First International Symposium
FIELD SCREENING METHODS FOR
HAZARDOUS WASTE SITE
INVESTIGATIONS
October 11-13, 1988
Symposium Proceedings
-------
FIRST INTERNATIONAL SYMPOSIUM
FIELD SCREENING METHODS FOR
HAZARDOUS WASTE SITE
INVESTIGATIONS
October 11-13, 1988
Co-Sponsors
U.S. Environmental Protection Agency
U.S. Army Toxic and Hazardous Materials Agency
-------
DISCLAIMER
Although this Proceedings Document reports the oral and poster presenta-
tions and discussions that occurred during this Symposium funded by the
United States Environmental Protection Agency, the contents represent
views independent of Agency Policy. This Document has not been subjected
to the Agency's peer review process and does not necessarily reflect the
Agency views. No official endorsement should be inferred.
-------
SYMPOSIUM ORGANIZATION
Symposium Chairman - Llewellyn Williams, EPA/EMSL-Las Vegas, NV
Vice-Chairman - Vernon Laurie, EPA/OADEMQA-Washington, D.C.
Vice-Chairman - John Koutsandreas, EPA/OADEMQA-Washington, D.C.
Exhibit Chairman - Joseph Roesler, EPA/EMSL-Cincinnati, OH
Poster Session Chairman - Donald Gurka, EPA/EMSL-Las Vegas, NV
ACKNOWLEDGMENTS
This symposium has been arranged through an Environmental Protection Agency contract with
ICAJJR, Life Systems, Inc. (as a subcontractor to Acurex Corp.) Mr. Gregory Schiefer and
Ms. Jo Ann Duchene managed the Project. Mr. Charles Tanner served as Exhibit Coordinator;
Mr. Jack Lanigan coordinated the oral and poster presentations.
-------
FOREWORD
This International Symposium was initiated to respond
to the need for a specialized forum to address hazard-
ous waste site investigations and the opportunities
afforded by new and emerging technologies to reduce
costs, reduce turnaround time of data and increase the
scientific confidence in decisions based upon site
investigation data.
The objective of this meeting was to bring an interna-
tional view to problems of hazardous waste site char-
acterization and monitoring.
• To discuss available and developing technology
for rapid, low-cost detection and monitoring of
toxicants on site.
• To address new opportunities for Federal/private
cooperative ventures to develop and commercial-
ize field monitoring technology.
• To inform Symposium delegates and scientists
through open discussions, technical sessions,
exhibits and peer-reviewed publications of new
approaches to solve site investigation problems.
The papers and discussions that follow represent three
days of intense communication and cooperation among
a variety of communities - regulatory, academic,
industrial and user. It is my hope that the products of
this symposium will find many widespread uses and
will provide the impetus for new initiatives in field
screening methods.
Llewellyn R. Williams
U.S. Environmental Protection Agency
Environmental Monitoring Systems Laboratory
Las Vegas, Nevada
-------
CONTENTS
SESSION 1:
Opening Plenary Session
Introductory Remarks - Llewellyn Williams, Environmental Monitoring Systems Laboratory, Las Vegas 1
Welcoming Address - Robert Snelling, Acting Director, EPA, Environmental Monitoring Systems Laboratory, Las Vegas 1
Keynote Address-John. Skinner, EPA, Director, Office of Environmental Engineering and Technology 3
Field Monitoring Methods and Use for Supeifund Analyses, I — Joan Fisk, EPA, Office of Emergency and Remedial Response 7
Field Monitoring Methods and Use for Supeifund Analyses, II — Carla Dempsey, EPA, Office of Emergency and Remedial Response 11
Field Monitoring Methods and Use for Supeifund Analyses, ///-Scott Fredericks, EPA, Office of Emergency and Remedial Response .. 15
Screening Environmental Pollutants and Biomarkers: The Analytical Challenge - Tuan Vo-Dinh, Oak Ridge National Laboratory 17
SESSION 2:
Fiber Optics and Chemical Sensors (I)
Chairman: Larry Eccles - EPA, Las Vegas
Monitoring of Gasoline Vapor and Liquid by Fiber Optic Chemical Sensor (FOCS) Technology
S.M. Klainer, K. Goswami, D. LeGoullon, O.K. Dandge, ST&E, Inc.; J.R. Thomas, Fiber Chem, Inc.; S.J. Simon,
Lockheed-EMSCO; L. Eccles, EPA 25
The Suitability of Surface Enhanced Raman Spectroscopy (SERS) to Fiber Optic Chemical Sensing
of Aromatic Hydrocarbon Contamination in Groundwater
M.M. Carrabba, R.B. Edmonds, PJ. Marren and R.D. Rauh, EIC Laboratories, Inc 31
Fiber-Optic Surface-Enhanced Raman System for Field Screening of Hazardous Compounds
T.L. Ferrell, E.T. Arakawa, R.B. Gammage, D.R. James, J.P. Goudonnet, R.C. Reddick, J.W. Haas and E.A. Wachter
Health and Safety Research Division, Oak Ridge National Laboratory 41
Porous Fiber Optic for Chemical Sensors
M.R. Shahriari, Q. Zhou and G.H. Sigel, Jr., Rutgers University and G.H. Stokes, GEO-Centers, Inc 43
Improved Luminescence Technique for Screening Aromatic Contaminants in Environmental Samples
R.B. Gammage, J.W. Haas III, G. H. Miller and T. Vo-Dinh, Oak Ridge National Laboratory 51
Detection of Solvent Vapors Using Piezoelectric Sensors
E.B. Overton, D. A. Gustowski, L.H. Grande, H.P. Dharmasena, P. Klinkhachorn, C.S. Milan, and G.R. Newkome,
Louisiana State University 57
SESSION 3:
X-Ray Fluorescence Spectrometers
Chairman: Harold Vincent - EPA, Las Vegas
Introduction by Harold Vincent 61
Application of Field-Portable XRF to Hazardous Waste Characterization
R.K. Glanzman, CH2M Hill 63
The Use of Transportable X-Ray Fluorescence Spectrometer for On-Site Analysis of Mercury in Soils
D.J. Grupp, D.A. Everitt, R.J. Bath, NUS Corporation; R. Spear, USEPARegion II 71
The Determination of Minimum Detection Limits for Inorganic Constituents in Soil Using Transportable Secondary Target X-Ray Fluores-
cence. 1. Arsenic in the Presence of Lead
D.A. Everitt, D. Grupp, R.J. Bath, NUS Corporation; R. Spear, USEPARegion II 73
The Application of X-Ray Fluorescence Technology in the Creation of Site Comparison Samples and in the Design of Hazardous
Waste Treatability Studies
JJ. Barich, III, EPA Environmental Services Division, Seattle, WA; R.R. Jones, EPA Quality Assurance Management Office,
Seattle, WA; G.A. Raab, Lockheed Engineering Management Services Co.; J.R. Pasmore,
Columbia Scientific Industries Corporation 75
Low Level XRF Screening Analysis of Hazardous Waste Sites
R. Perlis and M. Chapin, Ecology and Environment, Inc 81
-------
SESSION 4:
Fiber Optics and Chemical Sensors (II)
Chairman: Larry Eccles - EPA, Las Vegas
Wavelength Tunable Portable Laser for Remote Fluorescence Analysis
G.D. Gillispie and R. St. Germain, North Dakota State University 95
Field Screening for Aromatic Organics Using Laser-Induced Fluorescence and Fiber Optics
W. Chudyk, K. Pohlig, N. Rico, and G. Johnson, Tufts University 99
Second-Derivative Ultraviolet Absorption Monitoring of Aromatic Contaminants in Groundwaters
J.W. Haas III, E.Y. Lee, C.L. Thomas and R.B. Gammage, Oak Ridge National Laboratory 105
Hazardous Waste Analysis by Raman Spectroscope
C.K. Mann andT.J. Vickers, Florida State University Ill
Prototype Design and Testing of Two Fiber-Optic Spectrochemical Emission Sensors
K.B. Olsen, J.W. Griffin, D.A. Nelson and B.S. Matson, Pacific Northwest Laboratory;
PA. Eschbach, Washington State University 117
Porous Glass Fiber Optic Sensors for Field Screening of Hazardous Waste Sites
S.M. Finger, P.B. Macedo, A.A. Barkatt, H. Hojaji, N. Laberge, R. Mohr, M. Penafiel, Catholic University of America 127
Instrumentation and Methodology for Multicomponent Analysis Using In Situ Laser-Induced Fluorescence
J.E. Kenny, G.B. Jarvis and H. Xu, Tufts University 133
SESSION 5:
Soil Gas Analyzers
Chairman: Philip Durgin - EPA, Las Vegas
Influence of Naturally Occurring Volatile Compounds on Soil Gas Results
R.J. Nadeau, EPA, Environmental Response Team, Edison, NJ; J. Tomaszewicz, ERT Technical Assistance Team 141
An fn-Situ Technique for Measuring Soil-Gas Diffusivity
P.M. Kearl, T.A. Cronk and N.E. Korte, Oak Ridge National Laboratory 149
A Field Method for Determination of Volatile Organics in Soil Samples
T.M. Spittler, M.J. Cuzzupe, EPA Region I; J.T. Griffith,Goldberg, Zoino and Assoc. Inc 155
Soil-Gas Sampling at a Site with Deep Contamination by Fuels
H.B. Kerfoot, S.R. Schroedl, Lockheed Engineering and Sciences Co. and J.J. D'Lugosz, USEPA, EMSL-LV 159
Soil Gas Analyses to Delineate a Plume of Volatile Organic Compounds from a Hazardous Waste Site in Williamson County, Tennessee
R.W Lee, USGS, Nashville, TN;-M. Fernandez, USGS, Tampa, FL ". 171
Soil-Gas Screening: Its Theory and Applications to Hazardous Waste Site Investigations
L.M. Preslo, R. Pavlick and W.M. Leis, Roy F. Weston, Inc 179
SESSION 6:
Air Sampling Methods
Chairman: William McClenny - EPA, RTF
Atmospheric Analysis by Open Path Infrared Spectroscopy
PL. Hanst, Infrared Analysis, Inc 181
Development of the MINITMASS, a Mobile Tandem Mass Spectrometer for Monitoring Vapors and Paniculate Matter in Air
H.L.C. Meuzelaar, W.H. McClennen, N.S. Arnold, T.K. Reynolds, W. Maswadeh, PR, Jones and D.T. Urban,
Center for Micro-Analysis and Reaction Chemistry, University of Utah 195
The Preparation, Certification, and Use ofSumma Canister External Performance Evaluation Samples in Support of the TAGA 6000 E
Indoor Air Analyses During the Love Canal Emergency Declaration Area Hahitability Study
K.J. Caviston, R.E. Means, R.M. Harrell, B.J. Carpenter, Northrop Services, Inc.;
D. Mickunas and M. Bernick, Roy F. Weston, Inc.; T.H. Pritchett, USEPA Environmental Response Team 205
Unambiguous Identification and Rapid Quantitation in Field Air Monitoring Using a Fully Mobile Mass Spectrometer
F.H. Laukien and T.M. Trainor, Bruker Instruments, Inc 207
The Preparation ofSumma Canister Performance Samples and their Subsequent Analysis by the TAGA 6000E MS/MS
R.M. Harrell, R.E. Means, K.J. Caviston, NSI Technology Services Corp.; M. Bemick, D. Mickunas,
Roy F. Weston, Inc., T.H. Pritchett, USEPA Environmental Response Team; W.J. Mitchell, USEPA Environmental
Monitoring Systems Laboratory, Research Triangle Park 219
Results From the Environmental Response Team's Preliminary Evaluation of a Direct Air Sampling Mass Spectrometer
R.E. Hague, Rutgers University; K. Cho, Roy F. Weston, Inc.; T.H. Pritchett,
USEPA Environmental Response Team; B. Shapiro, formerly of Enviresponse, Inc 227
-------
SESSION 7:
Immunochemical Methods
Chairman: Jeanette Van Emon - EPA, Las Vegas
Introduction by Jeanette Van Emon 233
Delivery System for Rapid, Semi-Quantitative Analysis of Low Molecular Weight Contaminants and Residues
' RL.McMahon, R.Suva and C. Brooks, IDEXX Corp 235
Fieldable Enzyme Immunoassay Kits for Pesticides
PH. Duquette, RE. Guire, and M.J. Swanson, Bio-Metric Systems, Inc 239
Integrated Immunochemical Systems for Environmental Monitoring
J.M. Bolts, S.E. Diamond, J.F. Kolc, S.H. Lin, F.J. Regina, Allied-Signal Corporate Technology; RG. Koga, G.C. Misener and
J.C. Schmidt, Bendix Environmental Systems Division 243
Immunochemical Quantification ofDioxins in Industrial Chemicals and Soils
M. Vanderlaan, B. Watkins, and L. Stanker, Lawrence Livermore National Laboratory 249
Remote, Continuous, Multichannel Biochemical Sensors Based on Fluoroimmunoassay Technologies
J-N. Lin, P. Kopeckova * J. Ives, H. Chuang, J. Kopecek,* J. Herron, H-R. Yen, D. Christensen, J.D. Andrade,
University of Utah; *Institute for Macromolecular Chemistry, Prague, Czechoslovakia 251
A Micwhial Bioassay Developed for Rapid Field Screening of Hazardous Waste Sites
I.C. Felkner, B. Worthy, T. Christison, and C.F. Chaisson, Technical Assessment Systems, Inc 253
SESSION 8:
Portable Gas Chromatographs
Chairman: Steve Billets - EPA, Las Vegas
Monitoring Volatile Organics in Water by a Photovac Portable Gas Chromatograph With Multiple Headspace Extraction Method
J.S. Ho and J.F. Roesler, USEPA, Environmental Monitoring and Support Laboratory, Cincinnati, OH;
P. Hodakievic, Technology Applications, Inc 261
Hazardous Waste Site Measurements ofPPB Levels of Chlorinated Hydrocarbons Using a Portable Gas Chromatograph
A. Linenberg, Sentex Sensing Technology, Inc 271
Correlation Chromatography with a Portable Microchip Gas Chromatograph
E.B. Overton, R.W. Sherman, C.F. Steele, and H.R Dharmasena, Institute for Environmental Studies,
Louisiana State University 275
Development of a Field Portable Concentrator/Purge and Trap Device for Analysis of Volatile Organic Compounds in Ambient Air and
Water Samples
R.W. Sherman, E.S. Collard, M.F. Solecki, T.H. McKinney, L.H. Grande and E.B. Overton,
Institute for Environmental Studies, Louisiana State University 279
Ambient Air Sampling With a Portable Gas Chromatograph
R.E. Berkley, USEPA, Environmental Monitoring Systems Laboratory, Research Triangle Park, NC 283
A Portable System Under Development for the Detection of Hazardous Materials in Water
J.C. Schmidt, RG. Koga, G.C. Misener, Environmental Technologies Group, Inc 291
SESSION 9:
Expert Systems for Field Instrumentation
Chairman: Joseph Roesler - Cincinnati Engineers, Inc.
Design and Performance of a Mobile Mass Spectrometer Developed for Environmental Field Investigations
T.M. Trainor and F.H. Laukien, Bruker Instruments, Inc 299
Expert Systems to Assist in Evaluation of Measurement Data
D.G. Greathouse, Risk Reduction Engineering Laboratory, USEPA, Cincinnati, OH 311
A Positioning and Data Logging System for Surface Geophysical Swveys
J.E. Nyquist and M.S. Blair, Oak Ridge National Laboratory 315
Prototype Volatile Organic Compound (VOC) Monitor
J.D. Wander, Air Force Engineering and Services Center, Tyndall AFB, B.L. Lentz, L. Michalec, and
V. Taylor, S-CUBED, Corporation 319
Environmental Field Sampling Expert System Development of a Soil Sampling Advisor
R.A. Olivero, R.E. Cameron, KJ. Cabbie, M.T. Homsher, M.A. Stapanian, Lockheed Engineering & Services Company;
K.W. Brown, USEPA, Las Vegas, NV 325
-------
SESSION 10:
Other Advanced Field Techniques
Chairman: Ronald Mitchum - EPA, Las Vegas
Introduction by Ronald Mitchum 341
Evaluation of a Field-Based, Mobile, Gas Chromatograph-Mass Spectrometer for the Identification and Quantification of
Volatile Organic Compounds on EPA's Hazardous Substances List
A. Robbat, Jr. andG. Xyrafas, Tufts University 343
Ion Mobility Spectrometry for Identification and Detection of Hazardous Chemicals
J. Reategui, T. Bacon, G. Spangler, J. Roehl, Environmental Technologies Group, Inc 349
Utilization of Short-Term Bioassessments and Biomonitoring at Superfund Sites
D.W. Charters, USEPA, Edison, NJ 359
High-Performance Liquid Chromatograph as a Viable Field Screening Method for Hazardous Waste Site Investigations
V. Ekambaram and J.B. Burch, Woodward-Clyde Consultants 361
Specific Detection of Any Gas Chromatographable Element in Sediment Extracts
M. Szelewski and M. Wilson, Hewlett-Packard Company 367
The U.S. EPA Field Analytical Screening Project (FASP)
G.H. Chapman, Ecology and Environment Inc. and S. Fredericks, USEPA Hazardous Site Evaluation Division 375
CLOSING PLENARY SESSION
Concluding Remarks, Llewellyn Williams, Symposium Chairman 379
POSTER PRESENTATIONS
Quality Assurance Plan Used at the Love Canal Emergency Declaration Area Indoor Air Analyses by the TAGA 6000E Mass
Spectrometer/Mass Spectrometer
T.H. Pritchett, USEPA Environmental Response Team; D.B. Mickunas and N. Kurlick, International Technology, Inc 381
The Kwik-Skrene Analytical Testing System: Description of a Tool for Remediation ofPCB Spills
G.R. Woollerton, S. Valin and J.P. Gibeault, Syprotec, Inc 387
A New Method for the Detection and Measurement of Aromatic Compounds in Water
J.D. Hanby, Hanby Analytical Laboratories, Inc 389
Development of a Temperature Programmed Microchip, High Resolution GCI MS for VOC Analysis
E.B. Overton, E.S. Collard, H.P. Dharmasena, P. Klinkhachorn, and C.F. Steele, Institute for Environmental Studies,
Louisiana State University 395
Applications of the Pyran Thermal Extractor-GCIMSfor the Rapid Characterization and Monitoring of Hazardous Waste Sites
C.B. Henry, E.B. Overton, Institute for Environmental Studies, Louisiana State University; C. Sutton, Ruska Instruments 399
Field Deployable Instrument for the Analysis of Semi volatile Organic Compounds
E.B. Overton, C.B. Henry, Institute for Environmental Studies, Louisiana State University, C. Sutton, Ruska Instruments 407
Evaluation of Microwave Detection Techniques to Prepare Solid and Hazardous Waste Samples for Elemental Analysis
P.M. Grohse, D.A. Binstock, and A. Gaskill, Jr., Research Triangle Institute; H.M. Kingston, National Bureau of
Standards and C. Sellers, USEPA Office of Solid Waste 411
Rapid Screening of Organic Contaminants Using a Mobile Mass Spectrometer in the Field
M.C. Hadka, Walter B. Satterthwaite Associates, R.K. Dickinson, United Engineers 423
Determination ofChlordane in Soil by Enzyme Immunoassay
R.J. Bushway, J. King and B. Perkins, University of Maine; W.M. Pask, Purdue University; B.S. Ferguson,
ImmunoSystems, Inc 433
Development of a Protocol for the Assessment of Gas Chromatographic Field Screening Methods
M.T. Homsher, V.A. Ecker, M.H. Bartling, L.D. Woods and R.A. Olivero, Lockheed Engineering and Sciences Co.;
D.W. Bottrell and J.D. Petty, USEPA, Las Vegas NV 439
Cost Analysis for Using Mobile Laboratories vs. Fixed-Base Laboratories for Site Characterization at FUSRAP Sites
G. Ganapathi and D.S. Adler, Bechtel National, Inc.; M. Carkhuff, Weston Analytical Laboratory 463
Enzyme Immunoassavfor the Quantisation of an Alkaline Protease in Airborne Samples
L.S. Miller, V. Moore, A. Wardwell, M. Buchwalter, L.A. Smith, Battelle Biotechnology Section 459
Gas Chromatographic and Mass Spectrometric Analysis of Target Air Toxics at Remedial Hazardous Waste Sites
D.W. Hodgson, B.C. Miller, R.A. Ross and T.S. Viswanathan, NSI Technology Services Corp.; R.D. Kleopfer and
W.W. Bunn, USEPA, Kansas City, KS 475
-------
On-Site Soil Gas Analysis of Gasoline Components Using a Field-Designed Gas Chwmatograph-Mass Spectrometer
A. Robbat, Jr. and G. Xyrafas, Tufts University 48
Reflectance Specfroscopv (0.2 to 20 \Jjn) as an Analytical Method for the Detection ofOrganics on Soils
T.V.V. King and R.N. Clark, USGS, Denver, CO 4("
Field Use of a Microchip Gas Chromatograph
R.W. Sherman, T.H. McKinney, Institute for Environmental Studies, Louisiana State University; M.F.
Solecki, National Oceanic & Atmospheric Administration, Seattle, WA; R.B. Gaines,
U.S. Coast Guard R&D Center, Groton, CT; B. Shipley, USEPARegion IX 489
Rapid Assessment ofPCB Contamination at Field Sites Using a Specialized Sampling, Analysis and Data Review Procedure
W.W. Freeman and J. Karmazyn, Roy F. Weston, Inc 491
Participants'List 501
-------
INTRODUCTORY REMARKS
LLEWELLYN WILLIAMS
Welcome to The First International Symposium on Field Screen-
ing Methods for Hazardous Waste Investigations. I am very
excited about this area and about the response that we've had
since our initial call for papers last December. We have had
dramatic response from people who design, produce, and use new
technology for monitoring and for measurement.
The timing for this Symposium is right. We have brought
together some of the most outstanding representatives of acade-
mia, the private sector, and the federal sector. This program brings
together strong presentations from national and international
sources. One of its features is the combined reception, exhibit and
poster session, which is an opportunity to view some of the new
technologies that are available and to discuss your future needs
with the designers of those technologies. I've found that the
private sector is anxious to respond to the user's needs as long as
they are made clear. If you are such a user, be sure to let them
know that there is indeed a market for this technology not only in
the future, but now.
WELCOMING ADDRESS
ROBERT SNELLING
Welcome to Las Vegas. I think we have an outstanding
program for you. My job this morning is two fold. First of all, I'd
like to provide a few general words as to why we are here, and to
provide an overview of our program here at the Environmental
Monitoring Systems Laboratory.
The reality, in terms of hazardous waste site assessments, is
that we are in the dark. We are faced with the problem of trying to
feel our way through the darkness, to get some insight into the
problems associated with a particular hazardous waste site. That is
the thrust of this program. With the support and encouragement of
the Superfund Office within EPA, a program was initiated two
years ago which attempted to identify and establish rapid screen-
ing methods that could be applied to hazardous waste site
investigations.
What do we mean by screening methods? In academic terms, it
is the use of rapid, low-cost test methods to determine whether a
characteristic of interest is present or absent, above or below a
predetermined threshold at a given site, or in a concentration
within a predetermined range of interest. It is an attempt to define
the spatial extent of some specific characteristic.
We are interested in examining various screening methods
because we wish to gain a preliminary understanding of what
happens at a hazardous waste site. This can, in turn, guide us
toward a cost effective monitoring program, which is essential,
since monitoring, sampling and analysis are currently very expen-
sive. So the emphasis of the program will be on aspects of field
measurements, quick turn around, and low-cost screening
methods.
There are four primary aspects to the program which we have
defined in order to achieve these goals. The first is identification
of off-the-shelf technology, which is either already available on
the commercial market, or is at a stage in its development where it
can be made available shortly. Technologies are available that are
not used, or not fully characterized. We must develop a sense of
the validity and applications of a given method.
The second part of the program is to identify the needs in the
field which are being met by existing technology. In response to
these findings, a research and development program was initiated
to fill those needs. This program evaluates technologies that are
not quite ready for commercialization, but need some additional
research and development.
The third part of the program involves the demonstration of
both the commercially available technologies, and the technolo-
gies coming out of the methods development program. This
provides confirmation that the technologies perform as they are
intended to on hazardous waste sites.
The last aspect of the program is to transfer this technology to
the people who require it. We have struggled with this problem in
the EPA, because we have found that the knowledge gained in the
research and development community has not been communi-
cated to the regional, state and private sectors who are engaged in
the work itself. An important part of the program is to transfer this
technology to the people who need to use it. That, in part, is why
we are here.
The overall program strategy has a number of interesting
components; I would like to emphasize one in particular, which
involves three aspects of the program: the leverage of the private
sector, technology and matrix management within the Agency. We
felt that the evolving technologies required a mix of expertise. So
our program utilizes a number of our research laboratories
through a matrix management program. We have tried to incorpo-
rate the needs and inputs of the EPA Regional offices, who are our
primary clients. We are also very interested in leveraging other
agencies, where a lot of work is being done. For example, the
Department of Defense is working on field measurement tech-
niques, which can, in part, be transferred to the environmental
programs.
The aspect of the program we really want to emphasize,
however, is leveraging of the private sector, with special emphasis
on those technologies which are either currently commercialized
or will shortly be ready for commercialization. This topic will be
discussed in part, during the next speaker's talk on The Superfund
Innovative Technology Evaluation Program, in which we will
work with the private sector to evaluate technologies that might
be applied to hazardous waste site investigations.
Other aspects of our program attempt to focus on this specific
issue. We have found that a great deal of work is being done in
other programs, and we have tried to pull these together across
EPA's media programs, to focus on the technologies themselves.
There is a need in the field to gather information quickly and
cost effectively, rather than using old, time-consuming methods
such as collecting samples, sending them to a laboratory, and then
waiting three weeks to receive an analysis at the cost of thousands
of dollars. We need methodologies that will give us preliminary
insight while we are still in the field, which we can subsequently
act on. Those are the kinds of technologies that we want to focus
on in the coming days.
-------
KEYNOTE ADDRESS
Dr. John Skinner
Director
Office of Environmental Engineering and Technology
U.S. Environmental Protection Agency
I am going to discuss the program evaluating and
demonstrating field monitoring methods for hazard-
ous waste sites and the use of the Superfund
Innovative Technology Evaluation program or the
SITE program for this purpose.
This field is very important to EPA for a number
of reasons. We and the private sector spend
hundreds of millions of dollars every year evalu-
ating and analyzing samples for various chemical
constituents. Any reduction in cost or time
consumption which can be achieved in that process
should result in enormous savings, both for the
Agency and for all of the regulated parties. So
it should be obvious why this effort to develop
field screening and field monitoring methods is
important. There has been a considerable expansion
of sampling and monitoring requirements, due to
the passage of SARA, the Superfund Authorization
Amendments. Many more sites will enter the Super-
fund program, and much more sampling and analysis
will be required both at the early site characteri-
zation stage and in judging the validity or
acceptability of the clean-up.
However, the technologies that we will be discus-
sing today are not limited only to the Superfund
program. They are applicable also to the hazardous
waste regulatory program under RCRA, the Resource
Conservation and Recovery Act. These technologies
could apply to other Agency programs as well, such
as waste water discharges from industrial and
municipal sewage plants, or analysis of pesticide
residues; these technologies should be useful
across the Agency.
As our regulatory programs expand, and as Super-
fund itself expands to include more sites, the
current laboratory capacity will be taxed. Higher
cost of sampling and analysis and long delays in
getting the results will be the outcome.
It is important to define field screening and field
monitoring methods. These are methods that can be
taken to the field, to carry out—in a matter of
hours or days—screening activities to identify
the nature of contamination at that site. These
methods may not all be as accurate or precise as
the laboratory methods, although some of them are
very accurate and precise. But the idea behind
them is to allow priorities to be set at a site;
to allow identification of hot spots, and to quickly
establish either a more comprehensive monitoring
program for a given site, or to place the site in
some priority order relative to another site. We
believe that these field screening methods have the
potential to accelerate site clean-up, to improve
confidence in site clean-up, and to reduce costs.
Let me touch on each one of these.
With respect to accelerating site clean-up, the
critical time line elements generally are the
requirements for on-slte sampling, the shipment of
those samples to a laboratory, and the analysis of
those samples at the laboratory. Delays can be
caused throughout this entire process. So, if a
site can be quickly screened to eliminate some of
these delays, thereby moving on to more full scale
site monitoring sooner, it should be possible to
accelerate the site clean-up.
With respect to improving confidence in the clean
up, it should be possible to do more sampling at a
higher sampling density without excessive costs.
This would lead to more effective delineation of
contaminants on the site, detection of hot spots
or spots that should be cleaned up first or
quickly, and identification of which areas might
require even more intensive sampling in order to
be properly characterized.
In terms of cost reduction, these techniques could
reduce or minimize the laboratory analyses thereby
minimizing the stand-by time for sampling personnel
who await the results of the analysis in order to
determine whether more sampling is necessary. If
additional sampling is deemed necessary, it can
then be done right away. It is also possible to
minimize time for clean-up personnel, who may be
waiting for confirmatory sampling to determine
whether they really have cleaned up the site to
the required level. So, reduction of personnel
costs, both for sampling and clean-up, should
result in overall cost reductions.
The Las Vegas team performed some rough calcula-
tions to determine the potential of some of these
technologies. They looked first at a metals
analysis in soils. They assumed that at a particu-
lar site they had to analyze 500 samples for lead,
copper, and zinc. For example, if atomic absorp-
tion were used in a laboratory to do that, the
-------
sample turn-around time would be roughly one month.
To do those 500 samples would cost roughly $20,000
to $25,000. If the same process were done in the
field with X-ray fluorescence, the samples' turn-
around time would be less than one day. The results
could be found the same day, and the cost would be a.
few thousand dollars.
Even if these numbers are inaccurate by a factor
of two or three, the potential both for reducing
the turn-around time and for reducing the costs of
sampling are extremely significant. This only
pertains to the costs of doing the analysis; it
does not even account for the standby cost that I
just mentioned.
Similarly, with respect to organics, an analysis
was done which demonstrated that if 500 samples of
ground water or leachate were analyzed for penta-
chlorophenol using typical laboratory GC/MS, the
sample turn-around time would be 30 days. No
laboratory is able to turn around samples for
organics in 30 days. The time estimated is at
least double or triple that, and the cost is
approximately $50,000.
With the use of imniunoassay at a site, samples can
be turned around in one day with a cost of a few
thousand dollars. Thus, both for inorganics and
organics, the potential for cost and time
reduction with some of these field screening
methods is quite substantial.
I would now like to address the process of acceler-
ating the development and application of some of
these field screening methods, using the SITE
program. This program was established under the
Superfund amendments. This is not an area in which
the federal government would carry out extensive
technology development on its own. Rather, the
focus of this effort is to create a market for
private sector developments and to form a relation-
ship with the private sector which would encourage
the development of those technologies.
The approach will be to establish desirable
performance standards for these technologies.
Next, an evaluation or a demonstration of those
technologies will be carried out at an actual
Superfund site. In general, when such a demon-
stration is performed, the private developer is
expected to pay for the actual cost of running a
technique at his site. We would pay for the costs
of evaluating the technique, then publish those
results. Those technologies that were successful
would then be used in the Superfund program or in
the hazardous waste program, in place of more
traditional technologies. Thus, the demonstration
program would eventually lead to the potential
commercial use of these technologies in the actual
Superfund cleanup program.
The primary purpose of the SITE program, which was
established under the 1986 amendments to Superfund,
is to enhance private sector development of technol-
ogies through a demonstration or evaluation process
which establishes the commercial availability of
these technologies. This is accomplished through
a site demonstration which tests and validates
these field monitoring methods under one or more
real waste site conditions. The performance of
these technologies can be confirmed through
laboratory sampling and analysis, and the entire
demonstration effort can be coordinated with the
potential users of these technologies. Our
Regional offices, the REM/FIT contractors, who
carry out actual Superfund cleanups, would
document the performance of the technologies and
put out a report.
We will select these technologies in a number of
ways, both formally and informally. Formally,
this program has been announced a number of times
in the Commerce Business Daily. We have received
a fair response from the private sector to our
interest in evaluating these technologies.
Informally, we follow up on good ideas when they
present themselves to us. We actively pursue
information about these technologies. If anyone
here is a developer or knows someone who is,
please contact the individuals listed on the
sheets which were distributed to you today. We
believe that a large number of technologies will
be eligible for the program. Federal support is
not as limited for the development of the
technology as it is for testing and evaluation.
Examples of the types of technologies in which we
are interested, some of them already in the program,
are generally portable, transportable, and fieldable
instruments. These include portable GC/MS of suit-
case size, portable X-ray fluorescence systems,
chemical and immunochemical field kits, technologies
to detect in situ contaminants in soil, soil gas,
ground water, and other innovative sampling and
collection methods.
Before going ahead with a demonstration we would
like the following information. The developer
should define, from whatever data he has, the
bias, precision, the rate of false positive and
negatives, the detection limits, and the major
interferences of an instrument. Further, he
should define analytes and the matrices to which
he thinks the technology is applicable; provide
his set of standard operating procedures and
protocols; and provide whatever data he has from
his independent evaluation of the technology.
A demonstration plan would then be devised in
cooperation with the developer. This would vary
according to the technology, defining the role of
the developer, and our own role in the demonstration
program. Quality assurance would be emphasized in
the demonstration plan so that good data would be
produced. In the process, we would arrange for
active participation of our Regional offices, our
contractors, and our private sector representatives.
We want to ensure that these technologies ultimately
suit the needs of our clients, who will use them
after the demonstration projects have been completed.
The performance of these technologies must be con-
firmed through the contract laboratory program.
The products are expected to be a fair and objec-
tive evaluation of the technology in a '"real-world"
field situation. A demonstration report would be
issued, which includes the data that have been
-------
collected, and discusses the strengths, limitations,
and potential applications for the program. This
effort creates a number of opportunities for EPA to
work with the private sector, in the type of rela-
tionship that we think works very well.
The government is not in the business of commercial-
izing technologies. We feel that there are very
significant private, profit incentives, for doing
just that. We are trying to create the framework
in which private sector development can proceed,
and we feel that this is a genuine partnership. We
have a mutual interest with private developers in
getting better technologies and bringing them to
the marketplace. We hope that this will open up
some uew markets, both within the Superfund program
and outside of it for other programs in the Agency.
We believe there are mutual benefits, both to the
developer and to the potential user of the
technology. This is one area in which both the
regulator and the regulated party should benefit.
In conclusion, I wish you all success for a very
productive and informative conference, this First
International Symposium on Field Screening Methods.
I think the papers that will be presented over the
next three days offer some very exciting new oppor-
tunities for improved field screening techniques.
We hope that these efforts will eventually pay off
in reducing the delays in sampling and analysis of
chemicals at Superfund sites and will also result
in the creation of new markets for some of these
technologies.
-------
FIELD MONITORING METHODS AND USE
FOR SUPERFUND ANALYSES, I
Joan Fisk
Organics Section Chief
Office of Emergency and Remedial Response
U.S. Environmental Protection Agency
I will begin with an overview of the work of
Superfund and its contractors in the past and
present. Carla Dempsey will then discuss the QA
aspects and guidance for appropriate use. Finally,
Scott Fredericks will follow up with our success
story on field methods that have worked well
because of the procedures they have set up.
I am going to give some background to explain why
Superfund is seeking out the use of field methods
extensively. Superfund, The Comprehensive Environ-
mental Response, Compensation and Liability Act
was enacted by Congress in 1980. It gives the
President of the United States the authority to
clean up uncontrolled hazardous waste sites that
are a threat to public health or the environment.
A total of $1.6 billion was raised by Superfund
from taxes on chemicals and chemical companies to
effect this clean up. In addition, EPA was
permitted to recover costs from any potentially
responsible parties that could be identified.
Superfund was reauthorized as the Superfund Amend-
ments and Reauthorlzation Act in October 1986,
with changes and additions. Today, however, I
will only be discussing those changes in schedules
which set time limits for programs during the life
of Superfund. First of all, a list called the
Comprehensive Environmental Response and Liability
Information System (CERCLIS) was created, citing
28,000 potential sites which existed at the time
of Superfund's reauthorization. A deadline of
January 1 was set, at which time these Preliminary
Assessments all had to be completed. The EPA met
that deadline, leaving a backlog of about 10,000
Site Investigations which were also supposed to be
completed at that time.
The next item on the schedule demanded evaluation
of all sites on the CERCLIS within four years,
upon the recommendation of the Preliminary Assess-
ment and Site Investigations. In other words, the
scoring process called the hazardous ranking
system (HRS) must be employed at each site to
determine whether or not it gets placed on the
National Priorities List (NPL). I believe that
right now 797 sites are on the NPL and 380 are
proposed. This is a dynamic number which changes
as sites are cleaned up or discovered.
The next step in the Superfund dynamics is the
Remedial Investigation/Feasibility Study (RI/FS),
done on all sites which are on the NPL in order to
determine the extent of contamination. That is,
the concentrations and the boundaries of pollu-
tants at a site, whether groundwater, air, soil,
and so on are determined. No less than 275
RI/FS's were supposed to be completed by October,
1989. By the end of Superfund, a total of 750 are
to be completed.
In addition to identification of the extent of
contamination, the selection of remedy options is
also carried out during this period of time.
Ultimately a decision emerges, which takes socio-
economic factors into consideration. These
factors include proximity of the site to other
areas, resources available for doing the job, and
health risk assessment.
The last step of the Superfund process is the most
important: the design for remedial action at a
given site. One hundred and seventy-five of these
must be done by October 1989, and a total of 375
by the end of the existing Superfund law. Winston
Porter, in August of 1987, sent out a letter to
Regional Administrators suggesting they speed up
the remediation process, and complete a RI/FS in
18 months or less. Clearly, there is a big push
to use faster field methods.
All of this activity will result in an enormous
quantity of samples for analysis by some technique
or other. Traditionally, most Superfund analyses
were given to the Contract Laboratory Program
(CLP), which has about 100 contractor laboratories
nationwide performing analyses using standardized
protocols with standardized deliverables. Many
quality control (QC) requirements and criteria
must be met and there is a great deal of docu-
mentation on the QC, so that the data are of a
known quality.
This much care was taken because it was assumed in
the early days of Superfund that all data could
undergo the scrutiny of a court of law, either in
a settlement or in litigation. Later, it occurred
to some participants that such measures were not
necessary, and that alternatives were available.
-------
During this period, the concept of data quality
objectives emerged within the agency. These are
qualitative and quantitative statements made by
decision makers regarding the required level of
certainty acceptable in results of environmental
data collection. All of the decision makers
throughout the Superfund process are engaged in
up-front planning to determine the type of data
they will need through each one of these activi-
ties. The decisions made are the key to when and
how field screening can be appropriately used. In
1985 Superfund did make a commitment to connect
decision-making and data quality needs, through a
letter from Henry Longest.
In summary, the first two items, the schedules and
the directive, both show a desire to increase the
speed of activity. The last item, the data
quality objectives, suggests a need for more
ingenuity. Many people are interested in doing
work in the field.
These suggestions create a very basic, simple
definition of field technologies. "Portable"
equipment should weigh 50 or 60 pounds, so it can
be carried with minimal sample processing. "Field-
able" equipment is transportable in a van or small
truck and requires no more than a portable
generator for power. "Mobile" equipment is small
enough for a mobile lab or temporary trailer or
hut, with a generator and compressed gas cylinder,
which actually includes almost anything that can
be used in a fixed laboratory.
Discussions with people in the field have yielded
three major roles for using field methods of
analysis. The first is to enable "real time,"
on-site feedback, to aid in site characterization
by identifying hot spots and the extent of con-
tamination. The use of the word "aid" points up
the fact that this technique does not accomplish
the task alone. It also aims to direct ongoing
work such as redirecting sampling efforts, modify-
ing work plans, determining the depth of well
screen placement or making sure all contaminated
materials have been removed while the equipment is
on the site, so that the costly step of bringing
it back a second time is avoided. Field methods
can also prioritize samples through the CLP, which
is very useful. The majority of people tell us
that they do confirm the CLP analysis on about 10%
to 30% of all samples. Finally, field methods can
assure confidence in a clean up.
The second general reason for using field methods
is customized analysis. This is a site-specific
approach, which becomes increasingly important
during the later stages of the Superfund clean-up
process. Customized analyses can optimize methods
for dealing with the contaminants in question.
These can include the use of specific detectors,
screening, grab air samples, or soil gas sampling
for samples that have short holding times. We all
recognize that the sample that is analyzed in the
laboratory is not necessarily representative of
the one that was taken out in the field.
The last general reason for using field methods is
cost reduction. This is accomplished by minimizing
full CLP analyses; or more exactly, by making more
effective use of the resources to direct more of
the really contaminated samples through the CLP
labs thereby eliminating a lot of negatives.
Field methods can also minimize stand-by time for
sampling crews and clean-up personnel, since the
data they are waiting for now takes weeks to
generate and deliver.
There are many key organizations involved in this
process (Table I). These include EPA's Office of
Emergency and Remedial Response and the Hazardous
Site Control Division which are, among others,
responsible for the REM contractors; the United
States Army Toxic and Hazardous Materials Agency
(USATHAMA), which works with Oak Ridge National
Laboratories; and EMSL Las Vegas, which is part of
EPA's Office of Research and Development. The
Hazardous Site Control Division does extensive
field work using what they call close support
laboratories (CSLs).
