Seven tasks were performed as a part of this technology demonstration. These tasks are shown in Table 4.3.0.1. Initially, the formation was sampled by coring during the installation of wells and multilevel samplers. Descriptive logs of the wells cored are available in the database in the table labeled BoringLogs. The core samples were analyzed for a group of chemicals believed to be representative of the site NAPL and of regulatory concern. The selected contaminants represented several classes of chemicals. Following installation of the wells and monitoring equipment, the formation was allowed to re-equilibrate before any additional measurements were made. Ground-water samples were collected under both static and dynamic conditions. Injecting uncontaminated ground-water into the formation, simulating a forced gradient flush or water flood, created the dynamic flow conditions. Static samples were collected to assess the maximum concentration that would be anticipated in water samples from this location. Ground-water sampling was followed by a pre-remediation tracer study. The pre-remediation tracer study was designed to estimate the mass and distribution of NAPL in the formation and to characterize hydrodynamic performance. Once the pre-remediation tracer study was completed, the remedial technology was implemented and its performance monitored. After termination of the remedial demonstration, another tracer test was conducted, followed by post-demonstration ground-water sampling and core sampling. A timeline for these activities is shown in Table 4.3.0.1 and key operational parameters during the remediation demonstration and tracer tests are listed in Table 4.3.0.2. Analysis of these data permits an evaluation of the technology.
The tertiary butyl alcohol (TBA) used in this experiment was a gasoline-grade product donated by Arco Chemical Company. This grade of TBA is at least 96% TBA, with small amounts of water, methanol and other trace impurities. Approximately 24,000 l (7,000 gallons) of TBA and 1,150 l (330 gallons) of n-hexanol were used in the test. Selected physical properties of these alcohols are presented in Table 4.3.0.3. Initially, TBA was injected alone (3,800 l (1,100 gal.)). This was followed by an injection of 7,000 l (2,030 gal) of TBA and hexanol (16% hexanol by volume). The TBA-hexanol mix was followed by an additional 14,500 l (4,200 gal.) of TBA. Pure TBA freezes at about 20oC, but the freezing point rises quickly with addition of water. Laboratory tests suggested that freezing should not be a problem for the fuel-grade TBA at the temperatures expected in the field. The test was initiated in the afternoon with TBA as delivered with no apparent problems. However, during the first evening the lines froze suggesting the delivered TBA had a higher purity than the TBA used in the laboratory. Water was added to the TBA to raise the freezing point. Except for a small initial volume, purity of injected TBA did not exceed 95%. Tracer data collected from multilevel samplers showed that there was a highly conductive zone in the formation near the clay-sand interface. Tracer results also suggested that contaminant concentrations were relatively low in this conductive zone. Therefore, packers were used in the lower part of both the injection and extraction wells to help focus the remedial fluid in the most contaminated portion of the aquifer.
From the initial TBA injection through the TBA-hexanol injection, the
flow rate was maintained at 2.5 lpm (0.7 gpm). After all of the TBA-hexanol
mix had been injected, the flux was increased to 3.8 lpm (1.1 gpm) near
the end of the flood and the extraction pattern was modified to increase
flow through selected portions of the formation. Samples collected from
MLS in highly contaminated zones became cloudy as the alcohol front arrived;
suggesting the presence of an emulsion. This was typically followed
by the appearance of moderate to large amounts of NAPL in the samples.
In some cases, the NAPL floated in the sample and in some cases, it sank
(due to the low alcohol density). In a few cases, the NAPL split
into two separate phases, with part floating and part sinking in the sample
vials.
