Short answers, pulled from the story.
Roman concrete consisted of an aggregate mixed with hydraulic mortar. Builders used pieces of rock, ceramic tile, and lime clasts within the mixture. Tuff was often available as an aggregate in Rome itself.
Evidence suggests that coastal harbors around Baiae utilized this material before the end of the second century BC. The harbor of Caesarea in Roman Judaea rose from the sea between 22 BC and 15 BC. Some scholars argue that development began a full century prior to 150 BC.
Water seeping into cracks reacted with reactive calcium sources inside the stone. New calcium carbonate crystals formed automatically to reseal the damage. Modern concrete lacks this specific internal architecture found in ancient Roman mixes.
The upper region of the dome alternated layers of light tuff and pumice to reduce weight at the apex. Foundations used travertine aggregates with a much higher density near 2700 kilograms per cubic meter. This arrangement gave the concrete a density of approximately 1400 kilograms per cubic meter.
Seawater percolated within tiny cracks in Roman marine concrete over millennia. It reacted with phillipsite naturally found in volcanic rock to create aluminous tobermorite crystals. Usable examples of Roman concrete exposed to harsh marine environments are 2000 years old.