422 THIAMINE 



N=C— NHsHBr 9^' CH2CH2OH 



CH3— C C— CH2— Br + N 



II II ^- 



N— CH 



N=C— NH2HBr 9^' CH2— CH2OH 



-S 



CH3— C C— CHo— N 



II II r- 



N— CH Br 



Williams et al?^ did the first reported experiment with 150 mg. of 2- 

 methyl-5-bromomethyl-6-aminopyrimidine hydrobromide and got a yield 

 of 45 % when they heated the bromo compound with 4-methyl-5-i8-hy- 

 droxyethyl thiazole in the presence of 0.2 ml. of butanol for 15 minutes at 

 120°. Stein et al}^ got 67% for the 2-ethyl homolog of thiamine. If light 

 petrolatum is used as the medium and the heating is done under optimal 

 conditions, a nearly quantitative yield may be obtained. The petrolatum 

 is removed after the reaction with ether, and the crystalline product is 

 ready for the conversion to the chloride hydrochloride (thiamine). Andersag 

 and WestphaP^ follow the same procedure (no solvent, 30 minutes at 120 

 to 130°), but no yield is given in their paper. 



To replace the bromine ions by chlorine ions, Williams et al. use an 

 aqueous suspension of silver chloride and shake the bromide solution Avith 

 it for a half-hour. The ion exchange is quantitative, and no losses of thi- 

 amine occur at this step. After filtration of the silver bromide, the aqueous 

 solution of thiamine is concentrated to a small volume and alcohol is added ; 

 thiamine crystallizes in the form of needles. After one or two crystalliza- 

 tions from 90 % ethanol the crystals are dried and ready for delivery. From 

 the silver bromide the silver may be recovered by reduction with zinc dust 

 in aqueous suspension and the excess zinc removed by diluted acetic acid. 

 The carefully washed silver is then ready for conversion into silver nitrate 

 and in turn to silver chloride. The bromide hydrobromide can also be pre- 

 cipitated with an aqueous solution of picric acid and the crystalline thi- 

 amine picrate decomposed with 10% hydrochloric acid.^* The picric acid is 

 removed by filtration and the small amount in solution by extraction ^^•ith 

 ether. The process does not ofTer any advantages which would justify the 

 hazard of pi(!ric acid and an ether extraction. 



The conversion can be avoided if 2-methyl-5-chloi()m('thylpyrimidine 

 hydi-ochloride is reacted with the thiazole. The laboratories of the I. G. 



" G. A. SU'in, W. L. Sampson, .J. K. Cliiie and J. M. SU>V(>iis, J. Am. Chem. Sac. 63, 2061 



(1941). 

 ^^ F. Hoffmann-La Roche and Co., Ltd., Basle, Swiss Pat. 197,717 (August 1, 1938). 



