EFFECTS OF TEST DURATION AND SPECIMEN LENGTH ON DIFFUSION TESTING OF UNCONFINED SPECIMENS

Effective diffusion coefficients, D, of chloride and zinc diffusing i n saturated, unconfined specimens of a compacted sand-clay mixture are measured for three specimen lengths, L (2.91, 5.83, and 11.60 cm) and three test durations (7, 14, and 21 days). For a specimen length of 2 .91 cm, both the chloride and zinc D values tend to decrease with inc reasing test duration, possibly due to the measurement of concentratio n-dependent D values. For a 14-day test duration, no consistent trend in D with specimen length is observed, but the overall effect of spe cimen length on D is minor relative to the range of measured D* value s. A 21-day test duration provides the best correlation between the D values based on reservoir concentrations, D(Res) and the D* values b ased on soil concentrations, D(Soil), for chloride for a given test r egardless of the specimen length. The effect of test duration on the c orrelation between D(Res), and D*(Soil) for zinc is minor based on th e relatively narrow range of measured zinc D values. The observed eff ects of specimen length on the correlation between D(Res) and D*(Soil ) for a given test ale consistent with the more uniform final porosity distributions in the shorter specimens and the contrasting effects of the non-linear distributions in porosity and dry density that become less significant as the specimen length increases.