United States      Office of Research and      EPA/600/R-00/046
        Environmental Protection   Development        March 2000
        Agency	Washington, D.C. 20460	



   EPA  Environmental Technology


        Verification Report





        Explosives Detection Technology





        Barringer Instruments


        GC-IONSCAN™
               Environmental Security

               Technology Certification

                  Program






                oml

                Oak Ridge National Laboratory
ETV  ETY  ET

-------

-------
                 THE ENVIRONMENTAL TECHNOLOGY VERIFICATION
                                         PROGRA
       . En L'irMDUJ?Li.l;i I PJU IroLiuu A
                                         hnvsrrinnTnNnl Smir ty
                                        Trirhnnlnny ( nrlrirrrinn
                                                                        ornl
                                                                       Oak Ridge National Laboratory
                          Joint Verification Statement
    TECHNOLOGY TYPE:   EXPLOSIVES DETECTION

    APPLICATION:
MEASUREMENT OF EXPLOSIVES IN CONTAMINATED
SOIL AND WATER
    TECHNOLOGY NAME:  GC-IONSCAN™

    COMPANY:              Barringer Instruments

    ADDRESS:
    WEB SITE:
    EMAIL:
30 Technology Drive
Warren, NJ 07059

www.barringer.com
rdebono@bii.barringer.com
PHONE: (908) 222-9100
FAX: (908) 222-1557
The U.S. Environmental Protection Agency (EPA) has created the Environmental Technology Verification
Program (ETV) to facilitate the deployment of innovative or improved environmental technologies through
performance verification and dissemination of information. The goal of the ETV Program is to further
environmental protection by substantially accelerating the acceptance and use of improved and cost-effective
technologies. ETV seeks to achieve this goal by providing high quality, peer reviewed data on technology
performance to those involved in the design, distribution, financing, permitting, purchase, and use of
environmental technologies.

ETV works in partnership with recognized standards and testing organizations, stakeholder groups consisting
of regulators, buyers, and vendor organizations, with the full participation of individual technology developers.
The program evaluates the performance of innovative technologies by developing test plans that are
responsive to the needs of stakeholders, conducting field or laboratory tests (as appropriate), collecting and
analyzing data, and preparing peer-reviewed reports. All evaluations are conducted in accordance with
rigorous quality assurance protocols to ensure that data of known and adequate quality are generated and that
the results are defensible.

The Department of Defense (DoD) has a similar verification program known as the Environmental Security
Technology Certification Program (ESTCP). The purpose of ESTCP is to demonstrate and validate the most
promising innovative technologies that target DoD's most urgent environmental needs and are projected to
pay back the investment within 5 years through cost savings and improved efficiencies. ESTCP
demonstrations are typically conducted under operational field conditions at DoD facilities. The
EPA-VS-SCM-45
                      The accompanying notice is an integral part of this verification statement.
                                                                                      March 2000

-------
demonstrations are intended to generate supporting cost and performance data for acceptance or validation of
the technology. The goal is to transition mature environmental science and technology projects through the
demonstration/ validation phase, enabling promising technologies to receive regulatory and end user
acceptance in order to be fielded and commercialized more rapidly.

The Oak Ridge National Laboratory (ORNL) is one of the verification organizations operating under the Site
Characterization and Monitoring Technologies (SCMT) program. SCMT, which is administered by EPA's
National Exposure Research Laboratory, is one of 12 technology areas under ETV. In this demonstration,
ORNL evaluated the performance of explosives detection technologies. This verification statement provides a
summary of the test results for Barringer Instruments' GC-IONSCAN™. This verification was conducted
jointly with the Department of Defense's (DoD's) Environmental Security Technology Certification Program
(ESTCP).

DEMONSTRATION DESCRIPTION
This demonstration was designed to evaluate technologies that detect and measure explosives in soil and
water. The demonstration was conducted at ORNL in Oak Ridge, Tennessee, from August 23 through
September 1, 1999. Spiked samples of known concentration were used to assess the accuracy of the
technology. Environmentally contaminated soil samples, collected from DoD sites in California, Louisiana,
Iowa, and Tennessee and ranging in concentration from 0 to approximately 90,000 mg/kg, were used to
assess several performance characteristics. Explosives-contaminated water samples from Tennessee,
Oregon, and Louisiana with concentrations ranging from 0 to 25,000 |jg/L also were analyzed. The primary
constituents in the samples were 2,4,6-trinitrotoluene (TNT); isomeric dinitrotoluene (DNT), including both
2,4-dinitrotoluene and 2,6-dinitrotoluene; hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX); and octahydro-1, 3,5,7-
tetranitro-l,3,5,7-tetrazocine (HMX). The results of the soil and water analyses conducted under field
conditions by the GC-IONSCAN were compared with results from reference laboratory analyses of
homogenous replicate samples determined using EPA SW-846 Method 8330. Details of the demonstration,
including a data summary and discussion of results, may be found in the report entitled Environmental
Technology Verification Report: Explosives Detection Technology — Barringer Instruments, GC-
             EPA 600-R-00/046.
TECHNOLOGY DESCRIPTION
The GC-IONSCAN is a fully transportable field-screening instrument that combines the rapid analysis time of
ion mobility spectrometry (IMS) with the separation ability of gas chromatography (GC). The instrument can
be operated in IONSCAN mode or in GC-IONSCAN mode to detect explosives. The user can switch
between the two modes in less than 30 s through the instrument control panel. In the IONSCAN mode,
samples are deposited on a Teflon filter and thermally desorbed directly to the IMS, permitting the quick
screening analysis of explosives residues in 6 to 8 s. In the GC-IONSCAN mode, extracts are directly
injected onto the GC column and analysis occurs within 1 to 3 minutes, depending on the type of explosive.
The use of the IONSCAN mode permits rapid prescreening of samples with identification of the major
constituents of the sample and semiquantitative analysis, while the GC-IONSCAN mode permits full
characterization and quantitative analysis of the sample. This technology is capable of reporting quantitative
data for all of the Method 8330 analytes. The performance assessment described here is only for TNT and
RDX because a limited amount of data was available for evaluation of the other analytes. Reporting limits for
the GC-IONSCAN ranged from 0.3 to 10 mg/kg for soil and 25 to 1950 ng/L for water.
EPA-VS-SCM-45           The accompanying notice is an integral part of this verification statement.              March 2000

-------
VERIFICATION OF PERFORMANCE
The following performance characteristics of the GC-IONSCAN were observed:

Precision: For the soil samples, the mean relative standard deviations (RSDs) for RDX and TNT were 54%
and 51%, respectively. For water samples, the RSDs were significantly lower, at 20% and 26%, respectively.

Accuracy: For the soil samples, the median percent recoveries for RDX and TNT were 55% and 136%,
respectively. The results were generally biased low for RDX and biased high for TNT. For water samples,
only a few of the RDX and TNT results were reported above the reporting limits in the spiked samples. The
recoveries were significantly lower, with the highest recovery at 46%, indicating that the water results were
biased low for both analytes.

False positive/false negative results: Of the 20 blank soils, Barringer reported RDX in one sample (5%
false positives) and TNT in five samples (25% false positives). No false positives  were reported for RDX and
TNT in the 20 blank water samples. False positive and false negative results were also determined by
comparing the GC-IONSCAN result to the reference laboratory result for the environmental and spiked
samples (e.g., whether the GC-IONSCAN reports a result as a nondetect that the reference laboratory
reported as a detection, and vice versa). For the soils, 3% of the RDX results and  none of the TNT results
were reported as false positives  relative to the reference laboratory results. Significantly more samples were
reported as nondetects by  Barringer (i.e., false negatives) when the laboratory reported a detection (2% for
RDX and 13% for TNT).  Similar results were observed for water, where 2% of the TNT results and none of
the RDX results were false positives, and a higher percentage (39% of the RDX results and 21% of the TNT
results) were false negatives.

Completeness: The GC-IONSCAN generated results for all 108 soil samples and all 176 water samples, for
a completeness of 100%.

Comparability: A one-to-one sample comparison of the GC-IONSCAN results and the reference laboratory
results was performed for all samples (spiked and environmental) that were reported as detects. The cor-
relation coefficient (r) for  the comparison of the entire soil data set for TNT was 0.88 (slope (m) = 4.82).
When comparability was assessed for specific concentration ranges, the r value did not  change dramatically
for TNT, ranging from 0.71 to 0.85 depending on the concentrations selected. RDX correlation with the
reference laboratory for soil was similar (r values near 0.80), except for concentrations  greater than
1,000 mg/kg, where the correlation was lower (r = 0.28, m = 0.14). For the water samples, comparability with
the reference laboratory results for TNT was much lower than the soil comparison (r =  0.53). For RDX, the
correlation was much higher, at 0.95. Although the correlation was high, the slope  of the linear regression line
was 0.08, indicating that the GC-IONSCAN RDX results were biased low (see Accuracy).

Sample Throughput:  Throughput was approximately three samples per hour for soil and eight samples per
hour for water. This rate was accomplished by two operators and included sample preparation and analysis.

Ease of Use: Users unfamiliar with ion mobility spectrometry would require approximately two days of
training to operate the GC-IONSCAN. Training is provided by Barringer Instruments. No particular level of
educational training is required for the operator,  but knowledge of chromatographic techniques would be
advantageous.
Overall Evaluation: The  overall performance of the GC-IONSCAN for the analysis of RDX and TNT was
characterized as precise and biased low (both analytes) for water analyses, and imprecise and biased (low for
EPA-VS-SCM-45           The accompanying notice is an integral part of this verification statement.              March 2000

-------
RDX and high for TNT) for soil analyses. As with any technology selection, the user must determine if this
technology is appropriate for the application and the project's data quality objectives. For more information on
this and other verified technologies, visit the ETV web site at http://www.epa.gov/etv.
Gary J. Foley, Ph.D.                                            David E. Reichle, Ph.D.
Director, National Exposure Research Laboratory               Associate Laboratory Director
Office of Research and Development                           Oak Ridge National Laboratory
Jeffrey Marqusee, Ph.D.
Department of Defense
Director, Environmental Security Technology Certification Program
     NOTICE: EPA and ESTCP verifications are based on evaluations of technology performance under specific,
     predetermined criteria and appropriate quality assurance procedures. EPA, ESTCP, and ORNL make no expressed
     or implied warranties as to the performance of the technology and do not certify that a technology will always
     operate as verified.  The end user is solely responsible for complying with any and all applicable federal, state, and
EPA-VS-SCM-45            The accompanying notice is an integral part of this verification statement.               March 2000

-------

-------
                                            EPA/600/R-XX/YYY
                                               March 2000
Environmental Technology
Verification Report

Explosives Detection Technology

Barringer Instruments
GC-IONSCAN™
                         By
                      Amy B. Dindal
                    Charles K. Bayne, Ph.D.
                    Roger A. Jenkins, Ph.D.
                   Oak Ridge National Laboratory
                  Oak Ridge Tennessee 37831-6120
                       Eric N. Koglin
                 U.S. Environmental Protection Agency
                  Environmental Sciences Division
                 National Exposure Research Laboratory
                  Las Vegas, Nevada 89193-3478
            This demonstration was conducted in cooperation with the
                   U.S. Department of Defense
            Environmental Security Technology Certification Program

-------

-------
                                           Notice

The U.S. Environmental Protection Agency (EPA), through its Office of Research and Development (ORD),
and the U.S. Department of Defense's Environmental Security Technology Certification Program (ESTCP)
Program, funded and managed, through Interagency Agreement No. DW89937854 with Oak Ridge National
Laboratory, the verification effort described herein. This report has been peer and administratively reviewed
and has been approved for publication as an EPA document. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use of a specific product.