Table I Key Players in Field
Methods Analyses
• EPA, Office of Emergency and Remedial Response
(OERR), Site Assessment Branch (SAB) of the
Hazardous Site Evaluation Division (HSED) and
their Field Investigation Teams (FIT)
• EPA, OERR, Hazardous Site Control Division (HSCD)
and their Remedial Investigation/Feasibility
Study (RI/FS) contractors (REM)
• EPA, Environmental Response Team (ERT)
• U.S. Army Toxic and Hazardous Materials Agency
(USATHAMA) and Oak Ridge National Laboratories
(ORNL)
• EPA, Office of Research and Development (ORD)
• EPA, OERR, HSED, Analytical Operations Branch
(AOB)
Most of the people I have talked to in the REM
contractor community say that they firmly endorse
the use of data quality objectives to determine
the appropriate use of methods and the appropriate
methods themselves. Their process goes as follows.
First, they determine the list of indicated
parameters from previous data. Next, they
establish detection limits, precision and accuracy
requirements. Then, they determine the required
data completeness or the data deliverables, select
or develop a method that meets the above determined
data quality objectives, and validate that method,
providing performance information on it. They
then prepare a standard operating procedure des-
cribing the close support lab operating structure,
the sample handling, tracking, and the methods,
the QA/QC requirements, the data reporting
requirements, and health and safety requirements.
Last, they include this SOP in the quality
assurance project plan for review and approval.
The REM contractors use field screening techniques
for the purpose of supporting treatability studies,
-------
and supplementing health and safety decision
making. This includes levels of protection to be
worn out in the field; the type of equipment to be
used to monitor the problems in the field; and
supplementation of the data base, that is, generat-
ing additional analytical data for sites, allowing
for more flexible and cost effective use of the
CLP laboratories.
As an example, let me relate an actual event that
occurred in a trailer park in Region I. In this
park, the families had been moved to a hotel, at
the government's expense. Clearly this was not a
desirable, indefinite arrangement, so extensive
work was done right in the field to identify the
problems and clean up the park so the occupants
could be moved back in.
The Environmental Response Team (ERT) uses soil
gas analysis to determine emission sources and/or
determine the extent of contamination for
underground plumes. They also conduct on-site
soil gas analysis training seminars. They use
bio-assay analyses and they also teach the Regions
the use of bio-assays at hazardous waste sites.
Many people think it quite remarkable that they use
a Sciex TAGA^ ' 6,000 MS/MS for air sampling and
analyses, which is a half-million dollar piece of
equipment. However, they do obtain real-time in-
formation at very low sensitivities, although they
are not particularly suitable for mixtures. The
ERT uses these samples to identify plumes, and they
send them back to the laboratory for confirmation.
They also use atomic absorption, plasma emission,
HPLC, GC/MS, and possibly X-ray fluorescence.
The ERT is also well known for its evaluation of
commercially available technologies, such as the
Brooker MM1 Mobile Mass Spectrometer, remote
optical sensing by ,F£IR; Summa canisters for
air sampling; Tenax /carbonized molecular sieve
absorbent tubes; the TAGA and various portable GCs.
USATHAMA is providing chemical support to the
Rocky Mountain Arsenal, whose program manager was
charged with the responsibility of restoring the
Arsenal. The clean-up effort is directed toward
remedial excavation of areas that are contaminated
with specific toxic or hazardous compounds at a
defined concentration. USATHAMA has asked Oak
Ridge National Laboratory, operated by Martin
Marietta Energy Systems for the Department of
Energy, to author a document assessing various
technologies available for work in the field in
order to determine their applicability in charac-
terizing the Rocky Mountain samples. At this
point in time, I do not know whether anything but
the written assessment is available, or whether
there has been any testing of these various
methods in the field.
The document which was produced contains many
recommendations, including the use of purge and
trap with a portable GC/PID detector for volatile
samples; the use of heated head space solvent
microextraction or solid sorbent for semivolatiles
for less sensitive needs; testing of the Cole-
Parmer Mlxxor for water sample preparation
(which is for extraction and concentration and
supposedly uses very little solvent and.^h^s suf-
ficient recoveries) and use of a Soxtet ' device
which extracts and concentrates in less than
20 minutes with sufficient recoveries and using
minimal amounts of solvent. There are also many
inorganic recommendations in this document.
The Advanced Field Monitoring Methods (AFMM)
program seeks to identify, adapt, and field-
demonstrate field monitoring methods, and to
facilitate transfer and exchange of information.
There are two main components of the program.
First, basic research utilizes the competitive
process to seek out all of the new technologies
mentioned previously: they should be fieldable,
portable, qualitative, quantitative, sensitive to
the compounds of interests, rapid, and inexpensive.
Second, applied research utilizes readily
available technologies like X-ray fluorescence,
fiber optics, portable GCs, and immunoassays.
My last topic of discussion is the Analytical
Operations Branch, which will provide the link
between data quality needs and expectations for
the generated data quality. It is the focal point
for analytical field methods dealing with the EPA
Regions, our clients. The organic and inorganic
sections will address those technologies which
clients hope to have further evaluated or devel-
oped, and those compounds for which they wish to
use these technologies. As a first step in the
process of guidance, the Field Screening Methods
Catalogue was published. The catalog presents
field screening and analytical techniques being
used by the Regions.
In conclusion, Superfund and its contractors are
moving forward in a, commitment to providing more
effective remediation of Superfund sites by
utilizing the data quality objective process to
match data quality needs with data generated. This
action is leading to increased use of field analyt-
ical methods which is, in turn, streamlining the
process and making more effective use of resources.
(*) Registered trademark.
-------
FIELD MONITORING METHODS AND USE
FOR SUPERFUND ANALYSES, II
Carla Dempsey
Quality Assurance Coordinator
Office of Emergency and Remedial Response
U.S. Environmental Protection Agency
I am going to discuss some current considerations
in the field, which arose because of the new uses
of field analytical screening methods. Because
they have been used only as screening methods for
the past eight to ten years, these Issues were not
problematic up until now. Scott Fredericks will
follow my talk by explaining how a field screening
analytical methods program can be developed and
successfully implemented taking these considerations
into account. Such a program has been successfully
demonstrated for Superfund.
Many times we gather for group discussions to speak
about where a technology is, where we want it to
go, how it can be miniaturized, how methods can be
forced to lower detection limits. I would like to
discuss the use of these methods and to consider
their applications as we develop them. I will
discuss the current thoughts at Superfund regarding
the use, review, and oversight of field screening
analytical methods in the program. I use the term
analytical/screening because these methods are in
actuality used for both screening and analysis.
Field screening analytical methods have been used
by Superfund for the past eight to ten years, with
support from the REM, TES and FIT contracts, among
others. These methods are currently still being
used by Superfund. It is indeed appropriate to
question why their uses are being examined and
doubted after eight to ten years of successful
utilization. The answer is that in the past these
methods were used primarily for the detection of
qualitative differences in chemical concentrations,
not to yield "real" numbers for very rigorous
decisions. Field teams used the methods to assist
them in preliminary health and safety decisions; to
locate hot spots or very contaminated areas within
sites; and to prioritize samples to send back to
the CLP laboratories. Because such decisions were
strictly preliminary, there were very few horror
stories and many success stories related to the use
of these methods.
The SARA schedule for listing deadlines, completing
RI/FS's, and so on, has been greatly accelerated.
In the recent past, the need to acquire data
quickly in order to make site decisions has grown
tremendously. Since field screening methods were
very successfully used in the rapid acquisition of
qualitative data, they were chosen to quicken the
acquisition of quantitative data. It is important
to note that the acquisition of data does not
necessarily determine the role of any site investi-
gation or feasibility study. However, all avenues
to reduce the time involved in any part of the
Superfund process are now being examined, and all
time-saving steps are being utilized.
The transition of the use of field methods from
strictly screening tools to analytical tools is
occurring right now. This is the reason why we must
carefully consider their selection, the QA and QC
requirements, the data review requirements, and the
use of the resultant data. Because the data are
being used to make more demanding decisions, we
must be much more careful in our use of the methods.
How can field analytical methods be used? What
decisions can be supported by the data? Consider
the following statement: Field analytical/screening
data are just one type of Superfund analytical
data. Just as fixed lab analytical data are one
type of Superfund analytical data.
As for any type of Superfund analytical data, the
appropriate uses for such data must be defined in
order to use them correctly. Furthermore, the
definition must be made before the data are
acquired, so that the choices of method and the QC
requirements can be carefully considered. The
review of the data must be also carefully considered,
This will indicate what was accomplished in the
choice of method, and whether the data which were
acquired are usable for the decision they were
meant to support. Also, as with any type of
Superfund analytical data, the planning and review
procedures must be very clearly defined, in order
to make correct use of field screening analytical
data. The method, QA and QC, and review must be
stated in site-specific project plans to assure
that appropriate choices have been made. As for
any other type of Superfund analytical data,
planning must be done and must be documented in the
site-specific project plans and other Regional
sampling documentation.
The choice of which analytical resource to use is
always driven by the requirements of the decision
that the data will support. This concept, called
the data quality objective (DQO) process, has made
it possible for Superfund to use many different
11
-------
types of data. Early in the Superfund process, the
contract laboratory program and Regional laborator-
ies produced only one type of data, one type of
method, one type of very rigorous quality control,
one quality assurance program and a very strict
review procedure for all data produced for Superfund
decision making. Now, we are willing to state the
data's uses including the precision requirements
and the amount of risk allowable for a false pos-
itive or a false negative. This has enabled us to
start utilizing alternative methods, alternative QA/
QC and alternative review requirements based strictly
on the use of that data in the Superfund process.
Now, all types of decisions must be made before
data acquisition in order to ensure that the data
are suitable. Instead of procuring only one
specific type of data for Superfund, we can now use
many different types, all appropriately chosen to
support our decisions. There are minimum require-
ments for any kind of Superfund analytical data, no
matter what type. These requirements include
proper documentation, planning and review as listed
in the quality assurance project plans and the
site-specific Regional plans in each Region. In
addition, appropriate QA and QC must be specified
to support the decisions based on the intended use
of the data.
There are several reasons for these requirements.
Superfund must know the level of quality in any
kind of data. The quality of that data must be
documented somewhere, so that it can be confirmed
at a later date. Because it is documented and
planned for, the data can be assessed with regard
to its sufficiency and appropriateness for support-
ing each decision. In addition to these usual
requirements for all data, there are some special
requirements for field analytical screening methods.
Many field analytical methods are very sensitive,
because the detection limits are very low. They
are outstanding methods. If the object of your
search has been identified, it will manifest itself
through the use of these methods. We must, in
every case, avoid false negatives by identifying
that object before choosing a field method which is
very specific to one analyte or another. We have
some strictly preliminary screening methods to help
us in this identification. But if very close char-
acterization of a site is desired, by very rigorous
decisions, it is necessary to be certain that the
proper tool is being used to detect the chemical
which is present at the site. In order to properly
choose a fieldable instrument or field method, you
must know what you are looking for. The area must
be screened with a broad spectrum analysis such as
ICAP or GC/MS to look for everything.
The second special requirement is the verification
of removal of contaminants by a final broad-spectrum
analysis. This means a search for the contaminants
of concern. This broad spectrum analysis will occur
at an NFL delisting, when remediation has been com-
pleted and you need to determine that clean up has
occurred down to required protective levels for the
site. Many times a field screening tool will be
used in the search for indicator chemicals which
are easy to track at the site. In many cases, there
may not be an examination of all the contaminants
that have been found at the site. Therefore,
before delisting, there must be a search to confirm
that all contaminants of concern at that site have
been properly removed. This again will require
some kind of a broad spectrum analysis to make sure
the site has been cleaned to the protective levels.
The third special requirement is flagging the data
for the type of decision it was intended to support.
This is necessary because of the existence of such
a broad spectrum of field analytical techniques,
some of which are truly screening techniques, while
others are more analytical tools. It is important
that the data are used to support the kind of
decision for which it was gathered. In order to
appropriately use this data for other decisions,
its quality, its method, and the qualitative
procedures which were applied to it must be known.
In the future, SOPs, performance information and
QA/QC requirements for field analytical screening
methods will be much better defined. At that point
it will be much easier to specify their appropriate
use. If we do not have performance information on
these methods, it is almost impossible for a manager
to effectively choose the right method. So it will
be much easier in the future as we gather informa-
tion, and as our skills improve in the use of these
methods, to appropriately choose and utilize them.
In the interim, careful consideration must be given
to each proposed use of field screening analytical
data. It may work very well at one site with one
matrix, but not at another with a different matrix.
Until we get very good performance information and
until we have standard operating procedures which
can be applied from one site to another, each case
must be considered.
This does not mean that these methods should not be
used, but rather they should be used appropriately.
Because of our lack of knowledge, this will be
difficult until we have more information.
Many steps are being taken to assist EPA managers
in their choices of appropriate methods. For
example, a work group at EPA headquarters has been
formed to examine the uses of field methods, and
the producers of field methods (such as DSATHAMA),
and to examine the activities of groups such as ERT
at Edison. The work group is composed of various
users and producers of these methods, and is
expanding to include more Regional representation.
Out of this group some guidance will be issued on
appropriate use of these methods.
Perhaps our most important activity is the develop-
ment of performance information for these methods
based on their use on environmental samples,
preferably as demonstrated at Superfund sites.
We are attempting to use available data such as
that gathered by the FASP program, and to discuss
with users the appropriate functions for such data.
We will look at successful uses of the data to
support decisions, so that this information can be
utilized by other managers within EPA.
We are also updating the catalogue of field screen-
ing methods, which serves two purposes. It will
furnish more refined performance information, and
will serve to transfer the technologies among
12
-------
Regions. It is our link between the users and the
developers of the method. The catalogue will make
appropriate choices of field analytical methods
much easier. It will provide as much information
as possible about the use of these methods, includ-
ing positive and negative aspects of each: matrix
effects, cost, limitations of use, precision and
accuracy, quantitative or detection levels, the
availability of standards, the availability of
SOPs, and QA/QC requirements keyed specifically to
the uses of these methods for specific data
collection activities.
With broader representation in our work group, we
will be able to determine the real need for matrix
performance evaluation materials. We suspect, and
have been told, that in many cases the standards
for real matrix materials are not available,
although they are needed in order to actually
produce precision and accuracy information for
some of these field methods. But in talking to the
users from the Regions, we hope to get more
information and start production of more perform-
ance evaluation materials and standards keyed to
this kind of technology.
We will also produce guidance on the appropriate
use of this technology. The Regions and the EPA
require guidance which will lead managers to
appropriately choose this technology rather than
repeatedly choosing the CLP's fixed laboratory
method. The Field Screening Methods Catalog-User's
Guide is EPA publication number 540/2-88/005. It
will be transmitted within the next two weeks to
the EPA Regions, 50 copies to the Waste Management
Division Directors and 50 copies to the Environ-
mental Services Division Directors. A data base
accompanies the catalogue, which lists a phone
number for further information. We hope to mass
produce it very cost-effectively in the near
future, but all Regions will have a copy of the
data base in the interim.
This manual is a compendium of approximately 31
currently used field methods, including a list of
sites where these methods have been used success-
fully by EPA in the past few years. With every
technology, a technical contact is listed who can
provide user-support to supplement the limited
performance information in the catalogue.
The EPA contractors who would like to receive a
copy of the manual should call (513) 569-7562. It
is recommended that you specify your position, and
give your contract number. The catalogue will then
be mailed directly to you. It is also available
through NTIS (703) 487-4650, for non-EPA contractors
In conclusion, proper planning and oversight of the
use of field analytical screening methods, as with
any new technology, is the key to their successful
use now and in the future.
13
-------
FIELD MONITORING METHODS AND USE
FOR SUPERFUND ANALYSES, III
Scott Fredericks
Section Chief
Site Operations and Contracts Section
U.S. Environmental Protection Agency
I work in the Site Assessment Branch at EPA head-
quarters, which is responsible for the preremedlal
program. Our Branch performs site inspection and
listing inspections to support the listings of
sites on the National Priorities List. A field
analytical screening program has been developed
since the FIT contract is the main tool provided to
the Regions for performing site and listing inspec-
tions. This is a very large contract, exceeding
one million dollars ($1 million) per week. It
supports the performance of detailed site inspec-
tions on various waste sites by a team of workers
in each Region.
When the acronym FASP was first created, it denoted
the Field Analytical Screening Program. In retro-
spect, we realize that it is more appropriately
named the Field Analytical Support Program, because
it provides a number of different methods and
approaches which supplement the CLP's work and
benefit the Region as a whole.
Interest in the FASP before 1984 was somewhat
sporadic. Field screening methods within our
contracts and Regions mostly employed a "sniffer"
type of Instrumentation to increase safety and to
locate hot spots and well screens. Active interest
in the program began in 1984, when there was an
overload on CLP capacity. With a shortfall of REM
contracts (which came on line) and the shortchanging
of the preremedial program, we felt there was an
inappropriate use of resources. So we began to
investigate various methods which could be used for
our data needs, and which would be complimentary to
the CLP. We commissioned a number of studies during
that time which evaluated methods and existing
equipment and determined limitations and costs of
these methods through communication with manufac-
turers. T am quite convinced that utilizing the
field screening methods of Thomas Spittler, Rick
Spear, Tom Yates and others, is a viable approach.
The greatest success In the use of FASP has been
with the Environmental Services Division Directors
and their employees. The Regional people and the
contractors, working in tandem, have been respon-
sible for the program's success.
The FASP program began as a national effort, when
the FIT contracts were reawarded. The need for a
FASP-type of program was actually designed in the
proposal and a fixed budget was allocated. We
indicated that the use of dedicated FIT teams would
be necessary in performing these techniques.
The FASP program was designed to provide flexibility
in meeting Region-specific needs. In other words,
each Region has a great deal of input regarding the
design and operation of its particular FASP program.
We envisioned the use of the programs primarily for
listing site inspections and for support, where a
broad spectrum analysis has already been done, pri-
marily through the CLP. Once we have a fairly clear
picture of the compounds involved, we can select an
instrument and a method and clearly define our data
quality objectives. The FASP can also be used for
screening target compounds, for which the data must
be matched to data quality objectives.
The benefits of FASP are not perceived but realized
benefits: the program does provide quick turnaround
of data to aid real-time decisions. It is econom-
ical, and also possesses capabilities beyond that
of CLP for air monitoring, soil gas sampling and
screening. It enables Improved site characteriza-
tion and provides on-site and Regional managers
with the immediate information they require to make
better decisions and conserve resources. In
addition, we perceive the FASP program as a Regional
resource above and beyond the preremedial program:
It can actually assist in emergency response
actions as well as remedial actions.
I would like to give you a short overview of the
FASP capabilities in Region X, which, along with
Region I, has been our prototype. The FASP program
was developed faster in those particular Regions
than in others.
In Region X, we have 71% of the CLP's capabilities
for volatile compounds; 57% for semivolatile com-
pounds (acids, bases and neutrals); 70% of
pesticides and PCB compounds; and 67% of CLP
metals. This Region uses GC instrumentation with
dual columns, specialized detectors—it has no
MS capability.
We envision FASP facilities functioning as more
than just a mobile lab, in fact as an overall design
composed of base support facilities, support vehi-
cles and instrumentation. The instrumentation can
sometimes be moved from the base support unit into
a mobile unit, and taken out to the field. When
15
-------
the base units and support vehicles were designed,
they were intended to be functional as full-fledged
labs with interchangeable mobile instruments.
We hope that by this spring, the FASP program will
be formed within our Regions with basic facilities,
support vehicles and instrumentation which will
provide various levels of analytical capability.
To convey some of the cost benefits of using FASP,
we compared sample costs in Region X between 1984
and 1988, using capital equipment purchases,
facility modifications and upgrades, expendable
supplies, and labor costs. The labor costs include
documentation, packaging, shipping of samples and
QA review. The CLP costs do not include any FASP
costs for the analytical system preparation, soil
preparation or soil analysis.
In comparing our program with CLP, we do feel that
we have a fairly sophisticated level of precision,
accuracy, and reproducibility; we feel that some of
our savings are real, even though CLP's level of
sophistication and documentation are much greater.
For our decision making purposes, we always rely on
the CLP for a final confirmation. However, there
are the real-time and cost savings which are
responsible for the considerable support the
program receives.
I would like to briefly cite four case studies.
The first involved a former electrical transformer
salvage yard, for which we had determined that we
required about 236 soil samples, (surface and
subsurface) to screen for PCBs. The data was to be
used to guide subsequent sampling allocation place-
ments and to determine the extent of contamination
within the boundaries of this site.
The sample data turnaround ranged from less than
one hour to about one day, and the cost for FASP on
this particular site inspection was $16,538. This
included sample documentation, materials, and
labor. Five percent of the samples were sent to
the CLP for confirmation. The cost was about
$2,100. The FASP cost for work on this site is the
equivalent of 53 CLP samples, although the determina-
tion had been made that 236 were needed. It is
thus evident that our approach reduced the demand
on the CLP laboratories.
The second case involved a road oiling facility.
One hundred twenty soil gravel samples were needed
for PCB analyses. These samples were shipped from
Alaska to our facility in Region X. Daily telephone
reports provided the on-site officer with daily
information for his work. Sample turnaround was a
day or less, and the cost from the FASP was $8,409.
The cost of the CLP's 10% confirmation (which is
typical for our program) was $3,300.
The FASP cost equalled about 26 CLP samples, and we
had determined a need for 120 samples. Again, our
reliance and demand on the CLP were reduced to
about one-quarter of the normal analytical operation
cost envisioned for that particular site.
The third case study involved a former creosote
facility. The Region determined a need for 500
soil samples, screening them for pentachlorophenol
and polycyclic aromatic hydrocarbons.
The sample turnaround time for all of these
analyses was three days or less. The cost to our
contract through FASP was $91,000, and the CLP cost
was about $16,000 or approximately 10%. Yet again,
the FASP cost equalled 189 CLP samples, whereas 500
had been projected. Had we used the CLP exclu-
sively, (and received their level of data quality),
expenditures would have exceeded the above amount
by over $100,000.
Case study number four was a removal site. The
FASP benefited by supporting their need for
real-time data. In this case, we were required to
take 350 multimatrix samples, as well as screening
for PCBs, aromatic and chlorinated volatile organic
compounds. The FASP data were used to guide
removal activities and to verify clean-up action.
Sample turnaround time was 24 hours or less. The
cost to FASP was about $79,000, and these FASP
costs are equivalent to 138 samples, versus the 350
that we actually took.
There were additional savings at this site which
are difficult to quantify. The agency spent
approximately $1.3 million over two months at this
site, which is about $22,000 per day. By using the
FASP program, we were able to complete our work at
the site three weeks earlier than we would have
without the program's support. This yielded a
substantial savings, considering the $22,000 per
day cost.
In summary, the key factor in using the FASP
program is to use it appropriately.
16
-------
SCREENING ENVIRONMENTAL POLLUTANTS AND BIOMARKERS:
THE ANALYTICAL CHALLENGE
Tuan Vo-Dinh
Advanced Monitoring Development Group
Health and Safety Research Division
Oak Ridge National Laboratory
Oak Ridge, TN 37831-6101
U.S.A.
ABSTRACT
Detailed characterization of all chemical
pollutants in environmental samples from industrial
and waste sites is possible using analytical
techniques such as liquid or gas chromatography and
mass spectroscopy. But for many environmental
monitoring programs and assessment applications,
these analytical procedures would be needlessly
time-consuming and expensive. It is often desirable
to have a screening procedure to prioritize
environmental samples before detailed analyses are
conducted on a subset of samples to reduce the
total cost of monitoring programs and environmental
studies. This presentation describes the various
screening techniques such as synchronous
fluorescence (SF) , and room temperature
phosphorescence (RTF) and provides an overview of
advanced analytical techniques and instrumentation
such as surface-enhanced Raman scattering (SERS)
and antibody-based fiberoptics sensors for use to
detect trace levels of chemical pollutants and
related biomarkers in complex environmental
samples.
INTRODUCTION
The potential toxicity of many chemicals at
hazardous waste sites has created an area of great
concern. Analysis of complex environmental samples
is generally conducted using techniques such as
high-performance liquid chromatography (HPLC) or
gas chromatography/mass spectroscopy (GC/MS) .
These analytical techniques, however, are not
employed on a routine and systematic basis to study
all samples because of the high cost involved. To
reduce the total cost of process monitoring or
environmental assessment studies, it is desirable
to use a screening procedure to rank samples so
that a more detailed characterization can be
conducted on a select subset of the samples.
This presentation provides an overview of various
spectroscopic techniques and state-of-the-art
instrumentation for screening environmental
pollutants. An important challenge in chemical
analysis is the characterization of complex
mixtures . Whereas the identification and
quantification of a specific compound at trace
levels remain important goals, the ability to
screen complex mixtures has become a major focus of
current research efforts. The importance of
complex mixtures has arisen from the need to
identify, monitor and understand synergistic and/or
antagonistic effects of multi-component systems.
Methods such as synchronous luminescence (SL), room
temperature phosphorescence (RTF) and surface-
enhanced Raman scattering (SERS) are described.
The trade-off between selectivity, sensitivity and
cost-effectiveness in analysis is presented.
Advanced instrumental systems integrating
fiberoptics, laser technology and immunochemical
methods for potential applications in environmental
analysis are discussed. Another area of great
importance is the application of analytical
techniques to monitoring environmental biomarkers.
This is a challenging area for research and
development, and advances in analytical methodology
and instrumentation are critically needed to
analyze complex biochemical systems in the attempt
not only to monitor the presence of chemicals in
the environment but also to assess the ultimate
effects of these chemicals on global ecosystems and
on human health. Examples of analysis of chemical
and related biomarkers are given to illustrate the
usefulness and cost-effectiveness of screening
techniques for the analysis of complex
environmental systems and for the assessment of
human exposure to toxic chemicals.
LUMINESCENCE SCREENING TECHNIQUES
Fluorescence spectroscopy is one of the most
sensitive techniques for detecting polynuclear
aromatic (PNA) compounds, which are of particular
interest in environmental screening programs.
These pollutants, many of which are suspected to be
carcinogens (1,2), are produced during incomplete
combustion of organic materials and are found in
many industries, incinerators and waste dump sites.
Because of its higher sensitivity, fluorescence
analysis requires less raw material and shorter
sampling times than chromatography. This is
advantageous for environmental assessment since
only small quantities of PNA compounds are obtained
by standard extraction methods.
Despite the apparent advantages, fluorescence
spectroscopy has been limited in analyzing complex
mixtures because the emissions of the various
species tend to overlap, yielding spectra with
poorly resolved structures. This constraint can be
17
-------
overcome using synchronous luminescence (SL)
techniques to improve spectral resolution (3-6).
In conventional luminescence spectroscopy, either
the excitation or emission wavelength is varied.
In SL, the luminescence signal is recorded while
the excitation and emission wavelengths are scanned
simultaneously. This improved selectivity by
narrowing the targeted spectral bands to reduce
emissions from compounds that might otherwise cause
interference.
The fixed wavelength interval is (AA) is maintained
between the excitation and emission wavelengths
during the measurement (3,4). In the synchronous
fluorescence (SF) approach, the optimum wavelength
interval depends on the Stokes shift and is
typically 3 nm (4) For synchronous
phosphorescence (SP), the optimum interval is
determined by the singlet-triplet energy difference
and usually ranges from 100 to 300 nm (3) . The
conventional emission/excitation and the
synchronous fluorescence spectra using different AA
values are illustrated in Figures 1 and 2 for
benz o(a)pyrene.
SL offers instrumentation simplicity. Devices
intended for conventional fixed-excitation spectra
can often be employed for synchronous measurements
with little or no modification. Several
spectrometers are available with provision for
interlocking the excitation and emission
monochromators and the feature can be easily added
to many other units. A variety of environmental
samples have been analyzed to illustrate the
applicability of the SF techniques for screening
PNA compounds in waste water (6,7), air samples
from industrial (8) and residential environments
(9).
Room Temperature Phosphorimetry (RTP^:
Conventional phosphorimetry requires the use of
low-temperature matrices to reduce the collisional
quenching mechanisms and radiationless deactivation
processes. Due to the requirement of cryogenic
equipment and refrigerant, conventional
phosphorimetry has limited usefulness for routine
applications In field measurements.
Unlike conventional low temperature phosphorimetry,
RTF is based on detecting the phosphorescence
emitted from organic compounds adsorbed on solid
substrates at ambient temperatures (10) . The
general approach is to obtain a solution containing
the materials to be analyzed using standard
extraction procedures. A few microliters of the
sample solution are then spotted on a filter paper.
The spot is dried for about five minutes with a
heating lamp then transferred to the sample
compartment of the spectrometer. Measurements can
be performed with any commercial spectrofluorimeter
equipped with a phosphoroscope (10).
RTF sensitivity and selectivity can be enhanced by
mixing the sample of pretreating the filter paper
with a heavy-atom salt solution (11). Salts such
as thallium acetate, lead acetate, sodium bromide,
and cesium iodide are efficient in enhancing
§ 6
o
I 5
O
y- 4
to
LL)
(- 3
UJ
cr
o
_i
u_
i—i—i—i—i—r~
BENZO[Q]PYRENE
EXCITATION
SPECTRUM-
jl
II
II
r"
«!
. / i
/I
Ml
I II
I U
_>
1
1 1
II
300 400
WAVELENGTH ( nm )
500
Figure 1.
Conventional Fixed-Excitation and
Fixed-Emission Fluorescence Spectra of
Benzo(a)pyrene.
E 6
7^1 M n ^1 ' i I i ' i rr
SYNCHRONOUS FLUORESCENCE OF BENZO[o]PYRENE
(a) AX=3nm
A\=15nm
(c) AX=30nm
I I I
360 400 440 360 400 440 360 400 440
WAVELENGTH (nm)
Figure 2. Synchronous Fluorescence Spectra of
Benzo(a)pyrene. (The Line-Narrowing
Principle is illustrated using various
wavelength intervals: AA = 3, 15 and 30
nm) .
18
-------
phosphorescence quantum yields for most PNA
compounds.
Figure 3 illustrates the characterization of
fluoranthene, a PNA pollutant commonly found in
environmental samples, using the RTF techniques.
This figure shows the RTF spectra of a coal liquid
sample spiked with fluoranthene using thallium
acetate as the heavy-atom perturber. The efficacy
and cost-effectiveness of the RTF techniques for
screening complex environmental samples have been
demonstrated in previous studies (9-12).
Although most environmental assessment studies of
PNA have dealt with the solid phase, the compounds
are also found in the atmosphere as vapors. The
vapor phase explains inconsistencies in
conventional analysis (13) and creates a need for
more direct monitoring methods. The problem is
that low concentrations require higher sensitivity
than has generally been available.
A simple screening method for PNA vapors involves a
passive dosimeter that can measure time-weighted
average exposure for periods such as a day. The
proposed instrument would be a lightweight finger-
size badge, meant to be carried by a worker (14).
The sampler might also contain an organic absorbent
in a tube. This would set up a vapor concentration
gradient down the tube and induce transfer of PNA
by diffusion onto the filter paper substrate. The
PNA pollutants collected by the filter paper can be
detected by RTF.
The PNA dosimeter has been used to monitor
individual compounds as well as mixtures under both
laboratory and field monitoring situations.
Figure 4 illustrates the capability of the
dosimeter for detecting quinoline (QUI) and pyrene
(PYR) during the field evaluation. The figure
shows the RTF response of the dosimeter exposed at
different locations a synfuel production plant.
The RTF response of a dosimeter placed in a clean
room (blank) showed a broad emission.
NEW TECHNIQUES ON THE HORIZONS
Surface-Enhanced Raman Scattering (SERS):
Hazardous pollutants emitted from energy-related
technologies, chemical industries, or waste
materials are of increasing public concern because
of their potential adverse health effects. Many
pollutants have chemical groups of toxicological
importance that can be characterized and detected
by Raman spectroscopy.
Raman spectroscopy, however, has not been widely
used in trace organic detection, even though the
information contained in a Raman spectrum is most
valuable for chemical identification. One
limitation of conventional Raman spectroscopy is
its low sensitivity that often requires the use of
powerful and costly laser sources for excitation.
However, a renewed interest has recently developed
among Raman spectroscopists as a result of various
observations that indicate enhancements in the
Raman scattering efficiency by factors up to 10^
when a compound is adsorbed on or near special
| I I I I
545
595
I I I I I [
RTP SPECTRA
X,x= 365nm
SYNTHOIL
SYNTHOIL +
FLUORANTHENE
I I I I I I
J I
I
I
I
500 600 700
WAVELENGTH (nm)
Figure 3. Identification of Fluoranthene in Coal
Liquid (Synthoil) Using Room
Temperature Phosphorescence Screening
Technique.
10
9
<
z
CD
a.
l-
cc
T
T
PASSIVE DOSIMETER
AT SYNFUEL PLANT
QUI
I
FTR
FTR-BOT
VAC TWR
CLEAN ROOM
_L
_L
400 500 600 700
WAVELENGTH (nm)
Figure 4. Detection of Quinoline (QUI) and Pyrene
(PYR) Vapors Using a Simple Personal
Dosimeter at Various Locations in a
Synfuel Plant.
19
-------
metal surfaces. These spectacular enhancement
factors of the normally weak Raman scattering
process help overcome the normally low sensitivity
of Raman spectroscopy. The technique associated
with this phenomenon is known as Surface-Enhanced
Raman Scattering (SERS) spectroscopy. The Raman
enhancement process is believed to result from a
combination of several electromagnetic and chemical
effects between the molecule and the surface (15).
For the past few years, we have evaluated the SERS
technique for environmental applications using
practical SERS-active substrate materials based on
silver-coated microspheres deposited on glass and
filter paper (16-17). A wide variety of organo-
phosphorous chemicals including methyl parathion,
fonofoxon, cyanox, diazinon, formothion, and
dimethate have been investigated (18) . We also
report SERS analysis of chlorinated pesticides
including carbophenothion, bromophos, idichloran,
linuron, chlordan and 1-hydroxychlordene (19). The
detection limits for these pesticides were measured
at nanogram and subnanogram levels. The results
achieved with these chemicals are of great
analytical interest since these chemicals are
difficult to detect by other techniques such as
luminescence spectroscopy due to the weak
luminescence quantum yields of these compounds. A
mixture of structurally related compounds, and a
soil sample contaminated with pesticides were
analyzed by SERS to illustrate the selectivity of
this new technique as a screening tool for
environmental applications (18, 20).
A major advantage of Raman spectroscopy is the
spectral selectivity of the technique for the
analysis of complex mixtures because of the
sharpness of the Raman emission peaks. Figure 5
illustrates this spectral selectivity for the SERS
technique for the characterization of a synthetic
mixture containing benzo(a)pyrene (BaP) , 1-
nitropyrene and pyrene (21).
Immunological Techniques and Instruments:
Immunological methods, which offer the capability
of excellent selectivity through the process of
antibody-antigen recognition, have revolutionized
many aspects of chemical and biological sensor
technologies. Their high specificity and
sensitivity permit the measurement of many
important compounds at trace levels in complex
biological samples. Radiolmmunoassay (RIA)
utilizes radio-active labels and has been the most
widely used immunoassay method. Immunoassays have
been applied to a number of fields including
pharmacology, clinical chemistry, forensic science,
environmental monitoring, molecular epidemiology
and agricultural science (22). The usefulness of
RIA, however, is limited by several shortcomings,
including the cost of instrumentation, the limited
shelf life of radioisotopes, and the potential
deleterious biological effects inherent to
radioactive materials. For these reasons, there
are extensive research efforts aimed to develop
simpler, more practical immunochemical techniques
and instrumentation which offer comparable
sensitivity and selectivity to RIA.
New developments in sensing technology
instrumentation, laser miniaturization,
biotechnology and fiberoptics research have
provided opportunities for novel approaches to the
development of sensors for the detection of human
exposure to toxic chemicals and biological
materials. The development of fiberoptics chemical
sensors has been reviewed (23,24, and references
therein).
For the last few years we have devoted extensive
efforts to integrate immunological methods and
fiberoptics technology in order to develop advanced
in-situ monitoring instruments for chemical and
biological systems. The operating principle of
fiberoptics immunofluorescence biosensors has been
presented previously (25-27). Examples of
measurements will illustrate the application of a
laser-based fluoroimmunosensor (FIS) developed for
the detection of important biological compounds
such as carcinogen metabolites and DNA-adducts of
carcinogens. The FIS instrument derives its
analytical selectivity through the specificity of
antibody-antigen reactions (25,26). Figure 6 shows
a schematic diagram of the FIS device (26).
Antibodies are contained at the tip of the
fiberoptics sensor for use in in-vitro and in vivo
fluorescence assays. High sensitivity is provided
by laser excitation and fluorimetric detection. An
important PNA compound of great interest to
toxicologists and cancer researchers is
benzo(a)pyrene (BP). Studies have shown that BP is
metabolically activated to electrophilic
intermediates, which can bind covalently to DNA. A
specific diol epoxide derivative of BP, r-7,t-8-
dihydroxy t 9,10 epoxy-7,8,9,10
tetrahydrobenzo(a)pyrene (BPDE) was found as the
major carcinogenic metabolite involved in binding
to DNA. Metabolized BP is eliminated through the
urine and feces. Since the carcinogenic activity
of a compound might be associated with the degree
to which it binds to DNA, there has been a great
deal of interest in analytical techniques that are
capable of detecting DNA-carcinogen interactions
and thereby leading to a new approach to monitor
human exposure to PNA compounds. In this study a
new design of the FIS sensor tip was used. The
results of investigations employing a fiberoptics
FIS designed to measure the BP-DNA adduct product,
BP-tetrol (BPT) , indicate that the FIS is capable
of achieving a 40-attomole (10~18 mole) limit of
detection for BPT (28).
CONCLUSION
The development of rapid inexpensive screening
techniques and instrumentation is critical to
decrease the cost of environmental impact studies
and human health assessments. Extensive research
efforts and resources are required to develop and
apply advanced analytical techniques and state-of-
the-art instrumentation. However, current efforts
are often fragmented and constrained by limited
resources. It is our hope that the current
awareness for a clean environment will create a
strong focus, improved coordination and enhanced
collaboration between government institutions
(national laboratories, federal agencies), academic
20
-------
200 400 600 800 1000 1200 1400 1600 1800
RAMAN SHIFT (cm"1)
Figure 5. Analysis of a Multicomponent Mixture by
Surface-Enhanced Raman Spectroscopy
EXCITATION PATH
BEAM
SPLITTER
EXCITATION
\>
_ /
^
EMISSION PATH
OPTICAL FIBER
COUPLER
ANTIBODY
SENSOR
LENS
->• LENS
DETECTION SYSTEM
/\
Figure 6. Schematic Diagram of the Antibody-Based
Fiberoptics Sensor.
21
-------
institutions and the private sector in order to
achieve the ultimate goal of rapid development and
transfer these advanced screening technologies to
the user community.
ACKNOWLEDGEMENT
This research is sponsored by the Office of Health
and Environmental Research, U.S. Department of
Energy, under contract DE-AC05-840R21400 with
Martin Marietta Energy Systems, Inc.
REFERENCES
(1) Searle, C. E., Ed., Chemical Carcinogens,
ACS Monograph 178, Library of Congress,
Washington, DC, 1976.
(2) Gelboin, H. V., and Ts'o, P. 0. P., Eds,
Polycyclic Hydrocarbons and Cancer. Academic
Press, New York, 1978.
(3) Vo-Dinh, T., "Synchronous Excitation
Spectroscopy, " Wehry, E. L. Modern
Fluorescence Spectroscopy, Vol. 4, Plenum
Press, New York, NY, 1981, pp. 167-192.
(4) Vo-Dinh, T., "Multicomponent Analysis by
Synchronous Luminescence Spectroscopy,"
Analytical Chemistry. Vol. 50, 1978, pp.
396-401.
(5) Vo-Dinh, T., and Abbott, D. W., "A Ranking
Index to Characterize Polynuclear Aromatic
Pollutants in Environmental Samples,"
Environ. Intern. . Vol. 10, 1984, pp. 299-
304.
(6) Abbott, D. W. , Moody, R. L. , Mann, R. M. ,
and Vo-Dinh, T., "Synchronous Luminescence
Screening for Polynuclear Aromatic Compounds
in Environmental Samples Collected at a Coal
Gasification Process Development Unit,"
Amer. Ind. Hvg. Assoc. J.. Vol. 47, 1986
pp. 379-385.
(7) Vo-Dinh, T., Gammage, R. B., Hawthorne, A.
R. , and Thorngate, J. H. , "Synchronous
Spectroscopy for Analysis of Polyaromatic
Compounds," Environ. Sci. Technol.. Vol. 12,
1978, pp. 1297-1232.
(8) Vo-Dinh, T., Gammage, R. B., and Martinez,
P R., "Analysis of Workplace Air
Particulate Sample by Synchronous
Luminescence and Room Temperature
Phosphorescence," Anal. Chem.. Vol. 53,
1981, pp. 253-258.
(9) Vo-Dinh, T., Bruewer, T. J., Colovos, G. C.,
Uagner, T. J., and Jungers, R. H., "Field
Evaluation of a Cost-Effective Screening
Procedure for Polynuclear Aromatic
Pollutants in Ambient Air Samples," Environ.