Table 4.3.0.1. Study sequence
| Test | Activity | Fluid injected | Total flux rate (lpm)- volume (l) | Duration |
| Core and install instrumentation | Collect soil samples | None | 11/6/95 - 11/9/95 | |
| Ground-water sampling | Collect water samples | None | no data in data base | |
| Pre-remediation tracer | Establish flow field | Water | 4.54 (lpm) | ? |
| Inject tracer suite 1 | 4.54 (lpm)
1614 (l) |
5/15/96 | ||
| Maintain flow field | Water | 4.54 (lpm) | 5/15/96 - 5/24/96 | |
| Remedial technology | Establish flow field | Water | 2.65 lpm | 8/4/96 - 8/5/96 |
| Inject remedial fluid | 96% TBA 4% Water | 2.65 | 8/5/96 - 8/6/96 | |
| Inject remedial fluid | 81% TBA 15% Hexanol 4% Water | 2.65 lpm | 8/6/96 - 8/8/96 | |
| Inject remedial fluid | 96% TBA 4% Water | 2.65 - 4.2 lpm | 8/8/96 - 8/11/96 | |
| Inject remedial fluid | 50% TBA 50% Water | 4.2 lpm | 8/12/96 | |
| Remove cosolvent | Water | 5.7 - 9.5 lpm | 8/12/96 - 8/30/96 | |
| Post-remediation tracer | Establish flow field | Water | 4.54 (lpm) | ? |
| Inject tracer suite 2 | 4.54 (lpm)
1281 (l) |
8/31/96 | ||
| Maintain flow field | Water | 4.54 (lpm) | 8/31/96 - 9/9/96 | |
| Ground-water sampling | Collect water samples | None | no data in data base | |
| Core | Collect soil samples | None | 9/23/96 - 9/24/96 |
Table 4.3.0.2. Operational parameters
| Parameter |
|
||
| Pre-remediation Tracer | Demonstration | Post-Remediation Tracer | |
| Average Saturated Thickness (m) | 3.7 | 3.7 | 3.8 |
| Average Gradient Across Cell (m/m) | |||
| Average Influent Flow rate (lpm) | 4.5 | variable | 4.62 |
| Average Effluent Flow rate (lpm) | 4.5 | variable | |
Table 4.3.0.3 Properties of Cosolvents
| Property | tert-Butanol (TBA) | n-Hexanol |
| Density (g/cm3) |
|
|
| Viscosity (cP @ 25oC) |
|
|
| Flash point (oC) |
|
|
| Vapor Pressure (mm Hg) |
|
|
| Aqueous solubility (wt % @ 25oC) |
|
|
Prior to remediation, the contamination in the cosolvent mobilization cell was fairly tightly clustered with less variation in sample concentrations than observed for some of the other cells. There appears to be greater variability in contaminant concentrations after the cosolvent flood. This suggests that the treatment zone may not have been uniformly swept by the remedial fluid. The data have not been plotted as a function of travel path which might be used to evaluate this hypothesis. Since only a few pore volumes of remedial fluid were used, it is likely the performance would be improved significantly if more remedial fluid had been used or if the remedial fluid had been recirculated. However, even with less than optimal conditions, core data suggests that more than 80% of the initial mass of the monitored NAPL constituents were removed with the exception of 1,2-dichlorobenzene.
There is a reasonably consistent pattern between the number of samples
that are below the quantification limit after remediation and the water
solubility of the compound. The higher the solubility, the larger the number
of samples that are below quantification limit after remediation. The most
soluble chemical tracked, 1,1,1-trichloroethane, was essentially removed
during the study. This would suggest that if more remedial fluid had been
used, the remaining chemicals would have been removed by solubilization.
The results presented here should be considered as one point on a performance
curve.
Table 4.3.1.1. Summary of Core data based on boxcar averaged results
| Chemical | Pre-remediation concentration (mg/kg) | Post-remediation concentration (mg/kg) | Fraction removed |
| dichlorobenzene | 2.1 | 1.6 | 0.24† |
| 1,1,1-trichloroethane | 0.66 | 0.025 | 0.96 |
| toluene | 3.2 | 0.18 | 0.94 |
| o-xylene | 5 | 0.3 | 0.94 |
| m-xylene | 1.9 | 0.15 | 0.92 |
| naphthalene | 3 | 0.32 | 0.89 |
| trimethylbenzene | 7.4 | 0.47 | 0.94 |
| decane | 55 | 11 | 0.80 |
| undecane | 120 | 19 | 0.84 |
| ethylbenzene | 0.92 | 0.062 | 0.93 |
† Value significantly influenced by two samples near the bottom of the
test cell that were greater than any samples collected pre-remediation.
