AWI Final Report Phase II Treatablility Studies DNAPL Contaminated Soils.pdf

January 19, 2012

Mr. Peter Pellissier, PE Senior Project Manager EA Engineering, Science and Technology, Inc.

11019 McCormick Road Hunt Valley, MD 21031

Subject: Final Report, Phase II Treatability Studies for West Side DNAPL-Contaminated Soils, Atlantic Wood Industries (AWI) Superfund Site, Portsmouth, Virginia (Schnabel Reference 09150034)

Dear Mr. Pellissier:

In accordance with the approved Phase I Remedial Design Work Plan, Version 1 dated 3 April 2009, SCHNABEL ENGINEERING, LLC (Schnabel) was retained by EA Engineering, Science and Technology, Inc. (EA) to perform a two-phase stabilization/solidification (S/S) treatability study on the west side of the subject property for dense non-aqueous phase liquid (DNAPL) contaminated soils at the Atlantic Wood Industries (AWI) Superfund Site. The Request for Proposal 09 502 dated April 8, 2009, was contracted with Schnabel on May 14, 2009. The Phase I report was submitted March 8, 2010. The Phase II scope of work (SOW) was issued on April 1, 2010. This submittal constitutes the Technical Memorandum required by the SOW for the Phase II activities.

EXECUTIVE SUMMARY

The purpose of the Phase II treatability study for the AWI Superfund site was to determine the impact of using organophilic clay (OC) in soil-cement mix designs at doses of 1% to 4% to immobilize pentachlorophenol (PCP) and eight other key SVOCs found in the DNAPL at the AWI site. The bentonite and granular activated carbon (GAC) used as reagents in Phase I were eliminated in Phase II in lieu of OC, with the exception of the Phase II control mix (Mix C2), as reflected in the table below.

EA Engineering, Science and Technology, Inc.

Phase II Treatability Studies

January 19, 2012 Page 2 Schnabel Engineering, LLC Project 09150034 ©2012 All Rights Reserved

Phase II Treatability Study Mix Designs

MIX

ID

Total Reagent Dose

Portland- Slag Dose

Bentonite Powder Dose

GAC

Dose

OC

Dose

Basis

C2 8.55 8 0.5 0.05 0 Control

7 9 8 0 0 1 Low OC dose

8 10 8 0 0 2 Mid OC dose

9 11 8 0 0 3 Mid OC dose

10 12 8 0 0 4 High OC dose

11 16 12 0 0 4 Increased cement, High OC dose Note: All data in wt%

The main geotechnical tests performed on 28-day cure specimens from each mix design included the unconfined compressive strength (UCS) and hydraulic conductivity (k). The environmental testing suite emphasized SPLP testing and leaching behavior as evaluated by the semi-dynamic leaching (SDL) test

(ASTM C1308).

The UCS and k values obtained in Phase II were satisfactory (Table 1) though somewhat lower and higher, respectively, than Phase I, potentially due to cement quality differences between project phases and aging effects associated with storing the laboratory control samples in sealed buckets for approximately one year. S/S treatment promoted diffusion-controlled conditions in all mix designs which leads to significant reductions in mass transport of all constituents of concern (COCs). Under these conditions, the mass transport of a COC is primarily governed by its effective diffusivity (De). The lower the De, the higher the leachability index, the latter of which is benchmarked for acceptable beneficial use or disposal. This then allows for a comparison of mass transport of each COC out of the mix designs based on De through the solid matrix in comparison to its diffusion rate in bulk solution (Daw). The rate of reduction in the mass transport due to the mix design, or the retardation factor [RF], is the ratio of Daw / De. The higher the RF, the more successful the mix was in reducing the migration of the COC. Further confirmation that the mix design is diffusion-controlled is obtained by evaluating the slope of the COC mass release for each COC in each mix. Here, a slope of 0.5±0.1 is evidence of diffusion-controlled conditions. Also, the mix design controls (Mix 2 and Mix C2) from each project phase were determined to be similar from the perspective of the measured SPLP values (Table 3) and SDL test performance (Appendix B).

There were several ways that SDL test results were used to assess the performance of mix designs:

The primary methods were comparisons based on the measured effective diffusivity (De), leachability index (LX) and shape (slope) of the cumulative mass curve of each constituent of concern (COC) from each mix design specimen. All mix designs were determined to be diffusion-controlled based on the collective consideration of the leaching behavior of all SVOCs.

Mix design performance (Section 5) was evaluated by average rank, which ordered the mix designs in terms of attaining the lowest De (highest RF) value per COC.

Composite rank ordered mix design performance by the highest aggregate RF value across all SVOCs.

January 19, 2012 Page 3 Schnabel Engineering, LLC

Incremental improvement in mix design performance over Mix C2 (Phase II control) was evaluated on the basis of the highest RF difference between Mix C2 on the basis of PCP only and all SVOCs.

The LX for the mix designs was also compared to S/S industry guidelines for the beneficial use of treated soils or disposal of S/S treated wastes. From a remedial management approach, in-situ S/S treatment of organics-laden wastes corresponds to a minimally acceptable criterion of LX>8.

Based on the foregoing evaluation criteria and reagent cost, the following table was developed:

Summary Table for Phase II Mix Design Selection

Criteria (Parameter) Report

Reference Compounds

Recommended Mix

Rationale

Average Rank (De or RF)

Table 6 PCP only Mix 9 top performer

Table 6 All SVOCs Mix 9 top performer

Composite Rank (RF) Table 6 All SVOCs Mix 10 top performer

Improvement over Mix C2 (De or RF)

Table 6 PCP only Mix 9 top performer

Table 6 All SVOCs Mix 10 top performer

Leachability Index

(LX)

Table 7 PCP only Mix 7 Min. OC dose for LX>8

Table 7 All SVOCs Mix 7 Min. OC dose for LX>8

Cost ($/CY) see Note 1 All SVOCs Mix 7 Min. OC dose for LX>8

Note 1: Each 1% OC dose equals $34 to $42 per cubic yard (CY) installed (reagent cost only).