-------
                                  Table of Contents
   List of Figures  	    vi
   List of Tables  	  viii
   Acknowledgments 	    x
   List of Abbreviations and Acronyms  	    xii

1  INTRODUCTION  	    1

2  TECHNOLOGY DESCRIPTION  	    3
   Technology Overview   	    3
   Sample Preparation  	    4
   Analytical Determination  	    4
   Instrument Calibration and Quantification of Sample Results 	    4

3  DEMONSTRATION DESIGN  	    5
   Objective	    5
   Demonstration Testing Location and Conditions 	    5
   Soil Sample Descriptions  	    5
       Sources of Samples  	    5
          Iowa Army Ammunition Plant   	    5
          Louisiana Army Ammunition Plant  	    5
          Milan Army Ammunition Plant  	    5
          Volunteer Army Ammunition Plant  	    5
          Fort Ord Military Base 	    6
       Performance Evaluation Samples  	    6
       Soil Sample Preparation 	    6
   Water Sample Descriptions  	    7
       Sources of Samples  	    7
       Performance Evaluation Samples  	    7
       Water Sample Preparation  	    7
   Sample Randomization  	    7
   Summary of Experimental Design  	    7
   Description of Performance Factors  	    7
       Precision  	    8
       Accuracy  	    8
       False Positive/Negative Results  	    8
       Completeness  	    8
       Comparability  	    8
       Sample Throughput  	    9
       Ease of Use  	    9
       Cost 	    9
       Miscellaneous Factors  	    9

4  REFERENCE LABORATORY ANALYSES  	   10
   Reference Laboratory Selection  	   10
   Reference Laboratory Method  	   10
   Reference Laboratory Performance  	   10
                                             iii

-------
5  TECHNOLOGY EVALUATION  	   13
   Objective and Approach 	   13
   Precision 	   13
   Accuracy  	   13
   False Positive/False Negative Results  	   14
   Completeness  	   15
   Comparability  	   15
   Sample Throughput  	   18
   Ease of Use  	   18
   Cost Assessment  	   18
       GC-IONSCAN Costs  	   19
       Reference Laboratory Costs	   21
       Cost Assessment Summary   	   21
   Miscellaneous Factors  	   21
   Summary of Performance  	   22

6  TECHNOLOGY UPDATE 	   24
   Rapid Prescreening of Samples	   24
   Improvements to GC-IONSCAN  	   24

7  REFERENCES 	   25

   APPENDICES
       A: GC-IONSCAN Sample Soil Results Compared with Reference Laboratory Results   	   27
       B: GC-IONSCAN Sample Water Results Compared with Reference Laboratory Results	   33
                                           IV

-------

-------
                                    List of Figures
1  The GC-IONSCAN	   3
2  Comparability of reference laboratory TNT soil results with GC-IONSCAN results for vendor
   concentrations less than 500 mg/kg  	   17
3  Comparability of reference laboratory RDX water results with GC-IONSCAN results
   for all results reported above reporting limits  	   18
4  Range of percent difference values for TNT in soil and water  	   19
5  Range of percent difference values for RDX in soil and water  	   20
                                              VI

-------

-------
                                    List of Tables
1  Summary of Soil and Water Samples	    8
2  Summary of the Reference Laboratory Performance for Soil Samples   	   11
3  Summary of the Reference Laboratory Performance for Water Samples  	   11
4  Summary of the Reference Laboratory Performance on Blank Samples  	   12
5  Summary of the GC-IONSCAN Precision	   13
6  Summary of the GC-IONSCAN Accuracy  	   14
7  Number of GC-IONSCAN Results within Acceptance Ranges for Spiked Soils	   14
8  Summary of GC-IONSCAN False Positives on Blank Soil and Water Samples  	   15
9  Summary of the GC-IONSCAN Detect/Nondetect Performance Relative to the Reference Laboratory
   results  	   15
10 GC-IONSCAN Correlation  with Reference Data for Various Vendor Soil Concentration Ranges  ...   16
11 GC-IONSCAN Correlation with Reference Data for Various Vendor Water Concentration Ranges .   16
12 Estimated analytical costs for explosives-contaminated samples  	   22
13 Performance Summary for the GC-IONSCAN 	   23
                                            viii

-------

-------
                                  Acknowledgments

The authors wish to acknowledge the support of all those who helped plan and conduct the demonstration,
analyze the data, and prepare this report. In particular, we recognize Dr. Thomas Jenkins (U.S. Army, Cold
Regions Research and Engineering Laboratory) and Dr. Michael Maskarinec (Oak Ridge National
Laboratory) who served as the technical experts for this project. We thank the people who helped us to obtain
the samples from the various sites, including Dr. Jenkins, Danny Harrelson (Waterways Experiment Station),
Kira Lynch (U.S. Army Corp of Engineers, Seattle District), Larry Stewart (Milan Army Ammunition Plant),
Dick Twitchell and Bob Elmore (Volunteer Army Ammunition Plant). For external peer review, we thank Dr.
C. L. Grant (Professor Emeritus, University of New Hampshire).  For internal peer review, we thank Stacy
Barshick of Oak Ridge National Laboratory and Harry Craig of EPA Region 10. The authors also
acknowledge the participation of Yin Sun and Tri Le of Barringer Instruments, who performed the analyses
during verification testing.

For more information on the Explosives Detection Technology Demonstration contact

Eric N. Koglin                           Amy B. Dindal
Project Technical Leader                 Technical Lead
Environmental Protection Agency          Oak Ridge National Laboratory
Environmental Sciences Division           Chemical and Analytical Sciences Division
National Exposure Research Laboratory     P.O. Box 2008
P. O. Box 93478                         Building 4500S, MS-6120
Las Vegas, Nevada 89193-3478           Oak Ridge, TN 37831- 6120
(702) 798-2432                          (865) 574-4863
koglin.eric@epa.gov                      dindalab@ornl.gov
For more information on Barringer Instruments' GC-IONSCAN contact

Reno DeBono                          Andy Rudolph, Ph.D.
Barringer Instruments                   GC-IONSCAN Project Manager
30 Technology Drive                    Barringer Research Ltd.
Warren, NJ 07059                      1730 Aimco Blvd.
(908) 222-9100, ext 3017                 Mississauga, Ontario L4W 1V1
rdebono@bii.barringer.com               Canada
www.barringer.com                     Phone: (905) 238-8837
                                      Fax:(905)238-3018
                                      arudolph@barringer. com

-------

-------
                   Abbreviations and Acronyms
2-Am-DNT
4-Am-DNT
CRREL
2,4-DNT
2,6-DNT
DNT
DoD
EPA
ERA
ESTCP
ETV
fn
fp
GC
HMX
HPLC
IMS
LAAAP
MLAAP
NERL
ORNL
PE
QA
QC
RDX
RSD
SCMT
SD
TNB
TNT
USAGE
2-amino-4,6-dinitrotoluene, CAS # 35572-78-2
4-amino-2,6-dinitrotoluene, CAS # 1946-51-0
U.S. Army Cold Regions Research and Engineering Laboratory
2,4-dinitrotoluene, CAS # 121-14-2
2,6-dinitrotoluene, CAS # 606-20-2
isomeric dinitrotoluene (includes both 2,4-DNT and 2,6-DNT)
U.S. Department of Defense
U.S. Environmental Protection Agency
Environmental Resource Associates
Environmental Security Technology Certification Program
Environmental Technology Verification Program
false negative result
false positive result
gas chromatography
octahydro-l,3,5,7-tetranitro-l,3,5,7-tetrazocine, CAS # 2691-41-0
high performance liquid chromatography
ion mobility spectrometry
Louisiana Army Ammunition Plant
Milan Army Ammunition Plant
National Exposure Research Laboratory (EPA)
Oak Ridge National Laboratory
performance evaluation sample
quality assurance
quality control
hexahydro-l,3,5-trinitro-l,3,5-triazine, CAS # 121-82-4
percent relative standard deviation
Site Characterization and Monitoring Technologies Pilot of ETV
standard deviation
1,3,5-trinitrobenzene, CAS # 99-35-4
2,4,6-trinitrotoluene, CAS # 118-96-7
U.S. Army Corps of Engineers
                                       xu

-------

-------
                               Section 1 — Introduction
The U.S. Environmental Protection Agency (EPA)
created the Environmental Technology Verification
Program (ETV) to facilitate the deployment of
innovative or improved environmental technologies
through performance verification and dissemination
of information. The goal of the ETV Program is to
further environmental protection by substantially
accelerating the acceptance and use of improved
and cost-effective technologies. ETV seeks to
achieve this goal by providing high-quality, peer-
reviewed data on technology performance to those
involved in the design, distribution, financing,
permitting, purchase, and use of environmental
technologies.

ETV works in partnership with recognized standards
and testing organizations and stakeholder groups
consisting of regulators, buyers, and vendor
organizations, with the full participation of individual
technology developers. The program evaluates the
performance of innovative technologies by
developing verification test plans that are responsive
to the needs  of stakeholders, conducting field or
laboratory tests (as appropriate), collecting and
analyzing data, and preparing peer-reviewed reports.
All evaluations are conducted in accordance with
rigorous quality assurance (QA) protocols to ensure
that data of known and adequate quality are
generated and that the results are defensible.

ETV is a voluntary program that seeks to provide
objective performance information to all of the
participants in the environmental marketplace and to
assist them in making informed technology decisions.
ETV does not rank technologies or compare their
performance, label or list technologies as acceptable
or unacceptable, seek to determine "best available
technology," or approve or disapprove technologies.
The program does not evaluate technologies at the
bench or pilot scale and does not conduct or support
research. Rather, it conducts and reports on testing
designed to describe the  performance of
technologies under a range of environmental
conditions and matrices.
The program now operates 12 pilots covering a
broad range of environmental areas. ETV has begun
with a 5-year pilot phase (1995-2000) to test a wide
range of partner and procedural alternatives in
various pilot areas, as well as the true market
demand for and response to such a program. In
these pilots, EPA utilizes the expertise of partner
"verification organizations" to design efficient
processes for conducting performance tests of
innovative technologies. These expert partners are
both public and private organizations, including
federal laboratories, states, industry consortia, and
private sector entities. Verification  organizations
oversee and report verification activities based on
testing and  QA protocols developed with input from
all major stakeholder/customer groups associated
with the technology area. The verification described
in this report was administered by the Site
Characterization and Monitoring Technologies
(SCMT) Pilot, with Oak Ridge National Laboratory
(ORNL) serving as the verification organization. (To
learn more  about ETV, visit ETV's Web site at
http://www.epa.gov/etv.) The SCMT pilot is
administered by EPA's National Exposure Research
Laboratory (NERL), Environmental Sciences
Division, in Las Vegas, Nevada.

The Department of Defense (DoD) has a similar
verification program known as the  Environmental
Security Technology Certification Program
(ESTCP). The purpose of ESTCP  is to demonstrate
and validate the most promising innovative
technologies that target DoD's most urgent
environmental needs and are projected to pay back
the investment within 5 years through cost savings
and improved efficiencies. ESTCP responds to
(1) concern over the slow pace and cost of
remediation of environmentally contaminated sites on
military installations, (2) congressional direction to
conduct demonstrations specifically focused on new
technologies, (3) Executive Order 12856, which
requires federal agencies to place high priority on
obtaining funding and resources needed for the
development of innovative pollution prevention
programs and technologies for installations and in
acquisitions, and (4) the need to improve defense

-------
readiness by reducing the drain on the Department's
operation and maintenance dollars caused by real
world commitments such as environmental
restoration and waste management.  ESTCP
demonstrations are typically conducted under
operational field conditions at DoD facilities. The
demonstrations are intended to generate supporting
cost and performance data for acceptance or
validation of the technology. The goal is to transition
mature environmental science and technology
projects through the  demonstration/ validation phase,
enabling promising technologies to receive regulatory
and end user acceptance in order to  be fielded and
commercialized more rapidly. (To learn more about
ESTCP, visit ESTCP's web site at
http://www.estcp.org.)

EPA's ETV program and DoD's ESTCP program
established a memorandum of agreement in 1999 to
work cooperatively with ESTCP on  the verification
of technologies that are used to  improve
environmental cleanup and protection at both DOD
and non-DOD sites.  The verification of field
analytical technologies for explosives detection
described in this report was conducted jointly by
ETV's SCMT pilot and ESTCP. The verification
was conducted at ORNL in Oak Ridge, Tennessee,
from August 23 through September 1, 1999. The
performances of two field analytical techniques for
explosives were determined under field conditions.
Each technology was independently evaluated by
comparing field analysis results with those obtained
using an approved reference method, EPA SW-846
Method 8330. The demonstration was designed to
evaluate the field technology's ability to detect and
measure explosives in soil and water. The primary
constituents in the samples were 2,4,6-trinitrotoluene
(TNT); isomeric dinitrotoluene (DNT), including both
2,4-dinitrotoluene (2,4-DNT) and 2,6-dinitrotoluene
(2,6-DNT);hexahydro-l,3,5-trinitro-l,3,5-triazine
(RDX); and octahydro-l,3,5,7-tetranitro-l,3,5,7-
tetrazocine (HMX). Naturally contaminated
environmental soil samples, ranging in concentration
from 0 to about 90,000 mg/kg, were collected from
DoD sites in California, Louisiana, Iowa, and
Tennessee, and were used to assess several
performance characteristics. Explosives-
contaminated water samples  from Tennessee,
Oregon, and Louisiana with concentrations ranging
from 0 to 25,000 |jg/L were also evaluated. This
report discusses the performance of Barringer
Instruments' GC-IONSCAN™.

-------
                     Section 2 — Technology Description
In this section, the vendor (with minimal editorial changes by ORNL) provides a description of the
technology and the analytical procedure used during the verification testing activities.
Technology Overview
The GC-IONSCAN, which weighs approximately 70
Ib and is shown in Figure 1, is an on-site analytical
instrument combining the rapid analysis time of ion
mobility spectrometry (IMS) with the separation
capability of gas chromatography (GC). In IMS, ions
are generated by atmospheric pressure chemical
ionization and drift through a buffer gas under the
influence of an electric field. The rate of drift of ions
through the field is dependent upon both the physical
and electrical properties of the molecules, and can
be used to discriminate between compounds based
on size-to-charge ratio. In GC, components of
complex mixtures are separated by a stationary
phase; the separation occurs on the basis of the
relative affinity of the compounds for the stationary
phase.

The analytical process begins with the eluent
entering the IONSCAN inlet. The sample then
combines with makeup gas (air filtered with charcoal
and Drierite™) doped with reactant, and proceeds
into the ionization region, where the sample is
selectively ionized to form ions or ionic clusters of
specific mobilities (drift time). The gating grid opens,
allowing ions of the correct polarity (negative for
explosives) enter the drift region. The ions are then
focused and accelerated by the electric field along
the drift region of the IMS tube and arrive at the
collector electrode (typically 10-20 ms).  IMS
identifies individual explosives based on the unique
ion mobilities (drift time) of specific compounds.

The instrument can be operated in IONSCAN mode
or in GC-IONSCAN mode. The user can switch
between the two modes in less than 30 s  through the
instrument control panel. In the IONSCAN mode,
samples are deposited on a Teflon filter, allowed to
evaporate and then thermally desorbed directly into
the IMS, permitting quick screening analysis of
explosives residues in 6-8 s. In the GC-IONSCAN
mode, extracts are directly injected onto the GC
column and analysis occurs within 1-3 min,
depending on the type of explosive and the GC
column used. The use of the IONSCAN  mode
permits rapid prescreening of samples, with
identification of the major constituents of the sample
and semiquantitative analysis, while the GC-
IONSCAN mode permits full characterization and
quantitative analysis of the sample.