Sci. and Technol.. Vol. 18, 1984, pp. 477-
482.
(10) Vo-Dinh, Room Temperature Phosphorimetrv for
Chemical Analysis, Wiley, New York, NY,
1984.
(11) Vo-Dinh, T., Hooyman, R., "Selective Heavy-
Atom Perturbation for Analysis of Complex
Mixtures by Room Temperature
Phosphorimetry," Anal. Chem.. Vol. 51, 1979,
pp. 1915-1921.
(12) Vo-Dinh, T., "Air Pollution: Applications of
Simple Luminescence Technique," in
Identification and Analysis of Organic
Pollutants in Air. Edited by Keith, L., Ann
Arbor Science, Ann Arbor, Michigan, 1983,
pp. 259-270.
(13) Cautrells, W., Cauwenberghe, "Experiments on
the Distribution of Organic Pollutants
Between Airborne Particulate Matter and the
Corresponding Gas Phase," Atmospheric
Environment. Vol. 12, 1978, pp. 1133-1241.
(14) Vo-Dinh, T., "Development of a Dosimeter for
Personal Exposure to Vapors of Polyaromatic
Pollutants," Environ. Sci. Technol.. Vol.
19, 1985, pp. 997-1003.
(15) Chang, R. K., Furtak, T. E., Eds., Surface -
Enhanced Raman Scattering. Plenum, New York,
NY, 1982.
(16) Vo-Dinh, T., Hiromoto, M. Y. K. , Begun,
G. M. , and Moody, R. L. , "Surf ace-Enhanced
Raman Spectroscopy for Trace Organic
Analysis," Anal. Chem.. Vol. 56, 1984, pp.
1667-1672.
(17) Moody, R. L. , Vo-Dinh, T., and Fletcher, W.
R. , "Investigations of Experimental
Parameters for Surface-Enhanced Raman
Spectroscopy," Appl. Spectr.. Vol. 41, 1987,
pp. 966-971.
(18) Alak, A. and Vo-Dinh, T. , "Surface-Enhanced
Raman Spectroscopy of Organophosphorus
Chemical Agents," Anal. Chem.. Vol. 59,
1987, pp. 2149-2153.
(19) Alak, A. and Vo-Dinh, T., "Surface-Enhanced
Raman Spectroscopy of Chlorinated
Pesticides," Anal. Chim. Acta. Vol. 206,
1988, pp. 333-337.
(20) Vo-Dinh, T., Alak, A., and Moody, R. L. ,
"Recent Advances in Surface-Enhanced Ramar
Spectroscopy for Chemical Analysis,"
Spectrochim. Acta. Vol. 43B, 1988 pp 605-
615.
(21) Vo-Dinh, T. and Moody, R. L. , to be
published.
(22) Smith, D. S., Hassau, M. , and Nargessi, R.
D., "Principles and Practice of
Fluoroimmunoassay Procedures," in Modern
Fluorescence Spectroscopy, Wehry, E. L
Editor, Plenum, New York, NY, 1982.
22
-------
(23) Peterson, J. I., Vurek, G. G., "Fiber-Optic
Sensors for Biomedical Applications,"
Science. Vol. 224, 1984, pp. 123-127.
(24) Seitz, W. R. , "Chemical Sensors Based on
Fiber Optics," Anal. Chem., Vol. 56, 1984,
pp. 17A-34A.
(25) Tromberg, B. J., Sepaniak, M. J., Vo-Dinh,
T. , Griffin, G. D., "Fiber-optic Chemical
Sensors for Competitive Binding
Fluorimmunoassay," Anal. Chem., Vol. 59,
1987, pp. 1226-1232.
(26) Vo-Dinh, T., Tromberg, B. J., Griffin, G.
D., Ambrose, K. R. , Sepaniak, H. J.,
Gardenhire, E. M., "Antibody-based
Fiberoptics Biosensor for the Carcinogen
Benzo(a)pyrene," Appl. Spectr.. Vol. 41,
1987, pp. 735-738.
(27) Vo-Dinh, T., Griffin, G. D., and Ambrose, K.
R. , "A Portable Fiberoptic Monitor for
Fluorimetric Bioassays," Appl. Spectr., Vol.
40, 1986, pp. 696-700.
(28) Tromberg, B. J., Sepaniak, M. J., Alarie, J.
P., Vo-Dinh, T. and Santella, R. M.,
"Development of Antibody-Based Fiber-optic
Sensors for Detection of a Benzo(a)pyrene
Metabolite," Anal. Chem.. Vol. 60, 1988, pp.
1901-1908.
DISCUSSION
ROBERT SNELLING: I have a question, which I believe can be answered by
the audience itself. I spoke earlier about wanting to create a partnership
between EPA and the private sector, to develop and evaluate technologies
which can meet the needs that were discussed this morning. People in the
private sector, however, must also consider the economic aspects of the
problem. How does it profit the private sector to invest in this program? In order
to determine this, they need information similar to that which was presented
here. Where can private investors obtain the information they need to decide
whether an investment in a given program would be profitable?
SCOTT FREDERICKS: I have an immediate response. My upper manage-
ment is one hundred percent behind'this concept. Our REM and TES contracts
all contain options for developing this type of a capability. The FASP is accel-
erating, and we are going to utilize it within all of our new procurement actions
in the future. The private sector also has a strong need to perform a lot of its own
sampling and remedial action independently. In that type of arena, the contrac-
tors will be relying heavily upon the private sector to provide them with the
same cost savings, as it is an important component of their approach to clean-
up work.
With these demands on the private sector, the Agency has finally realized the
advantages of this sort of approach and is ready to make use of them. I think this
symposium is a result of this increased emphasis and acceptance of this
approach by all parts of the community, whether academic or governmental. I
really believe that there is a viable place in the market for this type of screening.
In fact, there is now a small company called Nutriclean, which promotes purity
in agricultural products. Apparently this is going to be a trend in the future,
among supermarkets, which will attempt to entice customers by claiming that
their products - whether fresh or packaged - have been tested for purity. In this
way, a lot of in-the-marketplace testing will occur. I think there is a big future
in this.
AVRAHAM TEITZ: I wish to know the legal defensibility of the FASP data
which was gathered.
SCOTT FREDERICKS: We have been concerned about this issue as well.
We have worked with the contractors who helped us to develop the HRS,
because we are concerned about how well it can be used for listing a site, for
going forth in the rule-making process. We have also worked with our Waste
Programs Enforcement Office to examine whether site data which originate
exclusively with FASP can in fact be used in a court of law. In both cases, the
answer was yes.
The key is to have a very well-documented approach, using trained personnel
who employ certain types of established methods. There must be a standard
chain of control or custody. If trained people are employed and standard oper-
ating procedures are followed, then nobody will have a problem with the data.
AVRAHAM TEITZ: 1 also wondered whether quality assurance oversight is
in place for the FASP program, similar to that used for the CLP program.
MR. FREDERICKS: Quality assurance is an integral part of FASP, as well.
In fact, one of our biggest problems in obtaining support for the FASP process,
is obtaining sample standards for field approaches. They are either not readily
available, or simply nonexistent, and in some cases, we have had to attempt to
produce standards for our own samples that can be used for the instruments
which we employ.
By and large we recognize that quality assurance is an important part of this
process. If you want more information on the technical approach, the last paper
presented in this symposium on Thursday is by Hunt Chapman, who is going
to describe the use of the FASP program on several sites, and to discuss this
point in detail. Also, Andy Hafferty from Region X in Seattle, who was one of
the chemists who helped pioneer this in our Regional office, can provide
additional information.
ALISSA HUDSON: Do you consider legal defensibility a mandatory criterion
for the use of a field screening method, or do you feel that there are other
applications where the method may be helpful, and therefore used, even though
the information may not be used in a legally defensible way?
SCOTT FREDERICKS: It is necessary to establish your intended purpose for
a given site before you take a sample from it. You may take three or four hundred
samples during the course of an overall investigation, hal f of which may not be
of legally defensible quality. However, if you make a final decision to proceed
with a listing, or with a negotiation with the PRP, it is essential that you possess
data which will support that decision.
RICHARD GAMMAGE: I would like to ask a rhetorical question. I have the
impression, which I hope is erroneous, that you have field screening methods
completely under control. You have reduced costs, you have an abundance of
case histories. What is left for the researcher to do?
ROBERT SNELLING: I think there are a lot of tools available to us, and it
has been pointed out that our concern with respect to existing technology is
appropriate use of these tools. We seek to develop protocols for their use so that
they will be used consistently.
But we are also aware that there are a number of emerging technologies which
are not yet commercially available, but offer advancements in our capabilities
to do on-site screening. Immunoassay, for example, is a technology emerging
from the pharmaceutical industry which offers tremendous potential for on-site
analysis. It is difficult to target those assays for the specific analytes in which
we are interested. It is a development effort. The technology that you will hear
about this afternoon, related to the use of fiber optics, optrodes and laser
stimulation, offers a whole range of new tools which could be applied to
hazardous waste sites.
Our first task was to identify what is available for immediate use and to eval uate
and standardize those methods. The second task is to identify those needs which
are not being met and to stimulate evolving technologies that can be applied to
the hazardous waste site characterization program.
23
-------
CARLA DEMPSEY: Managers are often asked at EPA headquarters how
good these field screening methods are. These methods are being used to make
site decisions, but many users are concerned about their precision, their level
of quality assurance. They consider stopping use until these questions are
answered completely, which would be a step backwards.
Right now, we are focusing a lot of attention on the users of these methods, to
find out what decisions are being made with this data, and to determine the
future uses for field screening or analytical methods. We are requesting infor-
mation about the program's current uses and suggestions for improvements to
the program, but we can also develop other applications for these methods from
the current information we receive. In gathering this information, we not only
define appropriate use for today, and target possible future uses, but attempt to
deliver that information back to the private sector and to the researchers who
can develop these areas.
24
-------
MONITORING OF GASOLINE VAPOR AND LIQUID BY
FIBER OPTIC CHEMICAL SENSOR (FOCS) TECHNOLOGY
Stanley M. Klainer, Kisholoy Goswami, Don Le Goullon, D. K. Dandge,
Johnny R. Thomas,3 Stephen J. Simon, and Lawrence Ecclesc
ST&E, Inc.
1214 Concannon Boulevard
Livermore, CA 94550-6002, USA.
ABSTRACT
Gasoline leakage from underground storage tanks
contaminate the drinking water system and poses
a severe threat to public health and to the
environment. To monitor any intrusions into the
ground water and vadose zone, a fiber optic
chemical sensor (FOCS) has been developed which
measures gasoline as a liquid, a. vapor,
dissolved in water and as a gasoline-water
emulsion. This sensor is based on a. special
coating which has a high affinity to gasoline.
The complete sensor system consists of a
portable spectrometer and the gasoline sensor.
The components of the spectrometer include a
tungsten-halogen lamp as the light source,
narrow band filters, a dichroic mirror, a
photodiode detector and associated electronics.
A fiber optic cable is utilized to direct light
into and out of the instrument. The chemical
sensing material is incorporated onto the side
of a short fiber optic core. One end of this
probe is coupled to the long cable and the other
end is impregnated with a fluorescent dye. The
fluorescence intensity of this dye is modulated
by gasoline in the sensing region of the probe.
This intensity variation provides quantitative
information.
Laboratory results indicate that the gasoline
FOCS is specific to gasoline in the presence of
such hydrocarbon mixtures 'as kerosene and jet
fuel. Sensitivity covers the range of < 10 /jL/L
to 100 percent liquid gasoline. Focus has been
on the measurement of gasoline vapors as well as
the vapors of the individual gasoline
constituents. Varying responses are indicated
to these components, the substituted aromatic
hydrocarbons being more responsive than the
aliphatics.
Key Words: Gasoline detection, fiber optic
sensor, underground storage tank
leak monitor
a. FiberChem, Inc., 3904 Juan Tabo NE,
Albuquerque, NM 87111
b. Lockheed-EMSCO, 1050 E. Flamingo Drive, Las
Vegas, NV 89109
c. Environmental Protection Agency, 944 East
Harmon Avenue, Las Vegas, NV 89109
1.0 INTRODUCTION
There is an existing requirement for a gasoline
sensor to monitor: (i) leaking underground
storage tanks, (ii) spills, and (iii) between
the walls of double liner storage tanks. This
capability is urgently needed because the
contamination of drinking and ground water by
gasoline leaking from underground storage tanks
presents a considerable health hazard in the
United States. As this directly relates to the
availability of potable drinking water, the need
to monitor is acute. Early detection of leaking
gasoline is imperative as it would not only
protect the nation's water supplies, but would
prevent costly clean up operations and avoid
heavy Government penalties and fines. Under the
U. S. Environmental Protection Agency'<= UST
(underground storage tanks) program, monitoring
these tanks is mandated [1]. In order for this
to be accomplished, the monitor should be able
to reversibly detect and quantify gasoline as:
(i) a vapor, (ii) a liquid, (iii) dissolved in
water and (iv) as a water emulsion. In
addition, to be practical the sensor system must
be: (i) specific to gasoline and not the
individual additives, (ii) sensitive over the
range of vapor to 100% liquid, (iii) reliable,
(iv) inexpensive and (v) easy to operate.
Several attempts have been made to develop a
gasoline sensor to meet these minimum
requirements [2-6] but none have been
successful. A FOCS that appears to meet these
criteria is described in this report.
2.0 DESCRIPTION OF THE GASOLINE FOCS
In the design of any FOCS the first criterion is
that the device must transmit light. Then it
must be able to interact with a target in a
known way to give a measurable quantity which
can be related to the information desired. For
the FOCS, this is governed by the selected
chemistry and, therefore, the physical design
must support the chosen chemical systems.
25
-------
Although, in the final analysis, a FOGS is a
rather simple device, its design is complex
because it involves many different disciplines
such as: (i) chemistry, (ii) fiber optics,
(iii) immobilization, (iv) optical spectroscopy
and (v) unique instrumentation concepts. Not
only are all of those a consideration in the
overall development plan, but they must be
integrated to make a reliable, rugged, long-
lived system which meets the needs for
environmental monitoring and, in some instances,
diagnostics. The starting point, therefore, is
a set of specifications and the end is a field-
hardened device.
2.1 REFRACTIVE INDEX FOGS [7.81
The gasoline FOGS is based on refractive
index matching. In these types of sensors
the amount of light refracted changes as
the analyte interacts with the coated
surface. This alteration in light
intensity can be directly related to the
concentration of gasoline present.
The refractive index sensor consists of a
bare fiber optic core with a thin clad of
an organic or inorganic compound on its
side. In one design the sensor has a
fluorescent tip formed with an immobilized
dye. An excitation signal is transmitted
through the fiber tip and the fluorescence
emission is used as a constant intensity
light source. This is detected as the
return signal. It is also possible to put
the excitation source and the detector at
the opposite ends of the fiber. A change
in refractive index of the medium
surrounding the fiber alters the
transmission characteristics and results
in a variation in the amount of light that
reaches the detector. Laboratory test
results indicate the ability to quantitate
gasoline over a wide dynamic range. The
schematic diagram of a gasoline FOGS is
shown in Figure 1.
The gasoline sensor consists of a
fluorescent dye, such as rhodamine B or
fluorescein attached to the tip of the
fiber. Thi is done with the help of a UV
curing glue. A 2-cm long coating of a
proprietary material, with selective high
affinity for gasoline is incorporated on to
the side of a fiber core at its distal end.
In the absence of gasoline, the returning
light has a high intensity because air (or
water) has a smaller refractive index than
the core, whereas, in the presence of
gasoline, the intensity of the return
signal becomes reduced because of the
higher refractive index of the analyte.
Figure 1: Schematic Diagram of a Gasoline FOGS
System
The intensity of the return signal
decreases proportionately to the amount of
gasoline present, due to the modulation of
the fluorescent light. This gasoline FOGS
has the following characteristics: (i) it
responds to liquids, vapors, dissolved
gasoline and gasoline-water emulsions;
(ii) it has wide dynamic range, i.e.
percents to <10 pL/L; and (iii) it operates
in either the alarm or quantitation mode.
Furthermore, this technology can be
extended to the measurement of other
species .
3.0 SUPPORT INSTRUMENTATION f9.101
The Fiber Optic Chemical Sensors are designed so
that specificity to a particular molecule or
class of compounds is relegated to the sensor
chemistry and, therefore, only the intensity of
light coming out of the fiber and its
wavelength need to be assessed. Consequently,
it is possible to use a very simple device to
make these measurements. The complete sensor
system consists of: (i) a tungsten-halogen lamp
for a light source; (ii) the gasoline FOGS;
(iii) a spectral sorter; (iv) a photodiode
detector; (v) signal collection, processing and
display electronics and (vi) a readout (meter or
recorder).
4.0 RESULTS AND DISCUSSION
The gasoline sensor was tested in the laboratory
against: (i) liquid gasoline, (ii) gasoline
vapor, (iii) gasoline dissolved in water and
(iv) some of the individual volatile gasoline
components. Figure 2 shows the response of the
FOGS to liquid gasoline and vapor. This
experiment was performed by using an argon-ion
laser as the excitation source (514 nm) and a
photomultiplier tube detector in combination
with photon counting equipment. It is important
to note that the sensor is completely reversible
26
-------
and this indicates that only a physical
interaction is taking place between the coating
and the gasoline. Further testing shows that
this sensor responds preferentially to gasoline
and not to kerosene and jet fuel.
Intensity
Vapor, L Liquid)
(L) (L)
1 .0
0.9
0 .B
0 .7
0.6
0 .5
©Chevron Regular Unleaded
• Chevron Super Unleaded
10
20
30
50
Figure 3:
Response of FOGS to Chevron Gasoline
Vapors (24-Hour Equilibrium)
Time (min.)
Figure 2: Reversibility of Gasoline FOGS
(Computer Trace of Actual Spectrum)
4.1 RESPONSE OF FOGS TO GASOLINE VAPORS
Several brands of unleaded, regular and
super unleaded gasoline samples have been
measured in the vapor state. Each brand
has its characteristic distribution of the
components and consequently slopes of the
response curves are different. Chevron
gasoline, however, has been measured in the
most detail. All samples were made up and
measured at ambient temperatures. The
results obtained for- Chevron gasoline
concentrations between 1 and 50 /iL/L are
shown in Figure 3. The term V/VO is the
normalized response of the gasoline sensor
when various concentrations (/iL/L) of
gasoline are present and V0 is the voltage
reading with no gasoline (7 volts) . V is
the final voltage reading. Figure 3 shows
the response of the FOCS in samples where
gasoline was allowed to equilibrate with
the air over 24 hours.
4.2 RESPONSE OF FOCS TO GASOLINE DISSOLVED IN
WATER
Preliminary experiments were undertaken to
determine if the FOCS could see gasoline
dissolved in water. This was considered <±
good test of the sensitivity of the FOCS
because the solubility of the key
constituents of gasoline is in the low fj.L/'L
range. Water and Chevron unleaded regular
gasoline mixtures (2:1) were made and then
shaken vigorously in a separatory funnel
to give a saturated solution. The
undissolved gasoline was removed and the
water fraction sealed and allowed to stand
for twenty four (24) hours. The center
portion of the water solution was then
analyzed by the FOCS. The results
indicated approximately 13 fj.L/L of gasoline
in the water when the response was
compared to that of the twenty four (24)
vapor phase response. Future testing will
include repeating the experiments with the
saturated solutions as well as those with
less gasoline in them while independently
analyzing the individual water-gasoline
mixtures with a mass spectrometer.
4.3 RESPONSE OF FOCS TO INDIVIDUAL GASOLINE
COMPONENTS
Since gasoline is a mixture of
hydrocarbons, it is very important to know
if the FOCS is responding preferentially to
one or more of the components of gasoline.
27
-------
To determine this, nine (9) of the key
constituents found in all gasolines were
measured using a FOCS. Each of the test
samples was prepared from compounds
purified to spectrographic standards.
Forty (40) /iL/L vapor samples were made for
each species and measured with the sensor
after they were allowed to come to
equilibrium for three (3) hours. The
results are listed in Table 1. The three
(3) hour equilibrium response for Chevron
gasoline is also shown in Table 1 for
reference. Those compounds whose
responses were (AV) more than gasoline are
the ones which respond best to the gasoline
FOCS, i.e. the substituted aromatics.
Figure 4 shows more detailed data for four
of the key aromatic hydrocarbons between 10
and 50 /iL/L.
- Initial Voll*gt
- rtn»l Voltage
TABLE 1
Response of a. FOCS Co 40 pL/L of Individual Gasoline Constituents
3 hrs.
10 15 20 25 30 35 40 45 50
• BENZENE D XYLENES
© TOLUENE A 1,2,4-TRINETHYLBENZENE
Figure 4: Response of FOCS to 10 to 50 /jL/L Vapor
of Aromatic Hydrocarbons
Constituent
2-Methylbutane
n-Pentane
n-Hexane
n-Octane
Cyclohcxanc
Toluene
Xylene
1.2, ti -Trirno thylbenzene
Gasoline
Vv
-------
[7] Le Goullon, D.; Goswami, K. ; Klainer, S. ;
Milanovich, F.; "Fiber Optic Refractive
Index Sensor Using a Metal Clad"; Patent
Pending 1988.
[8] Tanaka, M.; Ono, M. ; Degawa, S.; "Liquid
Leakage Detection System"; Patent #
4,270,049; May 1981.
[9] Milanovich, F.; Daley, P.; Klainer, S.;
Eccles, L. ; "Remote Detection of
Organochlorides with a Fiber Optic Based
Sensor II. A Dedicated Portable
Fluorimeter"; Analytical Instrumentation.
1986, 15, [4], 347.
[10] Kopola, H.; Kaijansaari, R. ; Myllyla, R. ;
"An eight channel fiber optical
sp ectrophotometer for industrial
applications"; SPIE. 1986, 586.
DISCUSSION
UNIDENTIFIED PARTICIPANT: You wrote you had ice.
STANLEY KLAINER: We can tell the difference between water, ice, and ice
crystals, that is, ice at its formation.
UNIDENTIFIED PARTICIPANT: There was really no interference regard-
ing your benzene detector?
STANLEY KLAINER: I can't answer that question completely. I suspect that
there is some interference, but we're not sure at this point. We're just beginning.
The sensitivity is where we want it, and now we'll start worrying about
interferences.
And of course, if the interferences are such that because of the specificity to the
benzene - the ability to absorb the benzene and not absorb other things as well
- we may overcome the interferences in that manner. That's one of the tricks
of picking the right coating.
UNIDENTIFIED PARTICIPANT: So we have two selective processes?
STANLEY KLAINER: Yes. The refractive index only detects the light leak.
A mechanical absorption technique is necessary to obtain specificity and
sensitivity.
JOHN SCALERA: I have two questions about this. First, what about the
stability of the instrument out in the field, since it is dependent upon an optic
system?
Second, do you use monochromatic radiation in a fiber optic system to be
species specific? I know you identify species on the outer coating on a fiber
optic. Do you also use monochromatic radiation to enhance that?
STANLEY KLAINER: First, about the stability of the optics. Obviously, if
there are wide temperature swings there are shifts, and you accommodate them
by using things such as feedback loops, which monitor the amount of light
going into the fiber all the time. That is where the problem is - at the light
transmission.
Regarding the question about the species specificity, you could use monochro-
matic light if you wanted to do some additional spectroscopy. We are not doing
it at the present time. We have been able to find sensing materials which seem
to be sufficiently specific.
The light is monochromatic, but it's chosen more for where it excites the
sample, or where we can detect it best, rather than for its spectroscopic
information content at this point.
JOHN SCALERA: Is there any problem with the focusing system on a box
system out in the field going out of alignment, or do you have solid state
technology?
STANLEY KLAINER: There is always the possibility over large temperature
swings of optics going out of line.
If you sample the signal in the right places, and compare them (essentially use
a double beaming system) you correct for it internally. You can then do some
tricks like chopping it. Or for instance, if you were going to run a system that
you knew was always going to be out in the desert, you might seriously consider
keeping the box at one hundred degrees the whole time, eliminating the tem-
perature swing.
You have options you can play with, you have to face the questions about what
the temperature does to the fiber optic, too. After all, some of the physical
sensors are based upon light transmission, which are what temperature sensors
are made of. You need to correct for the temperature effects in the fiber optic,
as well for the temperature effect in the instrument.
UNIDENTIFIED PARTICIPANT: We used a bulk optics device last year in
about one hundred and twenty degrees, and we saw no differentiation of signals
between the lab and the field. Properly designed bulk optics seem to work fine.
JOHN EVANS: I'm intrigued by that cyanide sensor. Have you determined
whether it is responsive to complex forms of cyanide, or simply free cyanide?
STANLEY KLAINER: I'm going to let Dr. Goswami, who made that sensor,
answer the question.
KISHOLOY GOSWAMI: Right now, these sensors are the irreversible,
integrating type.
MAHMOUD SHAHRIARI: Regarding the aromatic hydrocarbons sensors
you have made, I got the impression that the coatings you are using are actually
acting as chemical indicators?
STANLEY KLAINER: No, the coatings do two things. They have preferen-
tial mechanical sensitivity to the compounds we're looking at, so that makes
them reversible. They are also index matched, so that as the refractive index of
this aromatic mixture coating changes, more or less light leaks out of the side
of the fiber, and you get a change in signal.
MAHMOUD SHAHRIARI: So it is correct to say you are not using any
chemical indicators?
STANLEYKLAINER: No chemical indicators. One of the reasons that we're
not using chemical indicators is because we're really worried about bleaching
of fluorescent compounds.
There is a dye on there, but it is on the tip, and its only function is as a light
source. The tip of the fiber has a fluorofluoron it, in very large amounts that will
never bleach, whose purpose is to give us one color light in and one color light
out, so we can distinguish between the light color in and the light color out.
29
-------
THE SUITABILITY OF SURFACE ENHANCED RAMAN SPECTROSCOPY (SERS) TO FIBER OPTIC
CHEMICAL SENSING OF AROMATIC HYDROCARBON CONTAMINATION IN GROUNDWATER
Michael M. Carrabba, Robert B. Edmonds, Peter J. Marren and R. David Rauh
EIC Laboratories, Inc., Ill Downey Street
Norwood, Massachusetts 02062
ABSTRACT
A general need exists for chemical sensors to moni-
tor the presence, evolution and hydrological trans-
port of aromatic hydrocarbon contamination in oceans,
surface and ground waters. In the most favorable
configuration, a chemical sensor should operate
jjn situ and produce quantitative information in real
time with a high level of sensitivity. We are in-
vestigating the development of a prototype fiber
optic chemical sensing probe based on Surface
Enhanced Raman Spectroscopy (SERS) on electrodes.
The SERS signal is obtained from the Raman scatter-
ing of a molecule absorbed onto a roughened metal
substrate. The metal substrates that we have
investigated with the fiber optic SERS probe (FOSP)
are electrodes made of Ag. With this technique,
the SERS-related phenomena that are chemically
specific, such as the adsorption of organic mole-
cules on metal substrates and the potential depend-
ence of electrosorption, have increased the selec-
tivity. Our results indicate that the FOSP has the
capability of providing information about in situ
organic contamination that is both sensitive and
selective.
INTRODUCTION
"Universal" chemical sensors are needed for the
detection and monitoring of toxic substances in the
environment. Ideally, such a-sensor would produce
information in real time about their presence in low
levels and their chemical structures. In situ tech-
niques for chemical analysis of aromatic hydrocarbon
impurities in water are limited. Specific conduct-
ance measurement made in situ are useful in detect-
ing general levels of ionic contamination, but are
not species specific. Recent advances in the area
of chemically selective fiber-optic sensors or
optrodes have shown the usefulness of fiber optic
spectroscopic probes for the detection of ground-
water contaminations (1,2). But, optrodes are
usually specific for a particular compound. Fiber
optic probes have been suggested for laser-induced
fluorescence monitoring of some organics (e.g.,
benzenoid hydrocarbons, fluorescent dye "tracers")
(2,3). This technique, along with remote fiber
optic UV-visible absorption spectroscopy (4), is
useful in determining the presence or absence of
general classes of pollutants, but yield little
direct structural information. Thus, these spectro-
scopic methods are therefore most useful when a
known species or class is being sought which has
appropriate absorption or luminescence properties.
Our approach to this problem has been to sample
dilute chemical species by adsorption onto surfaces
and then to identify the adsorbates by Surface
Enhanced Raman Spectroscopy (SERS)(5). The combin-
ation of pre-concentration of dilute species due to
specific adsorption, and up to 1Q6 signal enhance-
ment of SERS over normal Raman spectroscopy, should
enable detection well below the parts per billion
level. Raman techniques utilize visible light to
obtain structurally unique vibrational spectra.
Thus, measurements can be made in media such as
water with high infrared absorption. In principle,
laser Raman excitation and scattering signals can
be transported through optical fibers for sampling
remote or hazardous environments.
We have previously reported the feasibility of
using the SERS technique for the detection of
organic water contaminations (5), but the adapting
of optical fibers to a truly remote fiber optic
SERS probe (FOSP) has not been reported. However,
there have been numerous reports of fiber optic
probes for normal Raman sampling (6-14). In this
paper^ we present results on configuring an optical
fiber delivery/collection system for conducting
remote SERS on electrode surfaces and the develop-
ment of remote FOSP.
EXPERIMENTAL
All results were obtained using a Raman instrument
incorporating a Spex Industries Triplemate spectro-
graph and an EG&G Optical Multichannel Analyzer for
detection. The excitation source was a Coherent
Model 70-4 argon ion laser which also was used to
pump a Coherent 599-01 dye laser. Unless otherwise
indicated, an excitation wavelength of 575 nm was
selected using the dye laser, and the intensity
leaving the exciting fiber was 100 mW. Electro-
chemical instrumentation and roughening of Ag elec-
trodes for SERS has been described elsewhere (5).
All spectra are shown at an electrode potential
of 0.6V vs. a SCE.
The fiber optic probes were constructed with all-
silica optical fibers (NA = 0.22) and microbore
polyimide tubing from Polymicro Technologies.
31
-------
Fiber Optic Probe Design
The design of the fiber optic probe incorporated a
small excitation fiber and four large collection
fibers. A theoretical calculation of Plaza et al.
(6) has indicated that a system would have a higher
collection efficiency if the excitation fiber was
small and the collection fiber was large. For
example, the efficiency of collection in water for
a 600 ym diameter fiber goes from 0.065 sr with a
600 ym excitation fiber to 0.132 for a 100 ym
diameter excitation fiber. They calculated an
optimal efficiency of 0.44 sr for a small excita-
tion/large collection (SELC) system that incorpo-
rated a 100 urn excitation fiber with four 600 ym
collection fibers. They have also indicated that
the efficiency could be increased to 0.7 sr if the
collection fibers were placed at an angle of 11
degrees.
The SELC designs are dependent on the overlap of
the cones of acceptance of the collection fibers.
Figure 1 shows the overlap of the acceptance cone
of the collection fibers with the cone of trans-
mission of the excitation fiber for the SELC design
of Plaza et al. using angled and non-angled collec-
tion fibers. The overlap begins at a distance of
750 ym from the collection fiber surface for the
non-angled design and at 380 ym for the angled
design. For efficient fiber optic SERS, the sub-
strate should be placed in the overlap region of
the collection fibers. The size of the illumina-
tion pattern (i.e., spot size) increases as the
distance from excitation fiber to substrate in-
creases, thus the spot size would be greatly
increased over a normal Raman excitation system
(spot size approximately 30 y). For the non-angled
system, the minimum spot size (i.e., the point where
the acceptance cones of the collection fibers over-
lap) would be ^500 ym while the angled system would
have a minimum spot size of <300 ym with an un-
terminated excitation fiber. In order not to waste
the excitation light, the area of the substrate
should be greater than or equal to the area of
illumination.
Experiments have indicated that SERS from a small
area (30 ym spot) of a surface is more than adequate
for sensor applications (5). A small surface area
in our probe should provide an. increase in collec-
tion efficiency, and it would also add to the
compactness of the probe. We attempted to reduce
the size of the illumination spot by placing a lens
at the end of the excitation fiber. One common
method is to mount a microsapphire ball at the end
of the fiber (1). This method would not be applic-
able to our probe, since the mounting mechanism
would interfere with our collection fibers. Instead,
we employed a more practical way of Tensing the end
of the excitation fiber, the laser microfurnace
technique of Russo and co-workers (15). With this
technique, laser light is passed through the fiber
to a target material which absorbs the light. The
light is then readmitted as infrared light which is
capable of melting the tip of the silica optical
fiber into a lens. These types of lenses are ca-
pable of producing spot sizes of approximately half
the fiber diameter. In our optimization experiments,
we evaluated the collection efficiency of the assem-
bly illustrated in Figure 1 using such lens-ended
excitation fibers.
SELC Probe Fabrication Details
In order to construct an optimum SELC probe for
SERS, several factors which complicate the measure-
ment of SERS using optical fibers were addressed.
First is the background fluorescence obtained from
polymeric fiber claddings. Second is an intrinsic
Raman scattering of silica arising from the exciting
fiber which is reflected back into the collection
fibers from the SERS substrate. In addition, an
unstructured background signal is always present in
SERS.
The background fluorescence was easily solved by
switching to silica cladding. Plastic clad silica
fibers were not used due to a broad fluorescence
signal that appeared with laser wavelengths greater
than 575 nm. The other complications have been
addressed by investigating the effect of distance of
the collection optics to the SERS surface, the posi-
tioning of the excitation fiber as well as the
effect of lens-ending the fiber. The placement of
the exciting and collection fibers can alter the
magnitude of the background signal, since it relates
to the intensity of the exciting source (both Ray-
leigh and silica Raman scattering) reflected back
into the collection optics.
The use of lens ended excitation fibers and angled
and non-angled collection fibers were studied for
the fiber optic excitation/delivery system. All of
the SELC probes reported in this investigation have
been silica clad. The SELC probes were constructed
with a central 100 ym excitation fiber (both lens
and non-lens ended) surrounded by four 600 ym
collection fibers. The four outer fibers were
sealed in opaque epoxy around a microbore polyimide
capillary tube large enough to accept the exciting
fiber. The array of collection fibers were con-
structed in configurations parallel to the central
fiber and at an angle of 11°. The probes have been
embodied into bundles that allow easy coupling of
our SERS probe to the spectrograph and to the
laser.
Optimization was carried out for a known SERS system
of an Ag electrode and 0.05M and 0.02M pyridine at
an excitation wavelength of 575 nm. A micrometer
was used to adjust the distance between the fiber
bundle array and the substrate. In addition, the
distance between the central fiber and the sub-
strate was also adjusted.
SELC Experimental Results
Figure 2 is the SERS spectra of pyridine as a func-
tion of the distance of the collection optics to
the SERS electrode surface for an angle type SELC
fiber probe. The results indicate that the optimum
position of the collection optic for the angle SELC
probe is 2.5 mm. The results for the non-angled
SELC probes indicate an optimum distance of 1.9 mm.
In comparing the angled and non-angled SELC (Figure
3), a significant increase in the SERS signal in-
tensity was realized by adjusting the collection
fiber acceptance angle. Depending on the other
variables, this gives up to a 15-fold increase in
the SERS signal level compared to a non-angled
orientation of delivery and collection fibers. At
the same time, enhancing the collection in this way
32
-------
increased the overall background level by a factor
of 3 and the silica Raman scattering level by a
factor of 7.
Enhancement of the overall resolution (ratio of
SERS signal to the silica background) in some cases
was accomplished by placing the excitation fiber
close to the SERS substrate. The levels of back-
ground scattering and silica Raman peaks were
evaluated with the excitation fiber position approx-
imately 0.1 mm above electrochemically roughened
SERS-active substrates. The results indicate that
the background Raman signal from the silica fibers
was reduced if the excitation fiber was placed
nearly touching the SERS surface, while the SERS
signal was relatively unchanged. The effect was
observed regardless if there was a lens on the end
of the excitation fiber. The reduction in the back-
ground silica signal was observed for all the SELC
probe configurations. Figure 3 compares the results
of the excitation fiber positioning for the SELC
probes.
The reduction in background Raman signal is probably
due to the way the light exiting the fiber is di-
rected off the SERS surface. When the fiber is close
to the surface, the directional light (both silica
Raman and Rayleigh) can be reflected back up the
excitation fiber, while the SERS signal is isotropic
and is thus only affected by the position of the
collection fibers.
Our results indicated that the 11 degree angle SELC
bundle placed 2.5 mm above the surface with or with-
out a lens ended excitation fiber nearly touching
the surface is favored over the other type of
probes. One problem with this design is the fabri-
cation of the probe. The assembly of the probe
requires bare (no jackets or mounts) fibers. In
order to have a good collection surface, the fiber
should be polished. But the polishing of bare
fibers is not easily accomplished and thus cleaved
fiber ends must be used. In addition, extreme care
must be taken during the assembly of the probe so
the cleaved surface will not be damaged by epoxy or
breakage. To reduce the assembly complication, a
fiber probe was constructed that had the collection
fibers at an 11 degree angle and the end completely
encased in epoxy. The end was then polished so that
the face of the bundle was flat. Figure 4 shows the
comparison of the 11 degree probe with and without
the end polished flat. This result indicates that
there is an insignificant reduction in collection
efficiency when the end of the angle probe is
polished flat. Thus, the 11 degree angle SELC can
be assembled in both a timely and simple fashion.
Experiments were conducted to determine problems
that might arise due to the length and type of
optical fiber. Two types (low OH and high OH) of
silica/silica clad fibers (Polymicro Corp.) were
investigated. Experiments using the Raman spectro-
scopy of a neat crystalline naphthalene sample at
625 nm and 46 meters of the low OH excitation fiber
indicated an extremely high background (Figure 5a).
The background from this fiber was so intense that
it obliterated the Raman spectrum of the naphtha-
lene. The high OH fiber; on the other hand,
produced a good spectrum (Figure 5), but large
silica bands were still present at this length.
The large silica background may be due to the silica
Raman bands which are travelling as cladding modes
in the optical fiber. The signal travelling in the
cladding can be easily removed by coiling the exci-
tation fiber around a circular rod. We determined
that coiling the excitation fiber around a 1/2"
diameter rod was sufficient to remove the cladding
modes while not dramatically reducing the total
power transmitted by the fiber. Figure 6 shows
that a few coils of this excitation fiber greatly
reduced the silica noise bands by 46% while reducing
the incident laser power and signal strength of the
naphthalene Raman bands by only 5%.
Fiber Optic SERS Probe (FOSP)
Two types of fiber optic SERS probes (FOSP) were
constructed with the housing material of 3/4" black
delrin rod. In the first design, the top half con-
tains the fiber bundle and a Pt counter electrode,
while the bottom half contains a Ag electrode sub-
strate and a reference electrode. The latter are
connected electrically to the top half by external
wires. The bottom half also contains a reservoir
for electrolyte which communicates with the outside
by a sieve-like array of holes. The electrochem-
ical control of the substrate enables repeated
renewal of the substrate surface, and also for
control of the adsorbate species, coverage and
molecular orientation by electrode potential. This
design was suitable for laboratory use, but it does
not allow for the easy interchangeable ity of the
SERS electrode. The second type of probe design
which used pin connectors to electrically connect
the electrode segment of the probe to the fiber
section was also investigated. Pin connectors were
attached to the working and reference electrode
leads in such a way that they can be separated and
connected again while remaining insulated from solu-
tion. The pin connector design is preferable
because the working and reference electrodes are
more easily changed. To maintain proper cycling
conditions, the fixed pin connectors must be iso-
lated from the solutions.
FOSP Experimental Results
Our experimental results on our FOSP are very en-
couraging. Figure 7 shows the SERS spectra of
pyridine as a function of electrode potential. The
electrode was cycled five times between the usual
oxidation/reduction cycling (ORC) limits of -0.6 to
+0.2V vs. SCE. We tested the renewability of the
SERS probe by monitoring the SERS signal from
several of the peaks, a function of repeated ORC
cycling. Figure 8 shows the signal intensity of
the 1006 cnr* as a function of 62 ORCs. The cycling
does not significantly affect the ability to observe
the SERS spectrum. In fact, we placed over 200
cycles on this SERS electrode with no significant
loss of signal. This result indicates the long-
term stability of our SERS probe is probably very
good.
A contamination problem with our FOSP intermittently
appeared at potentials of about -0.2V or higher when
cycling in certain solutions. This peak normally
disappared at lower potentials (about -0.6 to -0.7V)
The contamination is probably from the black delrin
plastic used to make the probe. The peak does not
33
-------
appear when other materials are used. Current probe
designs are incorporating Teflon.
The effect of changing the cycling conditions of an
Ag electrode was examined with the FOSP. The cycling
conditions are important especially in the case of
in situ renewability and durability of the electrode
"smrface. Electrodes were cycled in pyridine solu-
tions which contained either 0.1M KC1 or tap water.
The effect of preroughening (i.e., precycling) the
electrodes in 0.1M KC1 before placing it in the solu-
tion of interest was also examined.
