January - February 2001

A soil gas survey was performed at two adjacent AOCs located within the southwest quadrant of CSSA.  AOC 57 is approximately 65 acres in size and AOC 65 is approximately 5 acres in size.  AOC 65 consists of a potential VOC source area inside Building 90, and the study area extends approximately 100 feet from the building in all directions, whereas AOC 57 is long and narrow and extends over 3,600 feet north to south along the eastern edge of AOC 65.  AOC 57 is an area where temporary buildings reportedly used for gun cleaning and maintenance operations were located.  The initial survey grid covered approximately 70 acres.  The primary objective of the soil gas survey was to determine whether releases of VOCs had occurred into the underlying soil from these two AOCs.  Additional background information on the history and ongoing investigations for AOC 57 and AOC 65 are located in Volume 3-2, Areas of Concern, behind the AOC 57 and AOC 65 tabs.

The soil gas survey was established as one contiguous grid of sampling points that covered the entire survey area represented by AOC 65 and AOC 57.  The survey plan was established on a 100-foot grid throughout AOC 57, was tightened to a 50-foot grid in the vicinity of Building 90 (AOC 65) and was tightened to a 25-foot and adjacent to and inside Building 90.  The combined sampling grids were expected to yield approximately 400 individual sample locations during the first phase of sample collection.  An additional 200 samples were initially planned to tighten the grid in areas where higher levels of contamination was indicated.  Few portions of the initial grid (450 points) exhibited detectable levels of contaminants such that the initial grid provided adequate coverage, with most contamination found inside and near Building 90.  The remaining 150 soil gas points were utilized at other SWMUs or AOCs as directed by CSSA to assess possible releases of VOCs from these areas, including the wastewater treatment plant, SWMU B-3, SWMU B-4, AOC 55, and AOC 63.  Additional background information on the history and ongoing investigations for SWMU B-3, SWMU B-4, AOC 55, and AOC 63 are located in Volume 3-1, Solid Waste Management Units, behind the SWMU B-3 and SWMU B-4 tabs and in Volume 3-2, Areas of Concern, behind the AOC 55 and AOC 63 tabs.

Past soil gas surveys demonstrated that the soil/bedrock interface underlying the AOC 57 and AOC 65 sites is relatively shallow, and that the probes typically could not be driven consistently deeper than 4 feet below grade.  A geoprobe rig was used to drive (hammer) the probes into the ground, and to retrieve the probes after the soil gas sample had been extracted.  A pneumatic hammer and jack were used to drive and retrieve rods from sample locations inside Building 90.  Samples were obtained in Tedlar^� bags under vacuum as described in the SAP Addenda included in Volume 1-4, Sampling and Analysis Plan, DO5084 Addendum, and analyzed by an on-site mobile lab equipped with a gas chromatograph.  The laboratory tested each sample for benzene, toluene, ethylbenzene, xylenes (BTEX), vinyl chloride, tetrachloroethene (PCE), trichloroethene (TCE), trans-1,2-dichloroethene (DCE), and cis-1,2-DCE. In accordance with the SAP addenda, each sample was also field screened for oxygen and explosive gases (lower explosive limit) using a direct read instrument, an HMX 271 explosimeter.

Each of the grid points was surveyed with a Trimble asset-grade Global Positioning System (GPS) receiver with one-meter accuracy.  A Parsons and AFCEE chemist performed one on-site laboratory audit during the initial week of testing to evaluate the performance of the mobile laboratory.  In addition, electronic data was managed on a daily basis during the active testing periods.

The soil gas survey grids were established at each site prior to beginning each site specific survey.  The soil gas sampling grids were set up based on the shape and size of the AOCs or SWMUs.  These grids were established at intervals of 50 to 100 feet, and the grid spacing was modified based on the results from previous days� samples in the area.  Soil gas survey grids were established for all sites mentioned above for soil gas testing.  Two larger grid areas were further divided into separate but contiguous grids.  Three grids comprise AOC 57 and two comprise the AOC 65 survey area.  The locations of the soil gas survey grids are presented in figures included in Volume 3-1.1 and Volume 3-2, behind the site specific tabs.

