CALCULATION AND CONTROL OF THE MIX. 387 



Alumina. To the calcium aluminate of a cement are ascribed the 

 initial setting properties. Decrease in the alumina, therefore, tends to 

 make the cement slower setting, while high alumina affects it in the 

 opposite way. Though it is advisable to carry the alumina as low as 

 possible, so as to secure slowness of set and greater ultimate strength, 

 it is impossible to carry it below a certain minimum, for alumina aids 

 greatly in securing a low clinkering temperature, and a cement very 

 low in alumina will clinker only with great difficulty. Too much alumina, 

 on the other hand, will give a very fusible and sticky clinker, liable 

 to ball in the kiln. 



Le Chatelier considers that the aluminous compounds present in 

 Portland cement are the direct cause of its destruction by sea-water. 

 His theory to account for this disintegration is as follows: Free lime, 

 liberated during the hardening of the cement, reacts with the mag- 

 nesium sulphate always present in sea-water, to form calcium sulphate. 

 This in turn reacts with the calcium aluminate of the cement to form 

 a sulphaluminate of lime, which swells considerably on hydration and 

 thus disintegrates the cement mass. The extent of the disintegration 

 varies directly with the percentage of alumina present in the cement. 

 Cements containing 1 or 2 per cent of alumina are, for example, prac- 

 tically unaffected by sea-water, while in cements containing as high 

 as 7 or 8 per cent of alumina the swelling and consequent disintegration 

 are very rapid. 



If the alumina of a cement be replaced by an oxide not reacting 

 with calcium sulphate, the stability of the cement in sea-water is greatly 

 improved. Le Chatelier has demonstrated this by preparing cements 

 in which the alumina was replaced by oxides of iron, chromium, cobalt, 

 etc. All of these were more resistant than an alumina cement to the 

 disintegrating effect of lime sulphate. The best effects were obtained 

 when iron oxide was used, a cement corresponding in composition to 

 5Si0 2 ,Fe 2 3 ,17CaO being found to be not only stable in presence of 

 sea-water but to possess excellent mechanical properties. 



DevaPs researches * on the effect of direct addition of calcium sul- 

 phate to various cements confirm the above theory. Each of the finely 

 ground cements tested was completely hydrated by mixing with 50 

 per cent of water and storing the mixture under water for three months 

 out of contact with carbon dioxide. The mass was then dried, reground, 

 mixed with half its weight of calcium sulphate and 33 per cent of water, 

 and made up into rods which were kept moist and protected from car- 



* Abstract in Jour. Soc. Chem. Industry, vol. 21, pp. 971-972. 