A freshly polished 1 mm diameter Ag electrode was
preroughened by precycling 3 times in 0.1M KC1 with
an oxidation/reduction cycle (ORC) of -0.7 to 0.2
volts vs. a silver/silver chloride reference elec-
trode. After cycling, the solution was removed and
replaced with 0.02M pyridine in 0.1M KC1. Initially,
a weak SERS signal from pyridine appeared (Figure
9a) with the electrode at open circuit. When a
potential of -0.7V was applied, the signal improved
(Figure 9b). The FOSP was cycled again with the
same limits as the precycling and the SERS signal
dramatically improved (Figure 9c). This same elec-
trode was removed from solution and washed with
methanol and water. The Ag electrode was then
placed in a solution of 0.02M pyridine in 0.1M KC1.
The SERS signal appeared without recycling the elec-
trode (Figure 9d). After cycling with the previous
limits, the signal increased (Figure 9e) but was
slightly less than in Figure 9c.
It was determined that in the case of low ionic
strength solutions, stopping the cycle at 0.2V for
various times before cycling back down to -0.7V
improved the signal. The signal improved as the
time held at the upper potential increased (Figure
10). Times longer than 20 minutes were not inves-
tigated. The most practical time seems to be about
10 min.
A preroughened Ag electrode in a solution of 0.02M
pyridine in tap water was also examined. Tap water
was used to approximate a low ionic strength condi-
tion. The best signal was obtained after cycling
and holding at 0.2V for 10 min. An experiment to
determine if freshly polished electrodes could be
cycled (i.e., roughened) in the tap water solutions
of pyridine was also conducted. The SERS signals
obtained by cycling only in 0.02M pyridine in tap
water were compared with those obtained from a dif-
ferent freshly polished electrode which was pre-
roughened in 0.1M KC1. The results (Figure 11)
indicated that the non-preroughened electrodes gave
a signal which was half of the preroughened. Even
though the signal was half than that observed with
the preroughened electrode, the signal level was
stil 1 very useful.
The results of the cycling experiments indicate that
Ag electrodes can be very durable and sustain re-
peated ORC. The results also indicate that pre-
roughened electrodes would be an advantage in an
in situ application (low conductance).
CONCLUSION
Presently, no remote fiber optic chemical sensor
exists for the analysis of organic compounds in
environments which has both the sensitivity and
universal selectivity of the SERS technique. A
major limitation to the effective application of
SERS or any other Raman based technique is the
present stage of development of Raman based instru-
mentation. In order to apply the SERS technique
to a fieldable chemical sensing instrument, future
research should be directed at the development of a
small and low power consuming system that could
operate in a field or laboratory site.
ACKNOWLEDGMENT
This work was supported by the Office of Health and
Environmental Research and Ecological Research Divi-
sion of the U.S. Department of Energy under SBIR
Contract No. DE-AC01-86ER80333.
REFERENCES
(1) Angel, S.M., 1987, Spectroscopy 2_, 38 and
references cited therein.
(2) Malanovich, F., Hirschfeld, T., Miller, H.,
Garvis, D., Anderson, W., Miller, H. and
Kliner, S., 1985, "The Feasibility of using
Fiber Optics for Monitoring Groundwater
Contaminants. II. Organic Chloride Optrode,"
Lawrence Livermore National Report UCID-19774,
Vol. 2.
(3) Chudyk, W., Carrabba, M.M. and Kenny, J.,
1985, Anal. Chem. 57_, 1237.
(4) Seitz, W., 1984, Anal. Chem., 56^, ISA and
references cited therein.
(5) Carrabba, M.M., Edmonds, R.B. and Rauh, R.D.,
1987, Anal. Chem., 59, 2559.
(6) Plaza, P., Dao, N., Jouan, M., Fevier, H. and
Saisse, H., 1986, Appl. Opt., 25_, 3448.
(7) Dao, N. and Plaza, P., 1986, Analysis, M_, 119.
(8) Dao, N., Prod'homme, M., Plaza, P. and Joyeux,
M., 1986, C.R. Acad. Sc. Paris, 302, 313.
(9) Dao, N., Plaza, P. and Joyeux, M., 1986,
Analysis, U_, 334.
(10) McCreery, R., Fleischmann, M. and Hendra, P.,
1983, Anal, Oiem., J55, 148.
(11) Schwab, S. and McCreery, R., 1984, Anal. Chem.,
J16, 2199.
(12) Schwab, S., McCreery, R. and Gamble, F., 1986,
Anal. Chem., J58, 2486.
(13) Yamada, H. and Yamamoto, Y., 1980, J. Raman
Spect., J9, 401.
(14) McLachian, R., Jewett, G. and Evans, J., 1986,
U.S. Patent 4,573,761.
(15) Russo, V., Righini, G.C., Sottini, S. and
Trigari, S., 1984, Appl. Optics, 23, 3277.
34
-------
Substrate
(A)
77.5°
(B)
(C)
FIGURE 1. RELATIONSHIP BETWEEN FIBER ORIENTATION AND MINIMUM SUB-
STRATE SIZE FOR 100 fjm EXCITATION AND 600 /urn COLLECTION FIBERS: (A)
NON-ANGLED; (B) ANGLED AT 11.5°; (C) ANGLED WITH A LENS ENDED EXCITA-
TION FIBER.
3.0 rim
2.5 urn
1.9 rm
1600
1400
1200
1000
800
600
WAVENUMBER (cm-1)
FIGURE 2. THE SERS SPECTRA OF PYRIDINE (0.05M) AS A FUNCTION OF THE DISTANCE TO
THE SERS SURFACE FOR AN ANGLED SELC.
-------
SERS
Spectrum
A
B
c
D
E
F
Collection
Angle(deg)
0
0
11
11
11
11
Lens
no
no
yes
yes
no
no
Exc.
Distance
(mm)
2.5
0.1
2.5
0.1
0.1
2.5
80000 -
70000-
60000 -
^50000 -
£40000 -
oSOOOO -
20000 -
10000 -
1600 1400
1200 1000 800
WAVENUMBEH (cm-1)
600
FIGURE 3. SERS SPECTRA OF 0.025M PYRID1NE ON Ag ELECTRODE.
ANGLE PROBE (11°)
A NORMAL
B POLISHED FLAT
1600 1400 1200 1000
WAVENUMBER Ccm-1)
800
600
36
-------
(fl
Z
111
I-
z
LLJ
UJ
a:
46 METERS
LOW OH
1600 1100 1200 1000
WAVENUMBER (crrr
800
600
z
UJ
I-
I-
LU
a:
16 METERS
HIGH OH
1600 1400 1200 1000
WAVENUMBER (cm-1)
800
600
FIGURE 5. THE RAMAN SPECTRA OF CRYSTALLINE NAPHTHALENE USING
46 METERS OF HIGH AND LOW OH CONTENT EXCITATION FIBER AT 625
11000-
10000 -
9000-
„, 8000 -
H
1 7000-
o
6000-
5000 -
4000 -
3000 -J
UNCOILED
COILED
1600 1500 1400 1300 1200 1100 1000 900
WAVENUMBER (cm-1)
800 700 600
FIGURE 6. THE EFFECTS OF COILING THE EXCITATION FIBER ON THE RAMAN SPECTRA
OF CRYSTALLINE NAPHTHALENE USING 46 METERS OF HIGH OH FIBER.
37
-------
0.0
+0.2
-0.2
-0.4
10000 -
1600 1400 1200
1000 800 600
WAVENUMBER (CM-1)
-1.3V
1600 1400 1200 1000 BOO 600
WAVENUMBEH (cm-1)
FIGURE 7. THE SERS SPECTRA OF PYRIDINE (0.05M) FROM THE FIBER OPTIC SERS
PROBE (FOSP) AS A FUNCTION OF ELECTRODE POTENTIAL.
z
UJ
0.67 MIN
TIME
66.7 MIN
FIGURE 8. THE SERS SIGNAL FROM THE FOSP OF THE 1006 cm"1 PEAK OF PYRIDINE
AS A FUNCTION OF OXIDATION/REDUCTION CYCLING (ORC).
-------
80000
70000
60000
M50000
c-40000
030000
20000
10000
B
1600 1400 1200 1000 800 600
WAVENUMBER (CM-1)
FIGURE 9. THE SERS SPECTRA OF 0.05M PYR1DINE AT 625 nm (100 mW) .
20 MIN
10 MIN
1600 1400
1200 1000 800 600
WAVENUMBER (cm-1)
FIGURE 10. PREROUGHENED Ag ELECTRODE AT -0.7 in 0.02M PYR1DINE AFTER BEING HELD
AT +0.2V FOR VARIOUS TIMES.
39
-------
80000 -
70000 -
60000 -
z 50000 H
o
U
40000 -
30000 -
20000 -
10000 -
PREROUGHENED
1600 1400 1200 1000
WAVENUMBER (cm-1)
800
600
FIGURE 11. COMPARISON OF A POLISHED AND A PREROUGHENED ELECTRODE. SERS
SPECTRA OF 0.02M PYRID1NE IN TAP WATER AT 625 nm AT-0.7V.
DISCUSSION
JOHN SCALERA: Using the potential of the electrode as one of your
parameters to define the species in real sample situations, what difficulties do
you have with poisoning of the electrode surface and throwing the potentials
totally off?
Are you still able to pick up your species on the surface, even though it's
poisoned and the potentials have shifted? By poisoning, I don't necessarily
mean dramatic, but just throwing the potentials off so that you know it's
uncharacteristic of what you're looking for.
MICHAEL CARRABBA: Yes, that may be a problem. We are in the process
now of doing real-life samples in the laboratory, and that will be a question we
will address in the future.
CHARLES MANN: You mentioned that the silica signals in the fiber optics
change with time. What's the nature of the change and its origin?
MICHAEL CARRABBA: I wasn't talking about the fiber optics, the Raman
band itself, changing with time. What changes is the electrode substrate, which
will change the reflectants back into your collection system. If you can prevent
it from getting into your collection system, it's not going to be a problem. The
substrate will vary as an electrode potential, and it may roughen a little bit
differently, from one time to the next.
With all these oxidation/reduction cycles, we observe a difference of about five
percent in the signal intensities, which is pretty good for environmental
applications.
-------
FIBER-OPTIC SURFACE-ENHANCED RAMAN SYSTEM
FOR FIELD SCREENING OF HAZARDOUS COMPOUNDS
T. L. Ferrell E. T. Arakawa R. B. Gammage
J. P. Goudonnet R. C. Reddick J. W. Haas
Health and Safety Research Division
Oak Ridge National Laboratory
Oak Ridge, Tennessee 37831-6123
D. R. James
E. A. Wachter
ABSTRACT
Surface-enhanced Raman scattering permits
identification of compounds adsorbed onto
a metal microbase that is microlithograph-
ically produced with submicron resolution.
Less than one percent of a monolayer of a
Raman active target compound offers a high
signal-to-noise ratio. By depositing the
microbase on the exterior of a fiber optic
cable, convenient field screening or
monitoring is permitted.
By using highly effective microbases, it
is possible to reduce laser power require-
ments sufficiently to allow an economical,
but complete, system to be housed in a
suitcase. We shall present details of a
SERS system of this type and shall show
data on samples of interest in the
screening of hazardous compounds.
Key words: Raman scattering, hazardous
compounds
INTRODUCTION
There is a need for rapid and reliable
on-site qualitative analysis of aqueous
samples taken from aquifers, toxic waste
sites, industrial and agricultural areas,
and other environmentally sensitive
locations. For many targeted compounds,
the adsorption isotherm permits collection
on a surface with suitable filters.
Analysis of the spectra of the compounds,
including deconvolution analysis for
complex mixtures, can be carried out for a
variety of methods that are sensitive to
ultralow levels. Reliability of the
analysis is affected by sample degradation
for the relatively slower analytical
techniques, and the aqueous samples pose
limits for some methods such as infrared
absorption spectroscopy. Surface-enhanced
Raman spectroscopy (SERS) is a relatively
new tool for analytical chemistry which
fills a gap in the methods of attack
available for detection of concentrations
of parts per billion.
Raman spectroscopy identifies molecules by
their vibrational and rotational spectrum,
as does infrared absorption spectroscopy.
However, in Raman spectroscopy one uses
visible light, the Raman spectra appearing
as wavelength shifts from the wavelength
of the incident light. This requires that
the incident wavelength be taken out of
the scattered light by filtering or dis-
persion. Due to the fact that only a
small fraction of the light is Raman
shifted, one needs good filters or a
double monochromator or an optimized
combination. SERS provides signals
comparable to those obtained in Raman
scattering from solutions, but detects a
factor of over one million less in the
number of molecules. For instance, one
reason for this is that silver micro-
structures (upon which the molecules are
adsorbed) actually concentrate the
electric field of the incident light.
Since the scattering is proportional to
the square of the field, this produces a
scattering cross section that is ade-
quately large. The high reflectivity of
silver and the size and shape of the
microstructures are important in opti-
mizing the signal. We have produced
several different types of microstructures
using methods common to the semiconductor
industry. Our results demonstrate a high
performance level for silver microneedles
evaporated at near grazing incidence onto
an evaporated calcium fluoride surface.
Electron micrographs and optical absorb-
ance data have been taken in modeling our
samples. Good agreement has been obtained
with electrodynamical calculations which
model the silver microneedles as prolate
spheroids.
EXPERIMENTAL FEATURES
For on-site measurements it would be
desirable to utilize fiber optic cables
that can be used as extended probes or
left over a period of time for monitoring
purposes. We have demonstrated that SERS
can be carried out using a totally
41
-------
internally reflected laser beam and have
deposited several different types of
microstructures on a prism base and on
optical fibers. It now seems desirable to
attempt to perform stimulated SERS by
increasing the power density through
improvement of the microstructures and use
of higher power lasers. The data we have
obtained to date using internal reflection
show an adequate signal-to-noise ratio,
but further investigation is warranted.
Due to the importance of the shapes of the
microstructures, we have carried out
exhaustive analysis of the geometries
available for practicable systems. We
have employed scanning-tunneling electron
microscopy to obtain shape-effect data.
We have examined a number of compounds to
test the performance of SERS. While it is
more difficult to detect many of the more
symmetric molecules—and in certain cases
fluorescence contributes to the noise—a
wide range of compounds of interest have
demonstrated adequately large SERS cross-
sections. Reproducibility problems have
been encountered which need further
research.
SUMMARY
SERS can be utilized in a convenient,
rapid, and reliable form for a variety of
on-site applications involving ultralow
levels of targeted compounds. Some degree
of development remains to be carried out
in order to provide better reproduci-
bility.
Research sponsored by Div. of Facility
& Site Decommissioning Projects-, USDOE,
under contract DE-AC05-840R21400 with
Martin Marietta Energy Systems, Inc.
DISCUSSION
JOSEPH ANDRADE: Can you compare the enhancement factors you get
with all the various geometries you suggested? I'm particularly curious about
the silver post geometry, as opposed to the roughened electrode, which was
discussed by your colleague.
Second, regarding surface chemistry aspects: if I remember right, the enhance-
ment field drops off over about ten angstroms. What are you doing on the
surface of the silver to minimize water vapor absorption, for example, and to
treat the surface in such a way that you get the enhancement, but still minimize
some of the nonspecificity which will obviously be present?
ERIC WACHTER: To answer the first question, one area of tremendous
importance is to decouple the chemical effects from the physical effects. I think
of the electrode as a celebration of chemical effects, where we're putting a lot
of the physical emphasis for the SERS effect on preferentially absorbing
material.
With the substrate materials, without the application of the electrochemical
approach, we're looking primarily at the effects of the electromagnetic en-
hancement.
I don't think it's possible to directly compare the results from the electrochemi-
cal approach to the post spheres.
For the electrode approach, the surface is relatively uncontrolled, in terms the
microscopic topography.
MICHAEL CARRABBA: Once made, it's very controllable and reproduc-
ible to obtain the same surface over the oxidation/reduction cycle.
We can't controllably make a substrate with, say, ten angstroms, A to B ratio;
but once we have a surface, we can calibrate it.
The best enhancements that we have obtained on the electrodes are around 105.
We might be able to do a little better if we had some controlled electrode
substrates. These have been under discussion between EIC and Oak Ridge
National Labs.
ERIC WACHTER: We need to develop the electrode approach, with some
sort of a tailored surface, so we can combine both phenomena to result in the
highest enhancement.
GREG GILLISPE: What will be the laser for a field unit? Dr. Carrabba
showed that being able to choose the wave length to launch the surface plasma
is very important, and of course the effect is directly proportional to the laser
power. The more laserpower in, the more signal you're going to get, subject to
some background limitations. What are your thoughts on what the ultimate
laser in a field unit might be?
MICHAEL CARRABBA: We hope to have a prototype system to do some
field measurement sometime in mid-February 1989. The advancements in
microlaser technology have been fast and furious, and there are research scale
lasers available that will give you about 100 milliwatts of intensity at 532
nanometers. Those are diode pump YAGs.
One hundred milliwatts is more than necessary for this technique. Most of these
slides that I showed were done at about 100 milliwatts of laser intensity, and
there have been reports from Oak Ridge of 0.7 milliwatts of intensity.
I believe that the laser of choice will be a diode pump YAG, which will have
frequency doubling capabilities. It will give you multi-dimensional capabili-
ties, because you can use the 532 nanometer wavelength for silver, and
hopefully you could use the diode at 680 for copper and gold surfaces.
There has also been work reported by Mike Angel and Bruce Chase of doing
SERS at 1064. There seems to be a large enhancement at 1064, so that's another
wavelength of choice.
Those are three optimum wavelengths with a diode pump YAG system, and
that's the system that we are trying to acquire for our portable instrument.
GREG GILLISPE: You showed spectra with a dramatic difference between
the 575 and 620 excitation. Is that something that can be overcome?
MICHAEL CARRABBA: That's part of the selectivity of the technique,
because with silver, you want to use the 575, and on copper, you want to use the
red wavelength.
Copper gets even better towards 700 nanometers, as does gold. So there is that
advantage to using that diode pump system - taking some of the diode light and
using that for your analysis.
ERIC WACHTER: I think the results you're discussing are perhaps a little bit
deceptive. What one can show is that there will be an optimum wavelength for
a particular substrate material. And if we want to use both silver and copper for
their various chemical and absorption effects, then for optimal SERS enhance-
ment, we're probably constrained to using two different wavelengths.
GREG GILLISPE: But having 532 and 680 would be sufficient? Do you
really sacrifice very little flexibility that way?
MICHAEL CARRABBA: We could use 514 for silver, with no difficulty at
all. 532 is better than the argon wavelength and a little bit less than 575, but the
trade off is not that much.
42
-------
POROUS FIBER OPTIC FOR CHEMICAL SENSORS
M.R. SHAHRIARI, Q. ZHOU, G.H. SIGEL, JR.
RUTGERS UNIVERSITY, FIBER OPTIC MATERIALS RESEARCH PROGRAM, PISCATAWAY, NEW JERSEY
GRANT STOKES
GEO-CENTERS, INC., NEWTON CENTRE, MA 02159
ABSTRACT
A novel fiber optic sensor has been developed for
in line monitoring of chemical species in gas and
liquid systems. The sensor is based on a porous
optical fiber device developed at the Fiber Optic
Materials Research Program, Rutgers University.
In this approach, a small region of a fiber (-0.5
cm) is made porous and a chemical reagent is
immobilized in the pores. The porous section is
an integral part of the fiber. The preliminary
experimental results for detecting ammonia and
moisture in the gaseous state and pH in the liquid
state suggest excellent sensitivity, reproduci-
bility and a wide dynamic range. These results
also suggest the feasibility of developing a wide
range of sensors for on-line monitoring of ground
water contaminants at superfund sites using the
porous glass fiber approach.
INTRODUCTION
The use of optical fibers as components of chemical
sensors for in situ monitoring of different chemical
species is a relatively recent development. In
these sensors light propagating through the fiber
interacts with a reagent that in turn specifical-
ly and selectively interacts with the environment
to be sensed. The typical optical properties
monitored include absorption, fluorescence or
chemiluminescence. The reagents are normally
immobilized a membrane or porous polymer matrix
which is used to coat the fiber tip, or in the case
of evanescent sensors, the fiber cladding. One
of the problems encountered with the fiber optic
chemical sensors based on evanescent absorption
is the characteristic low sensitivity which is due
to limited depth of penetration of evanescent light
into the reagent cladding [1-4].
In this study, we describe a novel approach for
developing a generic, reversible and high sen-
sitivity chemical sensor based on porous glass
fibers. Figure 1 illustrates the principle of
detection for a porous glass fiber as compared with
a conventional evanescent sensor. In a typical
evanescent fiber optic sensor, the sensitivity is
limited both by the depth of penetration of
evanescent light into the reagent coated on the
fiber core as well as the number of internal
reflections. In a porous fiber, the analyte
penetrates into the pores and interacts with the
reagent which is previously cast into the pores.
Since the porous fiber has large surface area,
absorption is enhanced dramatically, leading to
the high sensitivity of the sensor. Another
interesting advantage of a porous glass fiber is
the elimination of problems associated with the
physical and optical coupling of the sensor probe
to the fiber due to the very small sensing region
(about 0. 5 cm in length and 250 microns in diameter)
which is an integral part of the fiber waveguide.
In addition, multiple fiber sensors can be deployed
from a single analytical unit and are expected to
be less expensive than conventional sensors based
on materials cost and ease of fabrication.
FABRICATION OF POROUS GLASS FIBERS
The material used in the porous fiber is an alkali
borosilicate glass with the composition of 60% Si02,
30% B203 and 10 (wt.%) alkali oxides. This type
of glass is chosen as it is a well characterized
system, producible at a low cost, and most impor-
tantly, it exhibits the phenomena of liquid/liquid
immiscibility within a certain temperature range.
The above composition Is melted in an electrical
furnace at 1400°C and cast into rods with a 20mm
diameter and 0.5m in length. The rods are then
drawn into fibers at approximately 700°C by a draw
tower equipped with an electrical furnace. Fibers
with a 250-300 micron diameter and many meters long
are produced in this manner. Fiber are broken into
strands of 5 cm to 10 cm in length. A portion of
fiber (about 0.5 cm) is heat treated in a tube
furnace at 600°C for about 3 hours. At this point,
the heat treated glass is phase separated into a
silica rich phase and a boron rich phase. The boron
rich phase is leached from the glass by placing
the fiber in a bath of 1 normal hydrochloric acid
at 95°C for 3 hours. The fibers are subsequently
washed with distilled water and rinsed with pure
alcohol. Figure 2 shows the flow diagram of the
processing steps for producing porous fibers.
CHEMICAL TREATMENT
Once the fiber is prepared, the porous segment is
cast with the sensing reagent (indicator). This
is accomplished by dissolving the reagent in a
solvent at a predetermined concentration and soaking
the porous fiber in the solution. The reagent then
is dried into the pores by heating the fiber. The
colorimetric Indicator used for ammonia gas sensor
was bromocresol purple (Fischer Scientific Co. B.
393). Bromocresol purple is generally stable at
room temperature and is resistant to photochemical
43
-------
degradation upon exposure to visible light. When
the indicator is exposed to ammonia gas a sig-
nificant absorption peak develops at 580 nm. Cobalt
chloride (CoCl2) was the reagent used for the
humidity sensor. In the presence of moisture CoCl2
can form a hydrated salt having 6 molecules of bound
water. When it is well dried, it appears bright
blue and has high optical absorption between 550
750 nm. When it is hydrated it appears pink and
the absorption peak shifts to 500 nm. Bromocresol
green and bromocresol purple were used in the pH
sensor. In this latter sensor, the reagents were
immobilized onto the pores by activating the surface
of the glass by silanization techniques and coupling
the reagents to the activated glass surface.
RESULTS AND DISCUSSION
Using BET and mercury porosimeter analysis,
micropores and macropores within the porous fiber
were investigated respectively. The distribution
of the pore size from BET is as follows:
Pore Diameter
> 80 A
10-20 A
Pore Volume
3.0 X 10"2 ml/g
14.2 X 10"z ml/g
The pore size results from the mercury porosimeter
is as follows:
80-400 A
2.4 X 10'2 ml/g
The BET results indicate that surface area of pores
is about 200 (m2/g) . Figure 3 shows the SEM
micrographs of cross sections of the porous fibers
before and after phase separation and leaching.
The micrograph (3C) shows clearly that the fibers
have a structure with interconnective porosity.
The preliminary experimental results for porous
glass fibers used as humidity and ammonia gas
sensors indicate excellent sensitivity, rever-
sibility, and reproducibility. The response curves
for ammonia gas and humidity are shown in Figures
4 and 5, respectively. The calibration curves
for ammonia gas, humidity and pH sensors are shown
in Figure 6.
SUMMARY
In summary, a new class of porous glass fiber optic
chemical sensor has been demonstrated. Gases or
liquids permeating into a suitably pretreated,
porous fiber optic core are detected by in-line
optical absorption. Although this paper focuses
on the monitoring of certain chemical species and
pH, the basic design principles of the device are
applicable to the monitoring of a wide range of
liquids and gasses with ground water contaminants
and biomedical sensors among the promising can-
didates for future studies.
REFERENCES
1. J.F. Giuliani, H. Wohltjen, andN.L. Jarvis.
Opt.Lett.8, 54 (1983).
2. A.P. Russell andK.S. Fletcher, Anal. Chem.
Actal, 170. 209 (1985).
3. D.S. Ballantine andH. Wohltjen, Anal. Chem.,
58, 883 (1986).
4. C. ZhuandG.M. Hiefttse, Abstract 606, paper
presented at the Pittsburgh Conference and
Exposition on Analytical Chemistry and Applied
Spectroscopy, Atlantic City, N.J., 1987
5. M.R. Shahriari, G.H. Sigel, Jr., andQ. Zhou,
Proc. of Fifth"International Conference on Optical
Fiber Sensors, Vol. 2, Part 2, 373, (January 1988) .
6. M.R. Shahriari, Q. Zhou, and G.H. Sigel, Jr.,
Opt.Lett. 13, 407 (1988).
44
-------
a) Porous configuration creates large surface areas for maximum absorption
X
b) Light penetration and hence absorption Is limited in evanescent configuration
FIGURE 1
Principle of detection for porous glass chemical sensor (a) as compared with typical
evanescent sensor (b). The porous glass sensor achieves high sensitivity over a
very short sensing region.
COMPOSITION DESIGN
MELTING AND CASTING
FIBER DRAWING
HEAT TREATMENT
LEACHING
SURFACE TREATMENT
FIGURE 2
Processing Steps for Producing Porous Glass Fibers
45
-------
Fiber As Drawn
FIGURE 3A
Fiber After Heat Treatment
FIGURE 3B
Fiber After Heat Treatment And Leaching
FIGURE 3C
SEM MICROGRAPHS OF CROSS SECTION OF A FRACTURED FIBER
46
-------
1.0 -
0)
u
CO
.a
0.5 -
0.0
35
5 10 15 20 25 30
Time (minutes)
FIGURE 4
Response Curve for Porous Glass Ammonia Sensor at Different Ammonia Concentrations.
so
o 5 10
Time (minutes)
FIGURES
Response Curve for Porous Glass Humidity Sensor at Different Humidity Levels.
47
-------
1.5
1.0 —
E
c
o
CO
in
03
0)
1 0-5
n
o
u>
0.0 -
O
I
I
0123
Ammonia Vapor Concentration (ppm)
FIGURE 6A
Calibration Curve for Ammonia Gas Sensor Based on Porous Glass Fiber.
60
50 --
40 --
.« w"
„ =
£ £• 30 - - I
20 --
10
10 mg/mp CoCI,
139 mg/mpCoCI2
213mg/mpCoCI
0 10 20 30 40 50
Relative Humidity at 25°C (%)
FIGURE 6B
Calibration Curves of Relative Humidity Sensor Based on Porous Glass
Fiber Treated with Different CoCI2 Concentrations.
48
-------
0)
o
I
o
in
1.0
0.8 -
0.6 -
0.4
0.2
0.0
Bromocresol
green
at 615 nm
Bromocresol
purple
at 580 nm
I
J_
I
I
8
10
234567
PH
FIGURE 6C
Calibration Curves for pH Sensor Based on Porous Glass Fiber with Immobilized Inidcators.
DISCUSSION
STEVEN SIMON: What pore size is on there?
MAHMOUD SHAHRIARI: We can control the pore size, depending on what
we want to sense, what kind of indicators we want to put there, and what the
molecular size of the indicator is. Regarding limitations, the smallest pore is
about 10 to 20 angstroms, and the largest is about 800 angstroms.
STEVEN SIMON: What if you went into a real-world environment? For
example, have you tried the pH sensor, particularly, or anything other than
laboratory standards? My concern would be clogging of the sensor.
MAHMOUD SHAHRIARI: Yes, clogging is obviously one serious problem
to be considered.
STEVEN SIMON: When you make the tip, it's part of the fiber optic cable.
What would you do in the case of nonreversible systems, or probes versus a
reversible sensor?
MAHMOUD SHAHRIARI: So far, we have only been considering the
reversible sensors, not the probes. So for humidity, ammonia gas, pH, and for
future applications, we would like to use reversible indicators, rather than ir-
reversible.
STEVEN SIMON: You're looking mostly at absorption base systems. Have
you calculated based on the fluoropher approximately what kind of sensitivity
you could get for something like ammonia?
MAHMOUD SHAHRIARI: Yes, we are currently working on a fluorescence
sensor for carbon monoxide and carbon dioxide sensors. Those are reversible.
CHARLES MANN: Given the effectiveness of the electrochemical pH
measurement, what is the advantage of the type of pH sensors that you are
suggesting? An electronic sensor would presumably have as great a physical
range as the fiber optic one, wouldn't it? It seems that the pH electrode does
essentially the same thing, and I was inviting some instruction as to why the
fiber optic system is superior.
MAHMOUD SHAHRIARI: I would say that the advantages of fiber optic
systems over electronic sensors include electromagnetic immunities, small
size, cost, and the like.
49
-------
IMPROVED LUMINESCENCE TECHNIQUE FOR SCREENING AROMATIC CONTAMINANTS
IN ENVIRONMENTAL SAMPLES
R. B. Gammage J. W. Haas, III
G. H. Miller T. Vo-Dinh
Oak Ridge National Laboratory
Health and Safety Research Division
P. 0. Box 2008
Building 7509, Mail Stop 6383
Oak Ridge, Tennessee 37831-6383
ABSTRACT
The sensitive spectroscopic technique of
synchronous fluorescence (SF) is readily
applied to screening environmental
samples for semi-volatile aromatic
contaminants. The technique is
applicable to water and soil samples,
either directly as water or as solvent
extracts. This screening technique
permits identification of semivolatile
EPA priority pollutants or groups of
pollutants, such as polynuclear aromatic
(PNA) hydrocarbons of common ring size,
in key environmental samples.
Conversely, negligibly contaminated
samples requiring no further analyses and
expense can be screened out. Examples
are given of coarse but direct screening
analyses of contaminated samples of
groundwater. The total synchronously
measured fluorescence has been expressed
in units of total PNA referenced to the
SF from a standard reference sample of
sixteen PNA compounds.
INTRODUCTION
Very large numbers of samples from
hazardous waste sites are being collected
and analyzed by conventional methods to
characterize near-surface and subsurface
contamination. Analyses for some
categories of organic pollutants, such as
semivolatile organics, are expensive,
often costing several hundred dollars for
a single class of compounds (1) . The
cost typically exceeds $2,000 for the
full battery of semivolatile,
PCB/pesticides and volatile organic
compounds.
Because current procedures for conducting
environmental surveying and monitoring
are extremely costly and time-consuming,
alternate or adjunct rapid screening
procedures are very attractive. For
example, chemical screening tests,
preferably ones having the sensitivity of
prescribed EPA standard procedures, can
be used to inexpensively identify
uncontaminated samples requiring only
limited or no further action and expense.
In addition, one can cost-effectively
identify samples with significant levels
of specific compounds or groups of
pollutants. One is then able to devote
more definitive and costly analytical
procedures only to samples deserving such
attention and expense. Screening heavily
contaminated samples is also valuable in
preventing excessive levels of organics
from entering a GC/MS instrument such
that down time of hours or days are
required to return the instrument to a
clean operating condition.
A relatively simple method exists that
allows large numbers of samples to be
analyzed for volatile organics in a
relatively short period of time. This is
the EPA Method 5020 or Headspace Method.
There is no analogous method available
for screening samples containing
semivolatile organic contaminants. The
need for new techniques suited for field
screening of semivolatile organic
contaminants was highlighted recently in
a guidance document for fundamental
research in subsoils and groundwater (2).
In this paper we describe a candidate
method for screening environmental
samples that contain luminescing
semivolatile aromatic contaminants; this
spectroscopic technique is based on
synchronous fluorescence (SF) (3) with
sensitivity capable of detecting ppb
concentrations of strongly fluorescing
compounds.
SPECTROSCOPIC TECHNIQUE
The use of luminescence techniques to
analyze complex environmental samples is
often limited in spite of inherently
excellent sensitivity. The problem is
usually caused by spectral overlap due to
large numbers of luminescing compounds in
real-life samples. Our aim for screening
51
-------
is to make improvements in analytical
selectivity without seriously
compromising simplicity in experimental
protocol or sensitivity.
Synchronous Fluorescence (SF)
In synchronous excitation spectroscopy
(4), the luminescence signal is recorded
while both emission and excitation
monochromators are scanned
simultaneously. The wavelength interval,
AA, between the two monochromators is
kept constant throughout the measurement.
The spectra of the individual components
are simplified and the bandwidth
narrowed. We are applying the concept
most often in SF because of its direct
application to the analysis of liquid
samples such as groundwater.
When dealing with a single fluorescing
compound, the optimum value of AA for
spectral simplification is determined by
the Stokes shift, i.e., the wavelength
difference between the 0-0 bands in
emission and absorption.
In the case of a sample containing a
complex mixture of fluorescing PNA, we
often find that the optimal selectivity
and sensitivity are attained when AA is
set at 3 to 5 nm.
The concept of synchronous excitation can
also be used with room temperature
phosphorimetry (RTF) (5,6). The optimum
value of AA is the singlet-triplet energy
difference of the compound, which is
usually between 150 and 250 nm. The RTF
technique can be used to provide
screening information complementary to
that obtained by SF. The particular
groundwater samples discussed in this
paper produced only barely measurable
RTF. Consequently the RTF screening
technique will not be discussed further.
Readers interested in RTF applied
successfully to the screening of
environmental samples are referred to
reference 3.
APPLICATIONS
Preliminary tests have been performed on
groundwater taken from wells at the Bear
Creek burial grounds located on the
Department of Energy reservation at Oak
Ridge, Tennessee. The hazardous organic
waste is composed principally of
degreasing solvents and transformer and
machine oils. Anthropogenic luminescing
constituents are associated with these
waste oils.
Synchronous Fluorescence of a Standard
Reference Sample
The SF was determined for a National
Bureau of Standards (NBS) reference
sample of 16 PNA (NBS SRM 1647) . The
PNA-containing solution of acetonitrile
was added to water to give dilutions of
1:100, 1:200, 1:600, 1:1000 and 1:3000.
At least three SF measurements were made
at each concentration. The SF intensity
was measured and integrated in steps of 1
nm over the wavelength range of 240-560
nm. Blanks were also run and their
integrated SF subtracted in order to
produce the 5-point calibration curve
shown in Figure 1.
Synchronous Fluorescence of Polluted Well
Waters
Withdrawn samples of well water were
stored in a refrigerator. Prior to the
SF screening, the water samples were
passed through a 0.45 pm membrane filter
to remove particulate matter. The
selectivity of the SF technique is
demonstrated in Figure 2. Individual
compounds have not been identified.
However, 2- and 3-ring PNA generally
fluoresce at 300-400 nm while 4- to 6-
ring PNA fluoresce at 400-600 nm.
Samples can thus be coarsely categorized
according to ring size using SF
screening.
These same samples were ranked by
measuring and integrating the SF
intensities at selected wavelengths and
comparing the integrated SF to that from
the reference sample of PNA. In this
manner the SF from the well water samples
shown in Figure 2 have been converted
into PNA equivalent units of the NBS
reference sample.
This type of direct measurement on water
samples is easy to make but gives only
limited qualitative information. One
drawback is that PNA bound to particulate
matter were removed in the filtration
clean-up step and, therefore, missed in
the SF measurement. A better and only
slightly more complex screening
measurement might be to enhance the PNA
solubility by addition of a water-
miscible solvent such as ethanol.
Another problem is fluorescence from
naturally occurring compounds such as
those produced by decaying vegetable
matter (7). Future studies will need to
address the problem of discrimination
between natural and anthropogenic
fluorescing compounds.
SF screening has also been used to
monitor temporal changes in the
fluorescing compounds of several well
waters. In the instance of two ground-
water samples collected from well GW 39
six months apart, Figure 3, the nature of
the fluorescing constituents remained
unchanged. Referenced to the NBS
standard, the PNA in the wellwater were
equivalent to between 1 ppm and 2 ppm.
52
-------
At the time that well GW 39 was drilled,
the core soil samples were analyzed by
EPA reference methods and found to be
free of any PNA contaminant. There is,
however, a nearby oil-retention pond. It
is conceivable that the traces of
fluorescing material we are detecting
have as their source the contaminated
pond water. An example of a fluorescing
component newly appearing in ground water
is shown in Figure 4. For well GW 15 the
new constituent fluorescing at about 500
nm has an equivalent concentration of 5
ppm. Examination of Figure 5 suggests
that the same fluorescing compound might
also have appeared recently in GW 23.
Otherwise, the SF spectrum for well water
GW 23 has remained qualitatively little
changed over a period of two years.
SUMMARY
The technique of SF has a previously
demonstrated utility for enhancing
selectivity in the analysis of
environmenta 11y-re1ated samples
containing luminescing organic
contaminants. Preliminary testing
indicates that this same technique can be
applied to the measurement and screening
of groundwater contaminated with
luminescing semivolatile organic
compounds. Measurement of the SF gives
an indication of the dissolved PNA
content in a groundwater sample. A
coarse ranking of a series of ground-
water samples can, therefore, be made
according to their PNA content. A semi-
quantitative estimate of the total PNA
content of several groundwater samples
was made by reference to the SF from an
NBS standard of sixteen PNA compounds.
Qualitative differences in the
composition of fluorescing constituents
could be discerned quite readily. This
property permits one to monitor ground-
water periodically and detect the
appearance or disappearance of specific
entities.
ACKNOWLEDGMENTS
Research sponsored by the U. S.
Department of Energy, under contract DE-
AC05-840R21400 with Martin Marietta
Energy Systems, Inc.
1985, "Groundwater
Sci. Technol.. Vol.
REFERENCES
(1) Dowd, R. M.,
Monitoring," Environ.
19(6), p. 485.
(2) U.S. Department of Energy,
1985,"Site Directed Subsurface
Environmental Initiative," DOE/ER/0344,
Office of Health and Environmental Health
and Office of Energy Research.
(3) Vo-Dinh T., Bruewer, T. J., Colovos,
G. C., Wagner, T. J,, and Jungers, R. H.,
1984, "Field Evaluation of a Cost-
Effective Screening Procedure for
Polynuclear Aromatic Pollutants in
Ambient Air Samples," Environ. Sci.
Technol., Vol. 18, pp. 477-82.
(4) Vo-Dinh, T., Gammage, R. B. ,
Hawthorne, A. R. , and Thorngate, J. H.,
1978, "Synchronous Spectroscopy for
Analysis of Polynuclear Aromatic
Compounds," Environ. Sci. Technol. . Vol.
12, pp. 1297-1302.
(5) Vo-Dinh, T., 1984, "Room Temperature
Phosphorimetry for Chemical Analysis,"
John Wiley, New York.
(6) Vo-Dinh, T. and Hooyman, J. R. ,
1978, "Selective Heavy-Atom Perturbation
for Analysis of Complex Mixtures by Room
Temperature Phosphorimetry" Anal. Chem..
Vol. 50, pp. 1915-21.
(7) Thurman, E. M., "Organic
Geochemistry of Natural Waters," Chapter
10. Aquatic Humanic Substances, ISBN 90-
247-3143-7, 1985.
53
-------
10 ,
1.0 .
0.1
0.01
AA = 5 nm
0.01
0.1 1.0
Concentration (ug/mL)
10
FIGURE 1 CALIBRATION CURVE OF SF INTENSITY FOR A NATIONAL BUREAU
OF STANDARDS REFERENCE SAMPLE OF 16 PNA COMPOUNDS
ORNL-DWG 88-13169
03
2
LU
LU
O
z
111
O
in
UJ
en
O
in
O
O
IT
I
O
I
300 400
WAVELENGTH (NM)
500
FIGURE 2 SYNCHRONOUS FLUORESCENCE FROM POLLUTED GROUNDWATER SAMPLES
REFERENCED TO A 16 COMPONENT MIXTURE OF POLYCYCLIC AROMATIC HYDROCARBONS
54
-------
GW39
i = 5 nm
•1-2ppm'totalPNA
Oot1988
May 1988
300 400 500
Emission Wavelength (nm)
GW15
AA = 5 nm
"5 ppm" total PNA
May 1988
October 1988
300 400
Emission Wavelength (nm)
500
FIGURE 3 EXAMPLE OF WELL WATER IN WHICH TRACES
OF FLUORESCING MATERIAL REMAIN RELATIVELY
UNCHANGED OVER FIVE MONTHS
FIGURE 4 EXAMPLE OF NEW FLUORESCING MATERIAL
APPEARING IN WELL WATER OVER A FIVE-MONTH PERIOD
October 1988
June 1986
-t-
300 400 500
Emission Wavelength (nm)
FIGURE 5 QUALITATIVE CHANGES IN THE SYNCHRONOUS FLUORESCENCE
SPECTRA OF A WELL WATER AFTER TWO YEARS
55
-------
DISCUSSION
MARTIN VANDERLAAN: I'm having a little trouble with the idea of large
polynuclear aromatic hydrocarbons being in water.
