The Forgotten History of Cement: From Roman Concrete to Modern Construction

Cement is one of those substances that fades into invisibility once it becomes ubiquitous. Modern concrete is poured into the foundations of every building, the substructure of every road, the bed of every bridge. Global production runs to roughly four billion tons annually, more than any other manufactured material by mass, and the carbon footprint of cement production accounts for around 8 percent of global CO2 emissions. The history of how this came to be, and how it almost did not survive the medieval period, is one of the strangest stories in the technological inheritance of civilization.

Roman concrete and the Pantheon ceiling

The Romans had a concrete that worked. Opus caementicium, developed in the second century BCE and standardized by the time of Augustus, combined slaked lime, water, volcanic ash (pozzolana, from the Italian town of Pozzuoli), and rubble aggregate into a hydraulic mortar that set underwater and hardened over decades. The Pantheon's 43-meter dome, cast in 126 CE, remains the largest unreinforced concrete dome ever built. Roman harbors at Caesarea, Cosa, and Portus used concrete that was poured into wooden forms below sea level and that gained strength specifically because of the seawater chemistry — a property modern Portland cement cannot match.

The recipe and the institutional capacity to deploy it at scale disappeared along with the western Roman administrative state in the fifth and sixth centuries. The chemistry was not secret — Vitruvius described the technique in some detail in De Architectura around 30 BCE — but the supply chains for high-quality pozzolana, the trained workforce, and the imperial coordination that made aqueducts and harbors possible all dissolved together. By the eighth century, large-scale concrete construction had effectively ceased in Europe. Medieval builders used lime mortars that were structurally inferior, prone to weathering, and incapable of underwater setting.

The recipe was occasionally rediscovered. Byzantine builders preserved some techniques; the Hagia Sophia (537 CE) used a concrete-like mortar in its construction. But the institutional architecture that had supported imperial concrete projects across the Mediterranean was not recreated for over a millennium.

The 18th century and Smeaton's Eddystone

The recovery began with John Smeaton, an English civil engineer commissioned in 1756 to design the third Eddystone Lighthouse on the rocks south of Plymouth. The two previous towers had been destroyed by storms; Smeaton needed a mortar that would set underwater and resist saltwater erosion. Working without any theory of cement chemistry, he experimented systematically with limes from different geological sources and discovered that limestones containing clay impurities produced hydraulic mortars while pure limestones did not. The Eddystone Lighthouse, completed in 1759, used Smeaton's hydraulic mortar in its joints and stood for 127 years before erosion of the rock below forced its replacement.

Smeaton's empirical work established the principle. Joseph Aspdin, a Leeds bricklayer, patented Portland cement in 1824 — a process for burning a specific mixture of limestone and clay until partial fusion, then grinding the result. The name came from the visual resemblance to Portland stone, a prestigious building stone from the Isle of Portland. Aspdin's original product was inferior to modern Portland cement; the calcination temperatures were not high enough to fully convert the raw materials. His son William refined the process in the 1840s, producing a cement that began to approach modern performance.

Industrial scaling

The 19th century scaled Portland cement from craft production to industrial output. The rotary kiln, invented by Frederick Ransome in 1885, allowed continuous production at temperatures above 1450°C, which is what is required to produce the calcium silicate compounds that give Portland cement its strength. Production rose from a few thousand tons annually in 1850 to several million tons by 1900, mostly driven by railway construction, canal locks, and the new reinforced-concrete buildings made possible by Joseph Monier's 1860s patents on combining concrete with steel rods.

Reinforced concrete was the technology that made cement structurally transformative. Pure concrete is strong in compression but weak in tension; steel is strong in tension but corrodes in atmospheric exposure. Embedding steel in concrete gives a composite that is strong in both directions, with the concrete protecting the steel from corrosion. This combination enabled the early 20th-century construction revolution: skyscrapers with reinforced-concrete frames, bridges with concrete arches, and the housing stock that defines most of the world's cities.

By 2026, global cement production is around 4.1 billion tons annually. China produces more than half of this. The United States, Europe, and India follow at much smaller scales. The infrastructure value of cement is essentially impossible to overstate: the substance is present in almost every piece of human-built infrastructure that has existed since 1900.

The climate problem

The chemistry that makes Portland cement so useful is also responsible for its carbon footprint. Producing cement requires two large CO2 emissions sources: the calcination reaction (CaCO3 → CaO + CO2) that releases CO2 directly from the limestone, and the fuel combustion required to heat the kiln to 1450°C. The first source accounts for about 60 percent of cement-related CO2 emissions; the second accounts for about 40 percent.

The fuel-combustion portion can be addressed with electric kilns, alternative fuels, and carbon capture. The calcination portion is fundamental to the chemistry: the calcium oxide that gives cement its strength comes from limestone, and freeing it requires releasing the CO2 that was originally captured when the limestone formed from marine biology hundreds of millions of years ago. No amount of process optimization changes this.

The current research surface includes several approaches. Calcined clay cements substitute kaolin clay (which decomposes at much lower temperatures) for a portion of the limestone. Magnesium-based cements use serpentine or olivine rocks that absorb CO2 during curing. Geopolymer cements use aluminosilicate feedstocks (often fly ash) activated with alkaline solutions. None of these is yet a complete drop-in replacement for Portland cement at industrial scale, but each addresses a portion of the emissions profile.

Roman concrete, with its pozzolana-based chemistry and seawater curing, is itself a partial answer. A 2017 Marie Jackson paper in American Mineralogist documented that aluminous tobermorite crystals grow in Roman marine concrete over centuries, strengthening rather than degrading the material. Modern Portland cement degrades over decades; Roman concrete in the right environment may have a service life measured in millennia. Whether the Roman approach can be industrially scaled to displace a meaningful portion of modern cement production is an open research question.

The deeper observation

The 1500-year gap between Roman concrete and Smeaton's lighthouse is a humbling case study in technological loss. The Romans knew how to build hydraulic concrete; the chemistry was not hidden; Vitruvius's text was preserved through the medieval period and printed widely after the 15th century. The reason no large concrete structures were built between 600 and 1700 CE is not that anyone forgot how — it is that the institutional capacity to organize the supply chains, trained workforce, and project management at imperial scale collapsed and took a long time to be recreated under different political arrangements. The Pantheon's dome is a continuous reminder that civilizational capabilities are usually contingent on institutional support, that the institutions are usually more fragile than the technology, and that the recovery from the loss of an institution can take very long indeed.