Deleting these two samples changes the fraction removed to ~0.9.
| Tracer | Pre-remediation | Post-remediation | ||||
| Concentration (mg.l) | Total Mass Injected (g) |
(g)* |
Concentration (mg.l) | Total Mass Injected (g) |
(g)* |
|
| bromide |
|
|
|
|
|
|
| methanol |
|
|
|
|
|
|
| tert-butanol |
|
|
|
|
||
| 2,2-dimethyl-3-pentanol |
|
|
|
|
||
| hexanol |
|
|
|
|
||
| 6-methyl-2-heptanol |
|
|
|
|
||
| Tracer Volume (l) |
|
|
||||
| Injection Time (min) |
|
|
||||
* Based on zero moment of extrapolated data.
Partitioning tracer tests (PTT) were performed on all cells at OU1, Hill AFB prior to and after completion of each technology demonstration. The purpose of the tracer tests was to estimate the pre- and post-treatment NAPL content and distribution within the saturated zone of the cell. Changes in NAPL content were used to assess the effectiveness of the technology. For the partitioning tracer tests, a flow field was established within the saturated zone by pumping water into and out of the cell at opposite ends. After steady state flow was established, a pulse of tracers was introduced through the injection wells. Breakthrough curves for the tracers were measured at extraction wells and intermediate multilevel samplers. A suite of both partitioning and non-partitioning tracers were used (Table 4.3.2.1), and the differences in arrival times were used to estimate NAPL content. Four different approaches were used to interpret the tracer data. Two of the approaches evaluate the data using moment analysis calculating the first and second moments. The first approach used all of the data, as it was available in the database. The second approach extrapolated the final 25% of the data assuming it followed a log-linear relationship until the concentration was 0.001 mg/l. The remaining two methods fit the data to simple one-dimensional models. The first model is a simple convective dispersive model as described in Section 4.1.2. The second model, also described in 4.1.2, is a stochastic one- dimensional model, which assumes dispersion is to be a result of the heterogeneity in the flow field. This approach assumes the flow field is log-normally distributed and that the contaminant is uniformly distributed within the flow field.
The experimental data collected prior to remediation from the extraction wells are shown in Figures 4.3.2.1, 4.3.2.6, and 4.3.2.11. Since the partition coefficient for the 2,2-dimethyl-3-pentanol is the largest of the tracers used, this tracer provides the greatest separation between the retarded and non-retarded tracer and should provide the most reliable data. Results are summarized in Table 4.3.2.2 for extraction well 1, and Table 4.3.2.3 for extraction well 2. The pre-remediation NAPL saturation (volume fraction) based on 2,2-dimethyl-3-pentanol data, ranges from 0.05 to 0.066 for the volume sampled by well 1, 0.07 to 0.122 for well 2 and 0.072 to 0.081 for well 3. The time of travel for bromide in extraction well 1 ranged from 0.96 days using the method of moments and the extrapolated data to 0.42 days(NOTE: the moment time reported in the table is not the travel time; the travel time is the reported time minus 0.5*the pulse duration). The modeling results suggest the residence time for bromide is approximately 1/2 the time suggested by the method of moments. This difference is likely due to the model not fitting the tail of the breakthrough curve. The descrepency between observed and fitted function is easily seen in Figure 4.3.2.7b Post-remediation residence time for bromide was not the same as the pre-remediation residence time. It is not reasonable to make conclusions about remediation based on changes in residence times of the non-retarded tracers since minor changes in the flow volume or flow rate can significantly influence the interpretation. Based on a visual examination (Figures 4.2.2.1 and 4.2.2.6), the one-dimensional-convective-dispersive model (CDM) appears to describe the data better than the stochastic model. The stochastic model suggests an earlier rise in concentration than is predicted by the CDM. Post remediation estimates are quite similar, however, the stochastic model still gives a poor fit to the data. The apparent problem is the assumed distribution function for the flow times. A log normal distribution does not appear to be correct.