Are these on suspended particulates in the water, and what do suspended
particulates do to your NBS standard? Do you get self quenching, or anything
like that if it is on particulates? Are you really measuring filtered water, or are
you measuring the paniculate phase?
RICHARD GAMMAGE: In the case of the NBS sample, it comes in a
solvent, and we take a little bit of it and dissolve it in pure water. Thatcalibration
curve is for pure material, dissolved in water at low concentrations, where
we're not getting any saturation and precipitation.
When we go to the field samples, we're taking the well water, as is, and making
measurements. So if there are problems with material being on colloidal
particles, as opposed to being in true solution, we're measuring the gross effect
of those two. We haven't done any work yet to distinguish between the two.
MARTIN VANDERLAAN: I don't have any trouble if you rank order
samples. I do have trouble if you compare to a standard curve, which is in
something that really isn't the same matrix that you're measuring. You just said
you didn't correct for things like pH, and we know pH is going to influence
fluorescence. Other things like that have to be controlled.
RICHARD GAMMAGE: You're absolutely right, and it's really just a scrude
screening technique at the moment.
TUAN VO-DINH: In fact, the polyaromatic hydrocarbons, the ones that have
no heterocyclics are very insoluble in water. For benzoin, I think the solubility
is about 10"6 molal. But what we did here is that in the standard, the trick is to
dilute with ethanol. One-to-one ethanol increases the solubility tremendously,
for one thing.
Secondly, what we measure in the well samples were filtered through. We don't
know whether it's full of hydrocarbons, or metabolites, or the hydroxide of this
compound. What we see here is a true screening. In fact, if you have a lot of
polyaromatics, and you prepare samples, you have to be sure that you find a
compatible solvent, and in that case we dilute in ethanol.
JOHN EVANS: What sort of apparatus are you using? Is this commercially
available equipment, or is it home grown?
RICHARD GAMMAGE: We used commercially available equipment. A
regular spectrofluorimeter, and the synchronous monochromate is now a
standard option that one can buy. And even with the complicated instruments
that we buy ($12,000-$ 15,000), 1 think one can take these sorts of measure-
ments with a much cheaper model, probably down in the $5,000 range.
-------
Detection of Solvent Vapors Using
Piezoelectric Sensors
E.B. Overton
D.A. Gustowski, L.H. Grande,
H.P. Dharmasena, P. Klinkhachorn,
C.S. Milan and G.R. Newkome
Institute for Environmental Studies
Louisiana State University
Baton Rouge, Louisiana
ABSTRACT
The ability to separate and quantitate components in
complex mixtures, such as air samples from the
environment, is found mainly in intricate analytical
instrumentation used in the laboratory. There arises
the need for portable and reliable on site equipment
for environmental testing. Such needs include
responses to hazardous chemical spills, waste site
cleanups and monitoring work areas for potential
exposure to toxic chemicals. The methods employed
in this instrument development project include
development of rugged, field deployable
instrumentation. Our approach in this case is to use
neutral organic host complexing agents as coatings
on quartz crystal microbalances (both bulk and
surface acoustic wave type) to develop molecular
size selective detectors for analysis of volatile
substances.
INTRODUCTION
Piezoelectric quartz microbalances (PQM) and
surface acoustic wave (SAW) devices have been of
interest for the development of analytical vapor
detectors. Since King (1) introduced the concept in
the mid-sixties, chemists have used oscillating
crystals as sensitive gravimetric detectors. When
stimulated by an external electronic circuit, these
crystals oscillate at a precise frequency due to the
formation of dipole moments in the crystalline
molecule. If any perturbation occurs at the surface of
the crystal, in particular a change in mass, a
corresponding change in the frequency of oscillation
will occur. This relationship between frequency and
mass change was first described by Sauerbrey (2) in
the equation
AF= 2.3 x106 F2 (AM / A
0)
where: AF = change in frequency; F = fundamental
resonance frequency of the crystal; A = area of the
coated crystal; and AM = change in mass caused
by sorption of vapors onto the coated crystal.
The high oscillation frequency of these crystals leads
to extremely high mass sensitivity. 10MHz quartz
crystals have approximately a 1 Hertz change in
frequency for every nanogram adsorbed onto the
crystal. The SAW vapor sensors are similar to the
bulk wave piezoelectric quartz crystal sensors but
have the advantage of substantially higher
sensitivity. This is due to their higher operating
frequencies (note that AF is proportional to the
square of its fundamental frequency). For example, a
158MHz SAW device can provide resonant
frequency shifts of about 400Hz for a one nanogram
change in mass. Analytical applications are
designed to use the high mass sensitivities of
vibrating quartz crystal microbalances to detect low
levels of volatile chemicals that can be sorbed*onto
the vibrating crystals' surface.
Though highly sensitive, the crystals are not selective
since they respond to any change of mass at their
surface. Based on polar interactions of molecules,
coatings have been used to couple selectivity with
the sensors sensitivity. Various coatings have been
utilized to distinguish different solvent vapors using
the piezoelectric quartz microbalance and the
surface acoustic wave device. Stationary phases
commonly used in chromatography, such as tenax
and OV-1, have been used (1).
In the present study, host-guest complexation was
investigated as a potential method for differentiation
of solvent vapors for the PQM and SAW devices.
Host-guest complexation occurs when a large
organic molecule (the host) combines with a smaller
solvent-type molecule (the guest). This phenomenon
is known as inclusion. In general, the host molecules
have two properties which enable them to form
neutral complexes. These properties are a cavity of
a specific size and shape, and the ability to establish
some type of attractive force between the host and
the guest which is reversible. For any molecule to
become a guest, it must be precisely oriented to and
of the same physical dimensions as the cavity in the
host molecule.
57
-------
The response of a liquid coated quartz crystal to
various volatile solutes can be mathematically
described by combining the Sauerbrey equation and
the partition equation (3)
AF= 2.3x106F2W|iq MS/M| P/P°
(2)
where: W|jq is the weight of the nonvolatile liquid;
M|jq is its molecular weight; Ms is the molecular
weight of the solute; P is the partial pressure of the
solute vapors; P° is the vapor pressure of the solute
at the temperature of adsorption and v is the activity
coefficient.
Examination of this equation reveals that the
frequency response of the crystals is proportional to
the term P / P° and the reciprocal of v. If the solute is
a relatively volatile substance, i.e. it has a high vapor
pressure (P°) at ambient temperature, the term P / P°
will be a small number. Consequently, the frequency
response for detection of a volatile substance will be
small. This factor limits the usefulness of quartz
crystal microbalances to the analyses of relatively
nonvolatile compounds, and of course, relatively
nonvolatile compounds are not generally considered
as inhalation hazards unless they have extreme
toxicities (i.e. chemical nerve gases).
By using host-guest complexation, the P° term may
be effectively lowered when the volatile component
is included by the molecular trap (host). This, we
believe, will allow enhanced sensitivity for detection
of specific volatile components. Additionally, since
the molecular traps are selective in their ability to
form stable neutral guest-host complexes, there
exists the distinct possibility of using them to develop
selective molecular size mass detectors.
EXPERIMENTAL
In this study, the frequency response of the coated
crystals was monitored under various combinations
of experimental conditions in order that the host-
guest interactions might be studied.
In order to attain high sensitivity, the crystal and its
oscillator circuit must be very stable with low drift
throughout the period of the experiment. The
frequency of crystal oscillation is not usually
measured directly but is mixed with a frequency
reference. The experimental apparatus housed
three piezoelectric crystals coated with host
molecules, [6.6.6]cyclophane hexalactam trimer
(Figure 1) and one crystal coated with a gas
chromatographic absorbent. A reference
piezoelectric crystal was used to give a beat
frequency between it and the coated crystals.
According to the manufacturers specifications (4), the
crystal oscillators have a frequency stability of
±0.0025% over an operating temperature range of 0
to 70°C. This allows the sensor to be operated over
a wide range of temperatures to accomodate the
characteristics of the analyte under investigation.
The temperature of the piezoelectric crystals within
the apparatus was controlled to ±0.03°C using a
microprocessor-controlled Peltier device (5). The
resulting analytical system has a frequency stability
of ±1 Hz/hour.
Nitrogen gas is introduced into the PQM apparatus
and a baseline frequency is obtained. Gaseous
guest molecules (CHCIs, CHCI2, etc.) are impinged
upon the coated crystals and the frequency response
is monitored (exposure time). After the appropriate
exposure time, the analyte gas is turned off and the
nitrogen gas is turned on for the desorption of the
solute vapors.
The experimental parameters varied during the
course of these experiments were: the amount of
coating, the concentration of the various guest
molecules, the temperature and exposure and
desorb times.
RESULTS and DISCUSSION
Figure 2 shows a plot of the frequency response of a
crystal sensor versus a reference crystal to a given
gas analyte over time. Sorption of the analyte is
observed during the exposure time. A permanent
increase in the baseline frequency of a piezoelectric
crystal was evidence of the formation of a host-guest
complex. This change in baseline frequency did not
occur for molecules which are known not to form a
complex with the host. The original baseline
frequency of a coated crystal was restored upon mild
heating within the apparatus which indicates the
inclusion process is reversible at elevated
temperatures.
Figure 3 shows a typical plot of the temperature
controlled by the Peltier device during the
experiment. The precise control of the temperature
will allow greater crystal stability and accurate
frequency measurement.
It must be emphasized that the potential is great for
development of truly portable vapor detectors for
volatile substances that have specific responses to
certain molecular structures. Preliminary data lead
us to conclude that the gas phase molecular trapping
reaction works. However, much additional research
must be done before the analytical advantage of
selective molecular traps can be realized. In addition
to studying the use of neutral organic complexing
agents as selective molecular traps, much
instrumental development work must be done on the
microbalances to achieve a truly portable analytical
device.
58
-------
REFERENCES 3.
1. King. W.H.. Anal. Chem.. Vol. 36. 1964. pp.
1735-1739. 4.
2. Sauerbrey, G., Z. Phys.. Vol. 155, 1959, pp.
206-222. 5.
Wohltjen, H., Sensors and Actuators. Vol. 5,
1984, 307-325.
Crystal Clock Oscillators Data Sheet, SaRonix
Nymph Products.
Shields, J.P., Radio Electronics. 1988, pp. 61-
62.
Figure 1. [6.6.6]Cyclophane Hexalactam Trimer
o
c
0)
cr
0)
ul
<
1200
1000-
800 -
600
400 -
200-
-200
20
120
Time (min)
Figure 2. Frequency Response for Lactam-Coated Crystal
59
-------
Q.
E
27.602-
27.598 -
27.594 -
27.590 -
27.586-
27.582 •
Temp (C)
20
—i—
40
60
—i—
80
100
120
FIGURE 3.
Time (min)
Peltier Controlled Temperature Profile
DISCUSSION
MARTIN SPARTZ: You were mentioning that you thought that the mecha-
nism for the sample was there. But did you check the frequencies in the FTIR
to see if they had been shifted, due to any binding that might give us specific
binding to a certain compound?
ED OVERTON: They are shifted very slightly. That's not a perfect overlay.
The molecular weight of chloroform is about 120, and the molecular weight of
lactam is 1,100. You don't really have much chloroform relative to the total
carbon material there.
We were seeing really small changes, and I'm sure we weren't getting 100% of
the molecule complex. So it was pretty hard to tell whether we were getting
much of a shift or not. Basically, what we were hunting for was evidence for a
gas phase reaction, and the spectra, in my opinion, indicated that yes, the bands
were in about the right spot. They were changed very slightly from the Saltier
spectrum.
60
-------
INTRODUCTION TO THE X-RAY
FLUORESCENCE SPECTROMETERS
SESSION
Harold Vincent, Chairperson
X-ray fluorescence is already being used in the field. What we're here to
address is to determine how much it's being used, the good things, the limits and
where we need to improve. What do we need? What does the client need, what
does the experimenter need, and what do the analytical samplers need?
What is the major challenge in application of XRF under field conditions? It
seems to be how we can use tools on the shelf better than we 're using them now.
A lot of these tools have been on the shelf for 10 years without much improve-
ment but we still aren't using them as much as we should. I think you're going
to see an acceleration of their use in the near future in field screening
application.
There are several questions to ask. What are the directions that we should take
in putting together modules that will do the job for us? Are we really using
modules off the shelf to our best advantage? Another question is, when do we
go in the field? You can go in the field as a screening outfit, prior to any kind
of sampling effort, and provide useful information. You can go in at the same
time that samples are being collected and sent back to the laboratory. You can
go in during a remedial effort. You can be part of the control system. You can
go into a hazardous waste site after the sample collection effort.
The information you provide by X-ray fluorescence may be a control, a final
wrap-up, or a traverse. It is usually provided in conjunction with samples that
are taken to a base laboratory. Many variations are possible depending on the
problem at hand.
Now what are some of the other needs? What equipment do we need to have
in hand to do this? I made a list of items that I think should be fulfilled, areas
that we should be looking at to help fulfill these needs.
The EPA Contract Laboratory Program (CLP) has a target list chosen to be
effective in toxic waste measurement - but it doesn't include very much of the
periodic table. It's a rather restricted list from an X-ray standpoint.
There are many elements that could be measured by X-ray fluorescence in the
field that are not measured currently, because they either haven't been a
problem, or they occur in very low amounts. Detection of low concentrations
could be a problem that we'll need to look into.
Many elements that are involved in manufacturing processes - and there are a
number of them not on the current CLP list - could become toxic problems in
the future.
Most of the heavier or high atomic number elements, could be measured by
some of the conventional methods now being used. ICP will measure most of
them. Gases, like radon, aren't likely to be measured by X-ray. We could do it
by X-ray if it were captured. You wouldn't be able to do that by ICP. We have
a unique advantage in measuring some of these by X-ray fluorescence.
X-ray excitation is a critical issue in the field, with portable a system. Safety is
a prime parameter. We can't put a person out on a hazardous waste site, or
possible hazardous waste sites, without that person being protected from us, as
well as from whatever environment he or she encounters.
If we use an instrumental source, we have off/on capability of removing any
excitation safety problems emanating from X-rays. If we use isotopes, we've
got to make sure that we can shield those isotopes. We've got to make sure that
we don't leave radioactive or toxic residue of any kind behind.
The intensity of the X-ray flux is an important parameter in terms of X-ray
detectability and in turn elemental detectability. The energy level is important.
It is necessary to exceed a threshold energy level for excitation for each of the
elements. This can be used as an advantage, and it can be a disadvantage at
times; but for the target element of choice, you have to have enough energy to
do this excitation, you have to have enough intensity to get a number of counts,
so that you can make the measurement.
In regard to the spectral band of the excitation, the choice is between a
monochromatic source or a polychromatic source. Most sources are broad-
band unless some kind of filtration, or secondary tartet is involved.
When we put in secondary targets, we usually have losses of intensity. We've
got to address these problems, because one of the other battles we're fighting
continuously in the field is detectability. We've got to be able to detect low
amounts of some of these inorganic materials.
The stability of the excitation is important. For radio isotopes, you never have
problems with power fluctuations. You may have problems with the short-lived
isotopes that must be corrected for output on the short term. If you use an
instrumental X-ray, a tube source, then you have power requirements and you
may have stability concerns. If the X-ray source is an isotope, you can be energy
selective, and there is a lot of advantage there.
The special characteristics for sources would include the size and shape, and
anything that allows you to get on site with a small package, do the job, and get
off with a small package intact.
Any time you put a large package on a hazardous waste site, you may have to
throw it away. If you contaminate the instrument on site, you may actually lose
it. Don't put $50,000 into a machine you might leave behind.
The efficiency of a detector is involved with detectability. The measurements
for many elements may be at a minimum level of detectability, where there is
uncertainty. We need to be able to measure small amounts of toxic materials.
That could be very, very important.
The geometry of the detector is related to detectability as are resolution and
discrimination. Can I tell lead lines from arsenic lines? The energy range may
limit detector application? If you have a detector that has a range cut off, it will
allow some discrimination.
Durability of a detector is very important at a hazardous waste site. We may
have one of these in the field. You can't easily replace it on site. If it's a crystal,
if it cracks, it may yield improper information. If it's a gas, and if you knock it
against something and it leaks, it's gone. If there is a hot wire in there and it
breaks, it's gone. You may have to leave it behind. You may not be able to afford
two of them on site at the same time. So cost is a very important thing.
If you have an instrumental source, high voltage is a requirement, and you
always have requirements for power to get signal transmission.
Does the system answer the question that was being asked in measuring on site?
Suppose you have a material that's water soluble and moving through the
ground, and you come in and measure the total amount of that material. How
do you discriminate between mobile amounts and the total? How sound is the
sampling schemes to answer the questions? Particle size is very important. It
has much the same effect as coatings do on X-ray emission.
How do we get a sample that represents a heterogeneous material? The more
heterogeneous or coarser it is, the larger sample we need, or the more samples
we have to take to represent a larger sample.
If we put an XRF machine on a robot, send it onto a hazardous waste site, and
the sample has to be mixed up in some fashion, how do we do that?
How do we present it (presentation to the excitation). Is it flat, is it curved, does
it have to be very smooth? A lot of this will depend on the penetration of the X-
rays and the energies of the exiting X-rays. Presentation excitation is part of the
sampling considerations for the site.
We need some knowledge about the sample make-up, which would let us know
something about matrix effects ahead of time. If we know that ahead of time,
we can be much more effective in designing standards and references, as well
as our whole plan of X-ray analysis.
61
-------
Howdowemarkthe location when we aren' t taking a handful of dirt out of the
soil for every sample that we're measuring. We're leaving it there, and
measuring it in situ.
Data have to be collected and presented, either to somebody for data handling
or to a client for whatever use he wants. How do we do that? The advantages
that we have in the last ten years in this field are advances in the data handling
area. The computers have gotten smaller, the memory is cheaper, and the trend
is continuing to move in those directions. I think we'll see more advantages and
more advances.
We need to make data management as automatic as possible. A person is
responsible for collecting data, for making sure that it gets the right tag, that it
gets transmitted, or recorded, and that he can verify that data are good.
How do you get hard copy out of a little black box that you carry out in the field?
You may not want to take your gloves off to write with pencil and paper. Maybe
you can punch buttons here. How clever are we going to be in getting that hard
copy?
How do we know when to reject data in the field and reject bad data. This is like
crossing off a page in your laboratory notebook. How do we know when to cross
off that page? One advantage is that taking these measurements is rapid enough
to allow for repetitive tests.
And a word about quality assurance. It includes a lot of buzz words - including
initial and continuing calibration. These are common buzz words with the CLP
program. They are no less important with field screening.
Suppose you are in the field, and somebody has just dug an acre of dirt down
to about five feet deep, and carted it off in trucks. Somebody comes back and
says that you carted off a lot of pretty good top soil. You say it was hazardous,
and he says, can you prove it to me? How do you know? How well do you know
that? You've got to have some pretty good quality assurance information to
prove you acted correctly.
We've got to have proper calibration, we've got to have it referenced to
standard materials if possible. We've got to keep good records. We've got to
verify that the samples were where they were.
We've got to be very sure of what we put down on paper, that it's accurate, that
it solves the problem, and that we can back it up.
These are just some of the parameters we can ask our speakers about in this
session.
62
-------
APPLICATION OF FIELD-PORTABLE XRF
TO HAZARDOUS WASTE CHARACTERIZATION
Richard K. Glanzman
CH2M HILL
6060 South Willow Drive
Greenwood Village, Colorado 80111-5112
ABSTRACT
Refinement and miniaturization in instrumentation
have allowed the development of field-portable,
x-ray fluorescence (XRF) instrumentation. This
XRF capability fills an important gap during field
sampling—the collection of representative samples
to define the extent of inorganic contamination.
Using a portable XRF, analyses can be performed
in the field, allowing for immediate screening of
many samples. Two instruments have been field
tested. Both have advantages and disadvantages.
Ideally, both are capable of detection limits in
the tens-of-parts-per-million range for most
metals, depending on the media and its physio-
chemical characteristics. In the field, both
instruments have been shown to be capable of non-
destructively and quantitatively determining con-
centrations of most metals in soils, sediments,
and rocks below the commonly applied action levels
for cleanups.
INTRODUCTION
Use of field-portable x-ray fluorescence (XRF)
instrumentation to determine the extent of metal
contamination reduces the cost while increasing
the probability of obtaining representative
samples from hazardous waste sites. Refinement
and miniaturation in electronics have led to the
development of several instruments that weigh
less than 25 pounds, can be placed into a back-
pack, and taken to the sample site. These instru-
ments allow field personnel to qualitatively to
quantitatively determine the concentration of a
suite of metals in many different types of media
at the site and to differentiate between back-
ground and contaminated media. Media may include
soils, stream sediments, tailing, slags, water,
vegetation, paint, landfill material, structural
steel, cement, etc. This allows the field person-
nel to obtain statistically representative samples
to define the nature and extent of metal
contamination.
Sampling design takes place in the field, based
on observed concentration rather than on an
assumed contaminant source and dispersive char-
acteristics. The spatial distribution of target
metals is defined in the field, allowing a more
accurate selection of media and metals to be
analyzed.
Measurement times are relatively fast (usually
one to two minutes), and analyses do not alter
the media being measured (nondestructive). There-
fore, integration of XRF instrumentation into a
field effort can reduce the number of samples to
as little as 10 percent of those required by
conventional field-sampling techniques in the
first (and sometimes only) sampling effort.
Field-portable XRF instruments have been utilized
in the minerals industry for approximately
10 years. One such instrument was developed and
used on the Mars Lander. Instrument size and
detection limits have decreased as the instruments
became more sophisticated and simpler to operate.
XRF instruments can be utilized in both Level II
and Level III analytical work as defined by
Furst et al. (1), depending on calibration and
sample preparation. Preliminary evaluation or
screening (Level III) involves minimal calibra-
tion and sample preparation. Remedial Investi-
gation (RI) and Feasibility Studies (FS)
(Level II) require calibration and may require
some sample preparation. The precision, accuracy,
and detection limits required for litigation and
enforcement support (Level I) exceed the analy-
tical capability of the field-portable XRF instru-
ments. The most cost-effective use of these
instruments is at Levels II and III to screen
samples for laboratory analysis. However, the
instruments can be effectively employed in defin-
ing metals concentrations for Emergency Response
and removal actions as well.
Documented use of field-portable XRF instruments
began in 1985(2) (3) . The Smuggler Mountain Site
near Aspen, Colorado, was the site of the first
published use of one of these instruments to
determine the boundaries of criteria levels of
1,000 milligrams per kilogram (mg/kg) lead and
10 mg/kg cadmium in soils and tailing (3). The
same site was used to evaluate a prototype
field-portable XRF instrument (4). A new
calibration technique (5) and published use on
nonmining-related media (lubrication oils) (6)
were reported in 1988. The use and range of
application have expanded to use in determining
the presence and amount of lead-based
63
-------
paint in homes and the areas of phytotoxic soils.
CH2M HILL is currently using the instrument at
four sites in the western U.S. in Level II and
III capacities.
PRINCIPLE OF THE XRF INSTRUMENT
Field-portable instruments are produced currently
by at least five companies. The same basic prin-
ciples are used in these designs (Figure 1}. The
instruments utilize the energy dispersive x-ray
(EDX) technique developed in the 1970s to produce
a light-weight field instrument. The technique
involves the use of low-level radioactive isotopes
to excite the elements in samples. The four com-
monly used isotopes are iron-55, cadmium-109,
curium-244, and americium-241. Selection of the
isotopes depends on the metals (elements) of
interest. For example, iron-55 is best suited in
determining the lighter elements (silicon through
vanadium) in the periodic table. Source activity
usually ranges from one to 100 millicuries.
The source excites the atoms of elements in the
samples, which give off an element-specific wave-
length of energy (fluoresces) that impinges on
the gas proportional tube (GPT). The GPT converts
energy into a signal that is amplified and proces-
sed through a microprocessor. The sample emits
the spectra that are recorded as counts at energy
levels specific for the elements making up the
sample. The number of counts at a specific energy
level is proportional to the amount of that
element exposed to the radiation source. The
technique dates from the early 1900s, but the
development of a field-portable instrument
required current technology.
The number of counts in a specific energy level
requires correction for background, interference,
etc. and is reported as an index number (or inten-
sity) . The index number is then compared with
calibration standards for each element of interest
to develop a calibration curve relating index
numbers to concentrations for each element.
VARIABLES IN THE ANALYSIS
The goodness of fit (linear regression line) of
the index number/concentration relationship for
each element is a function of radioactive source
strength, particle size, sample matrix, sample
surface characteristics, and other elements and
their abundance in the sample. The calibration
standard should match these sample characteristics
as closely as possible to produce a high level of
precision and accuracy. The more uniform and homo-
geneous these characteristics are in both the stan-
dards and the samples, the better the calibration
and the results will be.
Source strength is a function of both initial manu-
factured activity and the isotope's half life.
Year-to-year measurements of a specific element in
a specific sample would not change appreciably
(other factors being held constant) if the source
were an americium-241 isotope with a, half life of
433 years. However, a cadmium-109 isotope source
with a half life of 1.3 years has only half its
initial source energy after 1.3 years. Cadmium
isotope sources usually need replacement at least
every other year, but use also depends on initial
source level and concentration of specific elements
requiring analysis.
The sample particle size should be as homogeneous
as possible, with smaller particle sizes giving
better correlation between index numbers and con-
centrations. The 200- to 300-mesh (clay size)
particles give the best analytical results, with
correlation decreasing as particle size and varia-
bility of particle sizes increase. Qualitative-
to-semiquantitative correlations can be achieved
with 80-mesh (sand size) particles if the element
is a major element and/or is dispersed homogene-
ously throughout the sample.
The matrix of the sample can be expressed as the
average atomic number and range of atomic numbers
making up the sample. Matrix is the dominant
factor determining the depth of penetration and
the response from the sample. Depth of penetra-
tion can range from several centimeters to just a
few microns, depending on the matrix (Figure 2).
Water and hydrocarbons have the highest depths of
penetration and some of the best calibration
curves because of their low average atomic number
when a very narrow range of atomic numbered ele-
ments makes up the major proportion of the sample.
In dense minerals, such as pyrite (iron sulfide)
and galena (lead sulfide), the depth of penetra-
tion is in the tens-of-microns range because of
the minerals' high average atomic numbers, but
the precision and accuracy may still be excellent
if all the other factors are held constant. Sur-
ficial chemistry is particularly important on
these heavier minerals. The field XRF analyses
can be looked upon as complementing laboratory
work because the surficial layers analyzed by
the field instrument are most likely to be the
short-term dissolved portion from any solid
matrix.
Field media can be visualized in a fashion compar-
able to an average atomic number. Plants, contain-
ing mostly water and carbon compounds, with low
atomic numbers have the highest proportional depth
of penetration. The penetration decreases with den-
ser materials. The lack of sample preparation
necessary to analyze plants and fine-grained soils
means that the XRF analysis may be a more accurate
analyzing technique. Sample preparation is com-
monly difficult and can drive off parts of element
concentration from plants. Dissolution techniques
for preparing soils, rocks, and other solids for
analyses may not put all of particular elements in
solution. XRF is a total element analysis pro-
cedure that does not physically alter the sample
and does not depend on the compound chemistry con-
taining the element.
The smoother or more regular the sample surface,
the better the correlation between index numbers
and concentration. A painted surface is almost
ideal. Water and soils are commonly placed in a
cup with a thin polypropylene or mylar window
stretched across the base of the water or soil
64
-------
column. This window, too, gives very high corre-
lations. Rocks with very irregular surfaces will
give poor correlations.
Elements associated in a sample by overlap can
cause either enhancement, adsorption, or inter-
ference. Adjacent elements in nearly equal, low
concentrations can be difficult to analyze because
of overlap between their peaks. For example,
nearly equal, low concentrations of copper and
zinc would be difficult to separate and would
cause a high detection limit. Similarly, nearly
equal, low concentrations of arsenic and lead are
difficult to determine because the major peaks for
both elements have approximately the same energy
level; their concentrations must be determined by
using the much smaller, secondary peaks.
EXAMPLE CALIBRATION CURVES
Figures 3, 4, and 5 show calibration curves for
iron, lead, and cadmium, respectively, that were
developed using field-portable XRF instruments.
Iron is commonly present in percent-level concen-
trations, particularly in mining districts. The
iron calibration curve (Figure 3) represents iron
in soils and stream sediments in one such dis-
trict. Another curve was developed for pyrite-
rich tailing that resulted from milling to recover
galena. The lead calibration curve (Figure 4) and
the cadmium calibration curve (Figure 5) also
represent soils and stream sediments. In addi-
tion, the cadmium calibration curve indicates the
low detection limits (on the order of tens of
mg/kg) that can be achieved using the field instru-
ments without sample preparation.
APPLICATIONS AND RESULTS
The potential application of the field-portable
XRF instruments can be almost universal in that
virtually any media that contains elements heavier
in atomic weight than silicon can be qualitatively
to quantitatively analyzed. In a qualitative sense
alone, the instruments have been proven to be able
to:
o Determine background areas from contaminated
areas
o Differentiate uncontaminated soils from soils
containing phytotoxic concentration of metals (and
to identify the metals)
o Distinguish homes painted with lead-based
paints
o Distinguish and rank water samples containing
as little as 5 milligrams per liter zinc and
copper
o Determine volume of contaminated material
for removal
Figure 6 illustrates one example of using the
field-portable instrument in a qualitative sense.
Sampling drill cuttings for metals contamination
is commonly done by visual judgement or by
compositing equal lengths of core or cuttings. In
this case, cuttings from a 40-foot monitoring well
installation were qualitatively analyzed (scanned)
using the XRF instrument. Although the iron indi-
cates two well-defined peaks that were quite evi-
dent visually, lead, zinc, and copper did not
mimic the iron concentration. Only the upper
5-foot sample contained appreciably higher metal
concentrations. Copper and zinc were similar in
their distribution. Lead, although similar in the
upper part of the section, increased at the base,
where iron is lowest.
This information can be generated in about 15 min-
utes, allowing the sampler to focus on sampling
the elements and element concentrations where they
provide the most information. The samplers know
the metal distribution and relative analytical
results when they leave the field. Interpreta-
tions are made immediately in the field instead of
trying to make them from laboratory results, weeks
to months later, using field notes and memories as
a guide.
Qualitative results are developed quickly and
easily, but can be developed into semiquantitative-
to-quantitative results, increasing the analytical
description of the field setting with fewer labora-
tory samples. Samples are qualitatively analyzed
in the field to document the representative sample
sets, using statistical distribution calculations.
These calculations can be as simple as definition
of the "normal" distribution of metal concentra-
tion or as complex as geostatistical kriging—all
performed using index numbers. When a representa-
tive suite of samples has been analyzed, calibra-
tion curves can be developed and the index number
data can be assessed. Semiquantitative analytical
results are common. Quantitative results can be
achieved using relatively homogeneous media.
Figure 7 presents an example of using the field-
portable XRF instrument for quantitative determi-
nation of metals concentrations in mine wastes
occurring in residential areas. The lead data
were developed from mine wastes in a large mine-
waste pile adjacent to four homes. The data indi-
cate a very wide range in lead concentration in
the fine-grained mine waste (1,210 to
210,000 mg/kg). No galena was visible, even using
hand lens, so the percent-level concentrations may
be secondary lead sulfates, carbonates, and/or
oxides. Lead calibration curves, similar to those
in Figure 4, were developed from the laboratory
data that resulted from samples collected from
this and other mine-waste piles. Similar plots
were developed for copper, zinc, and other metals.
Figure 8 illustrates a second example of the quan-
titative determination of metal concentration:
analyses performed on a large slag pile. Slag is
a dark-colored, dense, glassy residue of the smelt-
ing process. Smelters in large mining districts
that continue to operate for a long time change with
improvements in technology, commodity prices, and
ores processed by the smelter. These changes are
reflected in the metals remaining in the slag, which
are slowly leached from the slags by the
65
-------
weathering process. The problem is similar to
that of the mine waste—collecting a documented,
representative sample from these homogeneous-
appearing piles. Collecting slag samples is diffi-
cult because they have to be chiseled from the
large glass mass. The XRF indicated that the
metals concentrations in this particular pile
were relatively homogeneous; however, the range of
lead concentrations almost covered an order of
magnitude (13,600 to 95,000 mg/kg). Also, the
north and east sides of the slag pile had lead
concentrations below 20,000 mg/kg, while most of
the mass contained between 20,000 and 30,000 mg/kg
lead. A grab sample from the western edge of the
slag pile would have exaggerated the lead concen-
tration by a factor of 4 to 5 (25,000 versus
95,000 mg/kg). On the other hand, grab samples
collected on the eastern or northern edges would
have underestimated the lead concentration by a
factor of about 2. The index numbers allowed the
samplers to collect fewer, but more representative
samples, and to obtain a better distribution of
sample data for interpretation. Other metals con-
centrations were simultaneously determined with
lead and plotted in the same manner.
CONCLUSIONS
The field-portable XRF is an under-utilized analy-
tical screening tool that allows determination of
the nature and extent of contamination in the
field, collection of representative samples, and
documentation that representative samples were
collected. The instrument simultaneously and non-
destructively gives a total element gualitative-
to-quantitative concentration of a suite of ele-
ments in the field in a matter of a few minutes.
The instruments have been used at inorganic conta-
mination sites involving metals, but applicability
can extend into halogenated organic sites as well.
ACKNOWLEDGEMENT
The author wishes to acknowledge the contributions
received from Alan Seelos with Aurora Tech Instru-
ments and John R. Rhodes with Columbia Scientific
Industries Corporation.
REFERENCES
(1) Furst, G. A., Tillinghast, V., and Spittler,
T., "Screening for Metals at Hazardous Waste
Sites: A Rapid Cost-Effective Technique
Using X-Ray Fluorescence," Proc. National
Conference on Management of Uncontrolled
Hazardous Waste Sites. Washington, D.C.,
1985, pp. 93-96.
(2) Mernitz, S., Olsen, R., and Staible, T., "Use
of Portable X-Ray Analyzer and Geostatistical
Methods to Detect and Evaluate Hazardous
Materials in Mine/Mill Tailings," Proc.
National Conference on Management of
Uncontrolled Hazardous Waste Sites.
Washington, D.C., 1985, pp. 107-111.
(3) Chappell, R. W., Davis, A. 0., and Olsen,
R. L., "Portable X-Ray Fluorescence as a
Screening Tool for Analysis of Heavy Metals
in Soils and Mine Wastes," Proc. National
Conference on Management of Uncontrolled Haz-
ardous Waste Sites. Washington, D.C., 1986,
pp. 115-119.
(4) Raab, G. A., Cardenas, D., Simon, S. J., and
Eccles, L. A., "Evaluation of a Prototype
Field-Portable X-Ray Fluorescence System for
Hazardous Waste Screening," EMSL, EPA 600/4-
87/021. U.S. Environmental Protection Agency,
Washington, D.C., 1987.
(5) Piorek, S., and Rhodes, J. R., "A New
Calibration Technique for X-Ray Analyzers Used
in Hazardous Waste Screening," Preprint of
paper presented at the Fifth National RCRA/
Superfund Conference and Exhibition on
Hazardous Wastes and Hazardous Materials,
1988.
(6) Johnson, G., Kalnicky, D. Wallendorf, B., and
Lass, B., "Analysis of ppm Levels of Additives
in Lubrication Oils Dsing a Portable XRF
Analyzer," American Laboratory, August 1988,
pp. 58-61.
66
-------
SOURCE
AMPLIFIER
MICROPROCESSOR
OUTPUT
V)
z
3
O
o
Fe
Mn
GPT
X-RAY ENERGY
FIGURE 1. SCHEMATIC INDICATING THE FIELD-PORTABLE
XRF ANALYTICAL PROCESS.
Z
O
UJ
Z
UJ
a
UJ
O
UJ
LL
u.
UJ
PLANTS
SOILS
cm -i
mm -
ROCKS
M
SLAGS 100 -I
MINERALS j /*•
\
10
/*
•• HYDROCARBON
• WATER
QUARTZ SAND
I
100
1 10
AVERAGE MATRIX ATOMIC NUMBER
FIGURE 2. APPROXIMATE RANGE OF EFFECTIVE
PENETRATION AND RESPONSE FROM SAMPLE
MEDIA.
IX
UJ
CD 5
Z 5
O O
CC Z
~ X
UJ
a
z
r = 0.95
1.0
2.0
IRON (PERCENT)
FIGURE 3. IRON CALIBRATION CURVE BETWEEN INDEX
NUMBER FROM A FIELD-PORTABLE XRF
INSTRUMENT AND LABORATORY RESULTS ON
THE SAME SAMPLE.
67
-------
2.0
0
<
UJ
QC
UJ
ffl
5
5 1.0
r = 0.92
0 10,000 20,000 30,000
LEAD (mg/kg)
FIGURE 4. LEAD CALIBRATION CURVE BETWEEN INDEX NUMBER FROM A FIELD-PORTABLE XRF
INSTRUMENT AND LABORATORY RESULTS ON THE SAME SAMPLE.
10 ,-
cc
ui
50
r = 0.99
50 100 150
CADMIUM (mg/kg)
FIGURE 5. CADMIUM CALIBRATION CURVE BETWEEN INDEX NUMBER FROM A FIELD-PORTABLE XRF
INSTRUMENT AND LABORATORY RESULTS ON THE SAME SAMPLE.
0
10
in
UJ
t 20
I
Q.
UJ
Q 30
40
Cu Zn Pb
I
T
ZrijCu Ptj Fe
I
I j I
0.5 1.0 1.5
METAL CONCENTRATION (INDEX NUMBER)
2.0
FIGURE 6. COMPARISON BETWEEN IRON AND THE METALS COPPER-LEAD-ZINC
IN DRILL CUTTINGS USING A FIELD PORTABLE XRF.
68
-------
7000
3700
XXX
LEAD CONCENTRATION (mg/kg)
FIGURE 7. LEAD CONCENTRATION IN A MINE WASTE PILE
DEVELOPED FROM ACOMBINATION OFXRF AND
LABORATORY DATA.
26000 2300Q. 18400.
•25000 .43000 13600
.24000
N
xxx
LEAD CONCENTRATION (mg/kg)
FIGURE 8. LEAD CONCENTRATION IN A SLAG PILE
DEVELOPED FROM A COMBINATION OF
XRF AND LABORATORY DATA.
69
-------
DISCUSSION
ALAN CROCKETT: What kind of sample preparation was used.
RICHARD GLANZMAN: The only sample preparation used on the mine
waste, was taking the big "goopers" out by screening down to an 80 mesh.
We found that if we can get an 80 mesh, for screening purposes, that's perfectly
adequate. The rest of the data were developed on soils and stream sediments
that were 80 mesh-type stuff, or less. There isn't much improvement below 80
mesh.
HAROLD VINCENT: You have three geologists sampling there. Did they
actually take those samples, were they grab samples, or pulverized samples?
RICHARD GLANZMAN: They represented both soil and rock samples,
because we were interested in some information on what was being wasted off
the outcrop, as well.
HAROLD VINCENT: On the lead mine waste, did you do any kind of
distinction or discrimination against soluble vs. insoluble lead.
RICHARD GLANZMAN: That's a good question. The XRF is a totalizing
instrument, so that the sample fluoresces from the element itself, and the form
of the element that is present within the substrate that you're using is not terribly
important, except from a matrix standpoint.
We use X-ray diffraction to determine the mineralogy, and on many of these,
we have done that, XRF being the totalizing type of instrument, it didn't
discriminate between the two, other than the calibration curves, when we
wanted to clean those up.
HAROLD VINCENT: You didn't run any solubility tests separately?
RICHARD GLANZMAN: Yes, we did that to see whether there was some
soluble lead. Although we generally consider lead to be fairly insoluble, we had
some samples from one waste pile that on bottle-roll leach tests with distilled
water were showing 44 ppb.
So it can be very soluble, and that's the reason we also did some X-ray
diffraction. We found that a complete suite of lead oxidation products were
there. We had galena, oxides, carbonates, and sulfates. They were all present,
and we did not think that it was going to be that high when we initially did the
test.
HAROLD VINCENT: Could you say something about the calibration and the
standards you used to back up your information?
RICHARD GLANZMAN: When we're out on the site, we do a lot of
scanning. When we use the instrument in a scan mode, we use it for, say 10-,
20-, 30-second counts, just to get an idea of the limits and what the dispersion
pattern looks like. Then we will go to the one-minute counts, to get serious
about what we're doing and what samples we're picking up.
We will collect the sample and use the XRF on the sample that we send to the
lab. When we get our laboratory results back, we do an index comparison with
a least-squares fit against their concentrations. In this manner, we do use
standards out in the field.
In the morning we run a suite of samples to make sure the instrument is
functioning. In the evening, we run that same suite of samples. The suite of
samples has been analyzed so that we know the statistics and the variability of
the concentration that we're after. So in that way, we're running an instrument
that we know is functioning. But we don't know that the matrix is exactly the
same. The instrumentation has some limitations, as little as I would like to
acknowledge that.
