Factors Governing the Reaction of the Benzyl Grignard Reagent. II

III. Formation of ortho and para products in reactions with alkyl sulfates via triene intermediates. Robert A. Benkeser , Thomas E. Johnston , Wen-Hon...
0 downloads 0 Views 953KB Size
2141 Cyclooctanecarboxylic acid and cyclodecanecarboxylic acid were prepared by carbonation of the cycloalkylmagnesiumchlorides with crushed Dry Ice: cyclocctanecarboxylic acid16 (4497,), bp 108" (0.12 mm), and cyclodecanecarboxylicacid1*(6797,), bp 122123'(0.11 mm), mp 53'. Electrolysis of Carboxylic Acids. The apparatus consisted of a 250-ml electrolytic beaker fitted with a large rubber stopper carrying a thermometer, a Dry Ice finger condenser, and two electrodes. The electrodes were rods 0.55 in. in diameter; the anode was graphite, the cathode, copper. During electrolysis, the mixture was stirred by a magnetic stirrer. The current source was a pair of 85-amp, 12-v batteries in series. A current density of 0.045 amp/cma was maintained for all electrolyses; this current density amounted to 0.8 amp with these electrodesand 100 ml of solution. The current efficiency was approximated by dividing the moles of acid consumed by the faradays of electricity passed as estimated from the volume of gas evolved. Three other cell designs were investigated. A platinum anode was used in place of the graphite one, and a circular copper cathode, perforated to facilitate stirring, was used to allow a uniform field and current density around each anode. An aqueous mixture consisting of 0.2 mole of cycloalkanecarboxylic acid, 0.075 mole of sodium hydroxide, and enough water to bring the volume to 100 ml was electrolyzed in the cell for 6 8 hr. The mixture was made strongly alkaline with sodium hydroxide, filtered to remove graphite which had crumbled from the anode, and extracted three times with petroleum ether. Cycloalkanecarboxylic acid was recovered from the aqueous layer. The organic extract was dried and distilled. The lower boiling hydrocarbon fraction was easily separated from the higher boiling alcohol and ester products. The hydrocarbon fraction was analyzed by gas chromatography; components in the mixture were identified by comparison of retention times with those of authentic samples and standard mixtures. The data are summarized in Table I. The volume of the gas evolved was measured by its displacement of water and used to compute the current efficiency of the electrolysis. (15) M. Godchot and M. Caquil, Chim. Ind. (Paris), 29, 1019 (1933).

Preparation of Authentic Hydrocarbon Samples for Analysis. cis-Bicyclo[x.l .O]alkanes were prepared by methyleneation of the corresponding cis-cycloalkenes with methylene iodide and zinc-copper couple.16 The properties were as follows. cis-Bicyclo[4.1.O]heptane16 had bp 115",n% 1.4549; cis-bicyclo[6.1.O]nonane~~ had bp 71-71.5" (26 mm), n% 1.4682. cis-Bicyclo[7.1.0]decane had bp 83-84" ~ Anal. Calcd for ClOHl8: C, 86.9; H, (16 mm), n Z 4 1.4740. 13.1. Found: C, 87.05; H, 13.0. The nmr spectrum included multiplet signals centered at 0.4 (1 H), -0.6 (4.4 H), - 1.6 (12.6 H), and -1.8 to -2.2 ppm (2 H). cis-Bicyclo[8.1.O]undecane had bp 76-77" ( 5 mm), PD 1.4786. Anal. Calcd for CllH20: C, 86.8; H, 13.2. Found: C, 86.8; H, 13.2. The nmr spectrum included multiplet signals centered at 0.5 (1 H), -0.6 (3.4 H), 1.3 to - 1.7 (12.6 H), and - 1.9 ppm (3 H). Bicyclo[4.2.O]octane was prepared by catalytic hydrogenation of bicyclo[4.2.0]oct-7-ene,l7 prepared by photoisomerization of 1,3cyclooctadiene.l 7 The saturated hydrocarbon was obtained with platinum on carbon as catalyst but not with palladium on carbon as catalyst, which apparently isomerized either the bicyclooctene or the bicyclooctane or both. The nmr spectrum of bicyclo[4.2.0]Octane includes signals centered at - 1.47 (unresolved multiplet, 8 H), 1.82 (quartet with small side bands, 4 H), and -2.32 ppm (unresolved multiplet, 2 H). cis-Cycloheptene and cis-cyclooctene were available commercially. A mixture of cis- and trans-cyclodecenewas obtained from Columbian Carbon Co., Lake Charles, La., and another was prepared from cyclodecanol by treatment with ptoluenesulfonyl chloride in pyridine solution. A commercial sample of decalin was resolved into cis and trans isomers by preparative gas chromatography. Mixtures of other hydrocarbons were prepared by decomposition of the appropriate cycloalkanone tosylhydrazones;6 peak identities in the gas chromatograms were estab lished by reference to literature descriptions of the mixtures.6

