Results Discussion

The raw data presented for the Triton X-100 experiments have been presented previously (Deitsch and Smith 1995) in a modified form. The distributed rate model was fit to the CFSTR desorption data (Figures 2 - 5). The optimal model parameters and percent errors are given for each experiment in Table 2. The distributed rate model was able to simulate the CFSTR data well. In several of the surfactant CFSTRs, the model underestimated the aqueous TCE concentrations at later times. This suggests that for some of the surfactant CFSTRs the amount of TCE flushed from the reactor may be greater than the amount predicted from the simulation.

Figure 2. Distributed rate model simulations of the data obtained from the two CFSTRs flushed with the non-surfactant solutions.

TIME (DAYS)

Figure 2. Distributed rate model simulations of the data obtained from the two CFSTRs flushed with the non-surfactant solutions.

Figure 3. Distributed rate model simulations of the data obtained from the CFSTRs flushed with the 30 mg/L, 300 mg/L, and 3,000 mg/L Triton X-100 solutions.

TIME (DAYS)

Figure 3. Distributed rate model simulations of the data obtained from the CFSTRs flushed with the 30 mg/L, 300 mg/L, and 3,000 mg/L Triton X-100 solutions.

Table 2. Minimum percent error and corresponding simulation parameters for the distributed-rate model

3

% Error

Water A

0.114

173.9

3.223

Water B

0.156

37.13

2.864

30 mg/L Triton X-100 A

0.419

2.244

7.259

30 mg/L Triton X-100 B

0.500

1.820

9.669

300 mg/L Triton X-100 A

0.564

2.460

8.418

300 mg/L Triton X-100B

0.472

5.917

14.07

3,000 mg/L Triton X-100 A

0.249

6.517

9.158

3,000 mg/L Triton X-100 B

0.238

38.061

9.558

30 mg/L Tween 20 A

0.354

3.103

4.309

30 mg/L Tween 20 B

0.289

2.686

4.410

300 mg/L Tween 20 A

0.267

4.138

8.410

300 mg/L Tween 20 B

0.316

2.097

4.032

3,000 mg/L Tween 20 A

0.172

16.413

6.637

3,000 mg/L Triton X-405 A

0.165

15.611

5.429

3,000 mg/L Triton X-405 B

0.190

4.705

6.672

3,000 mg/L SDSa A

0.0894

274.96

8.975

3,000 mg/L SDS B

0.151

10.889

7.978

3,000 mg/L SDBSb A

0.0765

323.30

12.35

3,000 mg/L SDBS B

0.0938

50.1113

11.21

J sodium dodecylsulfate b sodium dodecylbenzenesulfonate

J sodium dodecylsulfate b sodium dodecylbenzenesulfonate

To determine whether the surfactant had enhanced the rate of TCE desorption, the amount of TCE flushed from each CFSTR as function of time was determined from the model simulations. The model simulations were averaged for each surfactant concentration. The model predictions for the Triton X-100 CFSTRs are shown in Figure 6. Based on Figure 6, it appears that the addition of Triton X-100 at all three of the concentrations increased the amount of TCE removed the CFSTRs when compared to the water CFSTRs. The results of this analysis are consistent with the results reported by Deitsch and Smith (1995). In the preceding reference, the same data were analyzed using a model that incorporated a time-varying masstransfer rate coefficient. In that study, the rate coefficient was manually altered during the course of the simulation to obtain a "good" fit of the averaged experimental data. The modeling procedure employed by Deitsch and Smith (1995) enabled the CFSTR concentration profiles to be numerically integrated, and thus to develop mass-removed plots analogous to Figure 6. The present analysis using the distributed rate model and an optimization routine indicate that the conclusions of the original study were not biased by the fitting technique employed in the original study.

Figure 6. Calculated percentages (based on distributed rate model optimal simulations) of TCE removed from the CFSTRs flushed with the non-surfactant solution, the 30 mg/L Triton X-100 solution, the 300 mg/L Triton X-100 solution, and the 3,000 mg/L Triton X-100 solution. The removal profiles shown are averages of the replicate experiments.

TIME (DAYS)

Figure 6. Calculated percentages (based on distributed rate model optimal simulations) of TCE removed from the CFSTRs flushed with the non-surfactant solution, the 30 mg/L Triton X-100 solution, the 300 mg/L Triton X-100 solution, and the 3,000 mg/L Triton X-100 solution. The removal profiles shown are averages of the replicate experiments.

The model predictions for the percent TCE removed from the Tween 20 CFSTRs are shown in Figure 7. In Figure 8, the model predictions for the percent TCE removed from the 3,000 mg/L sodiumdodecylsulfate CFSTRs, the 3,000 mg/L sodiumdodecylbenzenesulfonate CFSTRs, and the 3,000

mg/L Triton X-405 CFSTRs are shown. In contrast to the Triton X-100 CFSTRs, the surfactants tested appeared to have no effect on the rate of desorption (i.e., 30 mg/L Tween 20 CFSTR) or appeared to have inhibited the rate of TCE desorption (i.e., remaining surfactants). For example, the surfactants shown in Figure 8 reduced the amount of TCE flushed from the column by approximately 20%.

