most of the tests were done on these contaminants. Some studies showed a removal
of 99% (Pennell et al. 1994), while for others a removal of only 40–60% was
reached. It is difficult to find a clear explanation of these differences. In the field,
Ramsburg et al. (2005) reached a 99% removal of the initial PCE in 270 days of
Tween80
®
flushing. In experimental vertical tanks, Page et al. (2007) reported a
significant removal of PCE (70%). However, concentrations in water did not really
decreased due to an enhanced area of contact between NAPL and water. Through a
large comparison, Londergan et al. (2001) showed that the efficiency of the reduction in NAPL mass is not linked to the amount of surfactant. In this study, the
maximum reduction obtained was 98% in a controlled cell. Oostrom et al. (1999)
showed in a laboratory tank that only 60% of the NAPL TCE was removed after
65 PV of alternating water and surfactant injection, most of the remaining fraction
being in low permeability layers.
Concerning sparging, most of the studies were conducted on 2D vertical laboratory tanks. These experiments explored the preferential pathways of gas bubbles that
are often at the origin of the poor treatment efficiency (Braida and Ong 2001). X-ray
scanning also illustrated the preferential channels in column studies (Chen et al.
1996). This effect has been modeled by semiempirical approaches (Braida and Ong
2000). In a natural medium, there is no way to predict the sparging efficiency for a
given airflow rate.
Thermal remediation is known to be really efficient in the field (McGuire et al.
2006) but very few studies reported testing in use columns. Heating led to a fast and
complete removal in an unsaturated column (Kawala and Atamańczuk 1998). At
90
C, in saturated columns, fast and complete removal of PCE was reached
(Burghardt and Kueper 2008). Heron et al. (1998) showed removal of dissolved
TCE (at saturation) higher than 99% in an 1 Â 0.5 Â 0.1 m tank.
Almost no study compared several techniques neither in columns nor at the field
scale. Moreover, most of the results were obtained for chlorinated solvents while
BTEX-contaminated sites are numerous and remediating them can sometimes be
challenging. Direct comparison of remediation techniques on the same soil and for
the same contamination is quite important because across the literature, all experiments were conducted under different conditions and the extension to other conditions is challenging.
At the field or tank scale, the very large variations in the removal efficiency seem
to be mainly linked to the presence of heterogeneities in hydraulic conductivities.
Indeed, the least permeable zones tend to deviate the treatment fluid. However,
during the source evolution, the contaminants are preferentially leached out from the
high permeability zones but they finally remain in the least permeable ones. This
may also explain why the removal of PCE by surfactant dissolution was reported to
be much lower in 2D tank than in column experiments (Taylor et al. 2001).
In order to face the almost impossible prediction of the heterogeneity effect, an
option is to compare the different techniques within the same medium presenting the
same heterogeneity. To the best of our knowledge, it was done in only one study
(Conrad et al. 2002) where a 2D vertical experiment was conducted with emplaced
NAPL pools. The authors compared permanganate oxidation and surfactant
214
F. Jousse et al.
of 99% (Pennell et al. 1994), while for others a removal of only 40–60% was
reached. It is difficult to find a clear explanation of these differences. In the field,
Ramsburg et al. (2005) reached a 99% removal of the initial PCE in 270 days of
Tween80
®
flushing. In experimental vertical tanks, Page et al. (2007) reported a
significant removal of PCE (70%). However, concentrations in water did not really
decreased due to an enhanced area of contact between NAPL and water. Through a
large comparison, Londergan et al. (2001) showed that the efficiency of the reduction in NAPL mass is not linked to the amount of surfactant. In this study, the
maximum reduction obtained was 98% in a controlled cell. Oostrom et al. (1999)
showed in a laboratory tank that only 60% of the NAPL TCE was removed after
65 PV of alternating water and surfactant injection, most of the remaining fraction
being in low permeability layers.
Concerning sparging, most of the studies were conducted on 2D vertical laboratory tanks. These experiments explored the preferential pathways of gas bubbles that
are often at the origin of the poor treatment efficiency (Braida and Ong 2001). X-ray
scanning also illustrated the preferential channels in column studies (Chen et al.
1996). This effect has been modeled by semiempirical approaches (Braida and Ong
2000). In a natural medium, there is no way to predict the sparging efficiency for a
given airflow rate.
Thermal remediation is known to be really efficient in the field (McGuire et al.
2006) but very few studies reported testing in use columns. Heating led to a fast and
complete removal in an unsaturated column (Kawala and Atamańczuk 1998). At
90
C, in saturated columns, fast and complete removal of PCE was reached
(Burghardt and Kueper 2008). Heron et al. (1998) showed removal of dissolved
TCE (at saturation) higher than 99% in an 1 Â 0.5 Â 0.1 m tank.
Almost no study compared several techniques neither in columns nor at the field
scale. Moreover, most of the results were obtained for chlorinated solvents while
BTEX-contaminated sites are numerous and remediating them can sometimes be
challenging. Direct comparison of remediation techniques on the same soil and for
the same contamination is quite important because across the literature, all experiments were conducted under different conditions and the extension to other conditions is challenging.
At the field or tank scale, the very large variations in the removal efficiency seem
to be mainly linked to the presence of heterogeneities in hydraulic conductivities.
Indeed, the least permeable zones tend to deviate the treatment fluid. However,
during the source evolution, the contaminants are preferentially leached out from the
high permeability zones but they finally remain in the least permeable ones. This
may also explain why the removal of PCE by surfactant dissolution was reported to
be much lower in 2D tank than in column experiments (Taylor et al. 2001).
In order to face the almost impossible prediction of the heterogeneity effect, an
option is to compare the different techniques within the same medium presenting the
same heterogeneity. To the best of our knowledge, it was done in only one study
(Conrad et al. 2002) where a 2D vertical experiment was conducted with emplaced
NAPL pools. The authors compared permanganate oxidation and surfactant
214
F. Jousse et al.