours. The product was hydrolyzed, extracted, dried, and fractionated as described in the previous experiment. Reaction of ethylene ozide and Grignard reagent in a b:1 molar ratio. The calculated amount of ethylene oxide was mired with one hundred mi. of anhydrous ether and added t o the Grignard reagent according to the procedure described above. Since the ethylene oxide tended to polymerize, it was usually not possible to stir the mixture for an hour after addition was complete. The product was hydrolyzed immediately, extracted, dried, and fractionated as previously described. A n a l y s i s of the intermedinte compound f r o m the reaction of ethylene oxide and Grignard reagent in a 1:l molar ratio. The preparation of the Grignard reagent (C.HjMgC1) and addition of ethylene oxide was accomplished according t o the previously described procedure. Part of the reaction mixture was diluted with ether and centrifuged for ten minutes at 1800 r.p.m. The precipitate was washed with ether, transferred quickly to a weighing-bottle which was placed in a desiccator and evacuated with a water-pump for three hours and with an oil-pump at about 2 mm. for twelve hours. Total chlorine was determined by sodium peroxide fusion in a Parr bomb and subsequent titration by the Volhard method. Magnesium was determined as MgO after heating to constant weight with a Meker burner. Determination of the alcohol and ethylene chlorohydrin. The product was distilled and the alcohol and halohydrin fractions collected together. After determining the total weight of the distillate, five hundred ml. of 0.3 molar sodium hydroxide was added and the mixture refluxed with stirring for a half hour. The contents of the flask was cooled, placed in a separatory funnel and the water-layer drawn off into a liter volumetric flask. The solution was made up to volume and five-nil. samples taken for analysis. Chlorine mas determined by means of the Volhard method. The amount of ethylene chlorohydrin was calculated from the results of the chloride analysis and the weight of the alcohol obtained by difference.
ETHYLENE OXIDE W I T H ALKYLMAGNESIUM CHLORIDES
95
Preparation of dichlorodiethoxymagnesium. One mole of diethylmagnesium in solution was added, dropwise, to a stirred solution containing one mole of ethylene chlorohydrin in ether. The reaction mixture m s centrifuged and the precipitate washed with anhydrous ether. Calc’d for CdH&12MgOz: C1, 38.68; Mg, 13.27. Found: (21, 37.92; Mg, 13.87. The precipitate was suspended in anhydrous ether and Grignard reagent (C2H;MgCl) was added drop by drop until solution was complete. Hydrolysis of dichlorodiethoxymagnesium, either in ether suspension or in solution, gave ethylene chlorohydrin. SUMMARY
1. The main reaction between one mole of ethylene oxide and one mole of Grignard reagent formed by the treatment of alkyl chlorides with magnesium is between the epoxide and either dialkylmagnesium or alkylmagnesium chloride. 2. The intermediate in the formation of 1%-butylalcohol from diethylmagnesium is dibutoxymagnesium. From ethylmagnesium chloride it is butoxymagnesium chloride which by rearrangement gives dibutoxymagnesium and magnesium chloride, The high yield of both alcohol and chlorohydrin when excess of ethylene oxide is used indicates, in the main, dibutoxymagnesium. 3. Analytical results indicate that the composition of the precipitate formed when the reaction mixture of one mole of ethyl Grignard reagent and one mole of ethylene oxide is diluted with ether approaches (C4H90)2Mg.MgC12. When the ratio of Grignard reagent to ethylene oxide is 2: 1 it approaches the composition of (C4H90)2Mg.2MgC1~. 4. Evidence submitted indicates that the precursor of the ethylene chlorohydrin is ,dichlorodiethoxymagnesium. 5. Results obtained by heating the reaction mixture from equal molecular quantities of alkylmagnesium chloride and ethylene oxide do not indicate a high degree of reactivity between dichlorodiethoxymagnesium and dialkylmagnesium. 6. Addition of two moles of ethylene oxide to one of alkylmagnesium chloride greatly increased the yields of both the alcohol and ethylene chlorohydrin. EASTLANSING, MICH. REFERENCES (1) (2) (3) (4) (5) (6) (7)
SCHLENK AND SCHLESK, Ber., 62, 920 (1929). RIBASA N D TAPIA, Andes SOC. espafi.f i s . quim., SO, 944 (1932). HUSTON AND AGETT,J . Org. Chem., 6, 123 (1941). GILMAN,WILKINSON,FISHEL, AND MEYERS, J.Am. Chem. SOC.,46,150 (1923). SCHLENK, Ber., 64,734 (1931). SOLLER A.ND CASTRO, J.Am. Chem. SOC., 64,2509 (1942). WHITMORE,AND WALLINGPORD, J . Am. Chem. SOC., 66,4209 (1933). HOMEYER.,
