418 THIAMINE 



The normal way of bulling up a thiazole ring starts with a-halogenocar- 

 bonyl compounds, and all practical syntheses for the 4-methyl-5-jS-hy- 

 droxyethyl thiazole use the pentanone chain, where X is a halogen, and R 



CH3— CO— CHX— CHo— CH2R 



is a hydroxyl, ethoxyl, 0-acyl, or halogen. As the final thiazole derivative 

 requires a free /3-hydroxyethyl group and the exchange of an ethoxyl by 

 hydroxyl is not too easy, it is obvious that compounds with the free hy- 

 droxyl or 0-acyl are superior ; the us^ of 7-aceto-7-chloropropylethyl ether 

 as described by Clarke and Gurin^^ as early as 1935 is therefore out of the 

 question for technical purposes. 



If 7-acetopropylacetate is brominated in ether, bromine enters mainly 

 in the 7-position and 7-aceto-7-bromopropylacetate is obtained (Andersag 

 and Westphal-^). The bromination can be done under very mild conditions, 



CH3— CO— CHBr— CH2— CH2OCOCH3 



i.e., in glacial acetic acid with a mixture of bromine and pyridine hydro- 

 chloride.^^ Whatever the conditions of direct halogenation with chlorine or 

 bromine, the possibility of substitution on other carbon atoms of the chain 

 is given and any purification of these mixtures of halogenated ketones is 

 difficult. 



If a second carbonyl is present at carbon atom 3, however, the chlorina- 

 tion can be performed with sulfuryl chloride and only substitution in 3 

 position takes place. Buchman^^ has shown that a-acetobutyrolactone can 

 be chlorinated with sulfuryl chloride to a-chloro-a-acetobutyrolactone with 

 a very good jdeld (83 %). When this lactone was decarboxylated by boifing 

 with concentrated hydrochloric acid, Buchman expected 7-chloro-7-aceto- 

 propyl alcohol and reported a yield of 73 %. As Stevens and Stein*° have 



CH3— CO— CHCl— CH2— CH2OH or CH3— C(OH)— CHCl— CH2— CH2— O 



shown, only a small part of the reaction product is the free alcohol ; by far 

 the larger part forms an ether from 2 moles of the expected alcohol by the 

 elimination of 1 mole of water. The yield was about 13 % alcohol and 62 % 



OCH2— CH2— CHCl— CO— CH3 

 CH3— C— CHCl— CH2— CH>— O 



" H. T. Clarke and S. Gurin, /. Am. Chem. Soc. 57, 1876 (1935). 



'8 Roche Products, F. Bergel, and A. Cohen, British Pat. 554,428 (July 2, 1943); 



Roche Products and F. Bergel, British Pat. 550,197 (December 29, 1942). 

 " E. R. Buchman, /. Am. Chem. Soc. 58, 1803 (1936). 

 "> J. R. Stevens and G. A. Stein, /. Am. Chem. Soc. 62, 1045 (1940). 