70
-------
THE USE OF TRANSPORTABLE X-RAY FLUORESCENCE
SPECTROMETER FOR ON-SITE ANALYSIS OF MERCURY IN
SOILS
David.I. Grupp. David A. Everill. Raymond J. Bath, Ph.D.
NUS Corporation
Richard Spear. PhD.
U.S. Environmental Protection Agency, Region 2
X-ray fluorescence is gaining increased acceptance as an analytical tech-
nique for the determination of inorganic compounds in environmental samples.
X-ray fluorescence provides a fairly rapid nondestructive analysis with mini-
mum sample preparation, allowing results to be rapidly generated and used on-
site to make project related decisions. A transportable x-ray fluorescence
spectrometer was used to analyze over 500 soil samples for mercury to
determine the lateral and vertical extent of mercury contamination at a hazard-
ous waste site.
Preparation and analysis of each sample was performed on-site and the
results generated were immediately used by the site project manager to make
critical decisions on the number of samples to be collected, the depth of
subsurface contamination and the lateral extent of contamination. A percentage
of the samples collected and analyzed were also analyzed by a fixed base
laboratory to verify the results generated by the XRF technique.
The paper presents mercury results generated using the transportable x-ray
fluorescence spectrometer, and presents a comparison of the XRF results and
the fixed base laboratory results. The decision making process utilized at the site
is also presented along with an examination of how the transportable XRF spec-
trometer was used to make critical on-site sampling decisions.
DISCUSSION
AVRAHAM TEITZ: What standards did you send? Were those the ones that
you mixed yourself, or were those the ones that you had gotten from outside,
that you had sent to the various laboratories?
RAYMOND BATH: The only standards we sent to the laboratory were those
made up ourselves. We did not send any standards to the lab, because at that time
(a little over a year ago) there was only one standard available from the
Environmental Systems Laboratory here in Las Vegas. In the low ppm range.
the standards are just not available.
DOLPH CARDENAS: When you were drying your sample, didn't you drive
your mercury off by volatilizing it?
RAYMOND BATH: That was a real concern, but it does not since we air dry
it. This was the first way proposed by Kevex, if you read their manuals. We
produced our own manuals for Region II.
Kevex used an oven at 70°, so we compared it with an oven at 70°, with letting
it stand overnight and drying, and with a microwave pulse. You can pulse dry
it. You don't do it all at one time at full power but use a 50% or 80% power level.
We have not found any mercury loss.
DOLPH CARDENAS: I'm assuming, then, that the CLP Laboratory used
standard CLP techniques.
RAYMOND BATH: They used atomic absorption and analyzed the sample.
It was a specialized test just for mercury so I don't believe they used a standard
digestion procedure.
DOLPH CARDENAS: I believe it's just a leaching technique, and it's for a
total mercury. Perhaps if they had used a total digestion, they would have gotten
closer to your numbers. We have seen roughly a two-to-one relationship, as you
demonstrated.
RAYMOND BATH: There is a real problem with this. The site has been
investigated for a long time, and one of the questions that was originally
brought to me was how to distinguish organic mercury from inorganic mercury.
So we tried about tW'O and a half years ago at this site, before we had the XRF.
taking soil samples and trying to get organic analyses done - a total digestion
and a leaching digestion, and looking at the difference between the two. to see
if it w as organic mercury.
The results were very unsatisfactory. We couldn't tell anything. I have not
tested the microwave digestion, and I don't know what that would do w ith the
mercurv.
BOB NOLKOFF: What kind of data reduction program did you use on the
Kevex? Was it fundamental parameters, or did you use a least-squares fit?
RAYMOND BATH: It's a least-squares fit. Again, it's a Gaussian technique.
They have a full computer, we didn't have the time to go into the computer
program for it. There are a lot of variables that we played \\ ith for quite some
time to get our software program to work the way we wanted it to.
One of the problems was that Kevex updated their software level halfway
through the program, which was a major blow. So we elected to stay w ith the
old software, until the project was finished.
BOB NOLKOFF: Do you know if this is all in automatic files, or was it done
by hand?
RAYMOND BATH: the process is that it collects data. It's a 500-second run.
At 250 seconds, it has taken the spectra, and stored it. It's processed later on.
TONY HARDING: Were you linearly correlating mercury intensity to con-
centration i.e., measured mercury intensity to concentration on the calibration
curve?
RAYMOND BATH: No, you figure out with the secondary target using sort
of a ratio effect. Because the secondary target has a constant intensity, it's sort
of an internal standard.
TONY HARDING: That's only true if you ratio your mercury intensity to the
intensity of the back-scattered zirconium. Is that what you did?
RAYMOND BATH: We tried each one of those techniques to get dow n to this
level. We felt comfortable with this.
TONY' HARDING: In that way. by ratiomg the mercury intensity to the
zirconium measured intensity, would you be able to account for things like
packing density?
RAYMOND BATH: We didn't see problems w ith the packing density in these
samples, considering the way we prepared them. We thought we might, and a
considerable amount of effort went into soil preparation, to differentiate one
soil type from another soil type. In packing density, we did not see that effect.
TONY HARDING: Some of the data show that your standards correlate very
well between, specifically, the ESD data and the XRF data, whereas some of
the actual solids don't correlate as well. Do you think that could be a matrix
effect?
RAYMOND BATH: No. I think that relates to how \\ e prepare samples for the
ESD laboratory'. It's a little bit different. They get a small portion right out of
the homogenized jar. In the other one. we had to physically prepare that. So that
sample was a little better distributed.
TOM SPITTLER: Just a couple of comments on the standardization. You
were probably using the zirconium Compton and scatter peak for your
normalization, were you not?
RAYMOND BATH: Yes.
TOM SPITTLER: You don't have to worn then about a packing effect. The
only significant problem you might have seen in sending samples to other labs
is that in any soil sample, particle size has an effect in terms of the concentra-
tion of the element being maybe more predominant in the very fine particles.
as opposed to the coarse particles.
That's particularly true when you're looking at lead in soils. We have seen that
in thousands of soil analyses.
So if you send off a sample that was homogenized well, but gets shaken down
by the time it gets to the laboratory, and they scoop off the top part of the sample.
you can have a higher concentration in the low er portion of the sample, w hich
is the fine material.
The problem with having a laboratory properly homogenize a sample before
they take their test sample out and analyze it is something v. e al w a\ s cope w ith.
The data look very reasonable, on the whole, for field analysis versus lab
analysis, and the more care taken w ith the sample preparation, the better those
numbers are going to line up.
RAYMOND BATH: People underestimate the importance of sample prepa-
ration. When you're trying to do comparisons from one to the other, it's worse
than apples and oranges. It's more like apples and elephants.
-------
THE DETERMINATION OF MINIMUM DETECTION LIMITS FOR
INORGANIC CONSTITUENTS IN SOIL USING TRANSPORTABLE
SECONDARY TARGET X-RAY FLUORESCENCE 1. ARSENIC IN
THE PRESENCE OF LEAD
David A. Everill, David Grupp, Raymond J. Bath, Ph.D.
NUS Corporation
Richard Spear, Ph.D.
U.S. Environmental Protection Agency, Region II.
Secondary target energy dispersive x-ray fluorescence (EDXRF) is a rapid
nondestructive analytical tool that has great potential for on-site screening of
inorganic soil constituents at hazardous waste sites. In soil, lead and arsenic are
naturally found in varying concentrations ranging from low ppm to 900 ppm for
lead and 200 ppm for arsenic. When analysis is performed for arsenic contami-
nation above background levels, the minimum detection levels (MDLs) for
arsenic vary proportionally to the lead concentration in the soil. This change in
MDLs is caused by spectral interference and specific absorption-enhancement
effects between lead and arsenic. The use of MDLs and a review of spectra are
used to evaluate the extent of the arsenic-lead interactions in a soil matrix. This
presentation will identify the actual limitations of EDXRF for the determination
of arsenic contamination at hazardous waste sites and demonstrate solutions to
the arsenic-lead spectra] and matrix interactions.
Spectral interference between arsenic and lead is described as the superim-
position of first order lines of a different series. This problem can be overcome
by using an alternate emission line or various computer-generated deconvolu-
lion, and pure element stripping methods to separate the regions of overlap and
extract intensities. These methods are used after the spectrum has been acquired.
Specific absorption-enhancement occurs during spectral acquisition of the
sample. The lead L-alpha emission energy interacts with the arsenic spectral
lines causing loss in peak intensities for arsenic. This false-negative effect is
concentration-dependent. As the lead concentration increases the intensity of
the arsenic decreases. This phenomenon occurs when spectral lines of a matrix
element and the absorption edge of the analyte are in close proximity.
Graphs are presented showing MDLs for arsenic versus lead concentrations
and a comparison of peak separation techniques. Spectra showing matrix
enhancement and regions of overlap will be used to identify the problems. A
series of calibration matrix-correction curves are also presented. The data will
be used to demonstrate arsenic-lead interactions and determine the limitations
of EDXRF to arsenic screening of soils.
DISCUSSION
BOB MOLKOPF: Just a quick clarification. It can't be that lead MJines are
around 2.4 or 2.5 keV. It would have to be a lead L, or some other line. It can't
be anM-
RAYMOND BATH: It's a minor L.
HAROLD VINCENT: Did you do this all with a zirconium secondary target,
or did you try it with other targets?
RAYMOND BATH: We have not tried any other targets at this time.
HAROLD VINCENT: Do you feel that there would be enhancement with any
other targets?
RAYMOND BATH: I don't think you could get the low ppm range.
73
-------
THE APPLICATION OF X-RAY FLUORESCENCE TECHNOLOGY IN THE CREATION
OF SITE COMPARISON SAMPLES AND IN THE DESIGN OF
HAZARDOUS WASTE TREATABILITY STUDIES
John J. Barich, III
Environmental Engineer
USEPA, Seattle, Washington
Gregory A. Raab
Lockheed Engineering Management Services Company
Las Vegas, Nevada
ABSTRACT
Site Comparison Samples (SCS) and treatability studies
are contemporary tools used in the investigation and
remediation of hazardous waste sites. Each depends on
the development of large volume samples which are
characteristic of the most difficult conditions at a site
to treat. The use of X-ray fluorescence spectrometers
(XRF) to identify sample locations at a major Superfund
site is described. The subsequent processing of samples
into SCS materials and treatment samples is presented.
INTRODUCTION
As byproducts of a growing technological society
continue to find their way into the environment, the
Environmental Protection Agency (EPA) must face an
ever-expanding problem of how to handle and measure
the harmful byproducts. Before contaminants can be
removed or neutralized, they must be characterized for
type and quantity. Field-Portable X-ray Fluorescence
(FPXRF) instrumentation has been shown to be useful
as a screening tool for heavy metals in soils at
hazardous waste sites (1,2). Instruments are smaller
than their laboratory counterparts, transportable by a
single individual, hermetically sealed, and provide
immediate data from analyses completed with little or
no sample preparation. Analyses are either conducted
in a field laboratory or in situ.
The Bunker Hill Superfund Site is located in the Coeur
d'Alene mining district of northern Idaho. The site is 7
miles by 3 miles. Primary site contaminants are lead
and zinc associated with the mining, beneficiation,
smelting and refining of lead-zinc-silver ores. Lead
smelting commenced in 1917 and zinc refining
operations began in 1927. Operations ceased in 1981.
Over the period of operation of these facilities, metals
were emitted to the atmosphere from both point and
fugitive sources. Tailings from the beneficiation
operations were discharged to the Coeur d'Alene River
prior to the construction and use of tailings
impoundments. These emissions and discharges resulted
in widespread contamination of area with metals (3).
The management of large, complex Superfund sites
requires years of effort by many parties, and is
composed of a series of individual projects and
concurrent tasks. Each task requires development of
its own quality assurance plan. Quality control within
and between projects relating to the same site is an
Roy R. Jones
Quality Assurance Management Office
USEPA, Seattle, Washington
James R. Pasmore
Columbia Scientific Industries Corporation
Austin, Texas
important element of an overall quality assurance
program. Due to the size of the site (21 square miles),
the number of parties involved, and the length of time
until remediation is complete, the use of Site
Comparison Samples (SCS) as tools for applied quality
control allow quality assurance of data between
projects on the same site.
As a result, two requirements presented themselves
simultaneously:
(1) The need to develop large, homogenous
volumes of heavily contaminated soils for
treatability studies , and
(2) The need to develop large homogenous
samples of soils which should be processed as Site
Comparison Samples ("SCS project").
Field screening using FPXRF technology was selected
as the analytical tool to ensure that appropriate soils
were developed for both of these purposes.
FIELD ACTIVITIES
Over 500 kilograms of soil was required for the site
studies and the SCS project. The soils needed to be
heavily contaminated and as dry as possible.
Authorization to proceed was received in October
1987. Then current weather conditions in northern
Idaho were unusually dry for that time of year; hence,
any field effort had to be mobilized quickly or
postponed until the following summer. Postponement
was not acceptable. The high cost of the treatability
studies and the critical nature of the SCS project to the
long term quality control program at the site demanded
that soils of known concentrations with known data
quality be obtained; sample collection without
concurrent analysis was not acceptable. Field
activities needed to be supported, therefore, with
instrumentation that could be mobilized quickly, be
portable enough to be moved throughout a large site
and be capable of providing analytical responses to field
personnel on a "real-time" basis.
Equipment
The FPXRF used at Bunker Hill is the X-Met 840
manufactured by Columbia Scientific Industries
Corporation. A technical description highlighting its
applicability for use at hazardous waste sites is
provided by Piorek and Rhodes (4). The X-Met 840 is a
75
-------
self-contained, battery powered, microprocessor-based,
multichannel X-ray fluorescence analyzer weighing 8.5
kg. The surface analysis probe is specially designed for
field use. The X-Met 840 is hermetically sealed and
can be decontaminated with soap and water. The probe
includes a radioisotope source of Curium-244, a
proportional counter and the associated electronics.
The source is protected by an NRC-approved safety
shutter.
The electronic unit has eight calibration memories
called "models". Each model can be independently
calibrated for as many as six elements each. These can
be used to measure elements from aluminum up to
uranium assuming two probes with the associated
isotope sources are available. The unknown sample
intensities are regressed against the calibration curves
to yield concentrations. For the Bunker Hill site only
lead and zinc were investigated and only two models
were calibrated. Model 1 was calibrated from
background up to 4980 mg/kg Pb and 9791 mg/kg Zn.
Reference Soil Standards for Quality Control and
Standardization
The commercially available FPXRF systems use
standards to establish calibration curves for
comparison. Heretofore there has not been a demand
for FPXRF systems in hazardous waste screening.
Because of this low demand, there were no standards
commercially available until recently. Columbia
Scientific Industries Inc. (CSI) has produced the first set
of commercially available standards designed
specifically for hazardous wastes in soils. The primary
calibration curves are based on these standards, which
are listed in Table I as CSI. A description of a
calibration technique for X-Ray Analyzers used in
hazardous waste site screening is presented by Piorek
and Rhodes (5).
Sampling
Sampling was completed in two days. Formerly
acquired metals data was reviewed to identify several
potential areas for field screening. These were visited
in an attempt to limit the number of areas actually
screened with the FPXRF. Three areas ranging in size
from less than one to greater than 10 acres appeared to
be appropriate, i.e., existing data suggested heavy
contamination at those locations, the soil matrix was
typical of the area, the areas were accessible and dry,
and samples processing could be accomplished without
disrupting other activities.
FPXRF screening was accomplished in two steps. First,
a series of stations were staked and located on site
maps. A two-person crew was used, one to set stakes
and one to map the sample locations using a Brunton
compass and a 300 foot tape. Second, a two-person
FPXRF crew completed on site screening at each
station. One person operated the instrument and one
served as data recorder.
FPXRF data was acquired at each of the three target
areas at a rate which exceeded one data point per two
minutes. The rate limiting factor at each target area
was the time required to survey the sampling grid, not
to operate the FPXRF instrument. It might have been
possible to eliminate the second person on the FPXRF
crew without compromising the data acquisition rate.
More time was required to move between target areas
than to sample once the team was in an area. Typical
FPXRF measurement times were 20 seconds per data
point.
The levels of contamination as measured by the FPXRF
for stations within the three areas ranged from 2300 to
70,000 mg/kg for lead, and 750 to 27,000 mg/kg for
zinc. These values cannot be compared directly to
contaminant values as obtained by standard SW 846
methods or CLP methods because they use partial
digestions or extracts for analysis and FPXRF provides
total elemental (or bulk) analyses.
Based on a review of these data, bulk soils were
collected at two target areas between stations
exhibiting the highest contamination levels. Sixteen
samples, each with a field weight of at least 60 pounds
was collected. Prior to shipping , each of these was
analyzed in duplicate for lead and zinc by the FPXRF.
Lead contamination in the samples ranged from 15,000
to 67,000 mg/kg. Zinc ranged from 1900 to 28,000
mg/kg. Samples with this level of contamination were
adequate for both the SCS project and the treatability
studies.
SCS DEVELOPMENT
As analytical instrumentation has moved into the field
to complement laboratory instrumentation, so have the
inherent problems of quality assurance and the
application of field quality control to compare to data
produced by established "conventional" methods of
sample analysis. Given the problems of variability in
results caused by selection of sampling points on a site,
or by variability in relative large volume samples later
analyzed by small aliquot "high sensitivity"
methodologies, project officers and sample plan
designers have turned to two recognized QC procedures
to establish comparability; splitting samples between
analytical facilities and increased use of Standard
Reference Materials. With the increased use of
contract laboratory facilities, the problems have
increased disproportionately with each added analytical
facility introduced in the larger multiple party
sites.Cost and resource expenditure in time and
logistics increase.
Definition
"A Site Comparison Sample (SCS) is a site specific
reference material which is representative of the type
of problems encountered when analyzing or treating
materials from the site." SCS's:
• Contain key contaminants in the matrix of
the site;
• Are available in sufficient numbers to
satisfy numerous site management and
QA/QC purposes;
• Exhibit the lowest possible coefficient of
variation (cv);
• Are managed by an organization capable of
being a depository of analytical results,
providing a common management point for
quality assurance, inter- and
intra-laboratory studies.
SCS differ from Standard Reference Materials (SRM) by
virtue of being site specific, and not produced under a
protocol requiring the pre-release rigorous analytical
method specific, statistically validated
76
-------
characterization applied to SRMs. They also differ
from Performance Evaluation (PE) samples used in
studies to directly compare inter-laboratory results
under a defined methodology. A SCS stock could
conceivably provide the material for a SRM or PE, but
would require those protocols to be applied before so
identifying.
Quality assurance of data developed from multiple
sources presents a complex situation. One major
problem is the question of sample variability and
comparability caused by distribution of compounds of
interest on a site. A second is the variability inherent
in, and between, analytical methods, particularly due to
matrix interference effects. Two common techniques
for dealing with these problems are the use of "split"
samples and analyses of Standard Reference Materials.
Splitting increases the risk of magnifying the problem
due to distribution; standard reference materials
seldom reflect the matrix effects present in "natural"
site samples.
Late in 1984 and early in 1985, the concept of
manufacturing a homogenized bulk sample was
developed to provide vendors of propietary soil
stabilization services uniform materials for evaluation.
The use of screening techniques to define areas of
concern on a site was directly applied to statistically
choosing sources of material to provide a sample
representative of the more highly contaminated
material distributed in the matrix of the site. Mixing
methods were investigated from the viewpoints of cost,
available resources, and practicality. Separate
elements of the methodology were tested on available
materials at various sites. Protocols and standard
operating procedures regarding from where to select
the material, how to homogenize it, and how to fill the
bulk sample containers in a manner that would reduce
bias in the distribution of the material to the large bulk
containers were developed.
The question of how to mix bulk samples of site matrix
materials to achieve a relatively homogenized material
had to be answered empirically. Because of the wide
variety of particle sizes, moisture content, cohesive
characteristics and distribution of contaminants, it was
decided to thoroughly mix the material for the first
1400 pound sample by manually quarter piling through
several cycles; and then do a multiple random fill of
enough buckets (sixty-nine) to meet all projected
needs. It was labor intensive, and took 4 people most
of one day.
The sequence of events discussed in the creation of the
bulk reference materials led logically to the concept of
further treatment of the bulk material to provide a
"Site Comparison Sample (SCS)" for each major site.
Initially, approximately two dozen 8 oz. sample
containers were "broken out" of a bucket, and used for
comparative analyses to determine the degree of
mixing achieved. Some pressure was felt to supply
some of these for comparison analyses instead of
splitting samples. At that time, resources were not
available to so use the material; no statistically sound
evaluation of the material existed to back up any
results.
It cannot be emphasized too heavily that the SCS is not
be to considered a sample that represents the actual
concentration of a contaminant at any given point on a
site. Also, it cannot initially be considered as a true
SRM, although it may be possible to up-grade it's status
if a large number of SCS are generated, and enough
analytical resources are available to utilize a portion of
the banked samples for a statistically sound
standardization analyses. The concept of the SCS is to
produce a material that can be used in lieu of split
samples, and provide a data bank for both continuing
and retroactive analysis of variation due to differing
methods of sample acquisition, handling, and analyses.
As the discrete SCS will be archived in controlled
storage, the effects of holding time can be
demonstrated for each set by continuing
characterization analyses. The more SCS analyzed, the
stronger the statistical evaluation of all data generated
by analyses becomes; not only of the SCS bank itself,
but of the sample of record data and the laboratories
producing the data.
In Statistics there is the "The Central Limit
Theorem": It states:
"From an unknown distribution a random sample
size n is obtained. If n is allowed to become
larger, the sample mean will behave as if it came
from a Normal distribution, regardless of what
the parent distribution looked like."
John Webber, Statistician for EPA Office of Policy and
Planning, had provided a table illustrating how
Normality affects a sample population (Table II) taken
from a universe, and reverse logic suggests that very
low variances could be expected from discrete samples
of nn, especially if the discrete samples were
produced by actually filling the randomly selected
sample containers with a series of multiple portions
selected at random from the bulk n^ material. (The
"double random" referred to hereafter.)
Reasoning from this point, if n is sufficiently large, and
then thoroughly mixed or homogenized, multiple
random creation of n^ should result in a low variance
that approaches the "true" value of the concentration
of the mean of n. As the number of random selections
used to create n^ increases, the coefficient of
variation should decrease.
Through the balance of 1985 and into 1986, the
analytical results from the stabilization tests made on
the bulk materials were reviewed Protocols were
developed through experimentation to mix sludges of
water, sediment and hydrocarbon products. A protocol
for groundwater SCSs was developed
Finally, in late 1986 an opportunity presented itself to
produce an actual SCS for a large, established
Superfund site. This dovetailed with the trial of the
X-Met FPXRF equipment, and made it possible to more
soundly screen the bulk "raw material" for both
stabilization studies and two SCSs; one "high" range and
one "low" range. A fairly ambitious design was
proposed to produce between 300 and 500 8 oz. samples
in each range.
Experience with the homogenization of the original
stability samples suggested that it would be desirable to
utilize more efficient methods of mixing the bulk
sample material. Accordingly, a "drum roller" was
obtained, and 55 gal O.T. steel drums were modified
with two interior deflection vanes similar to those used
in industrial dry mixing of materials. The bulk sample
material was batched through this drum and then spread
out in a distribution box for the double random
selection of the SCS samples. The available quantity of
material dictated that only a single SCS be produced, so
the "high" and "low" bulk retains were incorporated into
77
-------
a single batch for processing.
The 600 aliquots have been "banked", and a master
random distribution list prepared. From the bank, an
initial set of 10 SCS (the first block on the list) were
supplied to the USEPA Environmental Monitoring
Services Laboratory, Las Vegas, NV. for preliminary
characterization analyses. At the same time, a
principle contractor was issued the next 30 samples for
release to their contract laboratories for the same
purpose. All analytical data results are to be reported
to Region 10, and a running control chart of results
developed.
As the number of samples analyzed increases, the data
will become progressively more refined, and amenable
to other statistical analyses to more closely define the
sources of variability, from laboratory, to method, and
to a certain extent, the effects of holding time. Data
currently available are presented in Figures 1 and 2.
Although the number of data points are limited, there is
a suggestion that inter-laboratory differences may be
important (Figure 1), and that overall ev's are low (less
than 30%).
As related, this is an ongoing developmental effort.
Preliminary data indicate the approach is sound. For
middle to large site hazardous waste operations, and for
long term ambient monitoring projects, the economies
of scale would apply. For improved data quality and
scientific credibility the concept is entirely appropriate
and defensible. The practical application awaits
resources and initiatives on the part of the user
programs.
REFERENCES
(1) Chappell, R. W.( Davis, A. O., Olsen, R. L.,
"Portable X-Ray Fluorescence as a Screening
Tool for Analysis of Hazardous Materials in Soils
and Mine Wastes," the 7the National Conference
of Management of Uncontrolled Hazardous Waste
Sites, Hazardous Materials Control Research
Institute, Silver Spring, Maryland, 1986.
(2) Raab, G. A., D. Cardenas, and S. J. Simon,
"Evaluation of a Prototype Field-Portable X-Ray
Fluorescence System for Hazardous Waste
Screening," EPA/600/4-87/021, U.S.
Environmental Protection Agency, Las Vegas,
Nevada, 1987.
(3) .Gulf Resources and Chemical Corporation,
"Bunker Hill Site Remedial
Investigation/Feasibility Study for Unpopulated
Areas," April 24, 1987.
(4) Piorek, S., Rhodes, J. R., "Hazardous Waste
Screening Using a Portable X-ray Analyzer,"
Symposium on Waste Minimization and
Environmental Programs within DOD, American
Defense Preparedness Association, Long Beach,
California, April 1987.
(5) Piorek, S., Rhodes, J. R., "A New Calibration
Technique for X-Ray Analyzers Used in
Hazardous Waste Screening"
Standard Elements:
Name
Table I
Concentrations of Standards
Pb Zn Cu As
(All values are in mg/kg)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
CSI IB
CSI 2B
CSI3B
CSI5B
CSI6B
CSI7B
CSI8B
CSI9B
CSI 10B
CSI 11B
CSI 12B
CSI 13B
CSI 14B
CSI 15B
0
0
4980
240
484
4760
1474
1990
2930
2440
3405
4126
0
0
4790
0
0
240
482
4900
983
2970
3910
6360
8270
9791
0
4950
4790
0
0
8160
6300
3810
2950
982
1960
490
243
96
4950
0
6970
11,340
0
7740
5590
11,070
4530
3390
2250
1140
565
224
0
0
78
-------
Table II
Illustration of How Normality Affects Samples
Let us phrase the question "How many samples do I need to be within Q
sigma "s" (Standard Deviations) of the true value?":
Confidence Confidence Confidence
90% 95% 99%
Q Sigma Normal Worst Normal Worst Normal Worst
"s"
2s
Is
0.75s
0.5s
0.4s
0.3s
0.2s
O.ls
from:
jft
Coefficient of Variation (%)
-» NJ N) GJ Ol 4
Oi O cn o tn C
1 1 1 1 1 1
1 n
I U
5 -
Case Case Case
13 15 2 25
3 10 4 20 6 100
5 18 7 36 10 178
11 40 16 80 22 400
17 63 25 125 34 625
31 112 43 223 61 1112
68 250 97 500 136 2500
271 1000 385 2000 543 10000
"Statistical Considerations in Sampling Hazardous Waste Sites", John Warren,
E.P.A./O.P.R.M.
Figure 1
BETWEEN LABORATORY COMPARISON
^
\
YY,
1
^
*S//
'// YY //^
/// YY/ /// ///
// // // \ ^ //
/y ^ty ^/y ^y
^ \s| ^ ^
As Cd Fe Pb Zn
Target Chemical
IX\I Laboratory A Y//A Laboratory B
79
-------
O
r
c
o
o
o
40
35 -
30 -
25 -
20 -
15 -
10 -
5 -
Figure 2
SCS COEFFICIENT OF VARIATION
I
As
Cd Fe Pb
All Laboratories by Chemical
Zn
DISCUSSION
HAROLD VINCENT: How were you going to apply the zeolites to the
problem?
JOHN BARICH: Our first step was to determine whether or not the zeolites
would be a useful soil amendment. If the answer to that was a strong yes, then
the application technique would have been the next thing we would have looked
at.
HAROLD VINCENT: That's in place of removal?
JOHN BARICH: In place of removal, yes. We had literally many square miles
of land whose condition needed to be improved. There was just not enough
secure landfill capacity, to do anything other than in situ.
80
-------
LOW LEVEL XRF SCREENING ANALYSIS
OF HAZARDOUS WASTE SITES
Randy Perils
Mark Chapin
Ecology and Environment, Inc.
Denver, Colorado
ABSTRACT
Recent field investigations have demonstrated the
successful use of XRF screening analysis for
metal contamination at various hazardous waste
sites.
Using minimal sample preparation and field
sampling methods the results were comparable to
laboratory results using conventional methods
such as AA and ICP. Multi-elemental analysis
was performed on soil samples with particular
interest in lead, arsenic, chromium, and copper
levels. Detection limits achieved for some
elements were 10 ppm. The XRF inorganic results
were used in mapping and contouring the extent
of contamination of a hazardous waste site con-
taining organic and inorganic contamination.
The lower detection limits and quick turn around
times proved the feasibility of the XRF in
screening of hazardous waste sites and environ-
mental monitoring.
INTRODUCTION
The Ecology and Environment, Inc. (E & E) Field
Investigation Team (FIT) was tasked by the U.S.
Environmental Protection Agency to initiate a
field analytical screening program (FASP) to
assist in site investigations and listing or
expanded site investigations. Field screening
is projected to enhance the pre-remedial
program by: 1) assisting the EPA in completing
the site inspection inventory in a timely manner,
2) decreasing the number of "non-detected"
samples, 3) supporting the revised Hazardous
Ranking System, and 4) accelerating scope of
remedial investigations and feasibility studies.
The increased sampling capability consequently
increases the chances of detecting a release
without compromising data quality, since rapid
turn-around of screening samples allows selected
split sample Contract Laboratory confirmation.
Part of this program was to develop a screening
analysis for metal contaminated solids such as
soils and sediments including mine tailings and
mining waste materials in EPA Region VIII. FIT
determined the best instrumentation for these
types of analyses would be an x-ray fluorescence
spectrometer (XRF). Previous successful opera-
tions with the XRF (Raab et al, 1987; Furst &
Spittler, 1985; Mernitz & Staible, 1985; Piorek &
Rhodes, 1987) indicated the XRF's usefulness in
screening analysis of metal contaminated solids
on potential hazardous waste sites. However,
lower detection limits below 100-200 ppm were
difficult to achieve.
The rapid turn around times available on a wide
variety of elements and minimal sample prepara-
tion made the XRF almost ideal for screening
analysis. As previously stated, one of the major
drawbacks associated with the XRF was the
relatively high detection limits. However with
the Tracor 6000 XRF, E & E is able to achieve
detection limits of 10 parts per million consis-
tently and confidently without liquid nitrogen
cooling of the XRF detector as needed for other
conventional low level XRF analyses. This
advantage greatly increases the mobility of the
instrument. These detection limits are more
than adequate for most soil samples from metal
contaminated sites.
The purpose of this paper is to summarize E & E's
experience with low level XRF analysis to date
and present the comparability of data of XRF
screening analysis and AA/ICP analysis from the
Contract Laboratory Program on co-located
samples. An example of how the XRF screening
analysis is used to characterize a hazardous
waste site with grid sampling and contour mapping
is presented.
This material has been funded wholly or in part
by the United States Environmental Protection
Agency under contract #68-01-7347 to Ecology
and Environment, Inc. It has been subject to
the Agency's review, and has been approved for
publication. Mention of trade names or commer-
cial products does not constitute endorsement
or recommendation for use.
INSTRUMENTATION
The Tracor Spectrace 6000 energy dispersive
x-ray fluorescence analyzer system includes the
81
-------
following specifications. The source is a 50KV,
0.35mA rhodium x-ray tube. The unit has two
available filter positions with a 0.13mm aluminum
filter and a 0.13mm rhodium filter. Also used is
a 0.13mm copper filter. The x-rays are filter
directed. The detector is a thermally cooled
(Peltier) detector with a 20 square millimeter
area. The detector is cooled to approximately
-76 degrees C. The analyzer is a multi-channel
analyzer with 1024 channels. The XRF unit is
controlled by a NEC Power-Mate II P.C. that
controls the spectrometer and receives the data
via an interface card in the P.C.
The lower detection limits are achieved on this
XRF unit due to the high flux x-ray tube,
thermally cooled high resolution detector, and
peak deconvolution software. Earlier XRF instru-
mentation work was done with instruments using
radioisotope source (low resolution) excitation
and low resolution proportional counter detectors.
The Peltier cooled detector exhibits relatively
little performance differences compared to
liquid nitrogen cooled detectors (Harding, 1988).
XRF OPERATION
Elemental identification and quantitation is
obtained using the "Fundamental Parameters" PC
software (PCXRF) integrated with the Tracer 6000
XRF (Leyden, 198A).
When elements present in a soil sample are
irradiated with a beam of x-rays, electrons in
the atoms lower lying energy levels are excited
to higher energy levels. The vacancies left in
the inner electron orbitals make the atom
unstable. Relaxation to the stable ground
state occur resulting in the emission of x-rays
characteristic of the excited elements (Figure 3).
Thus, by examining the energies of the x-rays
emitted by the irradiated soil sample, identifi-
cation of elements present in the sample is
possible. Comparing the intensities of the
x-rays emitted from a given unknown sample to
those emitted from reference standards with known
analyte concentrations allows quantitation of
the elements present in the sample.
During sample analysis a spectrum is acquired
as shown in Figure 1. Optimized for various
emission energy levels with different instrumen-
tal parameters and excitation conditions the XRF
is able to analyze for various elements.
Generally, elements are segregated for analysis
into groups having similar atomic numbers.
Currently we are analyzing for fourteen different
elements using three separate excitation condi-
tions. Figure 1 is a sample spectrum for the high
atomic number elements: manganese, iron, nickel,
copper, zinc, arsenic, and lead. Figure 2 is
identical to Figure 1, but has the mid atomic
number elements analysis superimposed on to it.
Elements of interest here include potassium,
calcium, and chromium. The superimposed
spectrum shows that the excitation conditions
employed for the mid-atomic number analysis
greatly enhance the spectrum for those elements.
As previously stated, peak position along the
spectral energy axis (horizontal axis) is
indicative of the metal it arose from, and is
therefore the primary basis of elemental identi-
fication. It should be noted that each metal
will exhibit several peaks in the spectrum,
since a separate peak will be observed for each
allowed electron orbital energy transition. For
example, peak A in Figure 1 is lead's L-alpha
line. It arises when electrons initially
excited to a lead atom's M shell return to the
lead atom's L shell giving off x-rays which
have an energy of 10.5 KeV. Peak B is lead's
K-beta line. When electrons in the lead atom
energetically relax from the N shell to the L
shell, x-rays at 12.6 KeV are emitted. Figure 3
shows a representation of this process.
The multiple linear least squares deconvolution
that is used is excellent method of unfolding
peak overlaps in a spectrum. The software
peak extraction routine can integrate any
emisison line in the spectrum.
Prior to running a series of samples, the instru-
ment is calibrated using a pure copper disk.
Basically, the instrument adjusts its spectral
energy axis until the copper x-ray emission
peaks fall at the correct energies. The
energies of other metal peaks are then determined
relative to the established copper peaks. This
peak position monitoring is performed at least
daily.
The area under each element's peaks, termed peak
intensity, is proportional to the concentration
of that element in a sample. Through peak
deconvolution and using multiple linear least
squares, integration is carried out and the
results evaluated using Tracer's "Fundamentals
Parameters" software (Leyden, 1984 and Leyden,
1988).
XRF STANDARDIZATION
Standardization in the PCXRF program, which is
propriety fundamental parameters routine that is
integrated into the SSXRF software that controls
the spectrometer functions, is computationally
complex but descriptively quite simple. The
PCXRF program proceeds by modeling the x-ray
tube output from the spectrometer and using
fundamental parameters, the standard concentra-
tions, and the measured intensities for the
standards to calculate pure element count rates
for the XRF measurable elements. Theoretical
standards are produced solely to compute alpha
coefficients which account for all matrix
interactions. The pure element count rates and
alphas on stored on disk.
To compute unknown concentrations, an estimate
of concentration is first made using the pure
element count rates and measured peak inten-
sities for the unknown. This is followed by a
calculation of expected intensities from the
predicted composition of the unknown and the
alpha coefficients. These new computed
82
-------
intensities are compared to the measured
intensities and in light of the result, a new
approximate composition is assumed. The
iteration proceeds through composition/expected
verses measured intensity/new approximation of
composition. If the two compositions disagree
by less than 1%, convergence is assumed and the
final composition is output as the result for
the unknown (Harding, 1988).
For initial standardization, a set of reference
standards with known analyte concentrations is
run. Currently, certified samples are available
from the U.S. National Bureau of Standards and
Canadian Certified Standards Center. These
certified standards are very well characterized
and employ up to seven different analytical
methods. In a typical XRF analysis the
standards are used to construct a calibration
curve by plotting measured x-ray intensities
against known concentrations. However, in
soil sample analysis, the varied composition
of the soils causes problems that can attenuate
the emissions from elements being analyzed. In
general, the absorbing properties of a soil
matrix, termed matrix effects, increase as a
function of the average number of the elements
in the sample increase. In addition to matrix
effects, there are inter-elemental effects.
Inter-element effects occur when an element in
the matrix can specifically absorb or enhance
x-ray photons emitted from another element. The
"Fundamental Parameters" program quantitatively
corrects for changes in the sample's matrix
and for inter-element effects (Criss & Birk
1986, 1978; Leyden 1984, 1988).
QA/QC
QA/QC for XRF screening analysis includes
duplicate samples, standard checks, and splits
with other laboratories. Sample duplicates
are run at a 10-20% frequency with the smaple
split before sample preparation. This will
indicate the precision of an analysis as well as
the homogeneity of the sample matrix.
National Bureau of Standards (NBS) or Canadian
Certified standards are run at a 10-20%
frequency to determine continued standardiza-
tion of the instrument.
Also splits of the solid sample materials are
sent to analytical laboratories for AA/ICP
comparison at a 10-20% frequency rate.
SAMPLE PREPARATION
Soil and sediment samples are collected with
the usual protocol, however not as large a
sample is required as with the acid digestion
AA/ICP analyses. However, the most homogeneous
sample possible is recommended.
No great differences have appeared as whether
grab or composite sampling is more suitable
provided the samples are well mixed. Grab
samples have shown a slight statistical
advantage in comparing with AA/ICP results.
Analysis of particulates collected on dust
filters is just now being tested. No sample
preparation is involved with air filters,
however accuracy of the results depends
greatly on sampling procedures and accurate
measurement of the amount of particulate
matter collected.
Sample preparation for XRF screening analysis
was designed to be kept simple. XRF sample
preparation procedures can be as complex as
pellet pressing or fluxing the sample.
Accuracy of XRF results and their relation to
sample preparation is described in detail by
Wheeler, 1987. However XRF sample preparation
procedures are still obviously quicker and less
hazardous than the acid digestion AA/ICP
methods.
The XRF screening sample preparation is minimal
to ensure rapid turn around. This sample
preparation includes air or mild oven drying of
the solid sample and mixing in a mortar and
pestle to homogenize the sample as much as
possible. No sieving is performed unless the
sample contains particles larger than 10 mesh.
COMPARISON OF RESULTS
As in any inter-method comparison, the more
alike the sample and the procedures, the more
valid the comparison. However, in dealing with
soils and solid matrix contaminants, the
homogeneity of the sample is always in question
and therefore a true duplicate or split is
extremely difficult to obtain. Also, comparing
a XRF method with an acid digestion AA/ICP
method is risky since both methods (including
standardization and sample size) are quite
different.
Since most comparisons are used beyond the
intent of the initial project, one must keep in
mind we are comparing AA/ICP litigation and
regulatory enforcement CLP data with XRF
screening data.
Figures 4,5,6, and 7 represent comparisons of
chromium, arsenic, copper, and lead XRF results
with typical AA/ICP laboratory results
(SW-846). Lead and copper comparisons had
excellent correlation coefficients of 0.97 and
0.98 respectively. Arsenic's correlation
coefficient was 0.89. However three outlier
points are noted and therefore the actual
correlation may be better than reported. The
correlation coefficient for chromium was 0.81.
Chromium appears to be biased high on all XRF
results. The relatively low correlation can
be attributed to the low measured intensities
of the chromium in the standard materials, hence
the accuracy with which pure element count
rates can be calculated is poor.
All XRF results appear to be biased high
compared to the AA/ICP results. No documented
83
-------
explanation is available at this time to explain
this, however current speculation attributes
this occurrence to the AA/ICP acid digestion
procedure, the total sample analysis with the XRF,
standardization procedures, and sample size.
INTERPRETATION OF RESULTS
Sampling points or grid layouts are critical for
proper interpretation of the XRF results. Usual
grid layouts are based on site size, detail of
investigation, turn around time, and economics
such as number of samples and man hours available.
Interpretation of XRF results should be restricted
to use for evaluation and assessing the results
for average pollutant exposure to humans and
animals. This level of analytical requirements
should show a precision of 10% and an accuracy
of 15% on samples in the calibration range of
the instrument.