Table 4.3.2.2. Summary of Extraction Well 1 tracer analysis
| Extraction Well 1
|
|
|
||||
| conservative (bromide) | 2,2-dimethyl
-3-pentanol |
NAPL Saturation | conservative (bromide) | 2,2-dimethyl
-3-pentanol |
NAPL Saturation | |
| Pulse duration (days) |
|
|
|
|
||
| zero moment data |
|
|
|
|
||
| zero moment extrapolated |
|
|
|
|
||
| first moment / zero moment data (days) |
|
|
|
|
|
|
| first moment extrapolated / zero moment (days) |
|
|
|
|
|
|
| Convective Dispersive Model |
|
|
||||
| dispersivity (m) |
|
|
|
|
||
| est. initial conc. (mg/l) |
|
|
|
|
||
| mean time of travel (days) |
|
|
|
|
||
| Stochastic model |
|
|
||||
| variance in travel time |
|
|
|
|||
| travel time (days) |
|
|
|
|
||
| est. initial conc. (mg/l) |
|
|
|
|
||
Figure 4.3.2.1 Extraction
Well 1 Pre-remediation tracer analysis
2,2-dimethyl-3-pentanol
Figure 4.3.2.1.b Extraction
Well 1 Pre-remediation tracer analysis 2,2-dimethyl.3.pentanol log
Figure 4.3.2.2 Extraction
Well 1 Post-remediation tracer analysis 2,2,-dimethyl-3-pentanol
Figure 4.3.2.2.b Extraction
Well 1 Post-remediation tracer analysis 2,2-dimethyl-3-pentanol log
Figure 4.3.2.3 Extraction
Well 1 Post-remediation tracer analysis hexanol
Figure 4.3.2.3.b Extraction
Well 1 Post-remediation tracer analysis hexanol log
Figure 4.3.2.4 Extraction
Well 1 Post-remediation 6-methyl-3-hexanol
Figure 4.3.2.4.b Extraction
Well 1 Post-remediation 6-methyl-3-hexanol log
Figure 4.3.2.5 Extraction
Well 1 Post-remediation methylheptanol
Figure 4.3.2.5.b Extraction
Well 1 Post-remediation methylheptanol log
Table 4.3.2.3. Summary of Extraction Well 2 tracer analysis
| Extraction Well 2 |
|
|
||||
| conservative (bromide) | 2,2-dimethyl
-3-pentanol |
NAPL Saturation | conservative (bromide) | 2,2-dimethyl
-3-pentanol |
NAPL Saturation | |
| Pulse duration (days) |
|
|
|
|
||
| zero moment data |
|
|
|
|
||
| zero moment extrapolated |
|
|
|
|
||
| first moment / zero moment data (days) |
|
|
|
|
|
|
| first moment extrapolated / zero moment (days) |
|
|
|
|
|
|
| Convective Dispersive Model |
|
|
||||
| dispersivity (m) |
|
|
|
|
||
| est. initial conc. (mg/l) |
|
|
|
|
||
| mean time of travel (days) |
|
|
|
|
||
| Stochastic model |
|
|
||||
| variance in travel time |
|
|
|
|
||
| travel time (days) |
|
|
|
|
||
| est. initial conc. (mg/l) |
|
|
|
|
||
Figure 4.3.2.6 Extraction
Well 2 Pre-remediation tracer analysis 2,2-dimethyl-3-pentanol
Figure 4.3.2.6.b Extraction
Well 2 Pre-remediation tracer analysis 2,2-dimethyl-3-pentanol log
Figure 4.3.2.7 Extraction
Well 2 Post-remediation tracer analysis 2,2-dimethyl-3-pentanol
Figure 4.3.2.7.b Extraction
Well 2 Post-remediation tracer analysis 2,2-dimethyl-3-pentanol log
Figure 4.3.2.8 Extraction
Well 2 Post-remediation hexanol
Figure 4.3.2.8.b Extraction
Well 2 Post-remediation hexanol log
Figure 4.3.2.9 Extraction
Well 2 Post-remediation methylheptanol
Figure 4.3.2.9.b Extraction
well 2 Post-remediation methylheptanol log
Figure 4.3.2.10 Extraction
Well 2 Post-remediation 6-methyl-3-hexanol
Figure 4.3.2.10.b Extraction
Well 2 Post-remediation 6-methyl-3-hexanol log
Table 4.3.2.4. Summary of Extraction Well 3 tracer analysis
| Extraction Well 3 |
|
|
||||
| conservative (bromide) | 2,2-dimethyl
-3-pentanol |
NAPL Saturation | conservative (bromide) | 2,2-dimethyl
-3-pentanol |
NAPL Saturation | |
| Pulse duration (days) |
|
|
|
|
||
| zero moment data |
|
|
|
|
||
| zero moment extrapolated |
|
|
|
|
||
| first moment / zero moment data (days) |
|
|
|
|
|
|
| first moment extrapolated / zero moment (days) |
|
|
|
|
|
|
| Convective Dispersive Model |
|
|
||||
| dispersivity (m) |
|
|
|
|
||
| est. initial conc. (mg/l) |
|
|
|
|
||
| mean time of travel (days) |
|
|
|
|
||
| Stochastic model |
|
|
||||
| variance in travel time |
|
|
|
|
||
| travel time (days) |
|
|
|
|
||
| est. initial conc. (mg/l) |
|
|
|
|
||
Figure 4.3.2.11 Extraction
Well 3 Pre-remediation 2,2-dimethyl-3-pentanol
Figure 4.3.2.11.b Extraction
Well 3 Pre-remediation 2,2-dimethyl-3-pentanol log
