284 Day and Shepherd — Lime-Silica Series of Minerals. 



it is possible to wash out as much as 10 per cent of lime. This 

 probably accounts for the absence of this mineral in nature. 

 Ammonium chloride solution even when cold decomposes all 

 the mixtures of CaO and Si0 2 . 



The tricalcic silicate, 3CaOSiO r — This silicate owes its 

 supposed existence mainly to those investigators who have found 

 it necessary to postulate such a compound in order to explain 

 the constitution of portland cement. So far as the literature 

 shows, no one has ever isolated and described a pure and homo- 

 geneous compound of this composition or defined its proper- 

 ties.* Many and varied attempts to make it have uniformly 

 resulted in mixtures in which poor optical properties have made 

 the conclusions insufficiently positive.f 



We began the investigation of this composition by fusing the 

 components in the proper proportions and examining the fused 

 product microscopically as others had done. Most previous 

 investigators, however, appear to have depended for microscopic 

 evidence upon the ordinary optical figures and interference 

 colors. Now, it so happens that this mixture when fused crys- 

 tallizes in an extremely fine structure in which the interference 

 colors are quite different from those of the orthosilicate to be 

 sure, but this is merely the result of the fine state of division 

 and the overlapping of the crystals, and not to another com- 

 pound. If one examines any of the compositions in which the 

 tricalcic silicate might be expected to figure, using the very 

 sensitive index of refraction as a test of homogeneity, he will 

 find that in every preparation containing more lime than 65 

 per cent (orthosilicate composition), there is an excess of free 

 lime which can be positively identified. We have fused the 

 tricalcic silicate composition, cooled it rapidly and slowly in 

 various ways, without once failing to find free CaO present in 

 quantity. Through the kindness of Dr. Clifford Richardson 

 we were also given an opportunity to examine some of the 

 tricalcic silicate prepared and described by him, and while its 



*It is sometimes described as "nearly homogeneous. " 



f A. moment's consideration should suggest that there is no real necessity 

 for assuming the existence of the tricalcic silicate in order to explain the 

 nature of portland cement. It is at least a three-component system with a 

 great number of possibilities. The real difficulty appears to have been that 

 crystallized lime is relatively inert and does not readily give the reactions 

 common to ordinary lime, consequently the tests which were thought to 

 demonstrate the absence of free lime in these preparations have proved very 

 misleading. For example, we have found that crystals of CaO are but very 

 slowly attacked by water (see p. 271). Another argument which is freely 

 offered — that there can be no free lime present "because if free lime is added 

 the cement dusts spontaneously," is obvious fallacy. Free lime does not 

 cause the dusting and if it did the fact that the addition of free lime caused 

 dusting would be no proof that none was present. (Cf. 52 per cent CaO, 48 

 Si0 2 .) 