Mixes 9 and 10 (3% and 4% OC) are the top performers in terms of retarding the migration of PCP and all SVOCs, respectively. However, this comes at considerable expense as the reagent cost alone for each 1% addition of OC is approximately $34 to $42/CY installed. As in-situ S/S treatment appears to be consistent with the criteria LX>8, Schnabel believes Mix 7 is the minimally acceptable choice if the entire S/S mass is to have a uniform mix design composition. The composite RF of Mix 7 is also approximately 4 times greater than Mix C2 (Table 6), and the benefits of increased OC addition diminish significantly thereafter except for PCP immobilization, as shown in Figure 2.

The Record of Decision (ROD) recommended remedy for the DNAPL-contaminated soil area was excavation and reconsolidation on the eastern side of the AWI site; however, the ROD also allowed for in-situ stabilization/solidification if EPA determines it is a more appropriate method to minimize the migration of DNAPL contamination. The decision was made to utilize in-situ stabilization/solidification in lieu of excavation. Consequently, the cleanup goals for this alternative technology is the ROD criterion for the DNAPL-contaminated soil modified as well to be the groundwater MCLs from Table 7 of the ROD, which includes PCP at 0.001 mg/L.

While significant PCP concentration reductions were obtained through the addition of OC (Figure 1), no SPLP-PCP result for any of the Phase II mix designs was able to satisfy the ROD criterion for PCP in groundwater (0.001 mg/L). However, as indicated in Section 5.4, the PCP concentration data for the SDL tests has the potential to be incorporated into a regional groundwater model that may ultimately demonstrate that the ROD-specified PCP concentrations are satisfied at the points of compliance (POCs)

January 19, 2012 Page 4 Schnabel Engineering, LLC established for the AWI site. So, while Mix 9 produced the lowest PCP concentrations emanating from the mix designs (e.g., near the edge of the S/S treated mass) (Figure 4), the edge of a treated S/S mass, modeling may show that lower OC containing mixes (Mix 7 or 8) may perform satisfactorily when PCP concentrations are evaluated at the AWI site POCs.

Alternatively, to cut down on overall reagent costs, it is also possible to have a graded or perimeter-based mix design approach to the field-deployed S/S mass, as shown in Figure 5. In such a scenario, it may be preferable to have the most contaminated (internal) portions of the site treated with Mix 9, then grade the mix composition to Mix C2 on the perimeter, or vice-versa. Given these general scenarios, other simulations can be modeled to optimize costs for specified levels of environmental performance.

CLOSURE

We appreciate the opportunity to have worked with you on this project and look forward to the possibility of working on future phases of work at this challenging site. If you have any questions, please contact me at (610) 696-6066.

Sincerely, SCHNABEL ENGINEERING, LLC

Michael C. Canino, P.E.

Principal

DGG:RAD:MCC:jlc

FINAL REPORT

Phase II Treatability Studies for West Side DNAPL- Contaminated Soils Atlantic Wood Industries (AWI) Superfund Site Portsmouth, Virginia

Schnabel Reference 09150034 January 19, 2012

January 19, 2012 Page i Schnabel Engineering, LLC

FINAL REPORT

PHASE II TREATABILITY STUDIES FOR WEST SIDE DNAPL-CONTAMINATED SOILS

ATLANTIC WOOD INDUSTRIES (AWI) SUPERFUND SITE

PORTSMOUTH, VIRGINIA

TABLE OF CONTENTS

1.0 PHASE II TREATABILITY STUDY SCOPE

1.1 Phase II Mix Designs

1.2 Treatability Study Approach

2.0 LABORATORY COMPOSITE RESULTS (L-4 TO L-6)

2.1 SPLP Results

2.2 Totals Results

3.0 GEOTECHNICAL TESTING RESULTS

3.1 Unconfined Compressive Strength (UCS) Testing

3.2 Hydraulic Conductivity (k) Testing

3.3 Geotechnical Property Differences Between Project Phases

4.0 ENVIRONMENTAL TESTING RESULTS — MIX DESIGNS

4.1 SPLP Testing Results of Mix Design Control (Mix C2)

4.2 SPLP Testing Results of Mix Design (M-7 to M-11)

5.0 SEMI-DYNAMIC LEACHING TEST RESULTS (ASTM C1308)

5.1 Effective Diffusion Coefficient (De)

5.2 Leachability Index (LX)

5.3 COC Cumulative Mass Release Curves

5.4 Day-averaged COC Concentration Profiles Developed from SDL Test Data

6.0 CONCLUSIONS & RECOMMENDATIONS

7.0 LIMITATIONS

List of Acronyms Tables Figures References

Appendices Appendix A: Geotechnical Testing Results Appendix B: Semi-dynamic Leaching (SDL) Test Procedures and Results Appendix C: Environmental Analytical Results

January 19, 2012 Page ii Schnabel Engineering, LLC

List of Acronyms

ANS American Nuclear Society ASTM American Society for Testing Materials AWI Atlantic Wood Industries BGS Below Ground Surface COC Constituent of Concern CY Cubic Yards De Effective Diffusivity (cm2/s) DI De-ionized DF Dilution Factor (from reagent dosing) DNAPL Dense Non-Aqueous Phase Liquid EA EA Engineering, Science and Technology ERA Environmental Risk Assessment foc Fraction (soil) organic matter content (%) GAC Granular Activated Carbon HDA Historic Disposal Area HHRA Human Health Risk Assessment k Hydraulic Conductivity (cm/s) LX Leachability Index MCL Maximum Contaminant Level MGP Manufactured Gas Plant OC Organophilic Clay PAH Polycyclic Aromatic Hydrocarbons PCP Pentachlorophenol PPL Priority Pollutant List QA/QC Quality Assurance and Quality Control RAO Remedial Action Objective RF Retardation (or reduction) Factor (on De) ROD Record of Decision SDL Semi-dynamic Leaching SPLP Synthetic Precipitation Leaching Procedure S/S Stabilization/Solidification SVOC Semivolatile Organic Compound TCLP Toxicity Characteristic Leaching Procedure UCS Unconfined Compressive Strength USEPA US Environmental Protection Agency VOC Volatile Organic Compound

January 19, 2012 Page 1 Schnabel Engineering, LLC

1.0 PHASE II TREATABILITY STUDY SCOPE

The purpose of the Phase II treatability study for the Atlantic Wood Industries (AWI) Superfund site was to determine the impact of using organophilic clay (OC) in the soil-cement mix design at doses up to 4% to immobilize pentachlorophenol (PCP) and other key semi-volatile organic compounds (SVOCs) in the dense non-aqueous phase liquid (DNAPL) at the site.