At the time of the demonstration, the cost of
purchasing the GC-IONSCAN was $60,000. The kit
included a laptop computer, a standard spare kit, a
standard maintenance kit, a standard  consumables
kit, a sampling kit, a swab sampler, IM software, and
an operator's manual; the price also included training
at Barringer for up to six people. The kit  is supplied
in a aluminum carrying case that can also be used to
ship the instrument. The instrument can be operated
in two modes: explosives detection and drug
detection. As with any instrument, the cost on a per-
sample basis would decrease with an increase in the
number of analyses performed.
Figure 1. The GC-IONSCAN.

-------
Sample Preparation
In the demonstration, minimal sample preparation
was performed. The preparation of soil samples
involved one-step solvent extraction with acetone.
Ten milliliters of acetone was added to 2 g of soil in
a 20-mL vial. The mixture was shaken in a vortex
mixer for 2-3 min. After the solution settled for
approximately 1  min, two dilutions using acetone (10-
and 100-fold) were prepared for each sample from
the acetone fraction. If the sample contained high
levels of explosives (based on IONS CAN mode
analysis), it was  further diluted.

The preparation  of water samples involved adding 2
mL of sample to 1 g of sodium sulfate in a 20-mL
vial and then adding 1 mL of acetone. The mixture
was shaken in a vortex mixer for 2-3 min. If the
sample was too concentrated, based on the GC-
IONSCAN response, the sample was diluted by 1:10
and reanalyzed.

Analytical Determination
In GC-IONSCAN mode, a 15-m MXT-1  column
(internal diameter of 0.53 mm and film thickness of
1.0 |jm), operated in the splitless mode with nitrogen
carrier gas, was used for the demonstration sample
analyses. The injector port temperature was 260°C.
The oven temperature program was 160°C; this
temperature was held for 60 s and then ramped at a
rate of 40°C /min to 240°C, for a 180-s analysis time.
Analyses not including HMX were performed
isothermally at 160°C, resulting in a 90-s analysis
time. A 0.5-m transfer line from the end of the
column to the detector was at 195°C.

For the soil analyses, extracts were screened in the
IONSCAN mode. After a 1-^iL aliquot of the
sample extract was  deposited on the Teflon filter,
the acetone was allowed to evaporate (taking
approximately 15-20 s); then the filter was thermally
desorbed. The Barringer team analyzed extracts (for
a given sample) in the following order: the 100-fold
dilution, the 10-fold dilution, the undiluted extract, and
4 jjL of the undiluted extract. If no explosives were
detected, a more concentrated extract was analyzed.
If no explosives were detected in the 4-|jL undiluted
extract, the sample was reported as less than the
reporting limit of the GC-IONSCAN. If explosives
were detected in the extracts using the IONSCAN
mode, the sample was analyzed in the GC-
IONSCAN mode.

For the analysis of aqueous samples, no
prescreening was performed in the IONSCAN
mode. A 2-|jL aliquot from the acetone fraction was
injected into the heated injector port of the GC-
IONSCAN. The sample was either quantified,
diluted, and reanalyzed, or reported as a nondetect,
as appropriate.  The Barringer team elected not to
analyze water samples for HMX.

Instrument Calibration and
Quantification of Sample Results
At the beginning of the day multiple l-|jL injections
of a standard containing all of the method analytes
were used to deactivate any active sites within the
instrument. The standard contained 4 ng/|jL for all
analytes except as  follows: HMX, 20 ng/|jL; 2,4-
DNT, 20 ng/|jL; 2,6-DNT 1,000 ng/nL. The GC-
IONSCAN instrument was initially calibrated by use
of eight calibration standards, ranging in
concentration from 0.5 ng/|jL to 10 ng/|jL for TNT,
RDX, HMX, TNB, 2-amino-4,6-dinitrotoluene (2-
Am-DNT), and 4-amino-2,6-dinitrotoluene (4-Am-
DNT); from 10 to  100 ng for 2,4-DNT; and from
1000 to 4000 ng for 2,6-DNT. Linear or logarithmic
expressions were used to generate the calibration
curves. The calibration was checked every 15
samples by analyzing a mid-level standard. If the
value was within 30% of the initial calibration value,
the calibration was considered still valid. If the
response had changed by greater than 30%, the
instrument was recalibrated. The concentration of
explosives in the sample was calculated by
comparing the response in the sample extracts to the
calibration curve. Reporting limits ranged from 0.3 to
10 mg/kg for soil and 25 to 1950 |jg/L for water,
depending on the analyte.

-------
                       Section 3 — Demonstration Design
Objective
The purpose of this section is to describe the
demonstration design. It is a summary of the
technology demonstration plan (ORNL 1999).

Demonstration Testing Location
and Conditions
The verification of field analytical technologies for
explosives was conducted at the ORNL Freels Bend
Cabin site, in Oak Ridge, Tennessee. The site is
somewhat primitive, with no running water, but the
vendors were provided with some shelter (porch
overhang) and electrical power. The temperature
and relative humidity were monitored during field
testing. Over the ten  days of testing, the average
temperature was 77°F, and ranged from 60 to 88°F.
The average relative  humidity was 67%, and ranged
from 35 to 96%.

The samples used in  this study were brought to the
demonstration testing location  for evaluation by the
vendors. Explosives-contaminated soils from Army
ammunition plants in Iowa, Louisiana, and
Tennessee and a former Army base in California
(Fort Ord) were used in this verification. In addition,
explosives-contaminated water samples were
analyzed from DoD sites in Oregon, Louisiana, and
Tennessee. Because  samples were obtained from
multiple DoD sites, the samples represented a
reasonable cross section of the population of
explosives-contaminated matrices, such that the
versatility of the field technology could be evaluated.
The vendors had the  choice of analyzing either soil
or water samples, or  both matrices. More specific
details about the samples are presented below.

Soil Sample  Descriptions
The primary constituents in the soil  samples were
TNT, DNT, RDX, and HMX.  The samples also
contained trace amounts of 2-amino-4,6-
dinitrotoluene (2-Am-DNT) and 4-amino-2,6-
dinitrotoluene (4-Am-DNT), which  are degradation
products of TNT. The total concentration of
explosives ranged from 0 to approximately
90,000 mg/kg. The following sections describe the
sites from which the samples were collected.

Sources of Samples
Iowa Army Ammunition Plant
Currently an active site, the Iowa Army Ammunition
Plant was constructed to load, assemble, and pack
various conventional ammunition and fusing systems.
Current production includes 120-mm tank rounds,
warheads for missiles, and mine systems. During the
early years of use, the installation used surface
impoundments, landfills, and sumps for disposal of
industrial wastes containing explosives.  The major
contaminants in these samples were TNT, RDX, and
HMX.

Louisiana Army Ammunition Plant
The Louisiana Army Ammunition Plant (LAAAP),
near Shreveport, Louisiana, is a government-owned
facility that began production in 1942. The facility is
currently an Army Reserve plant. Production items
at LAAAP have included metal parts for artillery
shells; the plant also loads, assembles, and packs
artillery shells, mines, rockets, mortar rounds, and
demolition blocks. As a result of these activities and
the  resulting soil and groundwater contamination,
EPA placed LAAAP on the National Priorities List
of contaminated sites (Superfund) in 1989. The
major constituents in the samples from this site were
TNT, RDX, and HMX, with trace levels of 1,3,5-
trinitrobenzene (TNB), DNT, 2-Am-DNT, and 4-
Am-DNT.

Milan Army Ammunition Plant
Currently active, the Milan Army Ammunition Plant
(MLAAP) in Milan, Tennessee, was established in
late 1940 as part of the pre-World War II buildup.
The facility still has ten ammunition loading,
assembly, and packaging lines. Munitions-related
wastes have resulted in soil contamination. The
primary contaminants in these soils were RDX and
TNT.

Volunteer Army Ammunition Plant
The Volunteer Army Ammunition Plant, in
Chattanooga, Tennessee, was built in 1941 to

-------
manufacture TNT and DNT. All production ceased
in 1977. Past production practices resulted in
significant soil and groundwater contamination. In
the samples from this site, concentrations of TNT
and DNT ranged from 10 to 90,000 mg/kg, with
significantly smaller concentrations of Am-DNT
isomers.

Fort Ord Military Base
Fort Ord, located near Marina, California, was
opened in 1917 as a training and staging facility for
infantry troops and was closed as a military
installation in 1993. Since then, several nonmilitary
uses have been established on the site: California
State University at Monterey Bay has opened its
doors on former Fort Ord property, the University of
California at Santa Cruz has established a new
research center there, the Monterey Institute of
International Studies will take over the officer's club
and several other buildings, and the post's airfield
was turned over to the city of Marina. The Army
still occupies several buildings.

An Army study conducted in 1994 revealed that the
impact areas at the inland firing ranges of Fort Ord
were contaminated with residues of high explosives
(Jenkins, Walsh, and Thorne 1998). Fort Ord is on
the National Priorities List of contaminated sites
(Superfund), requiring the installation to be
characterized and remediated to a condition that
does not pose unacceptable risks to public health or
the environment. The contaminant present at the
highest concentration (as much as 300 mg/kg) was
HMX; much lower concentrations of RDX, TNT,
2-Am-DNT, and 4-Am-DNT are present.

Performance Evaluation Samples
Spiked soil samples were obtained from
Environmental Resource Associates (ERA, Arvada,
Colo.). The soil was prepared using ERA's
semivolatile blank soil matrix. This matrix was a
40% clay topsoil that had been dried,  sieved, and
homogenized. Particle size was 60 mesh and
smaller. The samples, also referred to as
performance evaluation (PE) samples, contained
known levels of TNT and RDX. The  concentrations
that were evaluated contained 10,  50,  100, 250, and
500 mg/kg of each analyte. Prior to the
demonstration, ORNL analyzed the spiked samples
to confirm the concentrations. The method used was
a modified Method 8330, similar to the reference
laboratory method described in Section 4. For the
demonstration, four replicates were prepared at each
concentration level.

Blank soil samples were evaluated to determine the
technology's ability to identify samples with no
contamination (i.e., to ascertain the false positive
error rate). The soil was collected in Monroe
County, Tennessee, and was certified by ORNL to
be free of contamination prior to verification testing.
A reasonable number of blanks (N = 20) was chosen
to balance the uncertainty for estimating the false
positive error rate and the required number of blank
samples to be measured.

Soil Sample Preparation
A few weeks prior to the demonstration, all of the
soil samples were shipped in plastic Ziplock bags at
ambient temperature to ORNL. The samples were
stored frozen (<0°C) prior to preparation. To ensure
that the developers and the reference laboratory
analyzed comparable samples, the soils were
homogenized prior to sample splitting. The process
was as follows. The sample was kneaded in the
Ziplock bag to break up large clumps. Approximately
1500 g of soil was poured into a Pyrex pan, and
debris was removed. The sample  was then air-dried
overnight. The sample was sieved using a 10-mesh
(2-mm particle size) screen and placed in a 1-L
widemouthed jar. After thorough mixing with a metal
spatula, the sample was quartered. After mixing
each quarter, approximately 250 g from each quarter
was placed  back in the 1-L widemouthed jar, for a
total sample amount of approximately 1000 g.
Analysis by the ORNL method confirmed sample
homogeneity (variability of 20% relative standard
deviation or less for replicate measurements). The
sample was then split into subsamples for analysis
during the demonstration. Each 4-oz sample jar
contained approximately 20 g of soil. Four replicate
splits of each soil sample were prepared for each
participant.  The design included a one-to-one pairing
of the replicates, such that the vendor and reference
lab samples could be directly matched. Three
replicate sets of samples were also prepared for
archival storage. To ensure that degradation did not
occur, the soil samples were frozen (<0°C) until
analysis (Maskarinec et al.  1991).

-------
Water Sample Descriptions
Sources of Samples
Explosives-contaminated water samples from
Tennessee, Oregon, and Louisiana were analyzed.
The contamination in the water samples ranged in
concentration from 0 to about 25,000 |Jg/L. Water
samples were collected from LAAAP, MLAAP,
and Volunteer, described in the previous section (see
"Sources of Samples"). Water samples were also
obtained from Umatilla Chemical Depot, described
below.

Umatilla Chemical Depot is located in northeastern
Oregon. The mission of the facility recently changed
to storage of chemical warfare ammunition. Once
the chemicals are destroyed, the installation is
scheduled to close. Several environmental sites have
been identified for cleanup prior to base closure.
One site has explosives-contaminated groundwater;
the cleanup identified for this site is to pump and
treat the water with granulated activated carbon.
The major contaminants in these samples were
TNT, RDX, HMX, and TNB. According to a
remedial investigation conducted at the site, these
samples were not contaminated with any chemical
warfare agents.

Performance Evaluation Samples
Water samples of known concentration were
prepared by the U.S. Army Cold Regions Research
and Engineering Laboratory (CRREL) in Hanover,
New Hampshire. These samples were used to
determine the technology's  accuracy. The
concentrations of TNT and RDX in the spiked
distilled water samples were 25, 100, 200, 500, and
1000 |jg/L for each analyte; four replicates were
prepared at each concentration. Prior to the
demonstration, ORNL analyzed the spiked samples
to confirm the concentrations.

Distilled water obtained from ORNL was used for
the blanks. As with the soil samples, 20 blank
samples were analyzed.

