233
Remediation of PAHs and NSO-Heterocycles
Comparing the six activated carbons, an approximate order with respect
to the adsorption parameter K was as follows: AC(30N) ~ AC(Epibon) < AC
(ROW) < AC(F200) < AC(1240) ~ AC(F400). However, relative standard deviations of K vary in the range of 20%–25% only, surprisingly similar for all
compounds.
11.3.2.4 I sotherm Parameters from Column Runs: Comparison
with Batch Experiments
As already discussed, contaminated groundwater is a complex mixture of
compounds with highly variable intrinsic properties and concentrations.
To our knowledge, there is no available model that calculates breakthrough
curves in column experiments on activated carbon for such a complex mixture. This is especially true for mixtures consisting of the number of compounds investigated in this study, although they are well characterized with
respect to their concentration. Furthermore, numerous additional parameters (e.g., diffusion coefficients) have to be known to calculate the breakthrough curves. Another possibility is the approximation of breakthrough
curves using equilibrium models, neglecting time-dependent processes and
therefore kinetic parameters. Nevertheless, commercially available models
(e.g., AdDesignS [36] or LDF [41]) take both breakthrough and equilibrium
curves into account only for a limited number of compounds.
Artificial mixtures were used to verify the experimental methods and to
compare batch and column experiments. Two column runs should serve as
instructive examples.
In the first example, breakthrough curves of a simple artificial threecomponent mixture (benzene, toluene, and benzofuran), shown in Figure
11.9, were obtained from port 2 with the column described in the experimental section (AC Epibon Y12×40, m(AC at port 2) = 15.55 g, flow = 7.1 mL/
min, c 0 = 10 mg/L for each component, T = 11°C, and run time 63 days).
Breakthrough curves were calculated with the LDF-model [41], equilibrium
curves using the IAS-approach described above. Freundlich parameters as
the main input data for both models were taken from batch experiments,
as already presented in Table 11.3. Breakthrough curves from LDF were
adjusted slightly using the necessary (and estimated) diffusion coefficients.
Concentrations above the inflow concentrations for benzene and toluene are
typical for desorption processes due to stronger adsorbing compounds and
were found in all column runs.
In extension to this three-component mixture, breakthrough curves were
measured for an artificial six-component mixture (phenol, benzene, toluene, benzofuran, 2-methylquinoline, and benzothiophene, concentrations
were sampled at eight ports, AC Epibon Y12×40, m(AC at port 8) = 70.9 g,
flow = 6.9 mL/min, c 0 = 3–10 mg/L, T = 12°C, and run time = 80 days).
Concentrations from port 3 were used to evaluate Freundlich parameters.
From the LDF-model, the best consistency was found after manual adjustment
Remediation of PAHs and NSO-Heterocycles
Comparing the six activated carbons, an approximate order with respect
to the adsorption parameter K was as follows: AC(30N) ~ AC(Epibon) < AC
(ROW) < AC(F200) < AC(1240) ~ AC(F400). However, relative standard deviations of K vary in the range of 20%–25% only, surprisingly similar for all
compounds.
11.3.2.4 I sotherm Parameters from Column Runs: Comparison
with Batch Experiments
As already discussed, contaminated groundwater is a complex mixture of
compounds with highly variable intrinsic properties and concentrations.
To our knowledge, there is no available model that calculates breakthrough
curves in column experiments on activated carbon for such a complex mixture. This is especially true for mixtures consisting of the number of compounds investigated in this study, although they are well characterized with
respect to their concentration. Furthermore, numerous additional parameters (e.g., diffusion coefficients) have to be known to calculate the breakthrough curves. Another possibility is the approximation of breakthrough
curves using equilibrium models, neglecting time-dependent processes and
therefore kinetic parameters. Nevertheless, commercially available models
(e.g., AdDesignS [36] or LDF [41]) take both breakthrough and equilibrium
curves into account only for a limited number of compounds.
Artificial mixtures were used to verify the experimental methods and to
compare batch and column experiments. Two column runs should serve as
instructive examples.
In the first example, breakthrough curves of a simple artificial threecomponent mixture (benzene, toluene, and benzofuran), shown in Figure
11.9, were obtained from port 2 with the column described in the experimental section (AC Epibon Y12×40, m(AC at port 2) = 15.55 g, flow = 7.1 mL/
min, c 0 = 10 mg/L for each component, T = 11°C, and run time 63 days).
Breakthrough curves were calculated with the LDF-model [41], equilibrium
curves using the IAS-approach described above. Freundlich parameters as
the main input data for both models were taken from batch experiments,
as already presented in Table 11.3. Breakthrough curves from LDF were
adjusted slightly using the necessary (and estimated) diffusion coefficients.
Concentrations above the inflow concentrations for benzene and toluene are
typical for desorption processes due to stronger adsorbing compounds and
were found in all column runs.
In extension to this three-component mixture, breakthrough curves were
measured for an artificial six-component mixture (phenol, benzene, toluene, benzofuran, 2-methylquinoline, and benzothiophene, concentrations
were sampled at eight ports, AC Epibon Y12×40, m(AC at port 8) = 70.9 g,
flow = 6.9 mL/min, c 0 = 3–10 mg/L, T = 12°C, and run time = 80 days).
Concentrations from port 3 were used to evaluate Freundlich parameters.
From the LDF-model, the best consistency was found after manual adjustment