Why Roman Concrete Gets Stronger After 2,000 Years in the Sea

In the first century, Romans poured concrete into the sea on purpose. Two thousand years later it is still there, and the water has not eaten it. In some places the concrete has gained strength.

That sentence sounds like folklore. It is not. Harbours built by Roman engineers — at Caesarea Maritima on the Israeli coast, at Baiae on the Bay of Naples — still have submerged concrete cores. A modern dock poured beside them in the twentieth century is already failing.

What they actually poured

Roman marine concrete is not a secret recipe. The ingredients were lime, volcanic ash and rubble, mixed with seawater and dropped into wooden forms that were floated out and sunk. The ash came from the Phlegraean Fields near Naples — pozzolana — and the same material was shipped across the Mediterranean because it worked.

Lime reacts with water and releases heat. The glassy component of the ash is attacked by the alkaline pore solution and forms cementitious phases. Seawater stayed inside the formwork, and that was not a problem to be prevented: it was part of the mix.

Modern concrete is built on the opposite premise. Portland cement is mixed with a carefully controlled amount of water, cured, and then protected from water for the rest of its life. We waterproof, drain, seal and coat. The Romans flooded theirs.

Why modern concrete dies at the coast

Two mechanisms do most of the damage. The first is the reinforcing steel. Concrete is strong in compression and weak in tension, so we embed steel to carry the tension — and steel rusts. Rust occupies more volume than the metal it came from, and that expansion splits the concrete from the inside. The crack admits more water, the corrosion accelerates, and the structure fails on a schedule set by the rebar.

The second is sulfate attack. Seawater carries sulfate and chloride; sulfate reacts with cement hydration products to form expansive phases, and the concrete slowly turns to mush.

Roman marine concrete had no rebar, so the first mechanism did not exist. But that alone does not explain two thousand years. Something else was happening in the pores.

The minerals that grow in seawater

In 2017 a team led by Marie Jackson reported what they found in Roman harbour concrete: aluminous tobermorite and phillipsite, growing in the pores and cracks. These are not ingredients anyone added. They precipitated in place, from the reaction between seawater and the volcanic ash.

That is the inversion at the heart of the story. Chloride and sulfate are the enemies of modern concrete. In Roman concrete they are reagents that dissolve the ash and grow new minerals, filling the pore network and making the material denser over time. The seawater is not kept out. It is invited in to react.

The lime clasts nobody mixed in

A second mechanism came out of MIT in 2023. Analysing Roman concrete, the team found lumps of lime — clasts — scattered through the material, apparently not fully mixed. That looked like sloppy workmanship until they tested what the clasts do.

The clasts are a reservoir of reactive calcium. When a crack opens and water reaches one, the clast dissolves and calcium carbonate recrystallises in the crack, sealing it. The concrete heals itself, but only where a crack lets water in — which is exactly the situation the material was designed to survive.

This does not mean Roman concrete never cracks. It means cracks are not automatically the beginning of the end.

The honest caveat

Roman concrete is not stronger than modern concrete. Measured in compressive strength, ordinary modern mixes beat it comfortably. Its advantage is durability: it does not fight its environment, so it does not lose.

It is also slow. Days and weeks show little; months and years show the difference. No modern client signs a contract for a two-thousand-year service life, and no contractor can float a formwork and wait a decade for the pour to finish curing.

Why it matters now

Cement is responsible for something like eight per cent of global carbon emissions. That number is the reason the recipe is being studied again rather than admired. If a material can be made to heal itself, then maintenance schedules, design life and the carbon budget of a structure all have to be recalculated — a bridge that repairs its own cracks does not need the same replacement cycle.

Startups are now working with lime clasts and volcanic ash, not to reproduce Rome but to recombine its mechanisms with modern engineering. The lesson is not "copy the Romans." It is that they asked a different question: are we designing a material, or a reaction that can last for centuries?

Next time you see concrete by the sea, ask whether it is fighting the water or aging with it. The Romans left no recipe book. They left walls still standing in the water, and they are still waiting for us to ask the right question.

Watch the full video on YouTube

▶ Watch the full video on YouTube

Exploring the strange questions about humans, science, history and the world around us. @TheQuestionArchive