1.1 Phase II Mix Designs

The Phase I treatability study report1 submitted by Schnabel recommended that six mix designs be considered for additional study (see summary table below). The motivation for incorporating greater concentrations of granular activated carbon (GAC) and the inclusion of organophilic clay (OC) were based on the desire to immobilize pentachlorophenol (PCP) and eight other key SVOCs associated with the DNAPL at the site. Collectively, these compounds were the main focus of Phase II, and are therefore referred to as the Phase II SVOCs.

Mix Designs Recommended for Phase II Treatability Study

MIX

ID

Total Reagent

Dose

Portland- Slag Dose

Bentonite Powder

Dose

GAC

Dose

OC

Dose

Basis

7 9.5 8 0.5 0 1 Low OC dose

8 9.5 8 0.5 1 0 Mid GAC Dose

9 10.5 8 0.5 0 2 Mid OC dose

10 10.5 8 0.5 1 1 Mid GAC/Low OC dose

11 12.5 8 0.5 0 4 High OC dose

12 16.5 12 0.5 0 4 Increased cement, High OC dose Note: All data in wt%

Subsequent to the Phase I Report, several discussions occurred with EA Engineering, Science and Technology (EA) and the United States Environmental Protection Agency (USEPA) Region III personnel.

The recommended mix designs from Phase I were revised to determine the impact of using OC in the mix designs at doses of 1% to 4%, as shown in the Mix Design Summary, Table 1. In Phase I, bentonite and GAC were used as reagents; these were eliminated in Phase II and replaced with OC, which has been shown to more effectively immobilize PCP in the remediation literature. The mixes from both S/S treatability study phases are shown in Table 1. Mix C2, which is the same as Mix 2 in the Phase I study, was used as the Phase II control.

1.2 Treatability Study Approach

Excess laboratory composite soil sample control materials from the Phase I study were available in sufficient quantities to be used in the Phase II study. This material was the homogenized, mixed soil

1 Final Treatability Study for West Side DNAPL-Contaminated Soils, Atlantic Wood Industries Superfund Site, Portsmouth, Virginia, March 8, 2010.

January 19, 2012 Page 2 Schnabel Engineering, LLC media comprised of the overlying DNAPL-laden fill materials blended with the underlying aquifer sand/silts/clays at a ratio of 1:2.

The excess Phase I laboratory control soil that was freshly mixed and homogenized in 2009 sat homogenized in sealed storage buckets in our laboratory for almost one year under controlled laboratory conditions (20 to 25C). This time duration may have allowed loss (volatilization) of contaminants as well as other phenomena to occur, several of which are specific to highly VOC/SVOC contaminated soils:

changes in wetting characteristics of the soil; changes in surface tension of pore fluids; shrinkage, cracking and the reduced reactivity of clay minerals (clays participate in cement reactions). Thus, it is possible that the Phase II laboratory control soil was somewhat more difficult to stabilize than the Phase I laboratory control soil.

Accordingly, three replicates of the laboratory composite soil were re-tested to establish a new contaminant baseline condition for the Phase II mix designs. After thorough re-mixing, the Phase II laboratory composites (L-4 to L-6) were analyzed for total SVOCs and synthetic precipitation leaching procedure (SPLP) SVOCs. As approved by USEPA, analytical testing in Phase II was limited to totals and SPLP testing for SVOCs of select UCS sample breaks, and dissolved SVOCs from the semi-dynamic leaching (SDL) test apparatus (ASTM C1308). The geotechnical testing suite in Phase II was limited to unconfined compressive strength (UCS; ASTM D1633) and hydraulic conductivity (k; ASTM D5084) testing on 28-day cured soil-cement cylinders.

This report is primarily organized by chronology. For quick reference, the key environmental data (SPLP) are contained in Sections 2.1, 4.1 and 4.2, and the SDL testing results begin in Section 5.0.

2.0 LABORATORY COMPOSITE RESULTS (L-4 to L-6)

In terms of presentation, the primary environmental data summary tables for the laboratory control and the mix design SPLP data (Tables 2 to 4) focus on the Phase II SVOCs. Summaries of the environmental testing suite (totals data) and the laboratory data reports are included in Appendix C.

After the analyte column, the next leftmost columns in Tables 2 to 4 show the ROD groundwater standards2 and the respective aqueous solubility limits (Ciw,sol) of the Phase II SVOCs. The adjacent columns show the reporting limits for each replicate series followed by the measured constituent of concern (COC) values for each replicate. SPLP concentrations that were both above the reporting limit and the ROD standards were yellow-highlighted for ease of reference. Both PCP and BaP were characterized by concentrations or detection limits occurring above their respective ROD standards, whereas the 2,4,5-trichlorophenol concentrations were below the ROD standard or non-detect (and thus were no longer given attention from a reporting perspective for this report).

2.1 SPLP Results

The baseline reference for the effectiveness of the stabilized soil mixes is the environmental data (leach, totals) associated with the Phase II laboratory composites (L-4 to L-6). Table 2 summarizes the SPLP- SVOC leaching results for the Phase I and II laboratory composites and illustrates that the leachable

2 Note that the ROD does not require that the SPLP test results meet the groundwater cleanup criteria.

January 19, 2012 Page 3 Schnabel Engineering, LLC concentrations of the SVOCs between the two project phases is comparable, but slightly higher PCP concentrations were measured in the Phase II (L-4 to L-6) samples.

All PCP concentrations were close to the aqueous solubility limit (14 mg/L). Because concentrations measured at or near the solubility limit can be difficult to pinpoint, the PCP concentrations shown in Table 2 are conceptually the same even though the magnitudes are different. The remaining Phase II SVOCs were characterized by SPLP concentrations that could support the existence of DNAPL.