Soil gas sampling analytical methodology, and quality control procedures are discussed in this Introduction and Methodology.  The soil gas survey results for each of the surveyed locations are presented Volume 3-1.1 and Volume 3-2, behind the site tabs.  Analytical results from the soil gas surveys are presented behind each site tab, as well.

Soil Gas Survey Methodology, January - February 2001

The purpose of this section is to summarize the activities performed during the soil gas survey completed at CSSA.  Soil gas surveys were conducted at SWMU B-3, SWMU B-4, AOC 55, AOC 57, AOC 63, AOC 65 and WWTP.  The soil gas survey described in this report was conducted at CSSA from January 2 through February 23, 2001.  A summary of the soil gas survey methods employed during the implementation of the soil gas survey is described below.  The scope of these activities is described in the Soil Gas Sampling and Analysis Plan Addenda (December 2000) prepared for this soil gas survey and located in Volume 1-4, Sampling and Analysis Plan, DO5084 Addendum.  Included in Volume 3-1.1 and Volume 3-2, behind the site tabs for SWMU B-3, SWMU B-4, AOC 55, AOC 57, AOC 63, AOC 65 and WWTP are how the sample locations were determined, sample collection procedures, analytical methodology, and soil gas analytical quality control procedures.

Prior to collection of any soil gas samples, all subsurface utilities and other man-made subsurface features, and surface obstacles (trees, buildings, roads, etc.), were identified and marked.  The soil gas samples were collected at locations within the grid layouts for each of the investigated areas.  The maximum depth of measured soil gas was approximately 12 feet, but most points were generally driven to less than 4 feet due to the thin soil layer overlying the limestone bedrock.  Soil gas samples were collected by driving a decontaminated, 1-inch diameter hollow steel sampling probe into the ground using a Geoprobe hydraulic/pneumatic hammer rig.  The probes were driven into the soil to a maximum depth of 6 feet or until refusal, except at SWMU B-3 where some points were driven into the base of the trench up to 12 feet below grade.  Soils up to 4.5 feet in thickness were encountered at each area investigated.  The minimum thickness of soil encountered was less than 1-foot of overburden at AOC 57.

After reaching the desired sample depth, a four-step process was implemented to obtain the soil gas sample from the appropriate depth interval.  First, the probes were pulled up slightly to detach the driving tip from the probe and to expose the open end of the hollow probe to surrounding soil gas from the sampling interval.  Next, a sampling adapter was placed on top of the probe with a Tygon� tube run from the adapter to a vacuum pump.  The vacuum pump was used to withdraw soil gases from the ground. For the third step in the process, the system was purged with 5 to 15 probe volumes prior to sampling to insure that a representative sample of soil gas was obtained.  The variance in purging volumes was dependent upon the permeability of the soil being tested.  In those areas where soils tended to exhibit greater permeability, an increased amount of soil gas could be purged prior to the actual collection of the sample.

Withdrawing the soil gas sample from the probe was the final step in the sample collection process.  After purging was completed, a new Tedlar bag was connected to the tubing from the sampling probe inside the dessicator.  A vacuum pump was then used to create a vacuum within the dessicator to withdraw soil gas from the ground.  After a sample was collected, the valve on the Tedlar bag was closed and the bag was removed from the dessicator and transported to a mobile field laboratory for analysis.  The vacuum pump was a 1 cubic feet per minute (cfm) rotary vane, oil-less, 1/6 horsepower model equipped with a vacuum regulator.

The samples were analyzed by a mobile laboratory field gas chromatograph (GC), operated by DHL Analytical, located at CSSA.  Samples were analyzed within 24 hours of collection.  After sampling, probes were decontaminated for use at another location.  The resulting hole was plugged and abandoned with dry, powdered cement where appropriate.

