PROPERTIES OF PORTLAND CEMENT. PART III. 209 



ical analysis. This conchision is strengthened by the work of E. D. 

 Campbell." 



This author heated calcimn oxide and marl in his experimental furnace, and 

 found that "when calcium oxide is heated with calCium aluminate, probably the 

 dicalcium aluminate is first formed, this formation being completed at about 

 1,200°. Further, that dicalcium aluminate which may be completely formed 

 above 1,200° then begins to combine with more alumina, if present, as the tem- 

 perature increases, until monocalcium aluminate has been formed at 1,400°. The 

 dicalcium aluminate when completely formed * * * has a quick initial and 

 a quick final set, even when the material has been partially fused. The propor- 

 tion of dicalcium aluminate and monocalcium aluminate existing at any given 

 temperature will be largely influenced by the relative masses of the two oxides." 

 Campbell's work also indicates that ferric oxide behaves very much as alumina, 

 both in chemical and physical properties, but in the light of Unger's more recent 

 researches we may carry the discussion still farther. 



The Portland cement which he fused in an electric furnace contained all of 

 the iron in the form of the lower oxide. This was contrary to Loebell's experience 

 with vitrified Portland cement, which showed that almost all of the iron was 

 present in the form of ferric oxide. 



W. Michaelis, jr.," concludes from these experiments that Unger's fused 

 Portland cement containing, as it does, the oxide of iron in the ferrous state differs 

 from the commercial product because the iron oxide acts as a base thereby 

 diff'ering from the oxide in the vitrified product where it has acid properties. 

 Unger's fused cement hardened very slowly and imperfectly whereas the same 

 mixture burnt only to the point of vitrification, hardened rapidly. 



As nothing more definite is known concerning the chemistry and 

 mechanics of this subject, the manufacturer, in the beginning, has no 

 recourse other than actual experimentation. Some raw mixtures if 

 burned to a hard, homogeneous clinker produce a cement of quick- 

 setting properties which it may be impossible to retard. Other raw 

 mixtures must be burned to a hard, homogeneous clinker in order to 

 produce a cement capable of a normal set. The very nature of the 

 burning processes employed renders it doubtful if the successful manu- 

 facture of Portland cement will ever depend upon anything more definite 

 than control by rule of thumb methods. However, if Portland cement 

 were to be ground from a perfectly fused mass, instead of from sin- 

 tered clinkers, there is little doubt that its chemical and physical 

 properties could soon be as accurately and scientifically controlled as 

 the manufacture of iron and steel. 



As it is, a number of factors, influencing the chemical and physical 

 conditions of the compounds involved in the setting of Portland cement, 

 must be taken into consideration. These are: (1) The amount of 

 water used. (2) The temperature of the water and the surrounding 



" Journ. Am. Chem. Soc. (1904), 26, 1152. 

 "Cement and Engineering News (1910), 22, 8, 327. 