2.2 Totals Results

The total SVOC concentrations for the L-series controls are shown in Appendix C, Table C.1-1. There was significant variation in the measured concentrations of all SVOCs from the L-5 and L-6 samples compared to both the L-4 samples and the Phase I (L-1 to L-3) samples. Accordingly, duplicates (D1, D2) were tested for L-5 and L-6, and the re-tested values for total SVOCs were found to have greater consistency with the other laboratory composite soil replicates despite the slightly extended hold time for the samples.

The SPLP leaching performance of the duplicates was not re-tested. Additional discussion on the L-series totals data is located in Appendix B, Section 1.2.1.

3.0 GEOTECHNICAL TESTING RESULTS

The geotechnical results for the mix designs are contained in Appendix A and include UCS testing results (ASTM D1633) and hydraulic conductivity testing (ASTM D5084) for 28-day cured soil-cement specimens. The reactivity of the Portland-Slag (PS) cement used in the mix designs was also assessed versus historic data on file with Schnabel.

Table 1 provides a summary of the mix design reagents, characteristics and geotechnical testing data for the two phases of the treatability study. The mix identification number appears in the first column. The next four columns (dry mix characteristics) show the total reagent dose by dry weight to the wet weight of the composite soil (~125 lb/ft3) and its breakdown by reagent (Portland/slag, bentonite, GAC, OC). For Mix C2, the reagents were administered as a bentonite-based slurry to the wet, composite soil, so the next three columns (slurry characteristics) are provided for convenience as operational parameters for potential full scale implementation. In the case of the OC-containing mixes, the cement was first slurried and the OC was then added to the slurry prior to mixing with the AWI soils. The next four columns (mix properties) show the average UCS values measured on 7 and 28-day cured samples, followed by the corresponding hydraulic conductivities for 28-day cured samples and whether or not the mix design was subjected to semi-dynamic leach testing (ASTM C1308). Results of the UCS and k tests for the Phase II mixes are discussed in Sections 3.1 and 3.2, respectively. A brief comparison of the geotechnical results between project phases is provided in Section 3.3.

3.1 Unconfined Compressive Strength (UCS) Testing

Table 1 summarizes the 28-day UCS testing results based on the average of triplicate specimens. The UCS test results for the Phase II mix design samples increased with increasing OC and cement dose from 66 to 185.5 lb/in2. For typical geotechnical and environmental S/S applications, a minimum UCS value of 20 to 50 psi is typically assumed to be adequate, depending on site and reuse considerations (Bates, 2010).

January 19, 2012 Page 4 Schnabel Engineering, LLC

3.2 Hydraulic Conductivity (k) Testing

Hydraulic conductivity tests were completed on all Phase II mix designs. Once saturated (2 to 4 weeks), the samples were consolidated to a cell pressure commensurate with a position approximately 10 ft below ground surface (bgs), with the water table assumed to be at 5 ft bgs. For typical S/S applications, a maximum hydraulic conductivity (k) of the mix is typically specified to be on the order of 10-6 or 10-7 cm/s.

Table 1 shows the measured k-values for the mix designs to be within the 10-5 to 10-7 cm/s range. These hydraulic conductivity values essentially render the soil-cement as a diffusion-controlled system. There was not a strong correlation between the OC content and the measured hydraulic conductivity, although an order of magnitude difference was observed when the OC dose was increased from 3% to 4%.

3.3 Geotechnical Property Differences Between Project Phases

Comparing the Mix 23 and Mix C2 strength (UCS) and conductivity (k) results from Table 1, Mix C2 is characterized by lower strength and greater conductivity values, which may suggest a loss in mix design performance. Part of the reason for this is that there were measured differences in cement quality between the project phases. In essence, the cement used in Phase II was approximately only half as reactive as the same Portland-slag cement used in Phase I, see Appendix A. However, it must be stressed that UCS and k values are only surrogates for promoting the durability and low leaching potential of SVOCs from soil-cement. Conversely, the semi-dynamic leaching (SDL) test provides a direct measurement of COC leaching. As shown in Appendix B, the PCP leaching behavior of the mix design controls between project phases was essentially identical (Figure B.1-1).

4.0 ENVIRONMENTAL TESTING RESULTS — MIX DESIGNS

After the 28-day replicates were broken, the UCS sample breaks for the weaker two of the three replicates were composited and subjected to environmental analysis (SPLP and totals). As additional samples of all mix designs were retained until all of the hydraulic conductivity tests commenced, it was also possible to submit 50-day cured soil-cement samples for environmental testing. The 50-day cured mix design samples were denoted by the suffix “-2.” The environmental testing results of the mix designs are presented in Tables 3 and 4 (SPLP) and Table C.1-2 and C.1-3 (totals), and are discussed in Sections 4.1 and 4.2 below.

4.1 SPLP Testing Results of Mix Design Control (Mix C2)

The comparison of the mix design controls (Mix C2 from Phase II versus Mix 2 from Phase I) in Table 3 shows that it was possible in Phase II to lower the SPLP detection limits for numerous SVOCs. For instance, the SPLP detection limit for BaP was lowered to 0.0004 mg/L (from 0.001 mg/L in Phase I) versus the ROD criteria of 0.0002 mg/L, mitigating concern regarding ROD compliance associated with BaP. PCP thus persisted as the only SVOC presenting compliance challenges with respect to the ROD standards.

3 Reference to Mix 2 denotes the blend shown in Table 1 without specific regard to the individual replicates used for geotechnical testing (A to G) or environmental analysis (D, E). The term Mix 2 is often used in the collective sense even though specific data is shown.

January 19, 2012 Page 5 Schnabel Engineering, LLC

Comparison of the other SVOC concentrations in Table 3 indicates that the measured SPLP concentrations in Mix C2 are slightly greater than Mix 2. As the total mass of SVOCs cannot increase between project phases (constant or decrease only), all SPLP-SVOC concentrations can only increase simultaneously if the Phase II laboratory controls were more thoroughly and evenly contaminated (aged) than the Phase I controls, or if sample aging resulted in the SVOCs being more leachable in Phase II.

4.2 SPLP Testing Results of Mix Designs (M-7 to M-11)

Table 4 presents the SPLP-SVOC concentrations for the mix designs for both the 28-day and 50-day (with suffix "-2") cured soil-cement samples. There is strong agreement between each of the mix design samples except for the naphthalene concentrations in Mix C2 and the PCP concentrations in Mix 11.

