Although the raw cement particles contain four different molecules eager to combine with water, only tricalcium and dicalcium silicates (C3S and C2S) precipitate into the crystals of calcium silicate hydrate (C-S-H) that form the magic glue around which our professional lives revolve.

In commercial portland cements, neither of these two molecules exists in pure form. Both are contaminated by various amounts of impurities, including the so-called alkali oxides of potassium and sodium that react expansively with certain silica-containing aggregates causing the concrete to disintegrate. With proper testing, such destructive alkali silica reactions may be avoided with the use of low-alkali cement, or by a 30% to 40% fly ash or 40% to 50% GGBS replacement.

The impure versions of tricalcium- and dicalcium silicate actually in the cement are called alite and belite. Because alite reacts with water to form C-S-H much faster than belite, its hydrolysis accounts for about 90% of the cement’s strength gain during the first 28 days. Thereafter, the alite’s reaction rate falls off, and the much slower hydrolysis of the belite into C-S-H drives the long-term strength gain. It takes about a year for the compressive strengths of both hydrates to become about equal.

The slaked lime, or dissolved calcium hydroxide, that makes cement caustic is a byproduct of both the alite and belite hydrolyses. It is a remanifestation of the quicklime, or calcium oxide, created in the kiln when the raw crushed limestone, or calcium carbonate, was heated initially to drive off its constituent carbon dioxide. This purposeful thermal decomposition of limestone is the process, in fact, that gives the manufacture of portland cement such a large carbon footprint. Adding in the carbon dioxide released in heating the kiln and generating the electricity, about 1 pound of CO2 gas is created for every 1 pound of portland cement produced. Concrete’s caustic character is certainly one of its fortuitous features, however, particularly with regard to the protection against rusting that it affords to any embedded steel.

The third bits of material on the cement particle’s surface—the tricalcium aluminate(C3A) molecules—are the ones most ready to combine with water. Though relatively few in number, accounting for only about one in 10, owing to their rapid, even violent reactions, they create serious set rate issues. If unchecked, these are the molecules that will cause the paste to flash set, or stiffen immediately after the water is added. It is precisely to prevent such flash setting that gypsum (calcium sulphate hydrate) is added to the cooled clinker in carefully controlled amounts before grinding. In solution, the gypsum prevents the C3A from hydrating directly by forcing the formation of unstable intermediate products that bind up the C3A and temporarily shield it from the water. But add too much gypsum and upon hardening, the cement will expand enough to fracture the paste. Add too little gypsum, and the C3A molecules will hydrate before the C3S molecules, creating a weak porous structure on which the later reaction products must form.

The fourth and final significant bits on the cement particle’s surface that react with water are the tetra-calcium aluminoferrite (C4AF) molecules. Also accounting for about 10% of the total, these aluminum and iron containing molecules are mostly just along for the ride, because they neither perform a useful function nor create any particular problems.

Portland cement is made from a cheap stew of commonly available calcium- and silicate-rich minerals: primarily limestone and clay. Because the desirable clays are all aluminum silicates, however, the generation of the troublesome C3A is practically unavoidable. Fortunately, the presence of the aluminum is not wholly without benefit; both it and the iron do act as fluxes in the kiln to significantly reduce the amount of heat required to sinter the clinker.