The Greeks made lime mortar, and special hydraulic cements and plasters were made at Laurion to waterproof the many cisterns and the ore washeries there.
In the C3 BC the Romans discovered that adding pozzolana to the lime mortar produced a concrete that would set under water. This was one of the Romans’ greatest inventions; discovered in all likelihood by accident. The first dated building definitely using it is the Temple of Concord, for the platform, 121 BC; this is poor quality concrete.
Basic lime mortar (v. old) is made of lime + sand, powder or straw for strengthening. To make lime one starts with limestone. Limestone (calcium carbonate) is calcined i.e. heated for long time at c. 900°C. Calcium carbonate is converted at this temperature to calcium oxide, otherwise known as ‘quicklime’ CaCO3 + heat becomes CaO + CO2. When ground up, quicklime is mixed with water; the water turns the calcium oxide into calcium hydroxide, ‘slaked lime’ CaO + H2O becomes Ca(OH)2 + heat. Anything added for strengthening, e.g. sand, powdered brick, is added now a/c to Vitruvius 2.5; we mix dry ingredients first. If nothing is added, the result is lime plaster not mortar. Slaked lime ‘dries’ to limestone again by evaporation of the water and absorption of CO2 from the air, to make CaCO3 again.
Health & safety issues were recognized, e.g. Theophrastos On Stones 66 points out that one can’t mix lime mortar by hand because of the heat; this implies that mixing was normally done by hand incidentally. The ‘magic ingredient’ in Roman concrete is pozzolana (~ Portland cement), a powdered volcanic ash, named after Pozzuoli (Puteoli), near Vesuvius. Later other similar deposits were found in Sicily, the Campagna, and Germany. It contains silica, aluminium & iron oxide. Make lime mortar as usual, but add pozzolana to the mix as the powder; the result will set under water, and (unlike traditional mortar) was stronger than the stones joined by it. The Romans called it caementum, whence ‘cement’. Amongst other things it was used as grouting on gravel road, thus producing the first concrete roads. Concrete was first used essentially as mortar was; to hold together the rubble infill of stone or brick walls, and it was still slightly suspect in Vitruvius’ time (C1 BC). Mortars and cements were poured over rubble, broken tiles and/or other aggregates to produce concrete foundations or cores to structures that were faced with other materials - the Romans did not like the appearance of concrete.
Concrete in building was always hidden from view by a facing of attractive materials (e.g. brick, mosaic, marble); Augustus’ claim to have ‘found Rome a city of bricks and left it one of marble’ meant concrete buildings with marble facings. Used as the infill on a semicircular brick arch created a great lintel that exerted little if any lateral pressure, so pillars didn’t need buttressing (horizontal thrust is brought down through arch to the abutments); arch thickness of >5.5% span will be stable, so long as the loading on the crown and haunches is apportioned properly (not too little on the crown, not too much on the haunches).
In the East where natural supplies of bitumen and asphalt appeared at the surface, e.g. DS 2. 12. 1, these were sometimes used to bond mud or clay bricks, e.g. DS 2. 7. 4, or as an alternative to lead to bond cramps in stone, e.g. DS 2. 8. 2.
References: Vitruvius De Architectura 2.5.1-3. See also C O'Connor Roman Bridges 1993.