412 



THIAMINE 



Williams' process grew in the hands of the chemists of Merck and Co., 

 Rahway, New Jersey; the ring closure of Todd and Bergel was adopted 

 industrially by Hoffmann-La Roche in their three plants in Basle, Nutley, 

 New Jersey, and Welwyn, Herts. Andersag and Westphal's method has 

 not yet reached the production in tons. 



The established structure of thiamine, opens three possibilities for the 

 synthesis : 



1. Building up the pyrimidine ring and the thiazole ring separately, then 

 connecting both by quaternization to the thiazolium ring. 



2. Building up the pyrimidine ring with the group — CH2 — NH2 in the 

 5 position, elongating this side chain, and forming the second ring by ring 

 closure to the thiazolium compound. 



3. Building up the thiazole ring, quaternization to thiazolium by elonga- 

 tion with a suitable side chain, and ring closure of the pyrimidine ring. 



All three ways are described in literature. The first method is used b}^ 

 Williams et al. and by Andersag and Westphal: as the two ring systems are 

 built up separately, the best conditions for the ring closure can be chosen, 

 namely, alkaline condensation for the pyrimidine and acid medium for 

 the thiazole nucleus. The quaternization goes easily and gives a nearly 

 theoretical yield, an important factor for the success of the method. 



Todd and Bergel follow the second possibility: the elongation of the side 

 chain is done with the simple and elegant step of thioformylation of the 

 amine, the ring closure to the thiazolium being performed in an acid 

 medium, both reactions giving excellent yields. 



The third possibility seems to have only slight chances, as pyrimidine 

 rings are best built up in an alkaline medium, a condition highly unfavor- 

 able for the thiazolium ring. This way is described for a thiamine homolog 

 in a French patent,^^ but no experimental data are given. There is no evi- 

 dence that the pyrimidine ring closure was performed. 



The following discussion of the industrial synthesis is based on the 

 technical point of view, with special emphasis on the reported yields. 



For practical reasons the molecule is divided into four groups. 



N=f=C-XH,-HCl 



I 13) I (4j I 



CH3-C"2jl (.■;)C CHo 



I (1)1 (6)1 ^7' 



N^^l— CH 



C 

 CH3 CH.-CHj-OH 





= C 



•5) 



N^3) 



CH S 



CI 



D 



" I. G. Farbenindustrie, A. G., French Pat. 816,432, p. 5 (May 3, 1937). 



