The Chemical Constitution of Portland Cement

The Chemical Constitution of Portland Cement, and the mechanism of the setting and hardening process, have afforded subjects for much discussion, and no generally accepted conclusions appear to have been reached as yet. Much has been done by petrographical study of cement clinker. Le Chatelier and Tornebohm introduced the microscopic examination of cement sections. Four different constituents were observed: "Alite," which is apparently the active principle of cement and forms rhombic crystals of composition represented by the formula 3CaO.SiO₂, with a small quantity of aluminate; "Belite," a finely striated substance, poorer in lime than alite, and sometimes found in a state of partial disintegration, suggesting the presence of unstable dicalcium silicate; " Celite," a very inert substance, resisting the action of both water and acids, and present in two varieties, white and grey, both probably solid solutions of dicalcium silicate in dicalcium aluminate; and, finally, "Felite," in the form of rhombic crystals, a constituent not very frequently observed except in blast-furnace slag, and probably the non-hydraulic form of dicalcium silicate.

Later investigators have also recognised the existence of some or all of these four constituents, but opinions differ considerably as to their composition. In the microscopic study of structure some experimenters have utilised the different adsorptive powers of the various constituents for certain dyes. This gives ground for regarding some portions of the cement clinker as colloidal.

The only finding upon which there seems to be general agreement is, that Portland cement clinker consists of a complex mixture of calcium aluminates and silicates, either as a more or less homogeneous solid solution, or as a heterogeneous mixture of different solid solutions or of single chemical individuals.

The setting of cement is a more complex phenomenon than the setting of plaster. The reaction with water is very incomplete, because even the most perfectly mixed cement still contains a large proportion of unchanged material after setting and hardening. Regrinding of set cement gives a product which is still hydraulic. Hydrolysis of the silicates and aluminates evidently takes place, because crystals of calcium hydroxide are observed in the set cement. There is no such certainty, however, as to the nature of the resulting silicates and aluminates.

There are, broadly, two theories as to the mechanism of the setting and hardening processes in Portland cement: -

1. The crystalloid theory which was originally put forward and is still maintained by Le Chatelier. According to this, the setting of cement is to be explained similarly to the setting of plaster of Paris. The anhydrous cement dissolves to form a solution supersaturated with respect to the hydrated compounds. These crystallise out in a confused mass of interlocking crystals, thus giving to the product its mechanical strength.
2. The colloid theory for which Michaelis is responsible. Michaelis regarded setting as due to the formation of gelatinous hydrated silicates and aluminates, and hardening to the subsequent drying-up of these hydrates forming hard, glue-like masses, impervious to water. Any slow crystallisation taking place in the midst of this colloidal mass is to be regarded as injurious rather than beneficial. In set cement only two constituents should be observed microscopically: a colloidal medium surrounding particles of unaltered clinker.

Le Chatelier admitted that, in the initial stages, the constituents may be in a colloidal condition. He considered, however, that there is no proof that this persists, even though the particles in the hardened cement may not be recognisable as true crystals, since they are too small to be detected.

Von Glasenapp observed a change in Portland cement after three years. He ascribed this to the slow conversion of colloids into crystalloids. Desch pointed out that Le Chatelier worked with dilute solutions and Michaelis with concentrated. This, according to Hatschek, would account for the difference between their views, because it is well known that slightly soluble substances may be obtained either as gels, as the familiar crystalline precipitates, or as sols, depending on the concentration of the reacting solutions. It is also possible for the originally formed gel to assume gradually a coarser crystalline structure. In gauging cement a comparatively small quantity of water is used, and, therefore, concentrated solutions are here under consideration.

A considerable amount of work has been done by the United States Geophysical Laboratory and the Bureau of Standards on the constitution, setting, and hardening of Portland cement, and the conclusions may be summarised briefly here. Lime, alumina, and silica are the only essential constituents of good cements. Well-burned clinker contains three compounds capable of independent existence: α-dicalcium silicate, tricalcium silicate, and tricalcium aluminate. If burnt at too low a temperature, or for too short a time, free lime and the compound 5CaO.3Al₂O₃ are also present. The initial set is probably caused by the hydration of tricalcium aluminate, any free lime being also hydrated to an amorphous product which crystallises later. The hardening, and the development of cohesive strength, are brought about by the slower hydration of tricalcium silicate which forms 30-35 per cent, of the clinker. This process continues for about a month. The dicalcium silicate hydrates still more slowly to a granular porous mass, free lime and gelatinous silica being formed. Subsequent partial crystallisation takes place in the gelatinous hydrated material. Gelatinous silica is probably the chief cementing agent. The value of the aluminate lies, no doubt, in its action as a flux in the burning of the clinker.

Small quantities of foreign salts influence the rate of setting of cement. Rohland regarded the effect as catalytic.

Fineness of grinding increases the rate of setting and also the tensile strength.

Portland cement should be of a bluish or greenish-grey colour, yellowness indicating insufficient burning.

A satisfactory Portland cement clinker should pass the following tests:

Fineness

The ground cement clinker should not leave more than 14 per cent, on a sieve of mesh 180 to the linear inch, and 1 per cent, on one of mesh 76 to the linear inch.

Chemical Composition

The proportion of lime to silica and alumina, after deduction of the proportion necessary to combine with the sulphuric anhydride present, when calculated (in chemical equivalents) by the formula , must not be greater than 2.85 nor less than 2.0. The insoluble residue must not exceed 15 per cent.; that of magnesia 3 per cent.; and the total sulphur content, calculated as sulphur tri-oxide, 2.75 per cent. The total loss on ignition must not exceed 3 per cent.

Setting-time

Unless an especially quick-setting cement is specified or required it must have an initial setting-time of not less than 20 minutes, and a final setting-time of not more than 10 hours. For a quick-setting cement the initial set should be not less than 2 minutes and the final 30 minutes. The time of initial set is the interval between gauging and partial loss of plasticity, and the time of final set is the time required to attain sufficient firmness to resist a certain definite pressure, usually applied by a loaded Vicat needle. Gary's method of determining the setting-time is to follow the changes of temperature. There is a considerable rise in temperature at the first set and a second rise at the final set.

Tensile Strength

This is determined by the force required to pull apart a dumb-bell-shaped briquette. It varies with the quantity of water, temperature, and time of setting. Briquettes kept in a damp atmosphere for 24 hours, and in water for 6 days, must have a tensile strength of not less than 400 lbs. per square inch, and, after 28 days, 540 lbs. per square inch.

Soundness

This test is made by Le Chatelier's method. The apparatus is a split cylindrical mould by which the expansion of the cement after treatment with boiling water can be measured. This expansion must not exceed certain defined limits.

Small quantities of gypsum, up to 2 per cent., may be used with safety as a retarder. In excess of this it produces unsoundness, apparently due to the formation of calcium sulpho-aluminate, 3CaO.Al₂O₃.3CaSO₄.xH₂O. The destructive action of sea-water on cement and concrete has been ascribed to the formation of this compound, but there seems to be reason to doubt this. The destruction of concrete, however, by surface water, peaty water, sewage, or sulphate solutions is probably to be attributed to it. This action may be prevented by the replacement of alumina by iron oxide.

Magnesia is a cause of unsoundness, because it hydrates much more slowly than the other compounds, thus causing changes in volume after the setting of the cement.

Mention has already been made of the unsoundness produced by excess of lime. According to Erdahl, however, unsoundness is due, not to free lime, but to dicalcium silicate, which disintegrates or "dusts." The destructive action of sea-water is probably due to the crystallising of alkali salts in the pores, and is best overcome by making impervious to water.