rnents resulted in the formation of 93.5 g . of n-propyl nicotinate (b. p. (1 mm.) 80-82"). This represented a yield of 56.7%. Nicotine and &picoline behaved in a similar manner and gave yields of the esters in direct proportion to their ease of oxidation t o nicotinic a c i d a U. S. DEPT. AGRICULTURE EASTERN REGIONAL RESEARCH LABORATORY CHESTNUT HILLSTATION I'IIII~ADELPHIA,P A . RECEIVED NOVEMBER 16. 191-1
sets in. Dihydropyrane was not the only product of the reaction but there was material boiling below and above this substance. At 450' about 3 0 water ~ ~was formed and most of the non-aqueo i l s product appeared as high boiling still residues. Ov-r thoria, tetrahydrofurfuryl alcohol was ;tible up to 450'. The low-boiling material collected during the 550' run polymerized on standing and had a sharp odor. These facts and its boiling point suggest that it may have consisted of pentenes and pentadienes. It should be Dehydration of Tetrahydrofurfuryl Alcohol1 pointed out that the recovery of tetrahydrofurB Y CHARLESHOWARD KLINE,J R . ,AND ~ JOHN TURKEVICHfuryl alcohol was high. The inactivity of the In the course of study of the catalytic synthesis thoria for the dehydration was not due to its genof pyridine from furfural derivatives, the dehy- eral Iack of catalytic activity. After revivificadration of tetrahydrofurfuryl alcohol was investi- tion following the 450' run, the catalyst could be gated. Paula has found that the dehydration of used to dehydrate anhydrous ethanol, One is tetrahydrofurfuryl alcohol over alumina results thus faced with the interesting fact that on i n the formation of I ,2-dihydropyranej CbHSO, but alumina the efficient dehydration of tetrahydrodid not investigate in detail the yields produced furfuryl alcohol poisons the alumina for the dedue to variation in temperature, contact body or hydration of ethanol, while on thoria the inei3cient dehydration of tetrahydrofurfuryl alcohol lifetime of the catalyst. The materials and experimental procedure were does not poison the dehydration of ethanol. the same as those described by the authors in the FRICK CHEMICAL LABORATORY UNIVERSITY publication of the catalytic synthesis of ~ y r i d i n e . ~PRINCETON RECEIVED J A N U A R Y 26, 1945 Over aIumina it was found that tetrahydrofur- I'RIVCETOI', KEV' JERSEY furyl alcohol is unattacked a t 250' but is readily dehydrated a t 300 and 350". The latter tempera- A Relation between Viscosity and Refractive :ure is the optimum temperature for the dehydraIndex tion and yields about 70% dihydropyrane. At BY R. T. LAGEMANN 400' and above, high boiling material is chiefly iormed and much of the feed alcohol is lost either During an attempt to estimate the viscosity of as cracked gases (19% carbon dioxide, ,5795 un- higher members of some homologous series, it was saturates, 13% hydrogen and 11% saturated noted that the viscosity and the refractive index hydrocarbons) or catalyst deposit. Condensafor liquids are functionally related. A linear ret ion of the tetrahydrofurfuryl alcohol residues appears to predominate a t these temperatures, lation is found to hold for every homologous series if the molecular refraction R is plotted against The alumina catalyst did not lose activity a t 300' and a t 350' the activity fell only moderately dur- the viscosity constant I , which is defined by ing four days of continuous operation. In all Souders' as cases carbonaceous material was left on the surface of the catalyst. It is of further interest that a t 300' the dehydration of tetrahydrofurfuryl alco- where q is the viscosity in millipoises. For such a 1101 poisoned the subsequent dehydration of ab- pIot and for determining the constants of the solute ethanol even though it did not poison equations, data for I may be obtained from further dehydration of the tetrahydrofurfuryl Souders' and values of R from Eisenlohr,2 H e n 3 alcohol. Undoubtedly the tetrahydrofurfuryl al- and the Landolt-Bornstein tables. In Table I are given the values of the constants cohol poisons the catalyst for ethanol dehydration by being strongly adsorbed on the surface. This a and b as calculated for several series by the L-I~W is confirmed by the fact that only fifteen method of least squares on the assumption that h. From this tableit may be seen that minutes after the start of the reaction, does the I = aK product coax out of the catalyst exit tube. the slopes are very nearly identical for all the Further ccltifirniation i, the iort y-degree rise i n series examined, while the intercepts vary. That tfvT-eraturc of the catdyst tic ti ')E the introduc- Iinear curves represcnt the relations very well may be deduced from the fact that use of the constants tion of the tetr,hytlrolurtur, i dL~ohcr1 This is tc be contrasted with a nine degree rise when the of Table I allows one t o calculate values of I-averaging within '/s of one per cent. of the experiammonia is simultaneously introduced. Over silica, tetrahydrofurfuryl alcohol is stable mental values. As a consequence, if the density up to 400°, at which point some decomposition and molecular weight of a liquid belonging to any of the series of Table I are known, the viscosity ( 1 ) Orimnal manu\