Water Sample Preparation
The water samples were collected in 2.5-gal carboys
approximately 7 to 10 days prior to the start of the
demonstration and shipped on ice to ORNL. To
ensure that degradation did not occur, the samples
were stored under refrigeration until analysis (~4°C)
(Maskarinec et al. 1999). Sample splitting was
performed in a small laboratory cold room, which
was maintained at 4°C. To prepare the water
sample, a spout was attached to the 2.5-gal carboy,
and the water sample was split by filling multiple
250-mL amber glass bottles. As with the soil
samples, four replicate splits of each water sample
were prepared  for each participant, and three sets of
samples were also prepared for archival storage.

Sample Randomization
The samples were randomized in two stages. First,
the order in which the filled jars were distributed
was randomized so  that the same developer did not
always receive  the first jar filled for a given sample
set. Second, the order of analysis was randomized so
that each participant analyzed the same set of
samples, but in a different order. Each jar was
labeled with a sample number. Replicate samples
were assigned unique (but not sequential) sample
numbers.  Spiked materials and blanks were labeled
in the same manner, such that these quality control
samples were indistinguishable from other  samples.
All samples were analyzed blindly by both the
developer and the reference laboratory.

Summary of Experimental Design
The distribution of samples from the various sites is
described in Table 1. A total of 108 soil samples
were analyzed, with approximately 60% of the
samples being naturally contaminated environmental
soils, and the remaining 40% being spikes and
blanks. A total  of 176 water samples were  analyzed,
with approximately  75% of the samples being
naturally contaminated environmental water, and the
remaining 25% being spikes and blanks. Four
replicates were analyzed for each sample type. For
example, four replicate splits of each of three Fort
Ord soils were  analyzed, for a total of 12 individual
Fort Ord samples.

Description of Performance Factors
In Section 5, technology performance is evaluated in
terms of precision, accuracy, completeness, and
comparability, which are indicators of data  quality
(EPA 1998). False positive and negative results,
sample throughput,  and ease of use are also
evaluated. Each of these performance
characteristics is defined in this section.

-------
   Table 1.  Summary of Soil and Water
             Samples
Sample
source or
type
Fort Ord
Iowa
LAAAP
MLAAP
Umatilla
Volunteer
Spiked
Blank
Total
No. of soil
samples
12
4
16
20
0
12
24
20
108
No. of water
samples
0
0
80
20
24
8
24
20
176
Precision
Precision is the reproducibility of measurements
under a given set of conditions. Standard deviation
(SD) and relative standard deviation (RSD) for
replicate results are used to assess precision, using
the following equation:

  RSD = (SD/average concentration) x 100% .
                                          (Eq. 1)
The overall RSD is characterized by three summary
values:
•  mean — i.e., average;
•  median — i.e., 50th percentile value, at which
   50% of all individual RSD values are below and
   50% are above; and
•  range — i.e., the highest and lowest RSD values
   that were reported.
The average RSD may not be the best
representation of precision, but it is reported for
convenient reference. RSDs greater than  100%
should be viewed as indicators of large variability
and possibly non-normal distributions.

Accuracy
Accuracy represents the closeness of the tech-
nology's measured concentrations to known (in this
case, spiked/PE) values.  Accuracy is assessed in
terms of percent recovery, calculated by the
following equation:

   % recovery = (measured concentration/
    known concentration)  x 100% .
                                          (Eq. 2)
As with precision, the overall percent recovery is
characterized by three summary values: mean,
median, and range.

False Positive/Negative Results
A false positive (fp) result is one in which the
technology detects explosives in the sample when
there actually are none (Berger, McCarty, and Smith
1996). A false negative (fn)  result is one in which
the technology indicates that no explosives are
present in  the sample, when there actually are
(Berger, McCarty, and Smith 1996). The evaluation
of fp and fn results is influenced by the actual
concentration in the sample and includes an
assessment of the reporting limits of the technology.
False positive results are assessed in two ways.
First, the results are assessed relative to the blanks
(i.e., the technology reports  a detected value when
the sample is a blank). Second, the results are
assessed on environmental and spiked samples
where the analyte was not detected by the reference
laboratory (i.e., the reference laboratory reports a
nondetect  and the  field technology reports a
detection). False negative results,  also assessed for
environmental and spiked samples, indicate the
frequency  that the technology reported a nondetect
(i.e., < reporting limits) and  the reference laboratory
reported a detection.  Note that the reference
laboratory results were confirmed by the ORNL
laboratory so that fp/fn assessment would not be
influenced by faulty laboratory data. The reporting
limit is considered in the evaluation. For example, if
the reference laboratory reported a result as
0.9 mg/kg, and the technology's paired result was
reported as below reporting limits (<1 mg/kg), the
technology's result was considered correct and not a
false negative result.

Completeness
Completeness is defined as the percentage of
measurements that are judged to be usable (i.e., the
result is not rejected). The acceptable completeness
is 95% or  greater.

-------
Comparability
Comparability refers to how well the field technology
and reference laboratory data agree. The difference
between accuracy and comparability is that whereas
accuracy is judged relative to a known value,
comparability is judged relative to the results of a
standard or reference procedure, which may or may
not report the results accurately. A one-to-one
sample comparison of the technology results and the
reference laboratory results is performed in
Section  5.

A correlation coefficient quantifies the linear
relationship between two  measurements (Draper
and Smith 1981). The correlation coefficient is
denoted by the letter r; its value ranges from -1 to
+1, where 0 indicates the  absence of any linear
relationship. The value r = -1 indicates a perfect
negative linear relation (one measurement decreases
as the second measurement increases); the value r  =
+1 indicates a perfect positive linear relation (one
measurement increases as the second measurement
increases). The slope of the linear regression line,
denoted by the letter m, is related to r. Whereas r
represents the linear association between the vendor
and reference laboratory concentrations, m
quantifies the amount of change in the vendor's
measurements relative to  the reference laboratory's
measurements. A value of+1 for the slope indicates
perfect agreement. Values greater than 1 indicate
that the vendor results are generally higher than the
reference laboratory, while values less than 1
indicate that the vendor results are usually lower
than the reference laboratory.  In addition, a direct
comparison between the field technology and
reference laboratory data is performed by evaluating
the percent difference (%D) between the measured
concentrations, defined as

%D  = ([field technology] - [reflab])/(reflab)
      x 100%                             (Eq. 3)

The  range of %D values is summarized and reported
in Section 5.

Sample Throughput
Sample throughput is a measure of the number of
samples that can be processed and reported by a
technology in a given period of time. This is reported
in Section 5  as number of samples per hour times the
number of analysts.

Ease of Use
A significant factor in purchasing an instrument or a
test kit is how easy the technology is to use.
Several factors are evaluated and reported on in
Section 5:

•   What is the required operator skill level (e.g.,
    technician, B.S., M.S., or Ph.D.)?
•   How many  operators were used during the
    demonstration? Could the technology be run by a
    single person?
•   How much training would be required in order to
    run this technology?
•   How much subjective decision-making is
    required?

Cost
An important factor in the consideration of whether
to purchase a technology is cost. Costs involved with
operating the technology and the standard reference
analyses are estimated in Section 5. To account for
the variability in cost data and assumptions, the
economic analysis is presented as a list of cost
elements and a range of costs for sample analysis.
Several factors affect the cost of analysis. Where
possible, these factors are addressed so that decision
makers can independently complete a site-specific
economic analysis to suit  their needs.

Miscellaneous Factors
Any other information that might be useful to a
person who is considering purchasing the technology
is documented in Section 5. Examples of information
that might be useful to a prospective purchaser are
the amount of hazardous waste generated during the
analyses, the ruggedness of the technology, the
amount of electrical or battery power necessary to
operate the technology, and aspects of the
technology or method that make it user-friendly or
user-unfriendly.

-------
                Section 4 — Reference Laboratory Analyses
Reference Laboratory Selection
The verification process is based on the presence of
a statistically validated data set against which the
performance goals of the technology may be
compared. The choice of an appropriate reference
method and reference laboratory are critical to the
success of the demonstration. To assess the
performance of the explosives field analytical
technologies, the data obtained from demonstration
participants were compared to data obtained  using
conventional analytical methods. Selection of the
reference laboratory was based on the experience of
prospective laboratories with QA procedures,
reporting requirements, and data quality parameters
consistent with the goals of the program. Specialized
Assays, Inc. (currently part of Test America, Inc.),
of Nashville, Tennessee, was selected to perform
the analyses based on ORNL's experience with
laboratories capable of performing explosives
analyses using EPA SW-846 Method  8330. ORNL
reviewed Specialized Assays' record of laboratory
validation performed by the U.S. Army Corps of
Engineers (Omaha, Nebraska). EPA and ORNL
decided that, based on the credibility of the Army
Corps program and ORNL's prior experience with
the laboratory, Specialized Assays would be  selected
to perform the reference analyses.

ORNL conducted an audit of Specialized Assays'
laboratory operations on May 4, 1999.  This
evaluation focused specifically on the  procedures
that would be used for the analysis of the
demonstration samples. Results from this audit
indicated that Specialized Assays was  proficient in
several areas, including quality management,
document/record control, sample control, and
information management. Specialized Assays was
found to be compliant with implementation of
Method 8330 analytical procedures. The company
provided a copy  of its QA plan, which details all of
the QA and quality control (QC) procedures for all
laboratory operations (Specialized Assays 1999).
The audit team noted that Specialized Assays had
excellent procedures in place for data  backup,
retrievability, and long-term storage. ORNL
conducted a second audit at Specialized Assays
while the analyses were being performed. Since the
initial qualification visit, management of this
laboratory had changed because Specialized Assays
became part of Test America. The visit included
tours of the laboratory, interviews with key
personnel, and review of data packages. Overall, no
major deviations from procedures were observed
and laboratory practices appeared to meet the QA
requirements of the technology demonstration plan
(ORNL 1999).

Reference Laboratory Method
The reference laboratory's analytical method,
presented in the technology demonstration plan,
followed the guidelines established in EPA SW-846
Method 8330  (EPA 1994). According to Specialized
Assays' procedures, soil samples were prepared by
extracting 2-g samples of soil in acetonitrile by
sonication for approximately 16 h. An aliquot of the
extract was then combined with a calcium chloride
solution to precipitate out suspended particulates.
After the solution was filtered, the filtrate was ready
for analysis. For the water samples, 400 mL of
sample were combined with sodium chloride and
acetonitrile in a separatory funnel. After mixing and
allowing the solutions to separate, the bottom
aqueous layer was discarded and the organic layer
was collected. The acetonitrile volume was reduced
to 2 mL, and the sample was diluted with 2 mL of
distilled water for a final volume of 4 mL. The
sample was then ready for analysis. The analytes
were identified and quantified using a high-
performance liquid chromatograph (HPLC) with a
254-nm UV detector. The primary analytical column
was a C-18 reversed-phase column with
confirmation by a secondary cyano column. The
practical quantitation limits were 0.5 |jg/L for water
and 0.5 mg/kg for soils.

Reference Laboratory Performance
ORNL validated all of the reference laboratory data
according to the procedure described in the
demonstration plan (ORNL 1999). During the
validation, the following aspects of the data were
                                                10

-------
reviewed: completeness of the data package,
adherence to holding time requirements, correctness
of the data, correlation between replicate sample
results, evaluation of QC sample results, and
evaluation of spiked sample results. Each of these
categories is described in detail in the demonstration
plan. The reference laboratory reported valid results
for all samples, so completeness was 100%.

Preanalytical holding time requirements for water
(7 days to extract; 40 days to analyze) and soil (14
days to extract; 40 days to analyze) were met. A
few errors were found in a small portion of the data
(-4%). Those data were corrected for transcription
and calculation errors that were identified during the
validation. One data point, a replicate Iowa soil
sample, was identified as suspect. The result for this
sample was 0.8 mg/kg; the results from the other
three replicates averaged 27,400 mg/kg. Inclusion or
exclusion of this data point in the evaluation of
comparability with the field technology (reported in
Section 5) did not significantly change the r value, so
it was included in the analysis.  The reference
laboratory results for  QC samples were flagged
when the results were outside the QC  acceptance
limits.

The reference laboratory results were evaluated by
a statistical analysis of the data. Due to the limited
results reported for the other Method 8330 analytes,
only the results for the major constituents in the
samples (total DNT, TNT, RDX, and HMX) are
evaluated in this report.
      Table 2. Summary of the Reference Laboratory Performance for Soil Samples
Statistic
Mean
Median
Range
Accuracy
(% recovery)
RDX
N = 20
102
99
84-141
TNT
N = 20
100
96
76-174
Precision"
(% RSD)
DNT*
NR=3C
56
32
14-123
HMX
NR=13
29
30
12-63
RDX
N= 13
25
21
4-63
TNT
NR=18
29
25
2-72
      "Calculated from those samples where all four replicates were reported as a detect.
      6DNT represents total concentration of 2,4-DNT and 2,6-DNT.
      CNR represents the number of replicate sets; N represents the number of individual samples
           Table 3.  Summary of the Reference Laboratory Performance for Water Samples
Statistic
Mean
Median
Range
Accuracy
(% recovery)
RDX
N = 20
91
87
65-160
TNT
N = 20
91
91
66-136
Precision"
(% RSD)
DNT*
NR=7C
30
30
8-80
HMX
NR=20
20
17
6-49
RDX
NR = 29
22
17
5-66
TNT
NR=28
24
20
5-86
           "Calculated from those samples where all four replicates were reported as a detect.
           6 DNT represents total concentration of 2,4-DNT and 2,6-DNT.
           CNR represents the number of replicate sets; N represents the number of individual samples
                                                   11

-------
The accuracy and precision of the reference
laboratory results for soil and water are summarized
in Tables 2 and 3, respectively. Accuracy was
assessed using the spiked samples, while precision
was assessed using the results from both spiked and
environmental samples. The reference laboratory
results were unbiased (accurate) for both soil and
water, as mean percentage recovery values were
near 100%. The reference laboratory results were
precise; all but one of the mean RSDs were less
than or equal to30%. The one mean RSD that was
greater than  30% (soil, DNT, 56%) was for a limited
data set of three.
Table 4 presents the laboratory results for blank
samples. A false positive result is identified as any
detected result on a known blank. The
concentrations of the false positive water results
were low (<2 |Jg/L). For the soil samples, one false
positive detection appeared to be a preparation error
because the concentration was near 70,000 mg/kg.
Overall, it was concluded that the reference
laboratory results were unbiased, precise, and
acceptable for comparison with the field analytical
technology.
       Table 4.  Summary of the Reference Laboratory Performance for Blank Samples
Statistic
Number of data points
Number of detects
% of fp results
Soil
DNT
20
0
0
HMX
20
0
0
RDX
20
0
0
TNT
20
2
10
Water
DNT
20
1
5
HMX
20
0
0
RDX
20
2
10
TNT
20
4
20
                                                 12

-------
                      Section 5 — Technology Evaluation
Objective and Approach
The purpose of this section is to present a statistical
evaluation of the GC-IONSCAN data and determine
the instrument's ability to measure explosives-
contaminated soil and water samples. The
technology's precision and accuracy performance
are presented for RDX and TNT only. Barringer
reported detectable data for other Method 8330
analytes (such as HMX, DNT, and TNB) in some of
the samples, but the amount of data available was
insufficient for evaluation.

