For more than 2,000 years, chemists, engineers and interested amateurs have been working to build a better concrete. The Romans started the process with their invention of a concrete made from quicklime, ash and pumice that enabled the construction of their fabulous--and long-lasting--architecture and infrastructure. Nearly two millennia later, John Smeaton--the father of civil engineering--improved this basic building material by improving the cement that held it together. Yet despite numerous improvements over subsequent centuries, concrete structures exposed to the worst conditions are not surviving for as long as expected. A new analysis of a long-running experiment on concrete, however, has yielded insights into what should make for a stronger building block.

Paulo Monteiro of the University of California, Berkeley, studied the results of tests conducted by the U.S. Bureau of Reclamation for more than 40 years. The bureau subjected concrete cylinders of varying water-to-cement ratios and cement compositions to a mild acid bath to test when they would fail. Because the liquid stone does not dry to become strong but rather solidifies from a chemical reaction between the cement and water, concrete is susceptible to sulfate attack. These salt solutions actually penetrate the concrete and change the chemical reaction between the cement and water both during and after the solidification process, ultimately causing structural failure.

Current techniques for fending off sulfates include cement mixtures low in tricalcium aluminate and with a low water-to-cement ratio, because it makes for a less reactive and less porous concrete. But the 40-year experiment proved that the water-to-cement ratio had little impact on the life span of the concrete. Rather it is the physical components of the concrete that proves the determining factor. Monteiro even worked out specific formulas that can more accurately determine a given concrete's endurance depending on its constituent parts.

"These models can be used to predict the remaining life of existing structures," Monteiro writes in the paper presenting the finding in this week's Proceedings of the National Academy of Sciences. "Such knowledge also can be used to determine which combinations of mixes are more optimal for use in aggressive environments, leading the way to more robust structures less prone to environmental degradation." In other words, our constructions might finally last longer than Roman ones.