408 Andersen — System Anorthile-Forsterite-Silica. 



Introduction. 



The present study was commenced as an examination of the 

 supposed binary system CaAl 2 Si 2 8 -MgSi0 3 . 



The preliminary work proved that MgSi0 3 (clinoenstatite) 

 was unstable at its melting point, breaking up into Mg 2 Si0 4 

 (forsterite) and melt, and the system had to be treated as a ter- 

 nary system with CaAl 2 Si 2 ? , Mg 2 Si0 4 (or MgO) and Si0 2 as 

 components. The study of this ternary system, however, 

 showed that, though all components were stable at their melt- 

 ing points, there was in certain mixtures a breaking up of 

 CaAl 2 Si 2 8 and Mg 2 Si0 4 in such a way that A1 2 3 combined 

 with MgO to form MgAl 2 4 , spinel. In fact spinel proved 

 to be the primary phase in a number of mixtures of Ca Al 2 Si 2 8 , 

 MgO (or Mg 2 Si0 4 ) and Si0 2 where no other phases but the 

 components might be expected. With MgAl 2 4 as a solid 

 phase, the composition of the liquid and the course of crystalli- 

 zation can no longer be expressed in terms of the three 

 components here considered nor, in fact, of any three com- 

 ponents. 



The relations of the system CaAl 2 Si 2 8 -Mg 2 Si0 4 -Si0 2 , there- 

 fore, can be properly explained only by considering this system 

 as a part of the quaternary system CaO-MgO-Al 2 3 -Si0 2 . This 

 great system embraces the ternary systems MgO-CaO-Al 2 3 , 

 MgO-CaO-Si0 2 , MgO-Al 2 3 -Si0 2 , and CaO-Al 2 3 -Si0 2 , 

 which, in a way form an exterior " skeleton " round the quater- 

 nary " body." The " body " itself is then made np of a num- 

 ber of binary, ternary, and perhaps quaternary systems. 



One of the "skeleton" systems, the 0aO-Al 2 O 3 -SiO 2 system, 

 has already been worked out in this laboratory* and others are 

 being studied. The general problem of the quaternary system 

 CaO-MgO-Al 2 3 -Si0 2 has not been undertaken hitherto, and 

 because of its complicated character may remain unstudied for 

 a long time to come. In the meantime, any study of the rela- 

 tione between some of the phases occurring in the system will 

 be of interest. 



The present paper deals with that part of the quaternary 

 system in which the composition of the mixtures can be ex- 

 pressed in terms of the components anorthite-forsterite-silica. 

 It is explained above that some of the mixtures of these com- 

 ponents show relations inexplicable in a ternary system. The 

 particular part of the system to which these mixtures belong 

 will here be only cursorily touched upon, emphasis being laid 

 on the larger part which can be treated as a ternary system. 



* E. S. Shepherd, G. A. Kankin, and F. E. Wright (Preliminary report on 

 the ternary system CaO-Al 2 3 -Si0 2 ; a study of the constitution of Portland 

 cement), J. Ind. Eng. Chem. iii, 1, 1911 ; G. A. Kankin and F. E. Wright, 

 this Journal (4), xxxix, 1-79, 1915. 