This section also evaluates comparability through a
one-to-one comparison with the reference laboratory
data. Other aspects of the technology (such as cost,
sample throughput, hazardous waste generation, and
logistical operation) are also evaluated in this section.
The Appendices contain the raw data provided by
the vendor that were used to assess the
performance of the GC-IONSCAN.

Precision
Precision is the reproducibility of measurements
under a given set of conditions. Precision was
determined by examining the results of blind
analyses for four replicate samples. Data were
evaluated only for those samples where all four
replicates were reported as a detection. For
example, for RDX, NR = 13 represents a total of 52
sample analyses (13 sets of four replicates). A
summary of the overall precision of the GC-
IONSCAN for both the soil and water sample
results is presented in Table 5. For the soil samples,
the mean RSDs for RDX and TNT were 54% and
51%, respectively. For water analyses, the RSDs
were significantly lower, at 20% and 26%,
respectively, indicating that the water analyses were
more precise than the soil analyses.

Accuracy
Accuracy represents the closeness of the GC-
lONSCAN's measured concentrations to the known
content of spiked samples. A summary of the GC-
IONSCAN's overall accuracy for both the soil and
water results is presented in Table 6.

For the soil samples, the recoveries for RDX were
highly variable, ranging from 24 to 675%. The mean
recovery of 92%, suggesting that the RDX results
are unbiased, is deceiving because the mean is highly
influenced by one extreme value of 675%. Without
this extreme value, the mean recovery was 62%.
The median recovery of 55% is a more robust
measure of the central tendency of the recovery
data. Overall, the RDX results were generally biased
low. For the TNT soil results, the mean recovery
was 220%. Because the mean and median
recoveries were both greater than 100%, most of the
TNT soil results were biased high. Based on the
                      Table 5.  Summary of the GC-IONSCAN Precision
Statistic
Mean
Median
Range
Soil RSD a
RDX
NR = 13 "
54
43
6-147
TNT
NR=13
51
42
22-133
WaterRSD fl
RDX
NR = 3
20
23
13-25
TNT
NR=12
26
27
7^6
                      " Calculated only from those samples where all four replicates were reported as a
                      detect.
                      b NR represents the number of replicate sets

                                               13

-------
                  Table 6.  Summary of the GC-IONSCAN Accuracy
Statistic
Mean
Median
Range
Soil recovery (%)
RDX
N = 20
92
55
24-675
TNT
N = 20
220
136
52-1620
Water recovery (%)
RDX
N=l
n/a
n/a
8a
TNT
N = 6
29
26
19-46
                  1 Recovery for the one RDX result reported for spiked samples.
            Table 7. Number of GC-IONSCAN Results within Acceptance Ranges for
                     Spiked Soils
Spike
concentration
(mg/kg)
10
50
100
250
500
RDX
Acceptanc
e range
(mg/kg)
8-11
38-57
76-113
190-283
379-566
No. of results
within range
0
0
1
0
3
TNT
Acceptanc
e range
(mg/kg)
7-13
35-63
70-126
174-315
348-630
No. of results
within range
2
2
2
0
1
performance acceptance ranges shown in Table 7,
which are the guidelines established by the provider
of the spiked materials to gauge acceptable
analytical results, 20% of the results (4 of 20) met
the acceptance criteria for RDX, while 35% (7 of 20
of the results) met the criteria for TNT.

For water analyses, the recoveries were low, with
the highest recovery at 46% for TNT. Only a few of
the RDX and TNT results were actually reported as
detects in the spiked samples. This was partially
because of the high reporting limits (up to 102 ug/L)
and also because the water results were biased low
for both analytes.
False Positive/False Negative
Results
Table 8 shows the GC-IONSCAN performance for
blank samples. The table includes DNT and HMX
results because there was a sufficient amount of
data available to evaluate false positive (fp)/false
negative (fn) results. Of the 20 blank soils, Barringer
reported DNT in 2 samples (10% of the samples),
RDX in 1 sample (5%), and TNT in 5 samples
(25%). No fp results were reported for DNT, RDX,
or TNT in the water samples. The fact that no fp
results were reported is not surprising, given the low
bias and the high reporting limits for the water
results.

Table 9 summarizes the GC-IONSCAN's fp and fn
results for all spiked and environmental samples by
comparing the GC-IONSCAN result with the
                                               14

-------
      Table 8. Summary of GC-IONSCAN False Positives on Blank Soil and Water Samples
Statistic
Number of data points
Number of detects
% of fp results
Soil
DNT
20
2
10%
HMX
20
0
0
RDX
20
1
5%
TNT
20
5
25%
Water
DNT
20
0
0
RDX
20
0
0
TNT
20
0
0
         Table 9. Summary of the GC-IONSCAN Detect/Nondetect Performance Relative to
                  the Reference Laboratory Results
Statistic
Number of
data points
Number of
fp results
% of fp
results
Number of
fn results
% of fn
results
Soil
DNT
88
3
3%
5
6%
HMX
88
1
1%
46
52%
RDX
88
o
5
3%
2
2%
TNT
88
0
0
11
13%
Water
DNT
156
0
0
0
0
RDX
156
0
0
61
39%
TNT
156
3
2%
32
21%
reference laboratory result. (See Section 3 for a
more detailed discussion of this evaluation.) For the
soils, 3% or less of the DNT, HMX, and RDX
results were reported as false positives relative to
the reference laboratory  results (i.e., the laboratory
reported the analyte as a nondetect when Barringer
reported it as a detect). Barringer reported a larger
fraction of the samples as nondetects (i.e., false
negatives) when the laboratory reported a detect.
 The highest number of fn results occurred for
HMX, where 52% of Barringer's results were
reported as nondetects. Similar results were
observed for the water samples, where 39% of the
RDX results and 21% of the TNT results were false
negatives.

For the water samples, the high number of fh results
appeared to be due primarily to the high reporting
limits, and because the GC-IONSCAN results were
biased low. For the soil samples, Barringer did not
report any false negative results for RDX when
RDX was present as the highest concentration (i.e.,
primary) analyte in the sample. However, false
negatives for the other nonprimary analytes were
observed. A similar trend was observed for TNT; no
false negatives were observed for TNT when it was
the primary analyte, but false negatives for the other
nonprimary components were reported.

Completeness
Completeness is defined as the percentage of
measurements that are judged to be usable (i.e., the
result was not rejected). Valid results were obtained
by the technology for all 108 soil samples and all  176
water samples. Therefore, completeness was 100%.

Comparability
Comparability refers to how well the GC-IONSCAN
and reference laboratory data agreed. A one-to-one
sample comparison of the GC-IONSCAN results
                                               15

-------
and the reference laboratory results was performed
for all environmental and spiked samples that were
reported above the reporting limits. In Tables 10 and
11, the comparability of the results are presented in
terms of correlation coefficients (r) and slopes (m).

The r value for the comparison of the entire soil data
set of TNT results was 0.88 (m = 4.82). Note that
including or excluding the unusual reference
laboratory value does not cause this value to vary
much. As shown in Table 10, if comparability is
assessed for specific concentration ranges, such as
isolating those values less than 500 mg/kg, the r
value for TNT does not change dramatically, ranging
from 0.71 to 0.85 depending on the concentrations
selected.

Figure 2 presents a plot of the GC-IONSCAN TNT
results versus those for the reference laboratory for
concentrations less than 500 mg/kg. The solid line on
the graph is a representation of a one-to-one
correspondence between the two measurements,
while the dashed line is the linear regression line. As
this figure indicates, the GC-IONSCAN soil
measurements were generally higher than the
reference laboratory results.
     Table 10. GC-IONSCAN Correlation with Reference Data for Various Vendor Soil
                Concentration Ranges
Concentration
range
All values "
< 500 mg/kg*
< 1,000 mg/kg
> 1,000 mg/kg
> 10,000 mg/kg
RDX
r
0.79
0.79
0.89
0.28
n/ac
m
0.54
0.52
0.27
0.14
n/a
TNT
r
0.88
0.71
0.77
0.85
0.85
m
4.82
0.76
0.96
5.32
5.55
     * Excluding those values reported as "< reporting limits" and including the one reference laboratory unusual value. (See
     Section 4 for more information on the unusual value.)
     6 Based on Barringer's reported values.
     c No RDX values above 10,000 mg/kg. were reported.
      Table 11. GC-IONSCAN Correlation with Reference Data for Various Vendor Water
                 Concentration Ranges
Concentration
range
All values"
<500|jg/LA
> 500 ng/L
> 1,000 ugL
RDX
r
0.95
0.82
0.87
-0.90
m
0.08
0.07
0.06
-0.08
TNT
r
0.53
0.67
0.32
-0.19
m
0.42
0.17
0.25
-0.11
      ' Excluding those values reported as "< reporting limits."
      ' Based on Barringer's reported values.
                                                  16

-------
     600
                                                           r=0.71
                                                                         one-to-one correspondence line
                     100           200           300           400          500

                             Reference Laboratory TNT Soil Results (mg/kg)
                                                                                        600
Figure 2.        Comparability of reference laboratory TNT soil results with GC-IONSCAN results for vendor concentrations
               less than 500 mg/kg. The slope of the linear regression line is 0.76 and the intercept is 55 mg/kg.
The correlation of the RDX values reported for the
soil samples with those reported by the reference
laboratory was similar to the correlation for TNT(r
near 0.80), except for concentrations greater than
1000 mg/kg, where the correlation was lower (r =
0.28, m = 0.14). A close examination of the data
from specific sites indicated that there were no
differences in performance for the various  matrices
(i.e., there were no matrix-dependent effects).

Overall, the GC-IONSCAN soil results were
generally lower than those of the reference
laboratory for RDX and higher than those of the
reference laboratory for TNT.

For the water samples, comparability of the GC-
IONSCAN results with the reference laboratory
results for TNT was much lower than for soil.  As is
shown in Table 11, the correlation coefficient was
0.53 (m = 0.42) when all results were considered.
For RDX, the correlation was much higher, at  0.95
with a linear regression slope of 0.08. While the GC-
IONSCAN correlation indicated a linear relationship
with the reference laboratory (see Figure 3), the
RDX results were lower than the reference
laboratory, as indicated by the ideal one-to-one
correspondence solid line shown in the figure, and
also by the y-axis scale maximum value of
2,000 |jg/L and the x-axis scale maximum value of
25,000 ng/L. Overall, the GC-IONSCAN water
results were generally lower than the reference
laboratory for both RDX and TNT.

Another metric of comparability is the percent
difference (%D) between the reference laboratory
and the GC-IONSCAN results. The ranges of %D
values for TNT and RDX are presented in Figures 4
and 5, respectively. Acceptable %D values would be
between -25% and 25%, or near  the middle of the
x-axis of the plots. For TNT, the  %D values for soil
were mostly greater than 25%, and %D values for
water results were mostly -25%  and below,
supporting the conclusions that the TNT soil results
were generally higher than the reference laboratory
results and the TNT water results were generally
lower than those of the reference laboratory.
                                                 17

-------
        2,000 -i
                                                          ._- '~r = 0.95
                                                              - one-to-one correspondence line

                                                              linear regression line
                            5,000           10,000           15,000           20,000
                                 Reference Laboratory RDX Water Results (jig/L)
                                        25,000
 Figure 3. Comparability of reference laboratory RDX water results with GC-IONSCAN results for all results reported
          above reporting limits. The slope of the linear regression line is 0.08, and the intercept is 51 |ig/L.
As shown in Figure 5, most of the %D values for
RDX were negative, supporting the conclusion that
the GC-IONSCAN RDX results, particularly for
water, were lower than the reference laboratory.

Sample Throughput
Sample throughput is representative of the estimated
amount of time required to prepare and analyze the
sample. Operating under the outdoor conditions, the
two-person Barringer team accomplished a sample
throughput rate of approximately three samples per
hour for soil and eight samples per hour for water.

Ease of Use
Two operators were used for the demonstration
because of the number of demonstration samples
and working conditions, but the technology can be
operated by a single person. Users unfamiliar with
ion mobility spectrometry would need approximately
two days of training to operate the
GC-IONSCAN. Barringer Instruments provides
training with instrument installation. No particular
level of educational training is required for the
operator, but skills in chromatographic techniques
would be advantageous.