No Phase II mix design satisfied the ROD criteria for PCP in groundwater of 0.001 mg/L. This is primarily due to the necessary reductions required (20 mg/L reduced to 0.001 mg/L equals factor of 20,000, or four orders of magnitude). As shown in Figure 1, each 1% OC dose corresponded to approximately a 4.0 mg/L (20%) reduction in the SPLP-PCP concentration, totaling approximately an order of magnitude reduction in SPLP-PCP concentrations at a 4% OC dose. Previous treatability studies with similar PCP contamination levels achieved concentrations on the order of 0.12 to 0.15 mg/L at high doses of OC and/or GAC (>2 to 5% individually or combined), and reagent (cement/fly ash) doses greater than 20% (Bates et al. 2000, 2002).

5.0 SEMI-DYNAMIC LEACHING (SDL) TEST RESULTS (ASTM C1308)

The SDL test yields an estimate of the effective diffusivity (De) of a COC from a soil-cement cylinder based on the radial diffusion of COCs from a mix design specimen (soil-cement cylinder) suspended in a water bath. SDL tests were used to evaluate the leaching potential of PCP from mix designs in Phase I.

In Phase II, the list of key SVOCs evaluated for their leaching potential from the Phase II mix designs was increased to include acenaphthene, anthracene, benzo(a)anthracene, fluoranthene, flourene, naphthalene, phenanthrene, and pyrene.

There are several ways that SDL test results can be used to assess the performance of mix designs: the primary methods being comparisons based on the measured De, leachability index (LX), and shape (slope) of the cumulative mass curve of each COC from each mix design specimen. Table 5 presents a summary of the Mix C2 SDL testing data for PCP and inputs that enable the estimation of the De, LX, and slope of the PCP cumulative mass curve. Fifty-four such SDL leaching tables were developed based on the number of mix designs (6) and SVOCs (9). The analysis procedure, data reduction process and raw data for the SDL tests and the SDL leaching tables are presented in Appendix B. The following sections address how these data sets and their comparisons can be used for decision-making and mix design selection.

5.1 Effective Diffusion Coefficients (De)

A main goal of S/S is to produce an intact soil-cement mass characterized by a hydraulic conductivity of 10-5 cm/s or less (~10-7 cm/s desired) which essentially renders the soil-cement as a diffusion-controlled system. Under these conditions, the retardation factor on contaminant transport effectively reduces to a comparison of the effective diffusivity of the COC versus its bulk diffusion coefficient (DAW) in water.

January 19, 2012 Page 6 Schnabel Engineering, LLC

Mathematically, the retardation factor (RF) on diffusive transport (leaching) for each COC from a soil-cement cylinder is expressed as DAW/De, and RFs much greater than 1 are desired. This makes two comparisons or ranking methods possible for mix design selection purposes:

The first ranking method orders the mix designs according to lowest De (and highest RF) value per COC. From this, average and composite ranks can be developed based on the individual rank per COC and cumulative (summed) RF values for all COCs, respectively. In essence, the average rank reflects how many times a mix design ranks first (e.g., attains the lowest De), which in some cases could occur by a very small margin (i.e., photo finish). Composite rank reflects how lopsided the aggregate RF total was due to strong incremental differences in mix design performance (a mix could rarely rank first, but wins by a large margin on one or more COCs). An average rank equal to 1 indicates that the mix design had the lowest De (highest RF) for all nine SVOCs tested, and was therefore the top performer on an individual COC basis. A composite rank equal to 1 indicates that the mix design had the highest overall RF for the entire group of COCs, and was therefore the top performer. In this way, the first ranking method facilitates the selection of a mix design based on performance with respect to a single COC or the entire group of COCs.

The second ranking method assesses the incremental performance of each mix over Mix C2, the

Phase II control. Mathematically, relative improvement over Mix C2 is expressed as RFmix / RFC2.

Likewise, this ranking method can be applied to an individual COC such as PCP, or all COCs (composite improvement).

Table 6 summarizes the estimated De values for each mix organized by SVOC, where the raw data was taken from Appendix B.2 with each data presentation resembling the format and content shown in Table

5. The color codings used in Table 6 denote the best (yellow) and worst (orange) rank in terms of mix design performance in terms of both De and its corresponding RF value.

Based on average rank, Mix 9 was the top performer, as illustrated by its average rank of 2.00, as shown in the center subtable located at the bottom of Table 6. Mix 9 was associated with the lowest De (highest RF) values for 4 of 9 SVOCs, including PCP.

Based on composite rank, Mix 10 was the top performer, as illustrated by its composite rank of 1

(RF~4.37x106) as shown in the center subtable located at the bottom of Table 6. Mix 10 was associated with the lowest De (highest RF) values for 2 of 9 SVOCs. Mix 10 provided the best composite performance, maintaining a slight advantage over Mix 9 principally due to lower RF values associated with pyrene.

Figure 2 shows the RF values (from Table 6) plotted as a function of OC dose for each COC, enabling a clearer visual understanding of the individual trends in incremental performance of the various mix designs over Mix C2 (0% OC). The key is ordered by the aqueous solubility limit of each COC.

Based on the best incremental performance over Mix C2 for PCP only, Figure 2 indicates that Mix 9 was the top performer. Mix 9 had the lowest De (highest RF) for PCP, and performed approximately 148 times better than Mix C2, as shown by the RFmix / RFC2 comparison in the lower left subtable located at the bottom of Table 6.

January 19, 2012 Page 7 Schnabel Engineering, LLC

Based on the best incremental performance over Mix C2 for all Phase II SVOCs, Mix 10 was the top performer. Mix 10 performed approximately 3.91 times better than Mix C2, as shown by the RFmix / RFC2 comparison in the center subtable located at the bottom of Table 6. As indicated above, Mix 9 was a close second.

Incremental performance over Mix C2 is related to COC solubility. Since Mix C2 contains the least organic carbon content, the steepest RF curves as a function of OC dose are associated with the most soluble COCs (PCP, napthalene and acenapthene).