The majority of XRF samples are run for field
screening purposes at lower analytical require-
ments and the results are used for screening,
preliminary evaluation, and on-site decision
making.
The figures presented show contamination zones
and relative amounts of contaminants of a
hazardous waste site. The intensity of the
sampling was employed to characterize the waste
present on the site and in the immediate areas
and to evaluate the on-site direct contact
pathway. The contouring program employed was
the Kriging contour method. Figures 8-15 present
results of a grid examination of a site using
XRF analysis. Figures 8 and 9 show overlays of
the contouring with the site map. Figure 8
represents the zinc contouring while Figure 9
represents the lead contamination. The
comparisons of these contour maps shows that
lead and zinc high contamination zones are not
related and the contaminaiton extends beyond the
site's boundaries. Figures 10,12, and 14 show
the grid layout of the site for different
elements with contouring. Figure 10 represents
a more detailed contour of the lead contamina-
tion as compared to Figure 9. Figure 10's
data points have the actual XRF concentration
values in ppm printed above the points.
Comparing Figures 10,12, and 14 (lead, arsenic,
and copper), the contour maps show differing
"hot spots" for each element indicating the
independent variables involved at this site.
Figures 11,13 and 15 show a three dimensional
contouring of the various elements. Comparisons
of the 3-D figures again reinforces the varia-
bility of the contamination zones and graphi-
cally illustrates the degree of contamination.
CONCLUSIONS
XRF screening analysis of low level metal
contamination is proving to be valuable in the
investigations of hazardous waste sites. XRF
screening analyses have been proven very
effective in establishing contamination boundaries
using contouring maps and very useful in
visualizing contaminated zones and amount of
contaminants in comparision with background
samples and on-site contamination. The cost
savings compared to usual inorganic analytical
services is estimated to be $80 per sample
after instrument payoff. The turn around
times with the XRF are conducive to field
screening analysis. The small amound of
sample necessary and minimal sample preparations
diminishes health and safety problems and
reduces the amount of sample disposal. With
the advancing technology EDXRFs are becoming
more mobile while maintaining low detection
limits, thus field screening analyses are
possible. Finally, the results obtained from
XRF screening analyses show good correlations
with other types of inorganic analyses and
basic trends and comparisons can be confidently
made.
ACKNOWLEDGEMENTS
The authors wish to recognize individuals who
contributed to this paper. We thank Les
Sprenger of the Region 8 EPA, Stuart Richardson
and Karl Ford of Ecology and Environment, Inc.,
for providing the support and resources for this
project. Greg Raab, LEMSCO, and Hunt Chapman,
E & E, for review comments. Also Anthony
Harding, Tracor X-Ray, Inc., for invaluable
help in technical accuracy.
REFERENCES
Raab, G.A., "Evaluation of a Prototype Field
Portable X-Ray Fluorescence System for Hazardous
Waste Screen", EPA Research & Development",
Aug. 1987.
Piorek, S. & Rhodes, J.R., "Hazardous Waste
Screening Using a Portable X-Ray Analyzer",
R&E Report #528, March, 1987.
Furst, G. & Spittler, T., "Screening for Metals
at Hazardous Waste Sites: A Rapid Cost-Effec-
tive Technique Using X-Ray Fluorescence",
Management of Uncontrolled Hazrdous Wastes Sites,
6th National Conference, Nov., 1985, pp. 93-96.
Mernitz, S. & Staible, T., "Use of a Portable
X-Ray Analyzer and Geostatistical Methods to
Detect and Evaluate Hazardous Metals in Mine/
Mill Tailings", Management of Uncontrolled
Hazardous Waste Sites", 6th National Conference,
Nov. 1985, pp. 107-111.
Wheller, B., "Accuracy in X-Ray Spectrochemical
Analysis as Related to Sample Preparation",
Spectroscopy, Vol. 3, No. 3, 1987, pp. 24-33.
Harding, A., TRACOR X-RAY, personal communica-
tion, 9/09/88.
Leyden, D.E., Bilbrey, D.B., Bogart, G.R.,
"Comparison of Fundamental Parameters Programs
for Quantitative X-Ray Fluorescence Spectromety",
X-Ray Spectrometry, Vol. 17, 1988, pp. 63-73.
84
-------
Leyden, D.E., Fundamentals of X-Ray Spectro-
raetry as Applied to Energy Dispersive Techniques,
TRACOR X-RAY Inc., 1984.
Criss and Birks, Analytical Chemistry, Vol. 40,
1968, pp. 1080.
Criss, Birks, and Gilfrich, Analytical Chemistry,
Vol. 50, 1987, pp. 33.
TUBE VOLTAGE : 3O KV
TUBE CURRENT : 0.17 mA
ATMOSPHERE : AIR
FILTER USED : FIVE
LIVETI ME : SOO SEC
F.S. = 6<+K
mi mi hm hm hm hm hm hi: ism i mil
ill li : Im hm -m iim i m m'; iim .';
P r.
b b
H | (B) L-beta
-
(A) L-alpha
1 ,
i
(
4
1 H
i
i
P R-1 F I: ZCPZ
b iri e i.i. nub n
-;<, - L }
in mi ri'i n n i in pi i" ' n i[ 11 rjn ' 'II
F
tj
P HFJ
bsd
j
'111 II
F'
b
P
b
P
b
' I
h rt | n i i 1 1 1 i i i 1 1 1 TT
3 1 2 3 4 5 6 7 8 9 18 1 1 12 13 1 4 1 5 16 1 7 IS 1 9 29
-------
TUBE VOLTAGE : 30 KV
TUBE CURRENT : O.I7
il
~T
,
"! Ml!
"^
P
b
r~'
mi
,
~n
1 ' i
~1 i 1 [
r
c /
1
1 1 1 1 j'l ITT
*
FM
F
A
C
[ r TI i | i
F
9
\-
Jl
j i i rj"i
H
i
J\.
111 1 TT
_£
1
ie
C ZCP ~
>3"i e u nub n
1
TT
4
III
M i'i
n | (
i i
III 1
1
r-
H
5 6 7
H
i
1
.'I
1 1
l".
r •' *
II
.'1(
i ••-
1 1
S 9
L
C
tl
P AP
b s
1
n HIT
18 11
H
F'
b
P
b
P
, b
1
,' 1 ' 1.
' V ,• S ^
II
III III N III 1 II III 1 II 1 II
12 13 14 15 16 17 13 19 29
KEV
Figure "i
86
-------
— Illustration of photoe/ection of K electrons by higher energy radiation, and L
electrons by lower energy radiation.
E 'Absorption edge
ENERGY-
ENERGY—
— Excitation of electron energy levels: taj transition between two quantized energy
levels and fbl ohotoeiection of electrons bv x-radiation
Figure 3
87
-------
CHROMIUM XRF vs AA/ICP RESULTS
•
350.00 -
300.00
250.00
200.00 -
150.00 -
100.00 -
50.00 -
n nn -
n
Correlation Coefficient = 0.81
n
n
T— i no
LJ
0.00 60.00 80.00 100.00 120.00 140.00 16O.OO 180.0O 2OO.OO 220.00
XRF (ppm)
Figure 4
ARSENIC XRF vs. AA/ICP RESULTS
1 «*-U.WU -
120.00
100.0O -
80.00
60.00 -
40.00 -
20.00 -
n nn -
n a
n
Correlation Coefficient 0.89
n
a
n D
a
n
tZnf ncr a
0.00 10.00 20.0O 30.OO 40.00 50.00 60.OO 70.00 BO.OO 9O.OO 100.00
XRF (ppm)
Figure 5
88
-------
COPPER XRF vs. AA/ICP RESULTS
1 OUU.WU
1600.00 -
1400.00 -
1200.00
1" 1000.00
^ 800.00 -
2
600.00 -
400.00 -•
2OO.OO -
n nn -
D
Correlation Coefficient 0.98
n
n
n
n
n n
n
n
n n
.00
500.00
1000.0O 1500.00 2000.00 2500.00
XRF (ppm)
3000.OO
f
1200.0O
1000.0O
800.OO
60O.OO
400.00 -
200.00
0.00
n
Figure 6
LOW LEVEL LEAD XRF vs. AA/ICP
Correlation Coefficient = 0.97
B
n
n
T 1 1 1 1 1 1 1
3.00 100.00 200.00 300.00 400.00 500.OO 600.00 700.00 800.00 900.00
XRF (ppm)
Figure 7
89
-------
L _J
\ C-4
UB-SO-26
400 —V.' ^ I'"' '""""'"" *"*J >^
^' G-3
F_3 AB-SO-J6
D
Olr A8-SO-8 Collet
1 800 (del SW ol American Barrel
Soil
XRF sample alias
FIELD INVESTIGATIONS OF UNCONTROLLED
HAZARDOUS WASTE SITES
3NC CONTOUR MAP
Conloui Interval : 200 ppm
F08-B802-09
ecology & environment, inc.
DENVER, COLORADO
I AB-SO-28 / AB-SO-29 I
" -"• " '
I ^ i
i Sit* boundary area
Soil umpla
XRF sample allai
FIELD INVESTIGATIONS OF UNCONTROLLED
HAZARDOUS WASTE SITES
TABH HCPOATTO THC I.P. *
LEAD CONTOUR MAP
Con I our Interval • 200 ppm
TOD, F08-8B02-09
ecology & environment, inc.
DENVER. COLORADO
., 04/8B_ 0,.wn br, RSM
90
-------
400
371 -
342 -
313 -
283 -
254 -
225 -
LEAD CONTOUR MAP
579
206
567
50 79 107 136 164 193 221 250 279 307 336
Figure 10
Contour Interval - 200
Cone, in ppm
Distance In Feet
LEAD 3-D MAP
Figure 11
91
-------
ARSENIC CONTOUR MAP
400
371 -
342 -
313 -
283
254 -
225 -
I I I ° I I I ? I
50 79 107 136 164 193 221 250 279 307 336
196
167 -
138 -
108 r
Contour Interval - 2
Cone. In ppm
Distance In Feet
ARSENIC 3-D MAP
Figure 13
92
-------
COPPER CONTOUR MAP
400
371 -
342 -
313 I-
283 -
254 -
225 -
129 295
107 136 164 193 221 250 279 307 336
Contour Interval - 100
Cone, in ppm
Distance In Feet
50 79
COPPER 3-D MAP
Figure 15
93
-------
DISCUSSION
HARRY McCARTY: You mentioned the sample disposal problem, and the
lack of homogenization, but you also said earlier that you were able to use other
tests on the sample, since it's a nondestructive method.
How do your sample sizes compare to what the CLP would use, and do you
avoid the problem of having to take splits? You could do XRF, and then send
the sample to an analytical laboratory for all the techniques, get two sets of
answers on the same sample, and give the lab the sample disposal problem.
RANDY PERLIS: Yes, you could do that, but when we laid out the grid
system, there were about 40 samples, and of those, we sent about eight to 10 to
the CLP.
This site had already been investigated, as have a lot of sites we do. When we
have some idea, we'll still send off the hottest samples that we get to the CLP
and try to avoid the sample disposal problem that way.
HARRY McCARTY: Are the sample sizes approximately the same? Would
you need the same material?
RANDY PERLIS: No, I believe we collect an eight- or four-ounce jar for the
CLP, and we collect about 1.5 mL.
HARRY McCARTY: CLP certainly doesn't use eight ounces for a metals
analysis?
RANDY PERLIS: No, but that's what they request.
HARRY McCARTY: Could you conceivably use the same sample size?
RANDY PERLIS: Yes, you could use it out of the same vial. You may run into
some sample custody problems that way, if you run it to lab first. Basically, we
have co-located samples. We'll take the eight-ounce jar and fill it up, and from
that one, we'll homogenize it and mix it up, take a split out of that for us, and
then send it on to the CLP.
HAROLD VINCENT: Concerning preparation for a very small sample, did
you pulverize before mixing? Did you affect the sample size?
RANDY PERLIS: Yes, we mixed it with a hand mixer, with a mortar and pestle
and then sieve.
HAROLD VINCENT: Did you consider briqueting, for both standards and
unknowns?
RANDY PERLIS: No, we haven't yet. We were thinking about some other
techniques, such as fluxing, but we 're try ing to keep the sample prep as minimal
as possible.
HAROLD VINCENT: I was thinking of briqueting in terms of stability, and
you could always relate the standards back to a powder. You could even mix
briquets with powders, I would think.
94
-------
WAVELENGTH TUNABLE PORTABLE LASER
FOR REMOTE FLUORESCENCE ANALYSTS
Gregory D. Gillispie and Randy St. Germain
Department of Chemistry
North Dakota State University
Fargo, ND 58105
ABSTRACT
Direct fluorescence is a sensitive technique
for in situ detection of many groundwater
contaminants. Details of a compact, high
performance tunable dye laser system suitable for
such remote measurements are reported. The second
harmonic (532 nm) or third harmonic (355 nm) of a
pulsed Nd:YAG laser pumps an oscillator plus one
or two amplifier stages. The dye laser output is
then frequency doubled to generate tunable
ultraviolet radiation. The current version of the
instrument provides pulse energies of up to 10 mj
in the 560-600 nm region and 300 pj between 280
and 288 nm with the laser operated at 10 Hz.
Better optic sets and optimization of the design
should eventually yield at least a factor of three
improvement in the pulse energies. With the
current version, naphthalene and fluoranthene in
aqueous solution have been detected at the ppb
level using a simple fiber optic probe.
KEY WORDS: Fluorescence, laser, fiber optic,
remote, in situ, analysis
INTRODUCTLON
The advantages of fluorescence as an analytical
technique are well established. For example,
detection limits of about one part per trillion
(pg/mL) and below for aqueous solutions of
polycyclic aromatic hydrocarbons have been
demonstrated in the laboratory setting (1). In
addition, fluorescence is a direct method which
eliminates the tedious and slow steps of sample
concentration, separation, etc. Thus, the
response of a fluorescence based instrument to an
analyte is virtually instantaneous.
The combination of speed, sensitivity, and
specificity makes fluorescence a good candidate
for field screening analysis. Moreover, as an
optical technique fluorescence can be combined
with fiber optic methodology for remote analysis.
Chudyk, Kenny, and coworkers at Tufts University
were among the first to explore this possibility
(2,3). Although incoherent light sources can be
used to launch light into an optical fiber for
remote fluorescence analysis, the properties of a
laser obviously make it the method of choice.
For maximum versatility and ability to distinguish
one species from another, the laser light source
should be tunable (wavelength selectable). In
this work we explore the performance capabilities
of a YAG-pumped dye laser suitable for remote
fluorescence analysis with fiber optic probing.
EXPERIMENTAL
The apparatus is schematically shown in Figure 1.
Key elements of the dye laser are discussed more
or less in the order they appear in the optical
train.
A. Pump Laser - The Nd:YAG pump source is a
Quanta-Ray model DCR-11 operated at 10 Hz. In the
Q-switched mode, pulse energies of up to 150 mj at
532 nm and 60 mj at 355 nm are available for dye
laser pumping. In the work so far, we have only
used the 532 nm output to pump Rhodamine 590 and
have had to keep the pump power low to avoid
damaging the currently available optics.
B. Energy Splitting to Oscillator and Amplifier
Cells - A microscope slide beam splitter (BS1)
splits off about 8% of the initial 532 nm beam to
pump the dye laser oscillator. BS2 is nominally a
half-silvered mirror acting as a 50% partial
reflector but even at modest YAG pump powers the
surface of the partial reflector degrades rapidly,
thereby reducing both its reflectivity and
transmission. The optimal reflection/transmission
ratio of BS2 will be determined by future
experiments with high quality dielectric
beamsplitters. A likely starting point would be
20% reflectivity of BS2 to pump amplifier cell AC1
with the remaining 80% going to pump the final
amplifier cell AC2.
95
-------
C. Focussing Optics to Oscillator and Amplifier
Cells - The doughnut-shaped beam profile of the
DCR-11 is somewhat inconvenient for dye laser
pumping. However, the combination of a spherical
lens and a meniscus lens satisfactorily flattens
the beam along its vertical axis and expands it
along the horizontal axis sufficiently to focus it
to a line image just inside the front face of the
flowing dye amplifier and oscillator cells.
D. Oscillator Cell, Amplifier Cell(s), and
Circulator - The oscillator and amplifier cell
bodies are machined from stainless steel with
front and side fused silica, windows epoxied to the
body. The same dye solution is used for both the
oscillator and first amplifier so only a single
circulator suffices for both cells. We have found
that an inexpensive epoxy-clad magnetic drive
centrifugal pump (Cole Farmer model J-7105-00)
with flow rate of up to 3 gpm works well.
Connections between the pump and the cells are
made with polyethylene tubing and Swagelok
stainless steel fittings.
a. Dye Laser - The dye laser oscillator is of
grazing incidence design (<4) and employs a 2400
grooves/mm holographic strip grating (PTR Optics)
and an aluminum tuning mirror. The dye laser is
operated in the so-called closed configuration
where the wedge prism feedback element (Melles
Griot, 3°52' angle) is also the output coupler.
The tuning mirror is affixed to a precision
rotation stage with motorized encoder drive (Oriel
model 13028) controlled from a personal computer.
F. Doubling Crystal - An angle tuned KDP crystal
in a gimbal mount is used to frequency double the
dye laser output. A red corex filter (of the type
commonly found on old Beckmann DU
spectrophotometers) removes the fundamental but
transmits the tunable ultraviolet with good
efficiency.
RESULTS
Pumping the oscillator with ca. 3 mj and the
amplifier cell with 15 mj of 532 nm radiation at
10 Hz repetition rate yielded tunable dye laser
pulse energies of about 3 mj above the Amplified
Spontaneous Emission (ASE) level at the peak of
the Rhodamine 590 gain profile. The ASE could
undoubtedly be reduced appreciably by optimization
of the dye concentration, splitting of pump energy
between oscillator and amplifier, and variation
of the delay time for the pump energy reaching the
amplifier cell. However, the presence of high ASE
levels is much less of a problem when the dye
laser fundamental is to be frequency doubled in a
subsequent step. Since as much as 150 mj/pulse of
532 nm radiation is available from the pump laser,
addition of a second amplifier cell and
optimization of the operating parameters (and the
incorporation of higher quality optics) ought to
yield very powerful laser pulses for the
generation of tunable ultraviolet radiation.
The collimated dye laser output from the amplifier
cell was passed into the doubling crystal without
any further focussing. Conversion efficiencies on
the order of 5% (on an energy basis) were
observed. The frequency doubled output can be
separated from the unconverted fundamental with a
dispersive prism or with a glass filter.
Our goal is to develop a robust and relatively
inexpensive dye laser suitable for field studies.
Little work has been done on actual analyses at
this stage. However, we did build a simple 2
meter probe of the type described by Scwab and
McCreery (5). The 200 urn Ensign-Bickford fiber is
not controlled for UV transmission and we only
included six collection fibers. Undoubtedly, the
efficiency of the probe could be greatly improved.
Nevertheless, we were able to detect both
naphthalene and fluoranthene at the ppb level,
even with dispersing the emission in a 0.3 m focal
length monochromator.
DISCUSSION
The purpose of the work reported here was more to
explore what excitation capabilities are available
with a dye laser based system of fairly simple
design than it was to show its superiority over a
fixed frequency excitation system. Only very
preliminary results are available to assess
detection limits, but these are very promising.
The ability to tune the laser wavelength is
important both for sensitivity and for being able
to distinguish different compounds simultaneously
present in the sample. The only possible fixed
frequency contender for direct fluorescence
methods on polycyclic aromatic hydrocarbons is the
266 nm fourth harmonic of Nd:YAG. For benzene,
toluene, xylenes, and some of the other one ring
aromatics, 266 nm lies quite close to an
absorbance maximum and little gain in sensitivity
would result from tuning capability. However, for
other benzene derivatives such as aniline, p-
cresol, or 1,4-dimethoxybenzene, 266 nm lies in a
valley of the absorbance spectrum such that a
sensitivity loss by a factor of up to ten results
from not being able to match the excitation to the
absorbance spectrum.
Probably the greater deficiency of fixed frequency
excitation is that it severely reduces the hope of
speciation, i.e., being able to resolve different
species in the sample. Some speciation would be
possible by time resolving the emission, thereby
exploiting the different fluorescence lifetimes,
but the ability to select the excitation
wavelength is unquestionably more valuable.
96
-------
The greatest drawback of the tunable laser system
is that more than one dye is necessary to cover
the region of interest. If we limit attention to
polycyclic aromatic hydrocarbons and their common
derivatives, then the spectral region of greatest
relevance is between 250 and 300 nm; almost every
aromatic compound has one or more strong
absorbance bands in this region. At a minimum,
two dyes would be necessary to cover this region.
Moreover, with 532 nm pumping the shortest
achievable dye laser wavelength is about 550 nm,
which gives a short wavelength limit of 275 nm
after frequency doubling. This is inadequate to
reach many mono-ring systems (e.g., benzene,
toluene, xylenes). To achieve fundamental
coverage in the A < 550 nm region requires
pumping of another dye, say coumarin 500, with the
355 nm Nd:YAG output.
Although changing dyes in the field is possible,
it is a tedious processs and would severely limit
how many samples could be measured in a given
time. A better alternative is to have separate
oscillator cell/amplifier cell/circulator
combinations for each one of the dyes to be used
so the dyes can be rapidly changed, although any
wavelength selective optics would also have to be
changed when the pumping wavelength is changed
between 532 and 355 nm. Even more flexibility
would be achieved if a separate dye laser were
built for each dye. Clearly this is only feasible
if the dye laser itself is very economical. With
commercial dye lasers costing $20,000 and up, it
would be out of the question. The design we have
followed makes multiple dye lasers feasible. A
instrument with two dye lasers, one for pumping
Rhodamine 590 with 532 nm and one for pumping
Coumarin 500 dye with 355 nm YAG output would give
coverage of the desired 250-300 nm region after
the frequency doubling.
REFERENCES
(1) Richardson, J. H. and Ando, M. E., "Sub-Part-
per-Trillion Detection of Polycyclic Aromatic
Hydrocarbons by Laser Induced Molecular
Fluorescence," Anal. Chem. Vol. 49, No. 7,
1977, pp. 955-959.
(2) Chudyk, W. A., Carrabba, M. M., and Kenny, J.
E., "Remote Detection of Groundwater
Contaminants Using Far-Ultraviolet Laser
Induced Fluorescence," Anal. Chem. Vol. 57,
No. 7, 1985, pp. 1237-1242.
(3) Kenny, J. E., Jarvis, G. B., Chudyk, W. A.,
and Pohlig, K. 0., "Remote Laser-Induced
Fluorescence Monitoring of Groundwater
Contaminants: Prototype Field Instrument,"
Anal. Instrumentation. Vol 16, No. 4, 1987,
pp. 423-445.
(4) Littman, M. G. and Metcalf, H. J.,
"Spectrally Narrow Pulsed Dye Laser Without
Beam Expander," Appl. Opt. Vol. 17, 1978, pp.
2224.
(5) Schwab, S. D. and McCreery, R. L.,
"Versatile, Efficient Raman Sampling with
Fiber Optics," Anal. Chem. Vol. 56, No. 12,
1984, pp. 2199-2204.
CONCLUSION
We have built a relatively simple dye laser which
provides significant power levels of tunable
ultraviolet radiation when pumped with a Nd:YAG
laser. The laser system is amenable to
incorporation into a field instrument and efforts
along these lines are underway. We believe that
detection limits of 1 ppb or better are feasible
for polycyclic aromatic hydrocarbons in
groundwater.
ACKNOWLEDGMENT
This work was supported by a cooperative agreement
between North Dakota State University and the
United States Geological Survey, which provided
the Nd:YAG pump laser. Financial support from the
North Dakota Water Resources Research Institute is
also gratefully acknowledged.
97
-------
RS: ROTATION STAGE
G/TM: GRATING/TUNING MIRROR
BS: BEAMSPLITTER
CL: CYLINDRICAL LENS
ML: MENISCUS LENS
OC: OSCILLATOR CELL
WP: WEDGE PRISM
AC: AMPLIFIER CELL
M: MIRROR
DC: DOUBLING CRYSTAL
IT)
in
ro
^
CM
ro
uo
RS
BS \
BS v
FIGURE 1 - LAYOUT OF DYE LASER
CL ML
CL ML
\
G/TM
OC
WP
AC
M
98
-------
FIELD SCREENING FOR AROMATIC ORGANICS
USING LASER-INDUCED FLUORESCENCE AND FIBER OPTICS
Wayne Chudyk, Kenneth Pohlig,
Nicola Rico, and Gregory Johnson
Civil Engineering Department
Tufts University
Medford, MA 02155
ABSTRACT
The prototype field instrument developed by our
research team uses remote laser-induced
fluorescence (RLIF) to measure aromatic organics
in-situ. Using a laser as light source, and fiber
optics to carry light to the bottom of a well and
back up again, a method for direct measurement of
fluorescent ground water contaminants has been
shown to be a useful field screening method.
Field results using sensor lengths up to twenty
meters are presented.
A typical measurement proceeds as follows. A
foul ing-resistant sensor made of fused silica
glass, PTFE, and stainless steel, of length
corresponding to well depth, is placed in the
water in a well. Sensors are inexpensive enough
to be dedicated to wells for long-term monitoring
applications. After well purging, fluorescence
measurements are taken and compared to previously
characterized standards. Results are expressed as
total aromatics concentration in ppb or ppm.
Usual sampling time, including setup and
decontamination, is on the order of fifteen
minutes per well.
Results to date indicate good correlation between
GC (Method 602) and RLIF; rapid response time; and
good sensitivity, with ppb-ppm results routine.
Analyses to date include primarily BTEX (from
gasoline), and aromatic solvents (phenol,
cresols). Due to the short analysis time,
independence from outside laboratories, and
comparable instrument capital cost, RLIF is shown
to be a cost-effective alternative to GC for
routine screening of aromatics. Where a large
number of wells need to be examined over a short
time, and aromatic organics are the suspected
contaminants of interest, the advantages of RLIF
become even greater.
Conclusions show that the prototype field
screening tests have been successful, indicating
the feasibility of the RLIF method. It is shown
to be applicable to sites where aromatic organics
are known or suspected problem compounds. Since
aromatic organics include the benzene, toluene,
and xylenes fraction of gasoline, as well as over
half of the organics on the EPA Priority
Pollutants List, RLIF is applicable to a variety
of hazardous waste sites. It should be useful for
monitoring and characterizing fuel spills (BTEX),
coal tar sites (phenols and cresols), and aromatic
solvent sites.
INTRODUCTION
In-situ field screening methods have become
attractive because they eliminate sample handling
and resulting changes in sample composition.
Analysis is also performed in real time, avoiding
decision delays resulting from the usual wait for
laboratory reports. Our method uses fluorescence
spectroscopy to measure aromatic organic compounds
in-situ. Fiber optics have been suggested as
useful tools for ground water monitoring, since
they can allow measurements to be made at distance
from the material of interest (1, 2, 3). By using
fiber optics and a laser as the light source,
measurements can be performed at distances of
practical application to monitoring wells.
Prior Work in RLIF Development and Use
Fluorescence analysis has proven to be an
attractive method for screening aromatic
components in water. For example, fluorescence
has been previously used in the laboratory to
identify and quantitate contamination of water by
petroleum products (4), as well as in studies of
the petrographic composition of the organic matrix
in soils and rocks (5). The RLIF application is
in using a similar approach with a portable field
unit.
The present configuration of the field prototype
forces it to be less sensitive than laboratory
instruments, so that RLIF is currently best used
as a screening tool. As advances in instrument
technology develop, we hope to expand the limits
of our method. Extensive testing has proven the
usefulness of RLIF in simulated well setups, and
limited field testing was performed with the first
RLIF field-portable prototype (6, 7, 8). The
second-generation prototype was constructed in
modules that could be easily temporarily mounted
in a mini van or four-wheel-drive vehicle for field
testing. Reduction in size of the RLIF instrument
has allowed easier access to field sites with a
99
-------
smaller vehicle. Further field testing on
existing contaminated sites, in parallel with GC
analysis, has answered questions concerning the
effectiveness of RLIF in measuring such
contamination. The results of this study
demonstrate its short analysis time and potential
for screening large numbers of wells versus
laboratory GC analysis.
FIELD SCREENING WITH RLIF
Field sampling using the second RLIF prototype has
progressed successfully. In the first two
quarters of 1988, sites sampled included twelve
gas stations, two manufacturing companies, and one
chemical company. Site geological characteristics
varied from fractured bedrock through glacial till
to sandy silt. Weather conditions ranged from
wind chill factors of -21°C (-5°F) through
freezing rain to balmy 25°C (77°F)
afternoons. Contaminant characteristics varied,
while most data concern gasoline and petroleum
product spills or leaks. For example, a typical
gas station site contained at least five wells at
varying depths. In field measurements, a series
of sensors ranging in length from five to thirty
meters was used. Field decontamination of the
sensors used detergent, methanol, and distilled
water. Field repair of sensors was also shown to
be practical.
Samples obtained from the same wells at the same
time as our tests were analyzed by an independent
laboratory (Groundwater Technology, Inc., Norwood,
MA).
RESULTS
RLIF response was calibrated using serial
dilutions of gasoline in the laboratory. To
eliminate effects of signal attenuation versus
sensor length, calibration was performed using
sensors of the same length as used in the field.
Figure 1 illustrates such a calibration curve for
gasoline ranging from tens of ppb to hundreds of
ppm using a ten meter sensor. A least-squares
curve fit was used to match the field RLIF
response to the calibration curve, yielding a
corresponding concentration. The laboratory
results from the same field sites were tabulated
and the total concentration of aromatics detected,
corresponding to what RLIF should detect, was
calculated. For each data set, this total
aromatics concentration was compared with RLIF
response for the same well. Figure 2 shows RLIF
response versus gas chromatographic (GC) response
for samples from six of the sites, with RLIF
response usually higher than GC values.
CONCLUSIONS
Field determinations of aromatic ground water
contaminants, such as the BTEX fraction of
gasoline, have been shown to be successful using
RLIF. Good correlations exist between RLIF
results and EPA Method 602 GC determinations. As
a rule, such correspondence between the methods
shows that RLIF is a useful screening method,since
its response is in real time. It is clear that
savings in time and convenience of RLIF over GC
make RLIF attractive as a field screening method.
ACKNOWLEDGEMENTS
The assistance of Groundwater Technology, Inc.,
and Goldberg-Zoino and Associates for providing
access to field sites and data is deeply
appreciated. The authors are thankful for the
support of the National Science Foundation, the
USEPA through the Tufts Center for Environmental
Management, the USGS, and the Alexander Host
Foundation.
REFERENCES
(1) Hirschfeld, T., Deaton, T., Milanovich, F.,
and Klainer, S., "Feasibility of Using Fiber
Optics for Monitoring Groundwater
Contaminants," Optical Engineering. Vol. 22,
No. 5, 1983, pp. 527-531
(2) Hirschfeld, T., Deaton, T., Milanovich, F.,
Klainer, S., and Fitzsimmons, C., "The
Feasibility of Using Fiber Optics for
Monitoring Groundwater Contaminants,"
Project Summary, USEPA Environmental
Monitoring Systems Laboratory, Las Vegas, NV,
April, 1984.
(3) Seitz, W.R., "Chemical Sensors Based on Fiber
Optics," Analytical Chemistry, Vol. 56, No.
1, 1984, pp. 16A-34A.
(4) Eastwood, D., in Wehrey, E.L., Ed., Modern
Fluorescence Spectroscopy, Vol. 4, Plenum
Publishing Corp., New York, 1981, pp. 251-275.
(5) von der Dick, H., and Kalkreuth, W., Advances
in Organic Geochemistry, Vol. 10, 1986, pp.
633-639.
(6) Kenny, J.E., Jarvis, G.B., Chudyk, W.A., and
Pohlig, K.O., "Remote Laser-Induced
Fluorescence Monitoring of Groundwater
Contaminants: Prototype Field Instrument,"
Analytical Instrumentation, Vol. 16, No. 4,
1987, pp. 423-446.
(7) Chudyk, W., Kenny, J., Jarvis, G. and Pohlig,
K., 1987c. "Monitoring of Ground-Water
Contaminants Using Laser Fluorescence and
Fiber Optics," InTech, Vol. 34, No. 5, 1987,
pp. 53-57.
(8) Chudyk, W.A., Carrabba, M.M., and Kenny,
J.E., "Remote Detection of Groundwater
Contaminants Using Far-Ultraviolet
Laser-Fluorescence," Analytical Chemistry,
Vol. 57, 1985, pp. 1237-1242.
100
-------
Co 'brat'on Curve
0.150
CD
O
c
CD
O
Ul
CD
k_
O
0.100-
0.050-
0.030
O
O
O
Distilled Water
0.001 0.10 10.0 1000.0
ppm Un eaded Gaso ine
Figure 1. RLIF Calibration Curve for Unleaded
Gasoline Using a Ten Meter Sensor
101
-------
RL F versus GC Response
total dramatics
—j 1000
cr
-R ioo.o|
oo
"D
0)
13
10.00--
1.000-
0.100-
o
^ 0.010
0
Typical Lab vs. Lab
Range
0.010 0.100 1.000 10.00 100.0
EPA Method 602 in Laboratory
1000
Figure 2. RLIF Response Versus EPA Method 602 for Total Aromatics
102
-------
DISCUSSION
JOE ANDRADE: Could you elaborate on the optical design of the sensor
itself? Your abstract you says that the sensor is designed to be fouling resistant.
Could you elaborate on that?
WAYNE CHUDYK: We used two fibers - one excitation and one collection
fiber. They are positioned in the sensor head so that they overlap - they would
have overlapping exit cones if you were to issue light out of both of them. The
excitation cone overlaps the collection cone from the emission fiber. This is
covered by a patent, so I don't want to go into too much detail.
The fouling resistance relates to the materials which include stainless steel,
silica glass, or Teflon. At some of the gasoline sites, we have a lot of iron
bacteria, that tend to grow on everything. We have not seen them interfere with
our sensors, even though they have been left in the ground for a period of up
to five weeks. We have more work to do in that area, so I didn 't report the data.
The idea, though, is to have something that is not amenable to supporting
biological activity, and we feel we have succeeded.
DELYLE EASTWOOD: The bar diagram you showed gave the fluorescence
of a number of species, some of which were mixtures. Mixtured fluorescence
will vary with where it's excited.
Secondly, you don't make aclear distinction between GC and GC/MS. I believe
you were actually talking about method 624, which is a GC/MS method, and
you should have indicated that.
Finally, you don't make a clear difference between humic and fulvic. The fulvic
is actually the soluble component. I have done work with humic and fulvic
acids, and they can fluoresce over a wider range, depending on what the source
is. They will fluoresce a little bit almost any place you excite them.
Do you intend to look at other fluorescing species besides gasolines and in
complex mixtures? Are you looking at UV fibers, in terms of being able to see
fluorescence under 300 nanometers?
WAYNE CHUDYK: The fibers that we have now are ultraviolet transmitting
fibers. We tend to cut off around 300 nanometers, because we have scatter from
the laser excitation (at 266). We also pick up a water Raman line around 293,
295. Those things interfere with what we're able to see. In general, we look
from 300 to 500 nanometers.
Regarding looking at other components, we are experimenting with different
filter combinations and also different monochromator types of approaches. The
biggest problem with have seen with monochromators relates to field reliabil-
ity. A monochromator in the back of a van, doesn't work after a while.
We have not been successful in taking monochromators out. So that has
restricted our use to the glass cut off filters.
JOHN KOUTS ANDREAS: Could you address the problem of not being able
to drive that station wagon into place? How portable is the equipment?
If the sensors are left in place over a period of six months, what kind of
degradation do you have?
And finally, what kind of eventual cost do you see for commercialization of this
unit?
WAYNE CHUDYK: In terms of portability, we have the system in modules,
each weighing a maximum of 40-50 pounds. There are two of them that are that
big, and two of them are much smaller.
We usually operate out of the back of the van, because of weather conditions
in New England. However, we have successfully carried the modules a few
hundred yards into a place that is not van accessible.
In lerms of making the instrument smaller, the instrument that we have now can
be carried in pieces, and can operate any place you can get a generator. The next
prototype we expect to be even smaller, and even more portable.
All of the components that we have now in the prototype are designed to operate
off 12 volts, and that's a step we expect to take in the future as well. So we'll
free ourselves from a generator.
The second question was the lifetime of the sensors. We have only been able
to leave sensors in the ground for a few weeks at a time. We would like to leave
them in the ground much longer. We don't expect problems, with the exception
of a couple of wells with very heavy bacterial growth. Those monitoring wells
had inert casings, and the iron bacteria were even sliming up those. Our sensors
were not in the ground long enough for us to accurately evaluate that.
I would expect in a "normal" situation, where you did not have that heavy type
of growth, that these sensors should stay relatively clean, and usable.
One of the things built into the system is the ability to measure power, both
before the laser light goes down and after it comes back. So we have the ability
to check the integrity of a sensor, even a sensor that's been left in place. The first
thing that's done on a site with a sensor that's been left in place is to check if
the power is going down by approximating what the power should be.
We have that ability to see if the sensor is okay. Did somebody pull it out and
snap a fiber, and shove the thing back in the ground, or is there something funny
happening to the end? We can check that relatively rapidly.
The instruments cost approximately $30,000 to construct, one at a time. The
single most important piece is the laser, which costs around $12,000 in single
quantities. They tell us there would be quantity discounts, but we haven't ap-
proached lots of 100 or whatever. I would expect the final cost to drop to the
twenties, if not lower, if this were to be commercialized.
GREG GILLISPE: When you were illustrating the dynamic range for
gasoline, changing concentration over a couple orders of magnitude, you were
only getting a change in signal of two, three, five, or something like that. Can
you explain why you're not getting an order of magnitude change in signal for
an order of magnitude change in concentration?
WAYNE CHUDYK: The way the signal is massaged after it comes out of the
PMT's, in terms of the conversion from current to voltage, gives us a number
that we could save. We could integrate over longer periods, and that will give
us a larger number.
GREG GILLISPE: What I am referring to is that an analytical criterion for a
method is that the signal be linear in concentration, and yours seems to deviate
very substantially from that. Whereas in the Richardson-Andall paper, for
example, their linearity went well down to the parts per trillion. So I am troubled
by the deviation.
WAYNE CHUDYK: So are we.
MIKE CARRABBA: You presented a slide that showed the GC/MS data
versus your data, and it is a little troublesome that there were some places where
there were maybe two or three orders of magnitude difference between your
data and the GC/MS. I'm quite confused by that. I could see an order of
magnitude, but three orders of magnitude is quite a large difference. Can you
explain?
WAYNE CHUDYK: Frankly, in some cases, we resampled and found errors.
There are, as you well know, many cases where things can slip between the field
and the lab.
The differences that concern us the most are where the numbers that we see
don't match the numbers that they see, and are off by at least an order of
magnitude.
I am not sure if that is a potential interference. In other words, are we seeing
something that they are not seeing, which is the most likely case, or is there
something that we're overlooking? We're still trying to figure some of that out.
That's the nature of research.
103
-------
SECOND-DERIVATIVE ULTRAVIOLET ABOSRPTION
MONITORING OF AROMATIC CONTAMINANTS IN GROUNDUATERS
J. W. Haas, III E. Y. Lee
C. L. Thomas R. B. Gammage
Health and Safety Research Division
Oak Ridge National Laboratory
P 0 . Box 2008
Oak Ridge, Tennessee 37831-6113
ABSTRACT
A po r t a
u I t r a v i
been ev
c o n t a m i
d e r i v a t
were f o
t e c h n i q
on t h e s
rugged,
down-we
begun.
success
When th
mon i tor
h a z a r do
b L e spect rom
olet absorpt
[ua ted for
nants in gro
i v e s , and po
und to be am
ue at concen
e encouragin
underwater
LI pollutant
A prototype
fully to obt
e new probe
benzene in
us waste sit
ete
i o n
the
und
Lye
en a
t ra
9 r
f ib
mo
of
a i n
i s
a g
e ,
r for
s pec t
scree
waters
y c I i c
b I e to
t i o n s
e s u I t s
e r opt
n i t o r i
the p
s pe c t
c ompIe
r o und w
s e c o
r ome
n i ng
B
a r om
ana
down
, c o
i c p
n g u
robe
r a i
ted
a t e r
nd-derivative
try (DUVAS) has
of aromatic
enzene, its
atic hydrocarbons
lysis by this
to 1 ug/mL. Based
nstruction of a
robe for direct,
sing DUVAS has
was used
n the laboratory.
it will be used to
well at a
INTRODUCT ION
Surface and subsurface water contamination is a
growing problem in the U. S. which can be
attributed primarily to the indiscriminate dumping
of hazardous chemicals into the environment. The
severity and widespread nature of this problem
require that polluted waters and their sources be
Located rapidly and cost-effectively. Gas
chromatography with mass spectrometric detection
has been a useful tool for identifying polluted
sites and the hazardous chemicals in them. This
method is costly, however, and other less
expensive, yet reliable, analytical techniques
should be considered when many samples are to be
tested for contamination.