Cost Assessment
The purpose of this economic analysis is to estimate
the range of costs for analysis of explosives-
contaminated soil and water samples using the GC-
IONSCAN and a conventional analytical reference
laboratory method. The analysis was based on the
results and experience gained from this
demonstration, costs provided by Barringer, and
representative costs provided by the reference
analytical  laboratories that offered to analyze these
samples. To account for the variability in cost data
and assumptions, the economic analysis is presented
as a list of cost elements and a range of costs for
sample analysis by the GC-IONSCAN instrument
and by the reference laboratory.
                                                 18

-------
                            Range of percent differences values
Figure 4. Range of percent difference values for TNT in soil and water.
Several factors affected the cost of analysis. Where
possible, these factors were addressed so that
decision makers can complete a site-specific
economic analysis to suit their needs. The following
categories are considered in the estimate:

•   sample shipment costs,
•   labor costs,
•   equipment costs, and
•   waste disposal costs.

Each of these cost factors is defined and discussed
and serves as the basis for the estimated cost ranges
presented in Table 12. This analysis assumed that
the individuals performing the analyses were fully
trained to operate the technology. (Note that
Barringer provides training with the purchase and
installation of the GC-IONSCAN instrument.) Costs
for sample acquisition and pre-analytical sample
preparation,  which are tasks common to both
methods, were not included in this assessment.

GC-IONSCAN  Costs
The costs associated with using the GC-IONSCAN
instrument included labor, equipment, and waste
disposal costs. No sample shipment charges were
associated with the cost of operating the instrument
because the samples were analyzed on-site.

Labor
Labor costs included mobilization/demobilization,
travel, per diem expenses and on-site labor.

•  Mobilization/demobilization. This cost element
   included the time for one person to prepare for
   and travel to each site. This estimate ranged
   from 5 to 8 h, at a rate of $50/h.
•  Travel. This element was the cost for the
   analyst(s) to travel to the site. If the analyst is
   located near the site, the cost of commuting to
   the site (estimated to be 50 miles at $0.30/mile)
   would be minimal ($15).  The estimated cost of
   an analyst traveling to the site for this
   demonstration ($1000) included the cost of
   airline travel and rental car fees.
•  Per diem expenses.  This cost element included
   food, lodging, and incidental expenses. The
   estimate ranged from zero (for a local site) to
   $150/day for each analyst.
                                                 19

-------
       30  -i
                               Range of percent difference values
Figure 5. Range of percent difference values for RDX in soil and water.
•   Rate. The cost of the on-site labor was
    estimated at a rate of $30-75/h, depending on
    the required expertise level of the analyst. This
    cost element included the labor involved with the
    entire analytical process, comprising sample
    preparation, sample management, analysis, and
    reporting.

Equipment
Equipment costs included mobilization/
demobilization, rental fees or purchase of equipment,
and the reagents and other consumable supplies
necessary to complete the analysis.

•   Mobilization/demobilization.  This included the
    cost of shipping the equipment to the test site. If
    the site is local, the cost would be zero. For this
    demonstration, the cost of shipping equipment
    and supplies was  estimated at $150.
•   Instrument purchase. At the time of the
    demonstration, the cost of purchasing the GC-
    IONSCAN was $60,000. The purchase included
    a laptop  computer, four kits (a standard spare
    kit, a standard maintenance kit, a standard
    consumables kit, and a sampling kit), a swab
    sampler, IM software, an operators' manual, and
    training at Barringer for up to six people. The kit
    is supplied in an aluminum carrying case. The
    instrument can be operated in two modes;
    explosives detection and drug detection. As with
    any instrument, the cost on a per-sample basis
    would decrease as the number of analyses
    performed increases.
•   Reagents/supplies. These items are consumable
    and are purchased on a per sample basis. At the
    time of the demonstration, the cost of the
    reagents and supplies needed to prepare and
    analyze explosives samples using the GC-
    IONSCAN was $1 per sample. This cost
    included the sample preparation supplies, assay
    supplies, and consumable reagents.

Waste Disposal
Waste disposal costs are based on the 1999
regulations for disposal of explosives-contaminated
waste. The analyses performed using the GC-
IONSCAN instrument generated approximately 70
Ib of vials (two 5-gal buckets) containing soils and
liquid solvents. ORNL's cost to dispose of the
explosives-contaminated waste at a commercial
facility was estimated at $90 per 5-gal bucket. There
are most likely additional costs for labor associated
with the waste disposal, but those costs are not
estimated here.
                                                20

-------
Reference Laboratory Costs
Sample Shipment
Sample shipment costs to the reference laboratory
included the overnight shipping charges, as well as
labor charges associated with the various
organizations involved in the shipping process.

•  Labor. This cost element included all of the
   tasks associated with the shipment of the
   samples to the reference laboratory. Tasks
   included packing the shipping coolers, completing
   the chain-of-custody documentation, and
   completing the shipping forms. The estimate to
   complete this task ranged from 2 to 4 h at $50
   per hour.
•  Overnight shipping. The overnight express
   shipping service cost was estimated to be $50
   for one 50-lb cooler of samples.

Labor, Equipment, and Waste Disposal
The labor bids from commercial analytical reference
laboratories that offered to perform the reference
analysis for this demonstration ranged from $150 to
$188 per sample. The bid was dependent on many
factors, including the perceived difficulty of the
sample matrix, the current workload of the
laboratory, and the competitiveness of the market.
This  rate was a fully loaded analytical cost that
included equipment, labor, waste disposal, and
report preparation.

Cost Assessment Summary
An overall cost estimate for use of the GC-
IONSC AN instrument versus use of the reference
laboratory was not made because of the extent of
variation in the different cost factors, as outlined in
Table 12. The overall costs for the application of
each technology will be based on the number of
samples requiring analysis, the sample type, and the
site location and characteristics. Decision-making
factors, such as turnaround time for results, must
also be weighed against the cost estimate to
determine the value of the field technology's
providing immediate answers versus the reference
laboratory's provision of reporting data within 30
days of receipt of samples.

Miscellaneous Factors
The following are general observations regarding the
field operation and performance of the GC-
IONSCAN instrument:
The system, which weighs approximately 70 Ib,
was easily transported by one person. The
instrument's aluminum carrying case has wheels
and a handle to allow the instrument to be
moved like a large piece of luggage. The case
can also be used to ship the instrument.
The instrument appeared to be rugged, as the
analysts were able to run the instrument during a
late afternoon storm that had strong winds.
The Barringer team completely disassembled
their work station at the close of each day. It
took the Barringer team less than a half-hour
each morning to prepare for sample analyses.
The instrument required 110 V of electrical
power for operation. The manufacturer
estimates that the GC-IONSCAN consumes
approximately 1200 W during initial startup and
approximately 900 W during normal operation.
The GC-IONSCAN was calibrated using a
multiple-point curve, as described in Section 2.
Initially, the operators injected several l-|jL
aliquots of the 4-ng standard to deactivate the
injector port to avoid adherence of the analytes
to the inside walls.
Data processing was performed with Microsoft
Excel rather than the software package used for
data acquisition.
Switching between the IONSCAN screening
mode and the quantitative GC-IONSCAN mode
was easily accomplished by pressing the
appropriate buttons on the screen display. For
the soil samples, the Barringer team screened
the samples in the IONSCAN mode until analyte
was detected. The team then switched to the
GC-IONSCAN mode, quantified the sample,
and switched back to the IONSCAN mode to
continue screening.
The screening operation of the IONSCAN was
simple and quick. Briefly, 1-4 |jL of solution
was added to the filter, and the filter was
desorbed. In approximately 8 s, the IONSCAN
indicated the presence or absence of analyte.
Note that the performance of the IONSCAN is
not evaluated in this report.
                                               21

-------
  Table 12. Estimated analytical costs for explosives-contaminated samples
Analysis method: GC-IONSCAN
Analyst/manufacturer: Barringer Instruments
Sample throughput: 3 samples/h (for soil)
8 samples/h (for water)
Cost category Cost ($)
Sample shipment 0

Labor
Mobilization/demobilization 250-400
Travel 1 5-1 ,000 per analyst
Per diem expenses 0-1 50/day per analyst
Rate 30-75/h per analyst
Equipment
Mobilization/demobilization 0-150
Instrument purchase price 60,000
Reagents/supplies 1 per sample
Waste disposal 180
Analysis method:
Analyst/manufacturer:
Typical turnaround:
Cost category
Sample shipment
Labor
Overnight shipping
Labor
EPA SW-486 Method 8330
Reference laboratory
21 working days
Cost ($)
100-200
50-150

Mobilization/demobilization Included"
Travel
Per diem expenses
Rate
Equipment

Waste disposal
Included
Included
150-1 88 per sample
Included

Included
   ' "Included" indicates that the cost is included in the labor rate.
Summary of Performance
A summary of performance is presented in Table
13. Precision, defined as the mean RSD, was 54%
and 51% for RDX and TNT soil sample results,
respectively. For water analyses, the RSDs were
significantly lower, at 20 and 26%, respectively,
indicating that the water analyses were more precise
than the soil  analyses. Accuracy, defined as the
median percent recovery relative to the spiked
concentration, was 55% and 136% for RDX and
TNT soil sample results, respectively, indicating that
the soil results were generally biased low for
RDX and biased high for TNT. For the spiked water
samples, only a  few values were reported as
detects.  The  water results were biased low, as the
highest percent recovery was only 46%. Of the 20
blank soils, Barringer reported RDX in one sample
(5% false positives) and TNT in five samples (25%
false positives). No false positives were reported for
RDX and TNT in the 20 blank water samples. False
positive and false negative results were determined
by comparing the GC-IONSCAN result to the
reference laboratory result for the environmental and
spiked samples. For the soils, 3% of RDX and none
of the TNT results were reported as false positives
relative to the reference laboratory results.
Significantly more soil samples were reported as
false negatives by Barringer (2% for RDX and 13%
for TNT). Similar results were observed for water,
where a higher percentage of the results were fn
than fp.

The demonstration found that the GC-IONSCAN
instrument was relatively simple for a trained analyst
to operate in the field, requiring less than an hour for
initial setup. The sample throughput of the GC-
IONSCAN was three samples per hour for soil
samples and eight samples per hour for water
samples. Two operators analyzed samples during the
demonstration, but the technology can be run by a
single trained operator. The overall performance of
the GC-IONSCAN for the analysis of RDX and
TNT was characterized as precise and biased low
(both analytes) for water analyses, and imprecise
and biased (low for RDX and high for TNT) for soil
analyses.
                                               22

-------
Table 13. Performance Summary for the GC-IONSCAN
Feature/Parameter
Precision
Accuracy
False positive results on blank
samples
False positive results relative to
reference laboratory results
False negative results relative to
reference laboratory results
Comparison with reference
laboratory results
Completeness
Weight
Sample throughput (2 operators)
Power requirements
Training requirements
Cost
Hazardous waste generation
Overall evaluation
Performance summary
Mean RSD
Soil: 54% RDX 51% TNT
Water: 20% RDX , 26% TNT
Median recovery
Soil: 55% RDX 136% TNT
Water: n/a (see Table 6); 26% TNT
Soil: 10% DNT, 0% HMX 5% RDX 25% TNT
Water: 0% DNT, 0% RDX 0% TNT
Soil: 3% DNT, 1% HMX 3% RDX 0% TNT
Water: 0% DNT, 0% RDX 2% TNT
Soil: 6% DNT, 52% HMX 2% RDX (0% if RDX was the highest concentration
analyte), 13% TNT (0% if TNT was the highest concentration analyte)
Water: 0% DNT, 39% RDX 21% TNT
r (all results) m (all results)
Soil: 0.79 RDX 0.88 TNT 0.54 RDX 4.82 TNT
Water: 0.95 RDX 0.53 TNT 0.08 RDX 0.42 TNT
Median %D range of %D values
Soil: -46% RDX 51% TNT -83 to 562% RDX -70 to 1500% TNT
Water: -90% RDX -59% TNT -97 to -82% RDX -88 to 1310%
TNT
100% of 108 soil samples
100% of 176 water samples
70 Ib
Soil: 3 samples/h
Water: 8 samples/h
110 V, 60 Hz
1200 W startup
900 W normal operations
Two days instrument-specific training
Instrument: $60,000
Supplies per sample: $1
Two 5-gal buckets filled with vials containing acetone extracts of samples,
empty vials, and pipette tips
(Total number of samples analyzed: 108 soils and 176 waters)
Soil: biased low for RDX and biased high for TNT; imprecise
Water: biased low for RDX and TNT; precise
Insufficient data to evaluate performance on DNT and HMX
                                         23

-------
                         Section 6 — Technology Update
In this section, the vendor (with minimal editorial changes by ORNL) provides comments regarding the
use and the performance of the GC-IONSCAN during verification testing and information regarding
new developments with the technology since the verification activities.
Barringer's past experience has been in the area of
rapid prescreening of surfaces for traces of
narcotics or explosives on a semiquantitative level.
This field trial was Barringer's first formal attempt
at the rapid quantitative screening of water and soil
samples. Barringer believes that future
improvements in the simple sample preparation
methodology used to extract explosives from water
and soil will significantly improve agreement with
laboratory-based data.

Rapid Prescreening of Samples
Part of Barringer's method includes using the
IONSCAN mode to rapidly prescreen all the soil
samples before GC-IONSCAN analysis. In the
IONSCAN mode the operator can screen samples
at a much higher throughput rate than in the GC-
IONSCAN mode, since the instrument analysis time
in IONSCAN mode is less than 10s per sample.
During verification testing, if no DNT, HMX, RDX,
or TNT was detected, then GC-IONSCAN analysis
was not performed. IONSCAN mode performance
was not verified by ETV, but Barringer observed
that rapid prescreening during the ETV trial
correctly identified the presence of explosives in
100% of the samples that contained explosives and
was able to screen out 35% of the blank samples
from full GC-IONSCAN analysis.1 With an
appropriate setting of threshold detection levels,
Barringer believes a larger number of blank samples
could have been screened out.

