DISTILLATION OF STUMPWOOD. 57 



proportion of the turpentine fraction is retained in the residual 

 pine oil. 



Because of these defects in refining processes, efforts were di- 

 rected in the beginning of this work to devise a laboratory method 

 for recovering a maximum of high-grade spirits from the crude oil 

 that would be applicable to the operation of a commercial plant, on 

 the basis of which yields of commercial turpentine from different 

 woods could be compared. It soon became apparent that sufficient 

 importance is not attached to the amount of alkali required and to 

 the manner of its application. Instructions merely to distil with 

 alkali or to treat with alkali until action ceases are entirely inade- 

 quate, because the amount of alkali used materially affects the pro- 

 portion of the light fraction, the sharpness of the fractionation, and 

 the quality of the turpentine as indicated by its odor, color, and boil- 

 ing point limits. The intimacy and period of contact of the alkali 

 has equal or greater influence. 



The alkali appears to serve a double purpose, aside from that of 

 neutralizing the free acid present in the crude oil. First, it brings 

 an apparent polymerization of the aldehydes whereby these are con- 

 verted into resinous, nonvolatile compounds, in which form their 

 elimination from the turpentine is effected on distillation. Second, 

 the action of the alkali, if used in sufficient quantity, results in the 

 formation of a soap with the tar and resin acids. This, in a man- 

 ner not understood, although it may be through formation there- 

 with of a so-called water-soluble oil, restrains the escape of the 

 pine-oil constituents, while the turpentine distils over, thus effect- 

 ing a materially sharper separation of the, two. The alkali solution 

 being immiscible with the turpentine and the polymerization process 

 partaking in its nature of a catalytic or surface reaction, the effec- 

 tiveness of the alkali depends on extremely intimate contact with 

 the turpentine for a sufficient length of time to permit the carrying 

 of the reaction to completion before beginning the distillation. It 

 is in this respect that refining methods as ordinarily carried out are 

 wrong in principle, for the reason that, with the alkali added in the 

 still, distillation begins before completion of the reactions that "fix" 

 the aldehyde bodies. These in part pass over with the turpentine 

 and are removed from the sphere of action of the caustic solution 

 before the reactions that render them nonvolatile have been com- 

 pleted. Agitation of the turpentine in a separate vessel and remov- 

 ing and distilling the oil thus separated from the alkali are also 

 wrong in principle, because advantage is not taken of the deterring 

 action of the soap solution on the distillation of pine oils. 



To combine the action of the two principles here set forth the 

 crude oil is agitated with caustic soda solution at boiling tempera- 

 ture in a return-flow condensing apparatus before distilling. Me- 