Figure 7. Calculated percentages (based on distributed rate model optimal simulations) of TCE removed from the CFSTRs flushed with the non-surfactant solution, the 30 mg/L Tween 20 solution, the 300 mg/L Tween 20 solution, and the 3,000 mg/L Tween 20 solution. The removal profiles shown are averages of the replicate experiments (except for the 3,000 mg/L Tween 20 profile that is from a single CFSTR).

TIME (DAYS)

Figure 7. Calculated percentages (based on distributed rate model optimal simulations) of TCE removed from the CFSTRs flushed with the non-surfactant solution, the 30 mg/L Tween 20 solution, the 300 mg/L Tween 20 solution, and the 3,000 mg/L Tween 20 solution. The removal profiles shown are averages of the replicate experiments (except for the 3,000 mg/L Tween 20 profile that is from a single CFSTR).

Figure 8. Calculated percentages (based on distributed rate model optimal simulations) of TCE removed from the CFSTRs flushed with the non-surfactant solution, the 3,000 mg/L Sodiumdodecylsulfate (SDS) solution, the 3,000 mg/L Sodiumdodecylbenzenesulfonate (SDBS) solution, and the 3,000 mg/L Triton X-405 solution. The removal profiles shown are averages of the replicate experiments.

TIME (DAYS)

Figure 8. Calculated percentages (based on distributed rate model optimal simulations) of TCE removed from the CFSTRs flushed with the non-surfactant solution, the 3,000 mg/L Sodiumdodecylsulfate (SDS) solution, the 3,000 mg/L Sodiumdodecylbenzenesulfonate (SDBS) solution, and the 3,000 mg/L Triton X-405 solution. The removal profiles shown are averages of the replicate experiments.

Without knowing the effect of the surfactants on the distribution coefficient for TCE sorption to the peat soil, it is not possible to determine what mechanism caused the reduction in TCE removal. It is possible that the addition of sodiumdodecylsulfate, sodiumdodecylbenzenesulfonate, Tween 20, or Triton X-405 may have enhanced the sorption of TCE to the peat soil. If this were the case, the reduction in TCE removal would be attributable to a decrease in the magnitude of the concentration gradient driving the desorption process. Another possibility is that the addition of these surfactants may have inhibited the diffusion of the TCE from the peat soil to the aqueous phase. This last hypothesis could result if the surfactant molecules are not compatible with the glassy phase of the soil organic matter. If the surfactant molecules do not transform the glassy region of the soil organic matter to "rubbery" soil organic matter, the presence of the surfactant would not increase the rate of TCE diffusion from the glassy soil organic matter. In contrast, if the surfactants accumulate within the rubbery phase of the soil organic matter, the polymer density of the rubbery soil organic matter may increase and thus reduce the rate of solute diffusion through the soil organic matter.

As described previously, Triton X-100 at concentrations of 30 mg/L and 300 mg/L should not affect the magnitude of TCE sorption to the peat soil. Therefore, for these two surfactant concentrations, the optimal distributions of rate coefficients can be compared to the distribution of rate coefficients from the water CFSTRs. The average distribution of rate coefficients for these Triton X-100 CFSTRs and the water CFSTRs are shown in Figure 9.

lxlO"5 lxlO"4 lxlO"1 lxl 0"2 1x10"' 1x10° lxlO1

Figure 9. Optimal distributions of mass-transfer rate coefficients for the labelled CFSTR experiments. The distributions shown are the average distributions of the replicate experiments.

lxlO"5 lxlO"4 lxlO"1 lxl 0"2 1x10"' 1x10° lxlO1

MASS-TRANSFER RATE COEFFICIENTS (1/DAYS)

Figure 9. Optimal distributions of mass-transfer rate coefficients for the labelled CFSTR experiments. The distributions shown are the average distributions of the replicate experiments.

It is apparent from Figure 9 that the addition of Triton X-100 has shifted the distribution of rate coefficients to larger values when compared to the water CFSTRs. For the 300 mg/L Triton X-100 CFSTR, the spectrum of rate coefficients is shifted approximately one order of magnitude higher than the water CFSTR rate coefficient distribution. Therefore, the addition of Triton X-100 to the system has increased the rate of diffusion of the TCE from the peat to the aqueous phase. It is hypothesized that the addition of the Triton X-100 has transformed regions of glassy soil organic matter to rubbery soil organic matter, thus increasing the rate of solute diffusion through the soil organic matter. Evidence supporting this hypothesis has been given previously.

The results presented in this section indicate that all surfactants may not be suitable for surfactant-enhanced desorption. Multiple factors may influence the suitability of a surfactant for surfactant-enhanced desorption. These factors have been discussed previously. It is clear that additional research is needed to better understand how surfactant molecular structure and soil composition/chemistry affect the rate of solute desorption. However, the results presented in this chapter and in other studies indicate that surfactant-enhanced remediation of aquifers is a promising technology that needs to be explored.

0 0

Post a comment