The permeability of hardened cement paste is reviewed with particular reference to the influence of pore structure. Permeability is usually modelled by applying D'Arcy's Law, although permeability measurements and pore size distribution determinations reveal the strong influence of large capillary pores (macropores) on flow through cement paste. The macropores form a continuous flow path within the paste. The effects of curing temperature, drying, and admixtures on permeability can be understood in terms of their influence on macropores. Paste containing fly ash shows anomalous behavior, which apparently arises from internal damage occurring during pore structure measurements. It appears that the presence of fly ash promotes the formation of a discontinuous macropore system that inhibits flow.

Study of permeability was made using six concrete mixtures ranging in water-to-cementitious material (w-c) ratio from 0.26 to 0.75. Concrete specimens were tested for permeability to water and air, permeability to chloride ions (rapid and long-term), volume of permeable voids, and porosity. Results confirm that permeability is a direct function of w-c ratio. The addition of silica fume results in even greater decreases in permeability than would be anticipated based solely on w-c ratio. A period of initial moist curing of at least seven days is essential for achieving low permeability. Results also indicate that rapid test procedures offer a reasonable alternative to more lengthy and complex conventional permeability tests.

A testing system which can accommodate up to seven samples simultaneously with computer-controlled data acquisition, analysis, and reporting is described. The system consists of seven core holders of the Hassler type which can handle cylindrical samples ranging from 1-1/2 to 4 in. in diameter and from 4 to 11 in. in length. Confining and driving pressures can be independently varied up to 4000 psi. The test medium can be either liquid or gas including brine, since all tubing and containers are stainless steel. Flow is determined by pressure increase in a collector tank for gas and change in liquid level in a pipette column for liquid. Four pressure transducers per core holder are used to monitor all pressure levels during a test. A computer-based data acquisition system is used to scan up to seven tests simultaneously and record all data on a disc. Upon termination of a test, flow and permeability are computed and plotted against time and a report is printed for the test. The data are saved permanently on the disk and a backup copy is transferred to a floppy disk for safe storage. Sample preparation, sealing, and testing procedures are explained. Data analysis and typical results are presented on salt cores and concrete samples.

Water, chloride-ion, and air permeability of two series of silica fume and non-silica fume concretes having water-cementitious ratios of 0.4 and 0.5 were studied as well as that of a 0.24 water-cementitious ratio silica fume concrete. Silica fume dosage varied from 5 to 20 percent by weight of cement. The water permeability of concrete samples having water-cementitious ratios lower than 0.5 is so low that they can be considered impervious whether they contain silica fume or not. The chloride-ion impermeability provided by silica fume rivals that of latex for water-cementitious ratios of 0.4 to 0.5 and polymer-impregnated concrete with a 0.24 w/c ratio. The two drying methods used in this research yielded a positive correlation between silica fume dosage and air permeability. Equal variations were observed for values of up to 10 percent, whereas at twenty percent, the increase was markedly sharper. The characterization of concrete permeability is not as simple as it appears. Sample preparation and fluid type can significantly affect the interpretation of the effect of an admixture such as silica fume.

This paper summarizes the results of permeability studies that have been undertaken since 1979. The research used a test procedure developed during the NCHRP Project 12-19A, "Concrete Sealers for Protection of Bridge Structures", which was reprinted in 1981 as NCHRP Report No. 244. This test method utilizes 10 cm concrete cubes, and chloride ion penetration is determined at 4 depths after 21 days exposure to 15 percent NaCl solution. The test results show that lowering of water-cement ratio in portland cement concrete or presence of superplasticizers, polymer admixtures, and silica fumes are able to significantly reduce concrete permeability.