8 



abstracts: physical chemistry 



orthosilicate and lime. Neither is there any evidence that the trical- 

 cic silicate takes up calcium orthosilicate or lime in solid solution. 



In addition to the tricalcic silicate, a new and probably unstable 

 form of the orthosilicate has been discovered which may prove to be of 

 importance in Portland cement. 



The general boundaries of the different phases occurring throughout 

 the entire diagram of the ternary system CaO— AI0O3— Si0 2 , have been 

 established, together with the course of the boundary curves. From 

 these it is possible to predict the compounds which will crystalline out 

 of a lime-silica-alumina cement of any composition, provided only that 

 the compounds are intimately mixed and the heat is sufficient to bring 

 the mixture to equilibrium during formation. The exact location of the 

 quintuple points, with the exception of points 13 to 17, and the isotherms 

 remains to be established. The location of these quintuple points is 

 as follows: 



Portland-cement clinker within the concentration limits set by Rich- 

 ardson, when in equilibrium, may exist in the following combinations, 

 depending primarily upon relatively small changes in the quantity of 

 lime present: 



Richardson's typical cement corresponds to class II. The relative 

 cement-forming value of the above mixtures has yet to be determined. 

 The more intimate the mixture of the raw material, and the more uni- 

 form the heat treatment, the closer will be the approach to equilibrium 

 and therefore to constant relations and the more uniform the behavior 

 of the final product. Classes III and IV are predicted from the general 

 direction of the boundary curves and the quintuple points 14 and 15. 

 ( 'hiss V will occur in cements low in lime and will differ only in the rela- 

 tive amounts of the different phases from pure slag cements whose 

 compositions lie below the line joining the calcium orthosilicate and 



