III. INDUSTRIAL PREPARATION 423 



Farbeiiindustrie'® used this process in 1943 (methyl acetamidc as a solvent 

 heating 1 hour at 115°), but the yield of pure thiamine in batches of 3 kg. 

 was only 45 %. The use of the bromo compoimd for the quaternization 

 followed by the ion exchange to the chloride hydrochloride seems to be 

 much superior. 



8. Thiazolium Ring Closure to Thiamine 



The model experiments of Todd ct al.*'^ for the thiazolium ring closure 

 e.g., X-phenylthioacetamide and chloroacetone (without a solvent at 80°), 

 gave quantitati\'e yields. But when they heated 7-chloro-7-acetopropanol 

 with 2-methyI-5-thi{)formylamine pyrimidine in dioxane (b.p. 101°), no 

 thiamine could be obtained; heating at 140° without a solvent gave con- 

 siderable resinification, darkening, and only a small yield. The results were 

 much better when 7-chloro-7-acetopropylacetate was used; after heating 

 at 115 to 120° for a few minutes, the ring closed in a smooth reaction. 



X=C— NHo CH3 CH2— CH2OCOCH3 



II II 



CH3— C C CH2— NHCHS + CO— CHCl -^ 



II II 

 X— CH 



thiamine monochloride + CH3COOH 



At this temperature the acetyl group was saponified and thiamine (as 

 chloride) was obtained directly. No indication regarding the jaeld is given. 

 By using 7-bromo-7-acetopropylbenzoate, Andersag and WestphaP could 

 lower the temperature for the ring closure to 100°, but here again figures 

 for the yield are lacking. Roche Products (Welwyn Garden City, Herts) 

 uses 7-bromo-7-acetopropylacetate for their actual manufacturing process. ^^ 

 The reaction between 2-methyl-2-ethoxy-3-chlorotetrahydrofurane and 

 the thio compound, as described by Klingenfuss,^^ used even milder condi- 

 tions : 95 % formic acid or 80 % acetic acid as a solvent and reaction tem- 

 peratures of 40 to 50° (10 to 20 hours). If calcium bromide was added to the 

 reaction mixture, a yield of 54% of thiamine was obtained. 



A similar yield (57%) is obtained when the thio compound is reacted 

 with a mixture of 2-methyl-2,3-dichlorotetrahydrofurane in formic acid 

 with pyridine; the pyridine splits off hydrogen chloiide, and the inter- 

 mediate reacts under ring closure. It is obvious that this process" was de- 

 veloped to circumvent existing patents, but it does not offer any advan- 

 tages. 



" A Factory lor Vitmniii li. Production, Tnd. Cheinisl, Juiio 1947, 350-68. 

 "M. KliiiKciifuss, U.S. Pat. 2,127,446 (August 16, 1938). 

 "Chinoin, Budapest, British Pat. 609,803 (October 7, 1948). 