5.2 Leachability Index (LX)

The LX4 is mathematically equivalent to –log(De), and the S/S industry uses the attained LX values to assess the potential for the subsequent beneficial use or ultimate disposal of the stabilized media. The classifications offered by Environment Canada (1991) recommend guidelines on the ultimate disposition of S/S treated wastes. In terms of LX, several minimum thresholds have been established for S/S performance (Stegemann and Côté, 1996):

For an LX>9, the S/S treatment is considered effective and “controlled utilization” (e.g., beneficial reuse) can be considered in such applications as quarry rehabilitation, lagoon closure, road bases, etc. This benchmark of performance is equated with less than 10% mass loss of a COC in 100 years.

For an LX>8, placement of the S/S treated waste can occur in a segregated landfill (monofill, with or without additional engineering controls) or sanitary landfill. These criteria and applications appear to be consistent with in-situ S/S applications. Benchmarks of performance have not been developed for LX>8 because a disposal scenario of some kind is the main outcome.

Guidelines for LX values less than 8 are absent from Stegemann and Côté (1996), the primary authors of the Environment Canada report (1991).

Regardless of the LX criteria, the "controlled utilization" of S/S treated wastes entirely disallows organics-contaminated media, as summarized by Stegemann and Côté (1996). Thus, due to the presence of organic DNAPL at the AWI site, this means that LX>8 effectively serves as the minimum criteria if LX is used as a decision making criterion for mix design selection for the AWI Superfund site.

LX>9 is usually easily accomplished for moderate to high levels of metals contamination, with or without the presence of organics. However, for organics, the ability to achieve elevated LX values correlates directly with the increasingly hydrophobic nature of the organic constituent. As such, the moderately soluble SVOCs, such as PCP and naphthalene, pose challenges for S/S mixes, depending on the OC and/or GAC content.

In a manner similar to Table 6, Table 7 summarizes the LX data for the mix designs as a function of COC.

The lowest (orange) and highest (yellow) LX values are indicated for each COC. Table 7 shows that for

4 However, because of how the incremental time data are averaged, the minimum De value does not always correspond to the maximum LX value.

January 19, 2012 Page 8 Schnabel Engineering, LLC several of the less soluble SVOCs, LX values can be in the range of 10 to 12. For PCP, LX>9 was achievable for several mixes, which would suggest the potential for beneficial use. For comparison, Vipulanandan (1995) evaluated the leaching of phenol (aqueous solubility 93,000 mg/L) from pure kaolinite (8.27<LX<8.71) and bentonite (8.79<LX<9.03) containing between 400 and 2,000 mg/kg phenol that were stabilized with 20% Portland cement and cured for 28 days.

In summary:

To achieve an LX>8, a minimum dose of 1% OC is required (Mix 7). Thus, all OC containing mixes (7 to 11) satisfy this criterion. However, Mix 7 has an LX value of 8 for PCP, which leaves very little margin for error on the OC content in the mix design during field implementation.

5.3 COC Cumulative Mass Release Curves

By evaluating the slope of the cumulative COC mass release curve (from the soil-cement cylinder), insight is gained on the dominant mass transfer phenomena occurring at the surface of the specimen e.g., whether the leaching of the COCs is surface-controlled or diffusion-controlled5, the latter being more preferable for long-term performance). The slope can vary significantly between COCs.

Figure 3 presents the cumulative mass release curve for PCP for all mix designs, illustrating how the slope and regression parameters were fitted to the data (with the exception of the initial 2-hour data) due to test start-up effects related to Mix C2. As would be expected from a total PCP mass perspective and the absence of OC, the highest release rates occur for Mix C2. Mathematically, a slope value of approximately 0.5±0.1 for the PCP cumulative release curve indicates that the release is diffusion-controlled. With respect to PCP, all mixes were diffusion-controlled at a minimum.

A summary of the slope and the regression parameters (R2) for the cumulative mass release curves for the Phase II SVOCs is presented in Table 8. Since the S/S treated matrix conditions are identical for all COCs in the same mix design, all should experience the same transport phenomena. If one COC is diffusion-controlled, they all should be. The key is evaluating the overall mix behavior based on the more soluble COCs which are characterized by detectable concentrations. As such, Mixes 7 to 11 all appear to be diffusion-controlled for the full range of COCs (Table 8) and their respective solubilities (Figure 2).

For Mix C2, it appears that the more soluble SVOCs are diffusion-controlled; whereas, the least soluble SVOCs are being transported at a slower rate. As shown in Figure 4, this is an artifact of the high initial release rate of the COC (in this case, benzo(a)anthracene) in Mix C2 relative to other mixes. Mix C2 was the only mix with detectable concentrations of benzo(a)anthracene. The plots for all other mixes are based on their detection limits, so the mere summation of concentrations based on the respective detection limit values produced a diffusion-controlled like result. Thus, in reality, the estimated De values for these compounds can be thought of as upper thresholds on diffusion.

5 These refer to whether or not the creation of films, precipitates, solid sloughing/dissolution, sorption and/or other phenomena occurring at the solid-water interface will impact the transport mechanism. In the event they do not, the system will be diffusion controlled.

January 19, 2012 Page 9 Schnabel Engineering, LLC

5.4 Day-averaged COC Concentration Profiles Developed from SDL Test Data

From a site management perspective, concern obviously arises about the actual SVOC concentrations emanating from the S/S treated mass, and their relationship to points of compliance (POCs) and impacts to local groundwater. The individual concentrations measured at specified time intervals during the SDL test are point values that can be averaged over their time intervals (day-averaged) to develop concentration profiles with time. As described in Appendix B, the developed "day-averaged" PCP concentrations can be taken to be loosely representative of conditions occurring near the exterior vertical surface of the stabilized media (i.e., near edge of treated mass). Moreover, if the COC concentrations level off, they can be considered as quasi-steady state concentrations, allowing their potential incorporation into a site groundwater model to evaluate impacts to local POCs.

In Figure B.1-1 and the Appendix B narrative (Section 1.4), the day-averaged PCP concentration profiles for Mix 2 and C2 were shown to be similar, implying continuity between the mix design controls in both project phases.