Figure 4.3.2.12 Extraction
Well 3 Post-remediation 2,2-dimethyl-3-pentanol
Figure 4.3.2.12.b Extraction
Well 3 Post-remediation 2,2-dimethyl-3-pentanol log
Figure 4.3.2.13 Extraction
Well 3 Post-remediation hexanol
Figure 4.3.2.13.b Extraction
WEll 3 Post-remediation hexanol log
Figure 4.3.2.14 Extraction
Well 3 Post-remediation methylheptanol
Figure 4.3.2.14.b Extraction
Well 3 Post-remediation methylheptanol log
Figure 4.3.2.15 Extraction
Well 3 Post-remediation 6-methyl-3-hexanol
Figure 4.3.2.15.b ExtractionWell
3 Post-remedaiton 6-methyl-3-hexanol log
The initial dichlorobenzene in the cell is approximately 175 g (mean of the estimated mass excluding the inverse distance squared method of calculation). Figure 4.3.3.1 dichlorobenzene shows the dichlorobenzene eluted from the cell. A total of 57 grams were quantified as removed. If we consider the standard deviation of the estimated mean concentration, the best we can estimate the initial mass is within 50% of the reported value. This makes estimating the fraction removed quite uncertain. The results are similar to the results observed from the core analysis (i.e., only a small amount was removed, if any). Nothing conclusive can be said based on the dichlorobenzene data.
The initial mass estimate for naphthalene was 240 g. Figure 4.3.3.2 naphthalene shows approximately 118 g of naphthalene removed during the elution phase of the project. This is a very small removal compared to the initial mass in the cell. The core data, on the other hand, shows significantly more removal. It is not likely that the difference can be attributed to variability in sampling, assuming the sampling volumes are consistent. One possible explanation might be that as droplets of NAPL were removed, they were not adequately sampled in the effluent stream. Since the objective of this demonstration was to promote mobilization, the cell effluent may have contained immiscible phases. Droplets of NAPL were observed in the effluent lines upgradient of the sampling locations, but they were not observed at the sampling points. It is believed that a portion of the remedial fluid bypassed the zone of contamination and that dissolution of the NAPL droplets took place in the flow lines. No attempt was made to confirm the presence of small organic droplets in the sample vials that may not have been sampled during chemical analysis. Any phase separation during sample handling prior to analysis would result in low estimates of mass removal.
Undecane should be the best choice for quantifying the mass removal with some certainty since the mass fraction of these chemicals is relatively high in the NAPL mixture. Initially, the core data suggests that there was 9.2 kg of undecane in the formation and Figure 4.3.3.4 undecane suggests that 3.2 kg of undecane were removed. This would only be about a 35% removal compared to approximately 90% removal based on the core data. This removal rate, like naphthalene, is significantly less than estimated using either the core data for specific chemicals or the tracer data for mass NAPL removal. The most likely explanation for this discrepancy is in the sampling of the fluid. Contaminants initially present as a separate phase or which separated from the aqueous phase between sampling and analysis may not have been included in the GC injection volume. Decane was also present in the core data in significant quantities at 4.3 kg. If a relationship were developed between either decane or undecane and tetradecane (e.g., Mole fractions of the compounds), then it might be possible to relate the elution curve shown in Figure 4.3.3.3 tetradecane that shows 314 g of tetradecane removal to one of the initial contaminants present.
Table 4.3.3.1 Fraction removed
| Chemical | Initial mass *(g) | Mass extracted (g) | Fraction removed |
| dichlorobenzene | 170 | 57.6 | 0.34 |
| decane | 4100 | 675** | NA |
| undecane | 9100 | 3178 | 0.35 |
| naphthalene | 240 | 118 | 0.49 |
