Page 135 - gas transport in porous media

Costanza-Robinson and Brusseau
128
the otherwise homogeneous column. At the highest velocities in the heterogeneous
column, rate-limited diffusion (e.g., nonequilibrium effects) between the macro- and
micropore domains became much more significant than dispersion processes.
Garges and Baehr (1998) simulated gas-phase miscible displacement breakthrough
curves using a one-dimensional advection-dispersion transport model and varying the
degree of dispersion, as represented by varying Peclet numbers. The paper provides
excellent plots displaying the effect of different magnitudes of dispersion on the
resulting breakthrough curves. As expected, larger dispersion coefficients results in
breakthrough curves with shallower slopes for both the arrival and elution waves.
Batterman et al. (1995) examined the influence of porous media properties and
the relative humidity of the gas-phase on diffusion rates and total dispersion coef-
ficients. Effective diffusion coefficients were measured for a number of dry natural
and synthetic media and were found to be consistent with predictions from empir-
ical correlations (see Section 2.3). Total dispersion coefficients were measured for
the same media with carrier relative gas humidities ranging from 0 to 90% (gener-
ally corresponding to very dry soils with gravimetric soil-water contents <1%). The
authors concluded that mechanical mixing and diffusion contributed about equally to
dispersion under these conditions.
Batterman et al. (1995) report that methane experienced greater overall disper-
sion than did trichloroethene (TCE) in column studies performed over a range of
soil humidities. This trend is expected due to the much higher diffusivity of methane
relative to TCE. However, additional analysis shows that the difference in reported
dispersion coefficients for TCE and methane is too large to be explained by differ-
ences in diffusion coefficients alone. After correcting the total dispersion coefficients
reported by Batterman et al. (1995) for diffusion, the absolute dispersion due to
mechanical mixing is almost three times greater for methane than for TCE. Theoret-
ically, mechanical mixing should be a solute-independent term. This indicates that
additional transport processes were likely occurring in the experiments that were not
considered in the original data analysis.
Costanza-RobinsonandBrusseau(2002)observedthatthelowestmolecularweight
compound studied, methane, had the largest diffusion contribution in a wetted homo-
geneous natural sand system, comprising approximately 60% of the total observed
dispersion. For higher molecular weight compounds (e.g., difluoromethane and TCE)
mechanical mixing dominated dispersion, contributing between 50 and 100% of the
dispersion, depending on soil-water content. Relative contributions from mechani-
cal mixing increased at higher soil-water contents due to the consequent increase in
tortuosity and decrease in diffusion.
The influence of specific system properties on experimentally-determined disper-
sivity values, including soil-water content and particle- and pore-size distributions
have not been fully studied. Furthermore, laboratory-measured dispersivity values
are often derived from data from nonreactive tracer tests and are assumed to be repre-
sentative of dispersivities for reactive compounds under various system conditions.
Because dispersivity is a measure of the heterogeneity of the physical system, this
assumption is theoretically justifiable. However, for this assumption to hold in a real