Improvements to GC-IONSCAN
The GC-IONSCAN unit evaluated in this trial was
one of the first prototypes built. Since this trial,
significant improvements have been made to the
GC-IONSCAN instrument and firmware. The
position of thermocouples in the oven has been
changed to ensure that the oven is uniformly heated
and to improve circulation in the unit. An additional
fan has added to allow faster cooling of the unit after
temperature has been ramped; this feature will
increase sample throughput. The carrier gas pressure
reading is now displayed to an extra significant digit
(i.e., 12 to 12.0 PSI) to permit more reproducible
settings of pressure. Upgrading of the firmware of
the unit to improve the peak finding routine for the
GC data should help minimize false positives for
DNT. The temperature controller for the transfer line
between the end of the GC column and the
IONSCAN has been upgraded to permit
temperatures above 190°C. This upgrade will help
minimize sample losses for HMX and decrease the
false negative rate for HMX. Note that samples can
be injected either split or splitless.
  Data analysis performed by Barringer and not ETV-verified.
                                              24

-------
                               Section 7 — References
Berger, W., H. McCarty, and R-K. Smith. 1996. Environmental Laboratory Data Evaluation. Genium
Publishing Corp., Schenectady, N.Y.

Draper, N. R., andH. Smith.  1981. Applied Regression Analysis. 2nd ed. John Wiley & Sons, New York.

EPA (U.S. Environmental Protection Agency). 1994. "Method 8330: Nitroaromatics and Nitramines by High
Performance Liquid Chromatography (HPLC)." In Test Methods for Evaluating Solid Waste: Physical/
Chemical Methods, Update II. SW-846. U.S. Environmental Protection Agency, Washington, D.C.,
September.

EPA (U.S. Environmental Protection Agency). 1998. EPA Guidance for Quality Assurance Project Plans.
EPA QA/G-5, EPA 600/R-98/018. U.S. Environmental Protection Agency, Washington, D.C., February.

Jenkins, T. F., M. E. Walsh, and P. G. Thorne. 1998. "Site Characterization for Explosives Contamination at a
Military Firing Range Impact Area." Special Report 98-9. U.S. Army Cold Regions Research and
Engineering Laboratory, Hanover, N.H. Available at http://www.crrel.usace.army.mil/

Maskarinec, M. P., C. K. Bayne, L. H. Johnson, S. K. Holladay, R. A. Jenkins, and B. A. Tomkins. 1991.
Stability of Explosives in Environmental Water and Soil Samples. ORNL/TM-11770. Oak Ridge National
Laboratory, Oak Ridge, Term., January.

ORNL (Oak Ridge National Laboratory). 1999. "Technology Demonstration Plan: Evaluation of Explosives
Field Analytical Techniques." Oak Ridge National Laboratory, Oak Ridge, Term., August.

Specialized Assays, Inc. 1999. "Comprehensive Quality Assurance Plan." SAL-QC-Rec 5.0. January 6.
                                              25

-------

-------
                 Appendix A

GC-IONSCAN Sample Soil Results Compared with
         Reference Laboratory Results

-------
      Table A-l. GC-IONSCAN Sample Soil Results Compared with Reference Laboratory Results
Is)
Sample site
or type
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Fort Ord
Fort Ord
Fort Ord
Fort Ord
Fort Ord
Fort Ord
Fort Ord
Fort Ord
Fort Ord
Fort Ord
Fort Ord
Fort Ord
Sample
no.
1
1
1
1
2
2
2
2
3
3
3
3
4
4
4
4
5
5
5
5
1
1
1
1
2
2
2
2
3
3
3
3
Sample
replicate
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
DNT«
(mg/kg)
Barringer
<10.0
<10.0
<10.0
<10.0
297.1
<10.0
<10.0
<12.5
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
1163.9
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
RefLab
<0.5
<0.5
<0.5
<0.5
<0.5
<51.0
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
0.5
<0.5
<0.5
<0.5
<0.5
0.5
HMX«
(mg/kg)
Barringer
0.6
O.6
0.6
O.6
O.3
0.6
O.6
O.3
0.6
O.6
0.6
O.6
O.6
0.6
0.6
O.3
O.6
0.6
O.6
0.6
DI»
22
DI
DI
DI
50.1
0.6
DI
DI
28.3
O.6
DI
RefLab
0.5
O.5
0.5
O.5
O.5
<51.0
O.5
O.5
0.5
O.5
0.5
O.5
O.5
0.5
0.5
O.5
O.5
0.5
O.5
0.5
370
252
259
264
278
248
322
185
300
185
392
214
RDX«
(mg/kg)
Barringer
0.6
O.6
0.3
O.6
O.3
0.3
O.3
O.3
0.3
O.3
0.3
18.8
O.6
0.3
0.3
O.3
O.3
0.6
O.3
0.3
O.3
0.3
O.6
0.6
0.6
O.3
0.3
0.3
O.3
0.3
O.3
1.9
RefLab
0.5
O.5
0.5
O.5
O.5
<51.0
O.5
O.5
0.5
O.5
0.5
O.5
O.5
0.5
0.5
O.5
O.5
0.5
O.5
0.5
0.6
0.5
O.5
0.5
0.5
O.5
0.5
0.5
O.5
0.5
O.5
0.5
TNT"
(mg/kg)
Barringer
0.6
1.9
0.6
O.6
O.3
0.9
1
O.3
52.2
O.6
0.6
O.6
O.6
0.6
0.6
O.3
O.6
0.6
1.5
0.6
O.6
0.9
<1.3
<1.3
<1.3
1.6
0.6
1
O.6
0.6
O.6
0.6
RefLab
0.5
O.5
0.5
O.5
O.5
70900
O.5
O.5
0.5
O.5
0.5
O.5
0.9
0.5
0.5
O.5
O.5
0.5
O.5
0.5
O.5
0.8
0.8
0.5
0.8
2.1
0.8
0.8
O.5
0.5
O.5
0.5
Barringer
analysis order6
1090
1077
1066
1101
1013
1032
1031
1001
1051
1027
1055
1036
1086
1026
1070
1007
1021
1095
1064
1048
1065
1025
1083
1078
1079
1030
1038
1060
1057
1017
1046
1087

-------
OJ
o
Sample site
or type
Iowa
Iowa
Iowa
Iowa
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Milan
Milan
Milan
Milan
Milan
Milan
Milan
Milan
Milan
Milan
Milan
Milan
Sample
no.
1
1
1
1
1
1
1
1
2
2
2
2
3
3
3
3
4
4
4
4
1
1
1
1
2
2
2
2
3
3
3
3
Sample
replicate
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
DNT"
(mg/kg)
Barringer
<12.5
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
1069.3
<10.0
<10.0
<10.0
<12.5
<12.5
<10.0
<10.0
<10.0
<10.0
<10.0
RefLab
<51.0
<0.5
<532.0
<50.5
<0.5
<0.5
<0.5
<25.0
<0.5
<0.5
<0.5
0.5
<0.5
<50.0
<50.0
<0.5
80
11.4
11.9
9.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<50.0
<200.0
<0.5
HMX«
(mg/kg)
Barringer
<0.3
0.6
O.6
0.6
69.6
0.6
DI
DI
0.6
O.6
O.6
0.6
O.6
0.6
DI
0.6
0.6
O.6
O.6
0.6
DI
O.3
0.6
O.6
0.6
O.3
0.3
O.6
O.6
88.5
0.6
DI
RefLab
<51.0
0.5
<532.0
<50.5
400
460
202
470
100
204
153
25.6
360
470
390
500
<5.0
O.5
O.5
0.5
22.2
23.2
16.5
42
8.3
3.6
3.5
4.3
330
420
400
343
RDX"
(mg/kg)
Barringer
O.3
0.6
O.6
0.3
2342.7
821.6
2164.8
2414.2
1378.3
1115.5
1177.1
455.4
1833.5
2169.4
3527.9
1195.4
2.1
5
5.5
2.9
41.9
52.3
36.8
85.6
7.4
9.2
50
9.5
3082.3
2775.5
2814.5
2735.6
RefLab
<51.0
0.5
<532.0
<50.5
3460
3520
2140
1900
1180
1450
1170
320
4300
3550
4650
5850
12
10.7
10.8
7.7
149
118
72.2
308
34.8
16.4
28
22.9
2350
1950
4080
3880
TNT"
(mg/kg)
Barringer
30195.6
50310
53214.9
75814.5
178.9
0.6
O.6
0.6
0.6
206
O.6
59.5
303.4
471.9
166.8
271.2
104
162.8
102.5
158.7
0.6
O.3
0.6
O.6
0.6
O.3
0.3
<1.3
319.7
624.8
204.3
205.4
RefLab
20400
0.8
33400
28300
109
120
111
125
50
51
51
10.6
205
170
300
400
89
78
81.5
67.5
2.7
1.1
1.4
1.7
0.5
O.5
0.5
O.5
190
270
320
273
Barringer
analysis order6
1008
1080
1076
1047
1029
1012
1098
1105
1094
1042
1022
1049
1073
1045
1091
1043
1099
1034
1096
1084
1062
1010
1053
1108
1059
1006
1002
1075
1005
1023
1102
1097
        Milan
<10.0
<50.0
42.7
240
893.9
2740
282.6
220
1024

-------
Sample site
or type
Milan
Milan
Milan
Milan
Milan
Milan
Milan
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Volunteer
Volunteer
Volunteer
Sample
no.
4
4
4
5
5
5
5
1
1
1
1
2
2
2
2
3
3
3
3
4
4
4
4
5
5
5
5
6
6
6
6
1
1
1
Sample
replicate
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
DNT"
(mg/kg)
Barringer
<10.0
<10.0
372961
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
566.2
<10.0
<12.5
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<12.5
<10.0
RefLab
<0.5
<0.5
<0.5
3.6
2.7
3.4
3.8
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<25.0
<0.5
<0.5
<50.0
<25.0
19
HMX"
(mg/kg)
Barringer
<0.6
DI
73.4
<0.6
DI
<0.6
<0.6
<0.6
<0.6
<0.6
<0.6
<0.6
<0.6
<0.6
<0.6
DI
<0.6
<0.6
<0.6
<0.3
<0.6
0.3
<0.6
DI
<0.6
<0.6
<0.6
<0.6
<0.6
<0.6
<0.6
<0.6
<0.3
<0.6
RefLab
260
405
222
0.5
O.5
O.5
0.5
0.5
O.5
0.5
O.5
1.5
1.1
1.2
1.4
0.5
0.8
1.1
0.8
0.5
O.5
0.5
O.5
5.9
6.3
8.8
6.9
<25.0
<25.0
<25.0
3.7
<50.0
<25.0
<5.0
RDX"
(mg/kg)
Barringer
3012.3
2005.1
869.3
0.7
O.6
O.6
0.3
0.3
O.3
0.3
O.3
54.8
96.5
149.3
23.8
14.4
32.6
13.5
27.4
5.1
5.4
4.5
3.5
119.3
560.4
382
528.3
126.6
1687
138.7
154
O.3
O.3
0.3
RefLab
2640
2600
3070
0.5
O.5
O.5
0.5
0.5
O.5
0.5
O.5
111
90.5
98
127
49.5
45
63.5
51
9.1
8.4
8.6
9.1
460
455
705
445
260
255
335
250
<50.0
<25.0
<5.0
TNT"
(mg/kg)
Barringer
0.6
207.9
467.5
14.2
5.2
5.6
15.2
180.4
107.9
93.8
143.4
0.6
0.6
O.6
0.6
16.4
17.2
12
11
56.3
119.3
41.2
65.3
597.6
335.8
129.1
343.8
1533
8104.3
793.6
465.4
361997
115876
898117
RefLab
260
80
162
11.5
10.2
11.3
10.6
81.8
104
90
124
0.5
0.5
O.5
0.5
8.4
7.6
10
8.5
47.5
48.5
48.5
47
230
205
435
205
535
505
675
510
108000
75500
117000
Barringer
analysis order6
1016
1081
1039
1074
1089
1104
1054
1033
1072
1015
1063
1035
1019
1092
1068
1061
1100
1071
1107
1003
1044
1011
1069
1058
1018
1088
1020
1040
1014
1082
1093
1037
1009
1050

-------
Sample site
or type
Volunteer
Volunteer
Volunteer
Volunteer
Volunteer
Volunteer
Volunteer
Volunteer
Volunteer
Sample
no.
1
2
2
2
2
3
3
3
3
Sample
replicate
4
1
2
3
4
1
2
3
4
DNT"
(mg/kg)
Barringer
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
RefLab
<250.0
<53.2
<538.0
<5.4
45.2
3.7
3
4.8
2.2
HMX"
(mg/kg)
Barringer
979.1
0.6
<0.6
<0.6
<0.6
<0.6
<0.6
<0.6
<0.6
RefLab
<250.0
<53.2
<538.0
<5.4
<5.4
<0.5
<0.5
<0.5
<0.5
RDX«
(mg/kg)
Barringer
64.9
0.3
O.3
0.3
O.6
8.9
O.3
0.6
O.3
RefLab
<250.0
<53.2
<538.0
6.5
<5.4
0.5
O.5
0.5
O.5
TNT"
(mg/kg)
Barringer
350830
36695.4
25550.9
31579.9
14381.3
9.8
10
21.6
20.9
RefLab
61000
11300
12600
26200
8920
12
10.3
13.8
10.4
Barringer
analysis order6
1004
1028
1052
1056
1106
1103
1067
1085
1041
         The data are presented exactly as reported. Note that the data are not consistently reported with the same number of significant figures.

        b These are the sample numbers, from which the analysis order can be discerned. For example, 1001 was analyzed first, then 1002, etc.