-

-

(16) H. E. Simmons and R. D. Smith, in Org. Sun., 41, 72 (1961). (17) S. F. Chappell, 111, and R. F. Clark, Chem. Ind. (London), 1198 (1962). A sample of 1,3-cyclooctadiene was generously provided by Columbian Carbon Co., Lake Charles, La.

Factors Governing the Reaction of the Benzyl Grignard Reagent. 11. Evidence for Triene Intermediates in the Reaction with Chloromethyl Methyl Ether Robert A. Benkeser and William DeTaIvo Contributionf r o m the Chemical Laboratories of Purdue University, Lafayette, Indiana 47907. Received August 8 , 1 9 6 6

Abstract: T h e reaction of the benzyl Grignard reagent with chloromethyl methyl ether has been studied in detail. There are at least five major products of this reaction: 2-phenylethyl methyl ether (I), 0-(11) and p-methylbenzyl methyl ether (111), and 0- (IV)and p(2-methoxyethy1)benzyl methyl ether (V). When tetrahydrofuran is substituted for diethyl ether as a solvent for the reaction, the yield of I is increased and the diethers IV and V are eliminated. The substitution of benzylmagnesium bromide for benzylmagnesium chloride causes a n increased yield of

XI. Benzyllithium forms only compound I in good yield. The major portion of compounds I1 and 111 is formed during the hydrolysis step with strong acid (e.g., aqueous HC1). When hydrolysis was effected with ammonium chloride, compounds I1 and I11 were virtually eliminated. The intermediate responsible for the formation of I1 was characterized in part by causing it t o react with maleic anhydride and diethyl maleate. Trimethylchlorosilane failed t o trap Grignard intermediates which could conceivably function as precursors of the diethers IV and V. Conclusions based on these results were vitiated by the observations that the benzyl Grignard reacts more rapidly with chloromethyl methyl ether than with trimethylchlorosilane. The intermediate responsible for the formation of I1 could be made to react with gaseous formaldehyde in the presence of anhydrous magnesium chloride in a Prinstype reaction. The same intermediate failed to react with chloromethyl methyl ether under the same conditions.

The

benzyl Grignard reagent possesses the intriguing capability o f reacting at the (Y position, to produce so-ca11ed products, as as at the Ortho and para positions to form "abnormal" products.

In the first paper in this series,2 it was shown that (1) See M. S. Kharasch and 0. Reinmuth, "Grignard Reactions of Non-Metallic Substances," Prentice-Hall, Inc., New York, N. Y., 1954, p 1133, for pertinent literature references.

Benkeser, DeTalvo J Benzyl Grignard Reaction with Ether

2142

when the benzyl Grignard reacts with acetaldehyde and trifluoroacetaldehyde, only a! and ortho products are formed but no para. In addition, ortho-substituted diols are produced in increasing quantities as the concentration of the carbonyl reagent is increased relative to the Grignard. Similar concentration effects had been noted earlier3 in reactions of the benzyl Grignard with benzaldehyde and citronellal, and mechanisms have been p r ~ p o s e dto~ explain ~ , ~ such effects. We now wish to report the results of a study involving a displacement reaction of the benzyl Grignard. It had been reported5 that treatment of benzylmagnesium halides with chloromethyl methyl ether produced both ortho and para coupling products along with the expected CY product. While a careful reexamination has substantiated this earlier r e p ~ r t , ~several a new products have been isolated from this reaction and some new observations made which tend to shed more light on the over-all process and the mechanism whereby it is occurring. We have found that at least five products are usually produced in the reaction between the benzyl Grignard reagent and chloromethyl methyl ether. The yields of some of these products are dependent to a small extent upon the relative concentrations of the two reactants.