Figure 5 shows the day-averaged PCP concentrations for all Phase II mix designs, which decrease with time to a quasi-steady state condition in about 10 days. The PCP concentration measurement from the first semi-dynamic test interval (2-hr) reflects the high initial mass release of COCs from the boundary layer that formed between the soil-cement cylinder and mold. This was a common occurrence for Mix C2 for all Phase II SVOCs. The measured 2-hr value for Mix C2 (20.1 mg/L) is very comparable to the SPLP value shown in Figure 1 and the Phase II controls (14 to 22 mg/L; see Table 2). Figure 5 shows that the PCP concentration was reduced to approximately 0.30 mg/L for Mix C2 at approximately 68 days, or a concentration reduction factor of approximately 67. To illustrate the important role of OC versus the control (Mix C2), Mix 9 achieved a much lower PCP concentration of 0.02 mg/L, or a PCP reduction on the order of approximately 1,000 compared to the initial conditions of Mix C2 (2-hr and SPLP values).

Effectively, the PCP concentrations occurring near the Mix 9 soil-cement cylinder are 10 times less than prior treatability study goals for PCP (0.2 mg/L) and achieved results (~0.15 mg/L) shown in the literature (Bates et al. 2000, 2002). While significant PCP concentration reductions were obtained for all of the Phase II mix designs, no SDL test results were able to satisfy the ROD criterion for PCP (0.001 mg/L).

However, the ROD criterion for PCP needs to be satisfied at the points of compliance (POCs) established for the site, not necessarily at the edge of the S/S mass. Accordingly, if the PCP concentration profile information (Figure 5) can be adequately incorporated into a regional groundwater model, PCP concentrations occurring at distant POCs (not the property boundary of the PPSD POC) may be simulated.

Lastly, as a group, the day-averaged PCP concentration profiles for Mixes C2, 7 and 8 have essentially leveled off by 60 days; whereas Mixes 9 to 11 still seem to be both dropping off and clustered as a group at lower concentrations (almost 10x lower). This suggests that due to the high levels of SVOC contamination, a threshold in performance may have been reached for mixes with OC doses less than 3%.

January 19, 2012 Page 10 Schnabel Engineering, LLC

6.0 CONCLUSIONS & RECOMMENDATIONS

The purpose of the Phase II treatability study for the AWI Superfund site was to determine the impact of using OC in the soil-cement mix designs at doses of 1% to 4% to immobilize PCP and eight other key SVOCs. The bentonite and GAC used as reagents in Phase I were eliminated in Phase II and replaced by OC, with exception of the Phase II control mix (Mix C2).

The UCS and k values obtained in Phase II were satisfactory (Table 1) though somewhat lower and higher, respectively, than Phase I, potentially due to cement quality differences between project phases and aging effects associated with storing the laboratory control samples in sealed buckets for approximately one year. However, the mix design controls (Mix 2 and Mix C2) were determined to be similar from the perspective of the measured SPLP values (Table 3) and SDL test performance (Appendix B).

There are several ways that SDL test results can be used to assess the performance of mix designs: the primary methods being comparisons based on the measured De, LX and shape (slope) of the cumulative mass curve of each COC from each mix design specimen. Mix design performance (Section 5) was broken out into average rank, which ordered the mix designs in terms of attaining the lowest De (highest RF) value per COC. Composite rank ordered mix design performance by the highest aggregate RF value across all SVOCs. Incremental improvement in mix design performance over Mix C2 (Phase II control) was evaluated on the basis of the highest RF difference between Mix C2 on the basis of PCP only and all SVOCs. The LX for the mix designs was also compared to S/S industry guidelines for the beneficial use or disposal of S/S treated wastes. From a remedial management approach, in-situ S/S treatment of organics-laden wastes corresponds to a minimally acceptable criterion of LX>8. All mix designs were determined to be diffusion-controlled based on the collective consideration of the leaching behavior of all SVOCs.

Based on the foregoing evaluation criteria, considerations and reagent cost, the following table summarizes the Phase II mix design results.

Summary Table for Phase II Mix Design Selection

Criteria (Parameter) Report

Reference Compounds

Recommended Mix

Rationale

Average Rank (De or RF)

Table 6 PCP only Mix 9 top performer

Table 6 All SVOCs Mix 9 top performer

Composite Rank (RF) Table 6 All SVOCs Mix 10 top performer

Improvement over Mix C2 (De or RF)

Table 6 PCP only Mix 9 top performer

Table 6 All SVOCs Mix 10 top performer

Leachability Index

(LX)

Table 7 PCP only Mix 7 Min. OC dose for LX>8

Table 7 All SVOCs Mix 7 Min. OC dose for LX>8

Cost ($/CY) see Note 1 All SVOCs Mix 7 Min. OC dose for LX>8

Note 1: Each 1% OC dose equals $34 to $42 per cubic yard (CY) installed (reagent cost only).

January 19, 2012 Page 11 Schnabel Engineering, LLC

Mixes 9 and 10 (3% and 4% OC, respectively) were the top performers in terms of retarding the migration of PCP and all SVOCs, respectively. However, this comes at considerable expense as the reagent cost alone for each 1% addition of OC is approximately $34 to $42/CY installed. As in-situ S/S treatment appears to be consistent with the criteria LX>8, Schnabel believes Mix 7 is the minimally-acceptable choice if the entire S/S mass is to have a uniform mix design composition. The composite RF of Mix 7 is also approximately 3 times greater than Mix C2 (Table 6), and the benefits of increased OC addition diminish significantly thereafter except for PCP immobilization, as shown in Figure 2.

While significant PCP concentration reductions were obtained through addition of OC (Figure 1), no SPLP-PCP results for the Phase II mix designs were able to satisfy the ROD criterion for PCP (0.001 mg/L). However, as indicated in Section 5.4, the PCP concentration data from the SDL tests has the potential to be incorporated into a regional groundwater model that may ultimately demonstrate that the ROD-specified PCP concentrations are satisfied at the POCs established for the AWI site. So, while Mix 9 produced the lowest day-averaged PCP concentrations (Figure 4), and these may be loosely taken to apply within a column diameter from the edge of a treated S/S mass at the site, modeling may show that lower OC containing mixes (Mix 7 or 8) may be sufficient when PCP concentrations are evaluated at the AWI site POCs.