An initial screening of each soil gas sample location was performed during the purging stage of the sampling process by scanning the exhaust from the vacuum pump with an explosimeter that reads oxygen and Lower Explosive Limits (LEL).  An Industrial Scientific Corporation, Model HMX 271 explosimeter was used to measure the levels of oxygen and LEL.  The explosimeter was calibrated daily for oxygen readings by setting the readout to 20.9 percent oxygen when held in ambient air.  For oxygen measurements, the explosimeter had a stated accuracy of + 1.2 percent oxygen by volume in the range of 5 to 30 percent.  If relatively lower levels of oxygen were indicated by the initial screening, then the field team would alert the mobile laboratory about possible high levels of organic contaminants present in the sample.  VOCs are commonly encountered in lower oxygen soils at CSSA.

Within the mobile laboratory, soil gas samples were analyzed on a Shimadzu model 14A GC equipped with a Hall Electrolytic Conductivity Detector (HECD) and a photo ionization detector (PID) with a 10.2 electron volt (eV) light source.  A Spectra-Physics model 4400 dual-channel integrator was used to plot the chromatograms, to measure the size of the peaks, and to compute compound concentrations.

The chromatographic column used for analysis was a 75-meter long, DB-624 (3μm), 0.53mm ID megabore column.  This column was selected for its ability to separate the compounds targeted (PCE, TCE, trans-, cis-1,2-DCE, vinyl chloride, and BTEX) for analysis within a relatively short analysis time.

Nine compounds were targeted for calibration and analysis.  These compounds were the fuel components benzene, toluene, ethylbenzene, and xylene (BTEX); tetrachloroethene (PCE); trichloroethene (TCE); trans-1,2-dichloroethene (trans-1,2-DCE); cis-1,2-dichloroethene (cis-1,2-DCE); and vinyl chloride.

Calibration standards used to calibrate the PID and the HECD were prepared in a two-step procedure for all compounds except vinyl chloride.  First, a primary gaseous standard was prepared by injecting a mixed neat liquid standard, obtained from Chem Services, into a pre-cleaned 1-liter glass dilution bottle.  The primary standard was allowed to equilibrate in a 40⁰C oven for a period of at least 1 hour.  In the second step, two working standards were prepared by injecting a fixed volume of the primary gaseous standard into two 1-liter glass dilution bottles to which vinyl chloride had been added.  Vinyl chloride was added to the clean, 1-liter glass dilution bottles from a certified gaseous vinyl chloride standard, obtained from Scott Specialty Gases, using a gas-tight syringe.  The working standards were allowed to equilibrate in a 40⁰C oven for a period of not less than 5 minutes.

The calculation of the primary standard concentration in a 1-liter glass dilution bottle and the calculation specifically for 1,2-Dichloroethene (1,2-DCE) in a 1-liter glass dilution bottle were performed as follows:

A fixed volume of one standard was then injected into the GC for each calibration point.  The concentrations of all standards were in micrograms per liter. The concentrations of the calibration standards were within the linear range of the detector.  The standards were analyzed immediately upon removal from the oven.  An example of the calculation of the concentration of TCE in the mid-point calibration standard is as follows:

Concentrations of the calibration points ranged from 0.51m g/L to 51.0m g/L for vinyl chloride and from 0.2m g/L to 20.0m g/L for all other compounds.

All data obtained during this soil gas survey were reported in micrograms per liter.  Data reported in m g/L can be converted to ppbv by the following formula:

The ppbv number calculated above represents the compound concentration on a volume basis at standard temperature and pressure.

A calibration verification standard was run at the beginning of each day to determine the retention time for each of the target compounds and to confirm that the system was operating within limits.  A calibration verification standard was also analyzed at the end of the analytical batch, or after every 20 samples, whichever was more frequent.  The calibration verification standards were injected directly into the gas chromatograph in the same manner as the soil gas samples.

