_ 
ON COLLOID CHEMISTRY AND ITS INDUSTRIAL APPLICATIONS. 99 
the total volume of cement and water diminishes during setting, a similar 
statement being true of the swelling of starch grains in water.® 
A gelatinous coating having once been tied around the particles 
of clinker, which are thus protected from direct contact with the solution, 
a further series of changes sets in. Water is withdrawn from the gel 
by the unchanged core of clinker, which thus becomes progressively 
more hydrated, although even after many years the proportion of cement 
which has escaped hydration is always considerable. At the same time, 
according to Michaélis, lime is adsorbed from the solution by the outer 
layers of the gel, which thus become harder and less permeable. On 
this view, the ‘initial set’ of cement corresponds with the hydration 
of the aluminates, forming crystalline masses comparable with plaster 
of Paris. The‘ final set’ is due to the formation of a silicate gel, traversed 
by crystals of calcium hydroxide, whilst the subsequent processes of 
desiccation and adsorption account for the progressive hardening of 
cement, which continues over a period of years. 
A modification of this hypothesis was proposed by G. Becker, according 
to whom the gelatinous membranes become stretched by the osmotic 
pressure, and then, through adsorption, undergo a change similar to that 
involved in the lignification of plant fibres. Becker also assigns a some- 
what greater share in the process to the crystalline products of hydration. 
Further evidence in favour of the colloidal hypothesis was adduced 
by Michaélis. If the calcium compounds in Portland cement be replaced 
by barium, a similar series of changes takes place, but the product is 
not hydraulic, owing to the much greater solubility of the barium meta- 
silicate, and the resulting compounds are entirely crystalline, so that the 
process of setting resembles that of plaster rather than of Portland cement. 
The presence of gelatinous material in briquettes of Portland cement 
alter setting and hardening was observed by E. Stern,!° opaque sections 
being examined by reflected light. Stern considered the gelatinous 
constituents to be alumina and calcium metasilicate, fine fibro-crystalline 
calcium aluminate also playing a part in the binding together of the mass, 
whilst the presence of calcium metasilicate in a crystalline form was 
doubtful. 
The interest shown in the chemistry of hydraulic cements shortly 
after the appearance of Michaélis’s paper was such as to justify the publica- 
tion of a special periodical devoted to the subject, which however only 
survived for eighteen months. 
Ah attempt was made to distinguish the various products of hydration 
by the application of organic dyes, which stain colloidal and zeolitic 
mineral substances selectively. By examining the behaviour of different 
substances, such as silica, alumina, calcium silicate, and calcium aluminate, 
when immersed in acid, neutral and alkaline solutions of such dyes 
as patent blue, anthrapurpurin, and methylene blue, and then testing 
hydrated cements with the same reagents, conclusions may be drawn 
as to the nature of the products contained in the cement. The result 
8 H. Rodewald, Zeitschr. physikal. Chem., 1897, 24, 193. 
® Tonind. Zeit., 1909, 38, 1493. 
10 Ber., 1908, 41, 1472; Zeitschr. anorg. Chem., 1909, 63, 160; Mitt. k. Material- 
Priif. Amt., 1910, 28, 173. 
1 Zentralblatt fiir Chemie und Analyse der Hydraulischen Zemente, ed. F. R. v. 
Arlt, Halle a. S., 1910-11. 
12'S. Keisermann, Kolloidchemische Beihefte, 1910, 1, 423. . 
H 
