Page 136 - gas transport in porous media

Chapter 7: Gas-Phase Dispersion in Porous Media
129
system, all relevant transport processes must be appropriately accounted for in the
analysis. For example, if additional transport processes, such as rate-limited mass-
transfer are occurring, but are not considered in the data analysis, the dispersivity
values will become lumped, solute-dependent values.
Costanza-Robinson and Brusseau (2002) performed experiments to examine these
issues. Diffusional dispersion contributions were explicitly accounted for using prop-
erties of the porous medium and literature tortuosity correlations and diffusion
constants. The total dispersion coefficient was obtained by fitting the experimen-
tal breakthrough curves with an equilibrium, one-dimensional, transport model. The
difference between the total dispersion coefficient and the diffusional contributions
was taken as the mechanical mixing contribution. The dispersivity values calcu-
lated in this manner were observed to be constant for the porous medium studied
at soil-water contents ranging from 2 to 14%. This indicates that it may be appro-
priate to use a single dispersivity value to represent a given porous media over a
range of natural conditions. Moreover, the calculated dispersivities were the same
for all compounds studied, which included a nonreactive (methane), a water-soluble
(difluoromethane), and a water-soluble and sorbing solute (TCE). This indicates that
in the 2 to 14% soil-water content range, the data analysis appropriately accounted
for transport processes, such that the dispersivity value was a true measure of the
porous medium heterogeneity, rather than a solute-dependent lumped term. At soil-
water contents greater than 14%, dispersivities became solute-dependent, indicating
that additional transport processes were being lumped into the dispersivity term. The
authors attributed this to rate-limited diffusion of the soluble solutes through water
films, which was not considered in the data analysis.

7.4 FIELD AND MODELING INVESTIGATIONS

The relatively few experimental field investigations of gas-phase dispersion have
focused on diffusion. Raney (1949) and Lai et al. (1976) have presented methods for
measuring effective diffusion coefficients in-situ that are applicable to surface soils.
Weeks et al. (1982) examined the use of atmospheric pollutants, fluorocarbons F-11
and F-12, in measuring effective vadose-zone diffusion rates, concluding that gas-
phase diffusion is likely the most important transport mechanism in regions where
groundwater recharge is small. As expected, soil tortuosity and the solubility and
sorption of the gases resulted in measured effective diffusion coefficients that were
much less than those estimated for free-air diffusion. Numerical-modeling results
indicated that the near-surface region had lower tortuosities, while deeper layers
contributed most significantly to reduced diffusion rates. Finally, Weeks et al. (1982)
concluded that the relative agreement between their optimized tortuosity factors and
tortuosity factors estimated via a number of theoretical and empirical approaches
lends support to the use of diffusion theory in predicting soil gas concentrations, even
in large-scale, heterogeneous natural systems.
Mathematical modeling of gas-phase transport has also focused largely on diffu-
sional processes, citing barometric pressure gradients and consequent advection as