and trifluoroacetaldehyde2 is obvious. It should be noted, however, that with the carbonyl additions,Z no para product corresponding to either TI1 or V was found. The identification of compounds IV and V was achieved by comparing the infrared spectrum of samples isolated by vapor phase chromatography from the reaction mixture with those obtained from authentic samples (see Experimental Section). As will be noted from entries 1-4 in Table I, the amounts of IV and V increased slightly as the ratio of the concentration of the chloromethyl methyl ether to Grignard was gradually decreased. At the same time, the percentages of I1 and I11 gradually decreased. Table 1. Reaction of Benzylmagnesium Halidesa-c with Chloromethyl Methyl Ether ReacGrignard : tion Run ether time, no. ratio hr 1 2 3

4 5 6 7 8 9 10 11

CICHnOCH, EtnO

6

+

+

@cH20cH3

I

I11

3

42

3 3 3

43

13 17 8 6 15 9

12 10 6

0.25 24 3

3 3 3 3

56 53

42 43

57 36 35 44 63

6

12 10

13

6

9

11 9 9 0

8 24 0

z IV Trace 1 5 5

0.5 0.5 0

V Trace 1 6 5 0.5 0.5 0

3

5

3 Trace 0

5 Trace 0

z

o + CH20CH3

I1

The initial Grignard concentration was always 0.4 M. In all runs except 7, anhydrous diethyl ether was employed as solvent. h All reactions listed in this table were hydrolyzed with 15 HCl. c In most instances, the percentages listed are the average values for duplicate runs. In this run, T H F was used as the solvent. e In this run, inverse addition was used, i.e., the Grignard solution was added dropwise to the chloromethyl methyl ether. f Inverse addition was employed (i.e.,same as 8), but 0.3 mole of anhydrous MgCll present in ethereal solution containing 0.3 mole of chloromethyl methyl ether. 0 Benzylmagnesium bromide was used in this run instead of the chloride. h Benzyllithium was used instead of the Grignard reagent.

CH3 I

IV

I

a

I1

b

3:2 1:l 1:2 1:4 1:l 1:l 1:ld 1:le 1:l' l:l0 1:2*

I

+

0 I

CHpOCH3

v As shown in the equation, in addition to 2-phenylethyl methyl ether (I), o-methylbenzyl methyl ether (11), and p-methylbenzyl methyl ether (111) already reported,ja we have detected varying amounts of the diethers, 0-(2-methoxyethyl)benzyl methyl ether (IV) and p-(2methoxyethy1)benzyl methyl ether (V). These two latter compounds are new and have never been reported as products of this reaction. The analogy between these diethers and the diols in the case of acetaldehyde (2) R. A. Benkeser and T. E. Johnson, J. Am. Chem. Soc., 88, 2220 (1966). (3) (a) J. Schmidlin and A. Garcia-Banus, Ber., 45, 3193 (1912); (b) W. G. Young and S. Siegel, J. Am. Chem. Soc., 66, 354 (1944); (c) S. Siegel, S. K. Coburn, and D. R. Levering, ibid., 73, 3163 (1951). (4) S. Siegel, W. M. Boyer, and R. R. Jay, ibid., 73, 3237 (1951). (5) (a) L. Malm and L. Summers, ibid., 73, 362 (1951); (b) H. Gilman and J. E. Kirby, ibid., 54, 345 (1932). See also A. C. Bottomley, A. Lapworth, and A. Walton, J. Chem. Soc., 2215 (1930).

An arbitrary reaction time of 3 hr was chosen for the study of concentration effects. A comparison of entries 2 and 5 (Table I) would indicate that this is a very rapid reaction and is almost complete within 15 min. Indeed, a comparison of entries 2 and 6 would indicate that prolonged reaction times (like 24 hr) have a deleterious effect on yield. The amount of compound I1 in particular seems to drop off in this period. We will return to a discussion of this point later. At least three other items of interest can be gleaned from the data of Table I. (1) Tetrahydrofuran (entry 7) is a propitious solvent for increasing the rate of CY coupling. In this run, the percentage of compound I was 13 higher than when diethyl ether was employed as solvent under otherwise identical conditions. (2) The use of benzylmagnesium bromide (instead of the chloride) increased the amount of compound I1 produced. The percentages of all other products in the reaction (compare entries 2 and 10, Table I), except for 11, are almost identical. (3) The use of benzyllithium instead of the Grignard reagents resulted in an exceed-

Journal of the American Chemical Society / 89:9 / April 26, 1967

Grignard : Run ether Mode of no. ratio hydrolysis

I

I1

z

111

IV

H CHZOCH,

V

VI

~~

1 2

1:l 1:l

AqHC1 AqNHaCl

43 45

17 1

10 2

1