         DI = detected by IONSCAN, but not GC-IONSCAN.
OJ
Is)

-------
                  Appendix B

GC-IONSCAN Sample Water Results Compared with
          Reference Laboratory Results

-------
Table B-l. GC-IONSCAN Sample Water Results Compared with Reference Laboratory Results
Sample site
or type
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Sample no.
6
6
6
6
7
7
7
7
8
8
8
8
9
9
9
9
10
10
10
10
5
5
5
5
6
6
6
6
Sample
replicate
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
DNT"
(|ig/L)
Barringer
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
RefLab
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
0.6
0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
0.5
RDX«
(|ig/L)
Barringer
<51.3
<25.6
<102.5
<102.5
<51.3
<51.3
<102.5
<51.3
<102.5
<102.5
<51.3
<51.3
<51.3
<25.6
<102.5
<102.5
<51.3
<102.5
<102.5
<102.5
<51.3
<102.5
<51.3
<51.3
<102.5
<102.5
<51.3
<51.3
RefLab
1.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
0.5
<0.5
<0.5
<0.5
0.5
0.5
0.5
O.5
0.5
O.5
200
159
180
158
210
238
160
256
TNT"
(Rg/L) J
Barringer
<72.5
<36.3
<58.0
<58.0
<72.5
<72.5
<58.0
<72.5
<58.0
<58.0
<72.5
<72.5
<72.5
<36.3
<58.0
<58.0
<72.5
<58.0
<58.0
<58.0
<72.5
<58.0
<72.5
<72.5
<58.0
67.6
<72.5
<72.5
RefLab
1.9
1.2
O.5
0.5
0.5
0.9
O.5
0.5
O.5
0.5
O.5
1.3
0.5
O.5
0.5
0.5
0.5
O.5
0.5
O.5
170
136
151
138
162
176
113
178
Jarringeranalysis
order6
2066
2007
2133
2132
2039
2063
2147
2071
2131
2153
2061
2097
2045
2012
2134
2142
2055
2168
2145
2121
2077
2146
2042
2052
2151
2172
2051
2032
Louisiana
<1950.0
0.5
<51.3
0.5
<72.5
0.5
2090

-------
Sample site
or type
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Sample no.
7
7
7
8
8
8
8
9
9
9
9
10
10
10
10
11
11
11
11
12
12
12
12
13
13
13
13
14
14
14
14
15
15
15
Sample
replicate
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
DNT"
(|ig/L)
Barringer
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
RefLab
<0.5
<0.5
<0.5
0.5
<0.5
<0.5
<0.5
<50.0
<50.0
<50.0
<50.0
6.8
7.1
8.9
7.6
1.7
0.5
O.5
0.5
152
6
<10.0
183
0.5
O.5
O.5
0.5
0.5
0.7
0.5
O.5
0.5
0.5
O.5
RDX"
(|ig/L)
Barringer
<25.6
<25.6
<102.5
<102.5
<51.3
<51.3
<25.6
209.6
156.4
140
<102.5
<51.3
<51.3
<51.3
<51.3
<25.6
<51.3
<51.3
<51.3
<51.3
<51.3
<51.3
<102.5
<51.3
<102.5
<51.3
<51.3
<51.3
<102.5
<102.5
<51.3
<51.3
<102.5
<102.5
RefLab
0.5
O.5
0.5
0.5
1.1
O.5
0.5
1760
1390
1410
1640
560
470
520
256
18.6
17.2
13
13.9
89
34
52
104
2.5
1.8
2
2.3
11.8
11.4
14
14
0.5
0.5
O.5
TNT"
(Rg/L) ]
Barringer
<36.3
<36.3
<58.0
<58.0
<72.5
<72.5
249.7
86.1
67.6
97.2
<58.0
<72.5
<72.5
<72.5
<72.5
<36.3
<72.5
<72.5
<72.5
<72.5
<72.5
<72.5
<58.0
<72.5
<58.0
<72.5
<72.5
<72.5
<58.0
106.2
<72.5
<72.5
<58.0
<58.0
RefLab
0.5
O.5
0.5
0.5
0.5
O.5
0.5
300
240
320
330
65
40
30
28
O.5
0.5
O.5
0.5
101
59
134
131
1.1
O.5
O.5
0.5
0.5
O.5
0.5
O.5
0.5
0.5
O.5
Jarringer analysis
order6
2005
2006
2154
2165
2108
2025
2021
2086
2009
2014
2161
2070
2028
2062
2091
2016
2112
2082
2064
2107
2120
2059
2162
2050
2158
2058
2037
2072
2141
2128
2095
2099
2143
2164

-------
Sample site
or type
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Sample no.
15
16
16
16
16
17
17
17
17
18
18
18
18
19
19
19
19
20
20
20
20
21
21
21
21
22
22
22
22
23
23
23
23
Sample
replicate
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
DNT"
(|ig/L)
Barringer
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
RefLab
<0.5
<0.5
<0.5
<0.5
0.5
382
340
340
320
<500.0
<500.0
410
190
<50.0
<50.0
1000
<50.0
<50.0
<10.0
<25.0
<5.0
<50.0
<5.0
<50.0
<50.0
<100.0
104
<50.0
90
81
60
60
32
RDX"
(|ig/L)
Barringer
<25.6
<51.3
<102.5
<51.3
<51.3
<102.5
<51.3
77.2
<102.5
1763.5
1992.9
1450.4
1646.6
431.8
504.9
621.4
343
<51.3
<51.3
<51.3
<25.6
122.9
106.8
<51.3
150.1
428
350.2
351.6
237.1
124.7
141.3
229.4
<102.5
RefLab
<0.5
<0.5
<0.5
<0.5
0.8
2160
2720
2600
1760
19600
16700
22800
18400
6100
3100
3500
4900
570
350
380
315
940
1180
1410
1130
3780
2960
2780
2680
2340
1430
1710
1930
TNT"
(Rg/L) *
Barringer
<36.3
<72.5
<58.0
<72.5
<72.5
346.4
576.1
784.7
573
3565.5
3319.7
3279.9
2993.1
911.6
995.6
1293.8
829.3
443.2
403.3
465.4
586.5
182.1
146.4
141.8
293.6
506.5
547.5
569.3
336.9
270.4
434.4
514.2
218.8
RefLab
0.5
<0.5
<0.5
0.5
0.7
1360
1620
1580
1240
6900
5800
8400
6000
3100
1600
2400
2500
1720
950
990
985
440
490
540
400
1280
1080
1210
1000
1520
850
1040
1260
Jarringer analysis
order6
2010
2074
2129
2033
2100
2140
2027
2001
2170
2096
2092
2022
2004
2127
2044
2076
2130
2056
2118
2068
2019
2041
2155
2103
2011
2003
2075
2036
2116
2157
2105
2078
2136

-------
OJ
oo
Sample site
or type
Louisiana
Louisiana
Louisiana
Louisiana
Milan
Milan
Milan
Milan
Milan
Milan
Milan
Milan
Milan
Milan
Milan
Milan
Milan
Milan
Milan
Milan
Milan
Milan
Milan
Milan
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Sample no.
24
24
24
24
6
6
6
6
7
7
7
7
8
8
8
8
9
9
9
9
10
10
10
10
7
7
7
7
8
8
8
8
9
9
Sample
replicate
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
DNT"
(|ig/L)
Barringer
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
16917.1
26142
22587
33924.5
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
RefLab
270
350
286
390
<5.0
<25.0
<25.0
<10.0
<50.0
<50.0
<10.0
<250.0
2840
19400
4300
19400
712
450
399
408
<0.5
<5.0
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
0.5
<5.0
<0.5
<5.0
<0.5
<10.0
RDX«
(|ig/L)
Barringer
118.5
<102.5
<51.3
<102.5
<102.5
<102.5
<102.5
<51.3
<51.3
<51.3
89.1
<51.3
<102.5
<51.3
<25.6
<51.3
<102.5
<102.5
<51.3
<51.3
<102.5
<51.3
<102.5
<25.6
<25.6
<51.3
<51.3
<51.3
<102.5
<51.3
<102.5
<51.3
<102.5
<51.3
RefLab
1770
3000
2260
1980
9
235
250
170
670
660
580
650
<50.0
<50.0
120
<50.0
36
<10.0
19
<10.0
93
91
84
96
83
88
88
65.5
17
19
22
19
<0.5
42
TNT"
Oig/L) I
Barringer
489.7
309.5
863.8
419.5
<58.0
106.2
<58.0
<72.5
982.3
1215.6
1806.7
1451.3
4521.1
5492
4280.4
4942.6
<58.0
<58.0
<72.5
<72.5
<58.0
<72.5
<58.0
66.4
<36.3
<72.5
<72.5
<72.5
<58.0
<72.5
<58.0
<72.5
<58.0
<72.5
RefLab
1260
2500
1860
1810
80
100
105
60
3600
3800
2960
2650
320
1610
540
2800
<10.0
<10.0
13
<10.0
154
149
150
167
19.8
22
20.5
17.4
72
77
90.5
66
185
244
tarringer analysis
order6
2125
2148
2110
2124
2139
2169
2173
2040
2035
2119
2018
2065
2167
2087
2020
2053
2175
2163
2098
2115
2160
2101
2122
2002
2024
2106
2034
2113
2135
2083
2137
2093
2171
2114

-------
Sample site
or type
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Spike/PE
Umatilla
Umatilla
Umatilla
Umatilla
Umatilla
Umatilla
Umatilla
Umatilla
Umatilla
Umatilla
Umatilla
Umatilla
Umatilla
Umatilla
Umatilla
Umatilla
Umatilla
Umatilla
Umatilla
Umatilla
Umatilla
Umatilla
Sample no.
9
9
10
10
10
10
11
11
11
11
12
12
12
12
1
1
1
1
2
2
2
2
3
3
3
3
4
4
4
4
5
5
5
5
6
6
Sample
replicate
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
DNT"
(|ig/L)
Barringer
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
RefLab
<0.5
25
<0.5
0.8
<0.5
<0.5
<50.0
<0.5
<0.5
<0.5
0.5
<50.0
<50.0
<0.5
<10.0
<0.5
<0.5
<0.5
<0.5
<0.5
0.5
O.5
O.5
<5.0
O.5
0.5
0.5
O.5
<5.0
0.5
O.5
0.5
O.5
O.5
0.5
O.5
RDX"
(|ig/L)
Barringer
<51.3
<25.6
<51.3
<102.5
<51.3
<51.3
<102.5
84.7
<51.3
<51.3
<51.3
<51.3
<51.3
<51.3
<25.6
<102.5
<51.3
<102.5
<51.3
<102.5
<51.3
<51.3
<51.3
<51.3
<102.5
<25.6
<102.5
<51.3
<51.3
<51.3
<102.5
<25.6
<102.5
<51.3
<51.3
<51.3
RefLab
0.5
O.5
188
320
146
210
650
1480
840
810
460
480
430
470
234
200
228
142
0.5
O.5
2.6
1.5
27
23
20
27
15
4.8
12
15
348
296
316
248
5.1
3.5
TNT"
(Rg/L) *
Barringer
<72.5
<36.3
<72.5
<58.0
<72.5
<72.5
123.9
94.1
<72.5
<72.5
398.9
463.2
279.1
189.4
<36.3
<58.0
<72.5
<58.0
<72.5
<58.0
<72.5
<72.5
<72.5
<72.5
<58.0
153
<58.0
<72.5
<72.5
<72.5
<58.0
<36.3
106.2
<72.5
<72.5
<72.5
RefLab
185
212
O.5
1.1
O.5
O.5
350
680
550
420
930
1020
930
910
42
34
32
20
0.5
0.6
1.3
1.3
146
117
109
127
57
27
83
96
O.5
0.5
O.5
O.5
28
22.5
Jarringer analysis
order6
2048
2015
2089
2156
2104
2057
2149
2109
2047
2102
2080
2085
2067
2029
2013
2152
2026
2166
2046
2138
2038
2084
2111
2054
2144
2023
2174
2060
2117
2069
2159
2017
2126
2094
2043
2049

-------
Sample site
or type
Umatilla
Umatilla
Volunteer
Volunteer
Volunteer
Volunteer
Volunteer
Volunteer
Volunteer
Volunteer
Sample no.
6
6
4
4
4
4
5
5
5
5
Sample
replicate
3
4
1
2
3
4
1
2
3
4
DNT"
(|ig/L)
Barringer
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
<1950.0
RefLab
<0.5
<0.5
<0.5
<0.5
<0.5
4.5
25
54
55
50
RDX"
(|ig/L)
Barringer
<25.6
<51.3
<51.3
<102.5
<51.3
<102.5
<51.3
<51.3
<102.5
<51.3
RefLab
3.3
5.9
<0.5
<0.5
<0.5
1.8
<5.0
<5.0
<5.0
<50.0
TNT"
(Rg/L) *
Barringer
<36.3
<72.5
<72.5
<58.0
<72.5
<58.0
667.3
585.2
420.4
364.6
RefLab
12.3
20.8
54
44.5
63
105
840
1290
1130
890
Jarringer analysis
order6
2008
2031
2088
2123
2073
2150
2081
2079
2176
2030
* The data are presented exactly as reported. Note that the data are not consistently reported with the same number of significant figures.
6 These are the sample numbers, from which the analysis order can be discerned. For example, 2001 was analyzed first, then 2002, etc.

-------