Soil gas samples were analyzed after the GC had been calibrated, the calibration verification sample run, and an ambient air (blank) sample had been analyzed.  The quantity of a target compound in each soil gas sample was determined by dividing the area of the peak registered on the chromatogram by the injection volume and inserting the calculated value into the linear equation generated by the calibration curve.  All compounds were quantified and identified using the PID.  The presence of the halogenated compounds (all compounds except BTEX) was confirmed using the HECD.  If the PID did not detect a halogenated compound or if the compound was not confirmed by (i.e., detected by) the HECD, the result was reported as non-detectable.

Identification of the target compounds in soil gas samples was based on a single column analysis and determined by a peak on the chromatogram being within the retention time window established for a particular compound.  If the concentration of target compounds in a soil gas sample was above the linear initial calibration (ICAL) range, the injection volume was decreased or the sample diluted to bring the sample concentration within the linear response range of the detector.

Method detection limits (MDLs) for target compounds in soil gas samples can vary depending on the volume injected and chromatographic interferences from adjacent peaks.  MDLs are defined as the minimum discernible peak divided by the typical injection volume and the response factor for the particular compound of interest.

Typically, a 10mL volume of each soil gas sample was injected into the GC giving MDLs that ranged from 0.02m g/L for benzene to 0.9 m g/L for vinyl chloride.

A number of QC procedures were followed to insure that acceptable and usable data were obtained during sample collection and analysis, as detailed in Attachment 1 of the Soil Gas Survey Sampling and Analysis Plan Addenda located in Volume 1-4, Sampling and Analysis Plan, DO5084 Addendum.

Probes were decontaminated with an Alconox and tap water wash, rinsed with clean water, and air dried before use at each sample point.  Syringes used to analyze samples were cleaned prior to each use by flushing the syringe with zero (clean) air a minimum of three times.

All analytical QC procedures were specified in the laboratory SOP included as an attachment to the SAP addenda for the soil gas survey.  The QC requirements for the soil gas survey analytical results were reviewed and considered acceptable within the context for which the data are to be used.

The QC analyses included analyzing calibration verification samples, analytical duplicates, and blank samples.  Calibration verification samples were analyzed at the beginning of every batch, after 20 samples, or at the end of every daily batch.  Analytical duplicates were performed at a frequency of one pair per analytical batch.  Analytical duplicates were performed by selecting one sample per batch and analyzing that sample in duplicate.  Blank samples (i.e., analysis of clean air) were run at the rate of one per analytical batch.  Blank samples were analyzed prior to the analysis of any sample to show the entire chromatographic system was free of contamination.

Accuracy and precision were demonstrated by the internal consistency of the analytical system.  Internal consistency is demonstrated in two ways: (1) the consistent response of the PID over the course of the day as shown by the results of the calibration verification samples (accuracy); and (2) consistent duplicate analyses of a soil gas sample as demonstrated by the results for duplicate analyses of the same sample having a relative percent difference (precision) of less than 30 percent.

Five-point curves were constructed to demonstrate the linear response range of the PID and electron capture detector HECD.  A measure of the linear response of the detector is the correlation coefficient.  The closer the correlation coefficient is to 1 the more linear the detector response.  The PID was used as the primary detector for all compounds.  The calibration curves for all compounds were required to have a correlation coefficient (r²) of 0.995 or greater for the PID.

Single point calibration verification was performed on a daily basis during the course of the soil gas survey.  These verifications were performed at a concentration equal to the mid-point of the calibration curve.

Results of Duplicate Sample Analyses

Duplicate sample results with a relative percent difference of 30 percent or less are considered within the normal range of analytical precision.  The relative percent difference (RPD) was calculated using the following equation:

Results from all laboratory duplicate samples were within the QC Criteria of 30 percent and met the necessary analytical precision requirement.  Laboratory duplicates were analyzed to meet DHL's internal QC requirements.  Laboratory duplicate data are not included in this technical report.

Each day, immediately following the analysis of the initial calibration verification standard, a method blank was analyzed.  The method blank consisted of zero air (clean air) injected using the same procedure as the samples.  The method blank was analyzed to identify any contamination present in the laboratory, chromatographic system, or glassware used.  All method blanks analyzed for this project were free of target analytes at the MDL.