Alternatively, to cut down on overall reagent costs, it is also possible to have a graded or perimeter-based mix design approach to the field-deployed S/S mass, as shown in Figure 6. In such a scenario, it may be preferable to have the most contaminated (internal) portions of the site treated with Mix 9 with the mix composition graded to Mix C2 on the perimeter, or vice-versa. Given these general scenarios, other simulations can be modeled to optimize costs for specified levels of environmental performance.

7.0 LIMITATIONS

This report has been prepared in accordance with Schnabel Contract No. 6866 with EA Engineering, Science and Technology dated May 14, 2009, and subsequent contract modifications. The limitations provided in the Phase I report likewise apply to the Phase II scope of work.

January 19, 2012 Schnabel Engineering, LLC

TABLES

Table 2 AWI Superfund Site, Portsmouth, VA

Stabilization/Solidification (S/S) Treatability Study Groundwater Contaminant MCLs/MCLGs vs. SPLP Data

Phase I and II Lab Controls Leach Comparison

L-1 L-2 L-3 L-4 L-5 L-6

(mg/l) (mg/l) (mg/l) (mg/l) (mg/l) (mg/l) (mg/l) (mg/l) (mg/l)

Pentachlorophenol 0.001 0.060 14.00 5.6 7.4 11.0 22.0 14.0 18.0 2,4,5-Trichlorophenol 0.05 0.200 - <0.200 <0.200 <0.200 0.013 <0.010 <0.050 Benzo(a)pyrene 0.0002 0.004 - 0.0051 <0.004 <0.004 0.0051 0.0042 0.005

Acenaphthene 0.040 3.42 0.170 0.320 0.380 0.310 0.220 0.290 Anthracene 0.040 0.05 0.049 0.060 0.069 0.060 0.049 0.059 Benzo(a)anthracene 0.004 0.01 0.013 0.0085 0.0099 0.011 0.0095 0.012 Flouranthene 0.040 0.21 0.066 0.066 0.064 0.081 0.068 0.082 Flourene 0.040 1.69 0.100 0.190 0.220 0.200 0.140 0.190 Naphthalene 0.040 31.70 1.600 2.700 4.600 2.800 1.500 2.700 Phenanthrene 0.040 1.00 0.190 0.270 0.290 0.290 0.210 0.260 Pyrene 0.040 0.13 0.044 0.040 0.041 0.051 0.042 0.052 SPLP - pH 9.70 9.86 9.68 9.11 9.87 9.63

Comments

RL1: SVOCs;

L1:

Anthracene;

RL7:PCP

RL1: SVOCs;

L1:

Anthracene;

RL7:PCP

RL1: SVOCs;

L1:

Anthracene;

RL7:

Naphthalene, PCP

RL7:

Acenaphthene, Flourene, Phenanthrene, Naphthalene, PCP

RL7:

Acenaphthene, Phenanthrene, Naphthalene, PCP

RL1: SVOCs;

RL7:

Naphthalene, PCP

Notes:

1) Ciw,sol = Aqueous solubility limit [USEPA 600/8-90/003 (1990). "Basics of pump-and-treat ground-water remediation technology," USEPA, Ada, OK, March (NTIS PB90-274549)]

2) Laboratory controls (L1 to L3) are 1/3 fill homogenized with 2/3 sandy clay aquifer media

Ciw,sol Analyte

ROD

Standard

SPLP Results

Not Applicable

Reporting Limit

L-Series Phase I Phase II

ROD-listed Semivolatile Organic Compounds (SVOCs)

Other SVOCs

2) Laboratory controls (L1 to L3) are 1/3 fill homogenized with 2/3 sandy clay aquifer media.

3) Yellow indicates ROD exceedence, including reported non-detects above ROD standard.

Comments:

L1: Laboratory Control Sample and/or Laboratory Control Sample Duplicate recovery was acceptable limits.

RL1: Reporting limit raised due to sample matrix effects.

RL7: Sample required dilution due to high concentrations of target analyte.

EA Project Number 1453011, Phase A Schnabel No. 09150034.00, Task 76, Phase II Mix Design

Table 3

Stabilization/Solidification (S/S) Treatability Study Groundwater Contaminant MCLs/MCLGs vs. SPLP Data

Mix Design Controls Comparison

Phase II

Mix 2-D Mix 2-E Mix C2

(mg/l) (mg/l) (mg/l) (mg/l) (mg/l) (mg/l)

Pentachlorophenol 0.001 0.150 0.300 14.0 18.0 21.0

2,4,5-Trichlorophenol 0.05 0.050 0.010 <0.050 <0.050 0.030

Benzo(a)pyrene 0.0002 0.001 0.0004 <0.001 <0.001 <0.0004

Acenaphthene 0.010 0.020 0.170 0.190 0.270

Anthracene 0.010 0.002 0.033 0.043 0.061

Benzo(a)anthracene 0.001 0.0004 0.0015 <0.001 0.0013

Flouranthene 0.010 0.002 0.015 0.023 0.034

Flourene 0.010 0.002 0.090 0.100 0.150

Naphthalene 1.000 0.020 2.600 3.100 1.400

Phenanthrene 0.010 0.002 0.095 0.120 0.140

Pyrene 0.010 0.002 <0.010 0.013 0.019 pH NA NA 11.69

Comments RL1: SVOCs;

RL7: Napthalene, PCP

RL7:PCP

RL7: PCP,

Acenaphthene, Naphthalene

Notes:

1) Blend C2 is a composite sample for the two lowest-valued UCS breaks at 28-days.

2) Yellow highlight indicates ROD exceedence, including reported non-detects above ROD standard.

Comments:

RL1: Reporting limit raised due to sample matrix affects.

RL7: Sample required dilution due to high concentrations of target analyte.

Analyte

ROD

Standard

SPLP Results

Not Applicable

Reporting Limit Mix C2

Phase I

ROD-listed Semivolatile Organic Compounds (SVOCs)

Other SVOCs

Reporting Limit Mix 2-D, -E

Table 4

Stabilization/Solidification (S/S) Treatability Study Groundwater Contaminant MCLs/MCLGs vs. SPLP Data

Phase II Mix Design Leaching Results

Reporting Limit

Mix C2 Mix C2 Mix C2-2 Mix 7 Mix 7-2 Mix 8 Mix…