Second-derivative ultr
spectroscopy (DUVAS) i
analytical technique w
screening water sample
analysis takes about o
wavelength can be moni
instrumental "dwell" m
is required prior to a
can be analyzed direct
with the relatively lo
render DUVAS a cost-ef
method is particularly
some of the most commo
hydrocarbons (includin
and polycyclic aromati
low as 1 ug/mL. Furth
aviolet absorption
s a well-established
ith distinct advantages for
s. Complete spectral
ne minute and a single
tored every few seconds in an
ode. No sample preparation
nalysis; even turbid samples
ly. These features, coupled
w expense of a spectrometer,
fective screening tool. The
well-suited for identifying
n pollutants, aromatic
g benzene, its derivatives,
c hydrocarbons), at levels as
ermore, the portability of
the DUVAS instrument allows for on-site screening
and monitoring of migrating pollutants.
The scope of recent research has been to determine
which priority pollutants are amenable to analysis
by DUVAS and to develop a new version of the
instrument for field analysis. Incorporation of
fiber optics into the new spectrometer was
undertaken to allow for direct analysis of
subsurface well waters.
POLLUTANTS AMENABLE TO DUVAS ANALYSIS
A review of the literature indicated tha
toluene, ethylbenzene, xylenes, chlorobe
phenol, and naphthalene have been the ar
contaminants detected in groundwater mos
frequently. Therefore, our DUVAS spectr
which was comprised primarily of polycyc
aromatic hydrocarbons found in synthetic
expanded to include these compounds. Sp
collected from 200-350 nm for the pollut
spectrum was unique, allowing individual
to be identified in the presence of the
species tested. Even the three xylene i
easily differentiated using DUVAS. A mi
detectable concentration was also determ
each pollutant using its predominant spe
(Table I). The results were similar fo
species, being about 1 ug/mL.
Table I
Minimum Detectable Concentrations
of Common Aromatic Pollutants
compound
benzene
toluene
ethylbenzene
o - x y I e n e
m - x y I e n e
p - x y I e n e
chlorobenzene
o-dich Lorobenzene
phenol
naphthalene
Uavelength of the determination
Minimum detectable concentration (S/N 2)
t benzene,
n z e n e s ,
o m a t i c
t
a I library,
I i c
fuels, was
ectra were
ants. Each
compounds
other
somers were
n i m u m
i n e d for
ctral peak
rail
u g / m L
254
268
268
270
272
274
271
277
277
31 1
2 .
1 .
2.
3 ,
1 ,
1 ,
4
3
2
0
, 4
.9
. 0
.8
.9
. 0
.5
.3
. 2
. 5
105
-------
Linear dynamic ranges for the compounds studied
were also nearly uniform, spanning over a 100-fold
concentration range.
The ability to detect benzene and toluene has a
useful application in monitoring fuel spills, a
frequent groundwater contamination problem. Both
benzene and toluene are major components of
gasolines and other fuels and can thus be used as
markers for these pollutants. Figure 1 shows
benzene and toluene spectra, and a spectrum for a
water sample contaminated with J P - 4 jet fuel.
Benzene and toluene were easily identified without
interference from other species in the sample.
Their presence was confirmed by gas chromatographic
analysis (photo ion izat ion detection) which also
showed that these two compounds were the
predominant aromatic hydrocarbons in both the water
sample and the jet fuel.
A shal I
into wh
been mo
was d om
30 u g/m
detecti
c ompa r a
o b t a i n e
have re
sample.
nume r o u
solvent
at c omp
offered
ow g roundwate
ich organic u
nitored with
inated by ben
L. Gas chrom
on confirmed
bIe I eve I (40
d with the ga
suited from i
The chromat
s overlapping
s and other a
arable levels
no interfere
r well at the base of a hill
aste was once poured has also
DUVAS. The spectrum (Figure 2 )
zene which was quantitated at
atographic analysis with FID
the presence of benzene at a
ug/ml). The higher value
s chromatographic method may
nterfering compounds in the
ogram was complex, with
peaks from chlorinated
liphatic hydrocarbons present
in the sample. These species
nee to the DUVAS analysis.
FIBER OPT I C DUVAS
A disadvantage of most method
groundwaters for pollutants i
samples and transport them to
Contamination and degradation
important concerns which our
reduced by allowing DUVAS ana
the field. However, samples
from wells, which maintained
contamination, required fully
the instrument, and precluded
monitoring. The solution to
considered to be the developm
fiber optic probe for the spe
configurations were considere
candidates for down-well moni
shown in Figure 3. In the fi
configuration, light from a m
passes to a transmission prob
and returns through a second
(D) located in the spectromet
design, the detector (a small
is integrated into the probe
signal is carried back to the
processing (ELECT). The latt
considered because of the poo
ultraviolet light through opt
twice the working distance of
s used to screen
s the need to collect
the laboratory.
of s amp Ie s are
portable spectrometer
lyses to be made in
still had to be drawn
the possibility for
attended operation of
continuous well
these difficulties was
ent of a down-well,
ctrometer Two probe
d as likely
tor ing and they are
rst probe
onochromator (MONO)
e through one fiber
fiber to a detector
er In t he second
, solid state device)
and the electrical
spectrometer for
er configuration was
r transmission of
i c a I fibers; it gave
the first design.
cuvette holder, and a lens. The lens focused
light emerging from the end of the fiber, through
the cuvette to the detector. The other end of the
fiber (600 urn, plastic clad hydroxylated silica)
was coupled through two lenses (collimating and
focusing) to the output of the monochromator A
150 U xenon lamp was used as the source of
ultraviolet light.
Figure 4 is a spectrum of 25 ug/mL phenol obtained
with the prototype instrument equipped with a 5 m
optical fiber The spectrum was the same as one
collected with the spectrometer prior to its
modification. Calibration curves were also
determined for phenol using 1, 5, and 44 m fibers.
As shown in Figure 5, the 1 and 5 m fibers gave
curves which varied little from the one obtained
without an optical fiber Ultraviolet light
transmission was attenuated greatly in the 44 m
fiber, however, resulting in a calibration curve
with a much smaller slope. Despite the
limitations for quantitative analysis posed by the
small slope, phenol was still detected down to
about the same minimum detectable concentration
with the long fiber as was attained using the
shorter fibers.
CONCLUSIONS
The results for phe
contaminant monitor
optic probe was fea
other aromatic poll
probe were also com
our library. On th
construction of a r
begun . The new pro
silica optical f i b e
transmission (up to
tested at the groun
benzene. The well
extended period of
benzene concentrati
polluted plume, mic
factors.
ACKNOWLEDGEMENT
Research sponsored by Division of Facility & Site
Decomissioning Projects, U.S. Department of Energy,
under contract DE-AC05-840R21400 with Martin
Marietta Energy Sy.stems, Inc.
nol demonstrated that down-well
ing using DUVAS with a fiber
isible. Spectra obtained for
utants using the prototype
parable to standard spectra in
e basis of these results,
ugged underwater probe has
be will incorporate an a I I -
r for improved ultraviolet
100 m is expected) and will be
dwater well containing 30 ug/mL
will be monitored over an
time to follow changes in the
on caused by movement of the
robial degradation, or other
A laboratory prototype of the fiber optic probe was
constructed according to the second configuration.
A light-tight box represented the probe housing and
contained a silicon photodiode detector, a sample
106
-------
HI
10
o
a.
BENZENE
TOLUENE
JP-4 JET FUEL
\
230 260
WAVELENGTH
o
a.
w
ui
cc
220
240 260
WAVELENGTH (nm)
280
Figure 2 - Groundwater Contaminated with Benzene
107
-------
Figure 3A - Dual Fiber Optrode for DUVAS
PROBE D
n
Figure 3B - Single Fiber Optrode for DUVAS
UJ
V)
z
o
Q.
V)
111
a.
235
265
295
WAVELENGTH (nm)
Figure 4 - Phenol Spectrum Using a 5 Meter Fiber
108
-------
1.2 —i
0.8 -
ill
w
O
0.
w
Ul
cc
0.4 -
0,1,6 meters
44 meters
I
20
I
40
I
60
CONCENTRATION (ug/mL)
Figure 5 - Phenol Calibration Curves Using
Different Lengths of Optical Fiber
DISCUSSION
DELYLE EASTWOOD: Why are you doing this on UV absorption, instead
of fluorescence? Fluorescence can also be used in the derivative mode.
O'Hagar did it, and Purcell did it, actually separating about 14 pinholes by
derivative fluorescence spectroscopy. Fluorescence is at least potentially much
more sensitive, and I believe all the compounds that you showed actually do
fluoresce.
JOHN HAAS: Benzene doesn't fluoresce very well, but the other ones do
better.
DELYLE EASTWOOD: It is a low yield, but it still fluoresces if you use a
good control.
JOHN HAAS: There are a couple of advantages to this technique. First, we had
the instrument sitting in the laboratory, and second, we were asked to monitor
this well for benzene. It's ideally suited for that. Also, the insensitivity to
turbidity in the sample will allow us to go in to muddy situations particularly
at this plant. This might be much more of a problem with fluorescence.
JOHN SCALERA: Do you have any problems with critical angle on your UV
construction, your fiber optic leads? In other words, do you keep them fairly
straight and parallel in all your designs, and monitoring?
JOHN HAAS: We tried curling the fibers around the room with the 44 meter
one. So long as it's fixed on the two ends, going into the detacher, and also
coupling to the output of the monochromator, we didn't see any difference in
the spectra. The end going into the detector isn't so crucial, because we use a
pretty large size. It's one centimeter across, so we can focus down, and we have
a little bit to spare there.
JOHN SCALERA: You didn't lose any intensity with the bending, using the
ultraviolet frequencies?
JOHN HAAS: You could see minor shifts, but when you take the derivative,
essentially, that washed out. So long as we can have 100 to 200 milliwatts of
intensity of light going through, then the derivative still ends up looking the
same.
MAHMOUD SHAHRIARI: Do you see the potential for using a UV
absorption technique in an evanescence mode, instead of the direct-gap
technique that you just mentioned?
JOHN HAAS: It's possible, but I think the porous approach is better, because
the intensity of the light getting out through the evanescence is not as great. We
don't have to worry about losses out the end of the fiber, or from coupling. We
are using it more as a light medium, so we get the full intensity that we can
possibly get through the fiber, and with UV, that's a major consideration.
MAHMOUD SHAHRIARI: Can you commenton what type of fibers are you
using for transmitting UV?
JOHN HAAS: We have actually used plastic clad fiber so far, and we look for
much better results when we go to the all silica fibers. Also, I will test some
liquid core fibers. But right now, we were very pleased to get anything through
these fibers, from what we had reported. These are 600 micron cores from
general fiber.
109
-------
HAZARDOUS WASTE ANALYSIS BY RAMAN SPECTROSCOPY
Charles K. Mann and Thomas J. Vickers
Department of Chemistry, Florida State University
Tallahassee FL 32306-3006
ABSTRACT
The application of ultraviolet resonance
enhanced Raman spectroscopy to chemical
analysis of low concentrations of organ-
ics in complex samples is described. The
physical observation consists of illumin-
ating a sample by a beam of ultraviolet
radiation and analyzing the light that is
scattered by the sample. The phenomenon
is applicable to solids, liquids, and
gases; however in this work, analyses are
limited to samples in the liquid state.
Two modes of operation are discussed. In
one, the analysis is applied to the efflu-
ent from a high performance liquid chrom-
atograph (HPLC). The sample is fraction-
ated completely or in part by the chrom-
atograph. Individual components are
identified and quantitated by means of
the spectroscopic signal. In the second
mode, the analysis is carried out by
direct Raman measurement upon the sample.
Discrimination of target substances from
the bulk is accomplished by using reso-
nance enhancement which allows the
response of specific classes of compounds
to be enhanced as compared with the bulk
of the sample.
To achieve limits of detection which can
be used in trace analysis, it is neces-
sary to perform the measurements with
ultraviolet (UV) exciting radiation. At
this time, lasers which produce adequate
power in the UV operate in the pulsed
mode. Although average power is not
especially high, the energy is delivered
in short bursts, producing a high photon
flux which can bleach the sample. To
avoid this, it is necessary to provide an
unconventional coupling between the radia-
tion and the sample. The apparatus is
discussed. The fact that absorption
peaks in the UV are not necessarily wide
as compared with Raman shifts leands in
some cases to a situation in which the
use of an internal standard does not
adequately correct for the effect of
sample absorbance on exciting radiation
power. The cause of this effect is
discussed and the necessary corrective
steps are outlined.
INTRODUCTION
The methods which currently are most used
for determination of organic compounds in
hazardous wastes, gas chromatography
(GC), HPLC, mass psectroscopy (MS), gas
chromatography-mass spectroscopy in
combination (GC/MS), and GC/FTIR, each
offer certain advantages and are con-
strained by certain limitations. GC/FTIR
and GC/MS offer very good selectivity
and, especially for GC/MS, excellent
sensitivity. They are contrained by the
necessity to have the sample in the vapor
state during the separation. HPLC pro-
vides excellent separating power and is
applicable to nonvolatile and thermally
labile compounds. However, the detectors
which are used with it do not provide a
very satisfactory combination of sensiti-
vity with generality of application.
Raman spectroscopy also offers certain
advantages and is limited by certain con-
straints. In the context of hazardous
waste management, some of the advantages
offered by Raman measurements permit the
handling of samples which cause problems
with other methods. The basic physical
principles upon which it is based are
firmly established. Raman scattering has
been applied to chemical analysis in
solid and liquid samples. It is parti-
cularly applicable to samples that are
dissolved or slurried in water The
actual measurement is carried out by
directing a beam of radiation onto the
surface of the sample. Measurement is
made by capturing and analyzing that
radiation which is scattered from either
the surface or the interior of the
sample. The technique is applicable to
nearly all of the substances on the
E.P.A. Priority Pollutants List, includ-
ing nonvolatiles.
Analytical uses of Raman are based upon
the high information content of its
111
-------
signals. Raman signals are one form of
molecular vibrational spectra, IR spectra
are another It has been recognized for
decades that vibrational spectra are char-
acteristic properties of molecular struc-
tures , very much as finger prints are
characteristic of humans. It is this
very high degree of specificity in type
of signal produced which serves as the
basis for applications to chemical analy-
sis of mixtures. Raman differs from IR
in two important aspects. It is well
suited to application to condensed phase
samples and it can provide high enough
sensitivity through resonance enhancement
to be useful in trace analysis.
Raman measurements can be made directly
on solids without regard for the sample
thickness. IR measurements on solids,
whether idspersed in a transparent matrix
or directly by diffuse reflectance, do
not produce the high accuracy and repro-
ducibility which is required in order to
do mixture analyses. Raman measurements
are well suited to accurate measurements
on liquid samples. Although chemists
have traditionally used IR for liquid
samples, attempts to achieve high accur-
acy fail because of the impact of refrac-
tive index changes on shapes and inten-
sities of absorbance bands. Whether the
measurement is made in a cell fitted with
two windows or with one having only one
fixed boundary, an attenuated total
reflectance cell (ATR), the details of
the shapes of peaks are appreciably
affected by refractive index changes.
(1,2) These effects are large enough to
interfere with attempts to make use of
subtle differences in band position and
shape in mixture analyses. This is im-
portant because it is necessary to have
linear performance in order to do general-
ized mixture analyses. Accordingly,
Raman measurements are appropriate to the
analysis of effluents from high perform-
ance liquid chromatography (HPLC) columns
and to direct determination of organic
compounds in natural samples.
The fundamental sensitivity of resonance
enhanced Raman spectroscopy has been
demonstrated (3-7) to be sufficiently
high to support analyses at the part per
billion level without the necessity for
concentration. Resonance enhancement is
an effect that occurs when the wavelength
of the incoming radiation corresponds to
an absorption band of a component in the
sample. The signal produced by that
component can be enhanced, by as much as
a million-fold, compared with its unen-
hanced signal. This provides a powerful
method for discriminating the enhanced
compound from other sample concomitants
which are not enhanced. The limits of
detection of unenhanced Raman measure-
ments are comparable to those of IR
measurements, usually between 0.05 and
0.10 percent.
If it is desired to carry out a sample
preparation step, such as extraction of
animal tissue, the extract would ordinari-
ly furnish a suitable sample. Successful
application to complex samples depends
upon the ability to elicit very intense
and very characteristic responses of
narrowly targeted compounds, or classes
of compounds, in the presence of much
larger concentrations of sample concomi-
tants. This is a very special capability
of Raman spectroscopy which occurs when
resonance enhancement is used.
Most compounds which cause concern owing
to pollution problems are colorless,
absorbing only in the ultraviolet. Ac-
cordingly, it is necessary to use UV
radiation. At this time UV laser radia-
tion is available either from visible
sources from which UV radiation is ob-
tained by frequency doubling or tripling,
or from light generated in the UV by an
excimer laser. We have been using an in-
jection-locked excimer laser as a source.
HPLC APPLICATIONS
The technology of HPLC is of course very
well developed. This research is con-
cerned with coupling the chromatographic
column to bring the optical signal to the
spectrometer and with data treatment
required to analyze the results. In
principle, a Raman spectrometer is cou-
pled to a chromatograph by directing the
light beam onto a transparent exit sec-
tion of the column. In doing this, allow-
ance must be made for the high photon
flux produced by. pulsed lasers, the need
to minimize detector dead volume, and for
the necessity to achieve high efficiency
in collecting the scattered radiation.
These questions must be considered
together in the design of an HPLC
detector cell.
HPLC Cell Design
In order to achieve adequate average
power to give useful sensitivities, the
laser must be operated at high pulse
energies. The resulting photon flux
density may be sifficient to bleach solu-
tions which contain absorbing samples.
This does not usually involve any perma-
nent chemical change in the system, sim-
ply a depletion of the concentration of
the ground state species which causes
nonlinearities in the relationship be-
tween concentration and signal intensity
It does not cause nonlinearities that
affect peak shapes. Two measures are
being taken to control sample exposure.
First, the laser intensity is monitored
and the output of the monitoring trans-
112
-------
ducer is used to control a beam attenua-
tor which determines the photon flux at
the sample.
The second measure involves coupling the
laser beam to the chromatographic efflu-
ent in a way that avoids focusing the
beam sharply in order to reduce the pho-
ton flux density. Several factors must
be considered. Performance of an HPLC
detector is critically affected by its
dead volume. Light coupling must there-
fore be arranged to produce a minimum
increase. There are two related factors
which must be considered. Conventional
spectroscopic optics do not efficiently
collect light produced from a diffuse
source. When light is collected and
taken to the spectrometer, the optical
constraints encountered in taking it into
the instrument are quite severe.
Accordingly, simply defocusing the beam
to reduce photon flux density gives unsat-
isfactory results owing to reduced collec-
tion efficiency. We reduce flux density
by introducing the light beam to the
solution through a bundle of silica opti-
cal fibers. Available fibers are approxi-
mately 100 micrometers in diameter.
Accordingly, a large number can be used
without drastic increase in the size of
the chromatographic outlet tube. Thus,
instead of focusing the entire beam at a
point in the sample, the beam power is
first attenuated according to the number
of fibers used. Each of these is brought
individually to the sample, achieving the
desired reduction in local intensity.
This arrangement permits efficient collec-
tion of the scattered radiation, since
the scatter produced by light emitted by
each fiber can be taken back through the
same fiber. This gives fl/ collection
efficiency. At the spectrometer, the
beam must be imaged on the entrance slit
in order to make it fall on the grating
where it can be used. This is accom-
plished by designing the fiber bundle to
match the slit cross section.
The use of fiber optic coupling introdu-
ces two additional signal components: the
Raman signals produced by silica and the
effects of solarization. The Raman spec-
trum of silica can be determined very ac-
curately and corrected for by the usual
methods that are used for removal of in-
terferences, e.g., by including it as one
of the references in a least squares fit.
Solarization is a gradual degradation of
the optical properties of silica which is
exposed to UV radiation which ultimately
makes it necessary to replace the fibers.
An internal standard corrects for the
effect in quantitative measurements. When
light is introduced into a sample for a
Raman measurement, the effective inten-
sity is affected by sample absorption
of both the incoming and scattered radia-
tion. Experimental variations are com-
pensated in emission spectroscopy by
using an internal standard. However, in
resonance enhanced operation, not all
effects of absorbance by either the ana-
lyte or sample concomitants are always
compensated. An internal standard is en-
tirely effective if the sample absorbance
is the same for the radiation scattered
by both the analyte and the internal
standard. The situation is illustrated
in Figure 1. Experimental error is plot-
ted against absorbance at the analyte
wavelength for values of analyte/standard
absorbance, X. When the ratio is unity,
there is no error. Positive or negative
error occurs when the ratio deviates from
unity.
A general purpose instrument operating in
the ultraviolet must be expected to en-
counter this situation because the absorp-
tion bands which occur are fairly sharp.
It is therefore necessary to provide a
compensating mechanism other than the
conventional internal standard. Two
factors are important: the solution absorb-
ances at the wavelengths scattered by
the analyte and internal standard and the
transfer function of the sample collec-
tion optics. We have demonstrated earlier
(8) that the sample absorbance can be cal-
culated from examination of the variation
in intensity of scattered light. Basic-
ally this is a matter of alternating
measurements with the laser and with a
continuum source and can be done under
computer control by swinging a mirror
into the laser beam to block it and admit
light from the other source. The trans-
fer function of the fiber optic collec-
tion system is set by the entrance cone
to the fiber, which is known, and the
sample absorbance, which is being deter-
mined. Accordingly all necessary vari-
ables are known.
DATA PROCESSING
If the chromatograph effects a complete
fractionation of components in the sam-
ple, identification and quantitation of
them by Raman spectroscopy is straight
forward. A reference spectrum is needed
and the instrument must have been cali-
brated. If fractionation is incomplete,
data processing often can be treated as a
problem in multicomponent analysis. (9)
The effluent from a chromatograph consti-
tutes a favorable medium for multicompo-
nent analysis and Raman measurements are
especially well suited to this applica-
tion.
When fractionation is incomplete and when
the information that is required to treat
the problem in terms of multicomponent
113
-------
analysis is unavailable, it is sometimes
possible to extract additional informa-
tion by means of factor analysis and
related chemometric techniques. Starting
with the pioneering work of Lawton and
Sylvestre (10), a considerable effory has
been made in this direction. (11-13)
For those situations in which no informa-
tion about the analyte is available,
unequivocal results are obtained only for
two component mixtures. However, when
some information is available, it is
possible to include it in the basis for
the calculations. The partial least
squares technique, which has been de-
scribed in detail by Haaland (14) , can be
applied to this type of problem. This
will involve furnishing calibration sets
which form the basis of the calculation.
DIRECT ANALYSIS
The background now exists to perform
a range of analyses on complex samples
without the necessity for prior separa-
tions. This basically involves making
use of spectroscopic and data processing
steps to cause the signals of specific
target compounds to be discernible
against the background produced by all
other components in the system. Its
success depends upon producing sensitive
signals from the target compounds and in
devising methods for discriminating them
from other signals. The benefits from
this operation are ability to perform the
analysis without altering the sample and
very rapid turn-around time.
At this stage we use samples in a liquid
form, either as solutions, extracts, or
as slurries of solid samples in a suit-
able solvent. We assume that analyses
are targeted at specific compounds for
which reference spectra are available.
It is assumed that, for a given class of
analysis, typical samples would be avail-
able upon which to base an analytical
design. These do not have to constitute
a training set, just some typical exam-
ples. This design would primarily in-
volve selection of conditions which would
be incorporated into custom software
which would in turn control the actual
analysis. Analyses would be targeted at
components in the range of parts per
million to parts per billion. Higher
concentrations could be handled by dilu-
tion.
The design of a sampling device for di-
rect analysis is quite similar to design
of an HPLC detector that is described
above. The major difference is that for
direct analysis, it is not usually neces-
sary to be greatly concerned about the
volume of the detector.
In contrast to the application in chroma-
tography, direct analysis depends upon
the spectroscopic detector for analytical
selectivity. The bulk of a sample such
as plant or animal tissue, soil or ground
water consists of materials which do not
absorb light and will therefore not show
enhancement. The signals produced by
these sample constituents consist of a
large number of individual components
which combine to produce a largely fea-
tureless background on which the reso-
nance enhanced signals are detected as
discrete peaks. A successful analysis
depends upon achieving a sufficient de-
gree of enhancement to make the peaks
stand out. It also depends upon being
able to distinguish target compound sig-
nals from those of interfering compounds
which happen also to show enhancement.
In general distinct functional groups
produce spectra which are different
enought to be measured separately. If a
target substance contains the same func-
tional group as a background component,
they will be detected together.
Operating in the direct analysis mode, a
determination takes from one to five
minutes, including time for data process-
ing. The actual process is largely done
under computer control, with the most
critical judgments taken by the person
responsible for configuring the soft
REFERENCES
(1) Funiyama, T., Herrin, John and Craw-
ford, B.L., Jr., "Some Systematic
Errors in Infrared Absorption Spec-
trophotometry of Liquid Samples,"
Applied Spectrosc. Vol. 24, 1970,
(2) Ingle, J.D. and Crouch, S.R., "Infra-
red Spectrometry, " Spectrochemical
Analysis , Prentice-Hall , Englewood
Cliffs, M.J. , 1988, pp. 432-434.
(3) Asher, S.A., "Ultraviolet Resonance
Raman Spectrometry for Detection and
Speciation of Trace Polycycllc Aro-
matic Hydrocarbons," Anal . Chem. Vol.
56, 1984, p. 720.
(4) Mann, C.K., Vickers , T.J., Mar ley,
N.A., and Ling, Y.-C., "Quantitative
Analysis of Low Concentrations of
Organic Pollutants by Raman Scatter-
ing," Adv. Instrum. Vol. 38, 1983,
p. 167.
(5) Marley, N.A., Mann, C.K., and Vick-
ers, T.J., "Determination of Phenols
in Water Using Raman Spectroscopy , "
Appl. Spectrosc. Vol. 38, 1984, p.
540T
114
-------
(6) Vickers, T.J., Mann, C.K., Marley,
N.A. and King, T.H., "Raman Spectros-
copy for Quantitative Multicomponent
Analysis," Amer . Lab . Vol. 16, No.
10, 1984, ppnS-WT
(7) Marley, N.A., Mann, C.K. and Vickers,
T.J., "Raman Spectroscopy in Trace
Analysis for Phenols in Water," Appl .
Spectrosc. 39. 1985, p. 628.
(8)
(9)
Englebreth, W.R. , Mann, C.K. and
Vickers, T.J., "Diode Array Spectro
photometry of Translucent Materials,
Appl. Spectrosc. Vol. 40, 1986,
pp. 1136-1141.
Tyson, L.L., Ling, Y.-C. and Mann,
C.K., "Simultaneous Multicomponent
Quantitative Analysis by Infrared
Absorption Spectroscopy," Appl .
Spectrosc. Vol. 38, 1984. pp. 663-
(10) Lawton, W.H. and Sylvestre, E.A.,
"Self Modelling Curve Resolution,"
Technometrics Vol. 13, 1971, p. 617.
(11) Ohta, N., "Estimating Absorption
Bands of Component Dyes by Means of
Principal Component Analysis," Anal.
Chem. Vol. 45, 1973, pp. 553-557.
(12) Gemperline, P.J., "A Priori Esti-
mates of the Elution Profiles of the
Pure Components in Overlapped Liquid
Chromatography Peaks Using Target
Factor Analysis," J. Chem. Inf.
Comp. Sci. Vol. 24, 19~M7 pp. 206-
TCZ:
(13) Sasaki, K., Kawata, S. and Minami,
S., "Optimal Wavelength Selection
for Quantitative Analysis," Appl.
Spectrosc. Vol. 40, 1986, pp. 185-
T91T
(14) Haaland, D.M. and Thomas, E.V., "Par-
tial Least-Squares Methods for Spec-
tral Analyses. Relation to Other
Quantitative Calibration Methods and
the Extraction of Qualitative Infor-
mation," Anal. Chem. Vol. 60, 1988,
pp. 1193-T2UI.
CT
O
o:
C£
LJ
LJ
O
(Y
LJ
CL
20
10
0
10
20
1 .5
= 0.5
0 0.5 1.0 1.5
ABSORBANCE
2.0
Figure 1. Effect Of Analyte/lnterna
Standard Absorbance Ratio, X
115
-------
DISCUSSION
GREG GILLISPE: How much power would you need in the UV to implement
the tunability? How much would you want for it to be a viable alternative to your
Raman shifter?
CHARLES MANN: I think what you were describing earlier was adequate.
That's why during the course of the morning, I have altered my view of what
may be necessary to do resonance-enhanced Raman. I was prepared when I
came here today to say that we will have to use an eximer laser, and a Raman
shifter, but after listening to what you had to say, I'm not quite so sure of that
now.
It seems to me that we might very well be thinking about the possibility of using
a dye laser if what you say about it being easy to change from one dye to another,
andits compactness is true. The diode-tunable YAG laser is a reality that sounds
like a very interesting possibility.
GREG GILLISPE: When you do have tunability, how much of an increase in
the resonance enhancement can that give you?
CHARLES MANN: Very big. The best results are if the excitation is right on
the top of the resonance bin, and then tapers off quite rapidly. There are some
other aspects of this that are very interesting. If you have real tunability, then you
can begin to use the different spectra produced as a result of doing your analysis
twice - once at one wave length, and once at another.
Tunability will allow you to discriminate not only against those compounds
which do not resonance enhance, but will also give us another handle on an
attempt to get interferences out of these mixtures that the environmentalists
keep asking us to analyze.
NELSON HERRON: I just encountered small portable nitrogen lasers, with
dye cells that are about $5,000. You just pull the cubette out, and they have a
battery pack for them. I think Laser Science is the name of the company. You
may know of it and want to comment on it.
GREG GILLISPE: It's a nice toy. It's a very nice laser, and you can get
tunability with it, but the power levels are pathetically inadequate for these
purposes. I believe the nitrogen pump laser, at 337, is in the vicinity of a few
hundred microjolts. The dye laser gives you 20 microjolts when you double
frequency. Perhaps you can double, but your power levels just become abys-
mally small. I don't think it can be applied in this sort of study.
JONATHAN KENNY: You can double those lasers, and get three nanosecond
shots. So it's really a very low powered device.
About your Raman shifter, you said you got 15 angstrom intervals. What kind
of fill gas were you using?
CHARLES MANN: I was using hydrogen, and deuterium, and methane.
JONATHAN KENNY: A mixture of all three?
CHARLES MANN: No, I don't think so. I was under the impression we were
going to have three of these pipes that we would put into the machine. As far as
I can see from looking at them, they are just pipes, and there is nothing especially
critical about that.
JOE ANDRADE: In our experience with the pulse laser systems and biological
compounds, the photo bleaching proves to be a horrible problem. Is there any
advantage to going beyond microjoules, or even antijoules per pulse, if you're
going to have photo bleaching problems? Why can't you simply use low power
pulses and integrate through a bunch of pulses to get a signal level.
CHARLES MANN: There isn't any point in going to larger pulses. But we
think, from the point of view of quantitative analysis, there is a real point in
going to parallel pulses. By using an optical fiber bundle, we could put easily
100 fibers in the bundle and take the light from each pulse 100 times to the
sample. That will reduce the flux density, obviously, by a factor of 100. It will
not completely eliminate the bleaching, but will increase the dynamic range by
roughly a factor of 100.
116
-------
PROTOTYPE DESIGN AND TESTING OF TWO FIBER-OPTIC
SPECTROCHEMICAL EMISSION SENSORS
Khris B. Olsen Jeffrey W. Griffin
Danny A. Nelson Bradley S. Matson
Pacific Northwest Laboratory
P. 0. Box 999, K6-81
Richland, Washington 99352
Peter A. Eschbach
Department of Physics
Washington State University
Pullman, Washington 99164
ABSTRACT
A unique radio frequency-induced helium plasma
(RFIHP) sensor and a spark discharge (SD) sensor
were designed, and prototype units were developed
and tested. Both sensors use an atomic excitation
source coupled to a fiber-optic cable and optical
spectrometer to monitor in situ the emission in-
tensity of selected elements of interest in the
ambient air. Potential applications include vadose
zone monitoring of volatile species. The RFIHP and
SD sensors were designed to measure in situ concen-
trations of chlorine-containing compounds. The
results of this research demonstrate proof of con-
cept of the theory, but suggest further refinements
are necessary to achieve detection sensitivities
sufficiently low to be useful for monitoring con-
centrations of selected elements in vadose zone
air.
Key words: Helium plasma, spark discharge,
spectrochemical sensors, fiber-optic sensor, in
situ monitoring.
INTRODUCTION
Since 1986, staff at the Pacific Northwest Labora-
tory (PNL) have been developing and evaluating new
chemical sensor concepts suitable for real-time,
multipoint environmental monitoring. The chief
impetus for this research has been the need to bet-
ter understand transport mechanisms of contaminants
in the subsurface environment at the Hanford Site
in southeastern Washington State. For these
measurement scenarios, fiber-optic sensors have
many attractive features, such as a small probe
size, a multiplex advantage (i.e., multiple probes
with one central detection and data acquisition
system), and the potential for fast response. In
addition, these sensors have other potential appli-
cations as monitors and alarms for a variety of
processes relating to other industrial and
government operations.
There is specific interest in developing sensors
capable of real-time, in situ monitoring to detect
selected elements in groundwater and chlorinated
hydrocarbon vapors in vadose zone wells at Hanford.
During past operations at Hanford large quantities
of process-related inorganic and organic chemicals
used in the production of special nuclear materials
were released to the environment. For example,
several hundred tons of carbon tetrachloride have
been disposed to the ground during past operations
at the Plutonium Finishing Plant.(1) Recent
groundwater monitoring efforts have detected carbon
tetrachloride in monitoring wells at concentrations
exceeding 1000 times the drinking water limit.
Furthermore, carbon tetrachloride has been measured
in the outgas from vadose zone wells at concentra-
tions exceeding 200 ppmv. Preliminary data col-
lected around the Hanford Site indicate that carbon
tetrachloride is migrating, both in the groundwater
in the aquifer and as a vapor through the vadose
zone. Currently, the areal distribution of the
carbon tetrachloride is reasonably well defined in
Hanford groundwater; however, the extent of areal
carbon tetrachloride contamination in the vadose
zone is unknown. Before serious consideration can
be given to remediation technologies, rapid, cost-
effective techniques must be developed to accu-
rately identify major vadose zone accumulations of
carbon tetrachloride and other contaminants present
in the subsurface environment at Hanford.
Two fiber-optic sensors based on optical emission
techniques using a radio frequency-induced helium
plasma (RFIHP) source and a low-voltage spark dis-
charge (SD) source are currently under development
at PNL. These two technologies were selected
because they interface well with fiber-optic tech-
nology and have the potential of measuring in situ
concentrations of selected metals and nonmetals in
vadose zone air and groundwater.
Atomic emission sources based on noble gas plasmas
interfaced to gas chromatographs (GC) and optical
spectrometers have been extremely useful to the
analytical chemistry community as element-specific
detectors since proof of concept was first demon-
strated by McCormack et al. in 1965.(2) Since
then, most of the work cited in the literature on
specific-element detectors focused on the
microwave-induced helium plasma (MIP) as the
preferred excitation source for GC detec-
tion. (3,4,5) The GC-MIP system provided suffici-
ently low detection limits and elemental speci-
ficity for most elements of interest. However, a
major drawback of the MIP source was the plasma's
inability to remain stable or, in some cases,
117
-------
ignite when air was injected into the carrier gas
stream. Operating a stable plasma with small but
measurable quantities of air without destroying its
excitation characteristic is a critical requirement
for a detector designed to measure total chlorine
in vadose zone air. In 1985, Rice et al. published
the results of a study using a low-frequency, high-
voltage, electrodeless-discharge plasma (RFIHP).(6)
They found that this plasma had detection limits in
the picogram range for several elements, including
chlorine, phosphorous, sulfur, and mercury. They
also found that the plasma was reasonably tolerant
of the presence of contaminants and air. Addi-
tional advantages included the compactness of the
plasma source and its low helium consumption. To
take advantage of all the favorable characteristics
of the RFIHP, PNL has designed an element-specific
prototype sensor capable of measuring carbon tetra-
chloride in air.
Electrical sparks have long been used as an excita-
tion source for spectrochemical analysis.(7) Con-
sideration of electrical spark excitation for the
SD probe was inspired by recent work by Cremers et
al. (7) describing the use of laser-induced break-
down spectroscopy (LIBS) for the analysis of liquid
samples.(8) The electrical spark offers unique
advantages for real-time environmental sensing.
First, the spark simultaneously provides sample
vaporization, molecular dissociation, and elemental
excitation. This makes the excitation mechanism
suitable for vapors, solid aerosols, and most
importantly, liquids. Second, the electrical spark
is essentially a generator of atomic emission
lines; that is, molecular species are generally
entirely dissociated (at some time during the spark
lifetime or immediately afterwards), and emission
spectra are observed for atoms and ions. Third,
sparks are relatively easy to generate, and they
are well suited for remote environmental probes.
Disadvantages of the spark excitation method
include a high level of variability in optical
emission from spark to spark (up to 50% in our
measurements), severe Stark broadening of spectral
lines, which effectively lowers wavelength resolu-
tion in the emission spectra, and the generation of
an intense continuum emmission. However, the
potential to supply additional information related
to in situ concentration of selected parameters
associated with the migration of contaminants sig-
nificantly outweighs any of the known drawbacks.
EXPERIMENT
Radio Frequency-Induced Helium Plasma
A schematic diagram of the prototype RFIHP probe
and related operating parameters is shown in Fig-
ure 1. This design is a variation of the design by
Rice et al.(6) Major differences from the pub-
lished design include 1) the use of a 6-mm-OD x
1.2-mm-ID quartz tube as the plasma chamber,
2) placing a 10-jjm-ID, 38-mm-long segment of
uncoated capillary column at one end of the plasma
tube (used as a critical orifice), 3) operating the
plasma chamber at subambient pressure, and 4) view-
ing the plasma axially. The overall experimental
system is shown in Figure 2. Fiber-optic input/
output coupling geometries are shown in Figure 3.
Because of inherent limitations in fused-silica
optical fibers, measurements are only possible over
the wavelength range of 400-1000 nm. A conceptual
design of a field probe is shown in Figure 4.
Nominal plasma cell and electro-optical detection
system operating parameters are summarized in
Table I. The probe concept was evaluated as a
detector for chlorinated hydrocarbons or fluoro-
chlorocarbons via detection of neutral chlorine
emissions at 8379.97 A. Detectability investi-
gations were performed with carbon tetrachloride,
1,1,1 trichloroethane (1,1,1 TCA), and dichloro-
difluoromethane.
Spark Discharge
A schematic diagram of the prototype SD probe is
depicted in Figure 5. The overall experimental
system is shown in Figure 6. The operating instru-
mental parameters used during this experiment are
summarized in Table II. Note that for the labora-
tory measurements, the trigger signal was provided
by monitoring spark emission from the top of the
probe. This was necessary because of the long
delay between the input to the trigger module and
actual spark initiation. This trigger mode also
minimized jitter-induced noise resulting from
variation between the spark time position and the
gate time position in the boxcar averager.
Calibration Apparatus
A Metronic Dynacalibrator Model 340 was used to
generate known concentrations of carbon tetra-
chloride to determine the detection limits of the
RFIHP sensor. The permeation chamber on the cali-
brator was operated at 50°C with a chamber flow of
171 mL/min. A 3.81-cm-long diffusion tube with a
cross-sectional area of 0.1963 cm2 was used to emit
known concentrations of carbon tetrachloride into
charcoal-purified air. The diffusion tube was
filled with 3 ml of Burdick and Jackson "Distilled
in Glass"-quality carbon tetrachloride. The diffu-
sion tube was allowed to equilibrate to temperature
for 1 h before calibrations began.
RESULTS AND DISCUSSION
Radio Frequency-Induced Helium Plasma Probe
Spectral wavelength scans near the C1(I) 8375.97 A
line for increasing 1,1,1 TCA analyte are repro-
duced in Figure 7- This analyte was introduced
into the RFIHP by placing a cotton swab soaked with
1,1,1 TCA adjacent to the capillary tube inlet.
Apparent in all the scans are the oxygen peak
located at 8446.38 A and the increasing peak
intensities at 8428.27 A and 8375.97 A caused by
neutral chlorine emission from the 1,1,1 TCA. The
RFIHP sensor responds to a variety of other chlo-
rinated species including dichlorodifluoromethane
(Freon-12) and carbon tetrachloride (Figure 8).
This behavior would be anticipated because this
detector responds to the chlorine atom contained in
the three aforementioned compounds. However, the
relative intensity of the chlorine emission for the
same concentration of a given component dependents
on the percentage of chlorine in the molecule being
measured. Thus, the RFIHP sensor would produce a
chlorine intensity twice as intense for carbon
tetrachloride as for dichlorodifluoromethane.
118
-------
Because the RFIHP detector is incapable of spe-
ciating any of the chlorine-containing compounds,
it is imperative to identify the chlorinated hydro-
carbon of concern so estimates of contaminant con-
centration can be made.
A calibration curve of sensor response versus
analyte concentration (carbon tetrachloride) in air
determined that the prototype probe had a lower
detection limit of 500 ppm. It is important to
note that several factors affect the intensity of
the spectral emissions of chlorine in the helium
plasma. These factors include 1) operating pres-
sure, 2) air bleed rate through the capillary
column (i.e., sample flow rate) into the plasma,
3) helium flow rate (which determines sample
dilution), 4) RF excitation power, and 5) the
location of the plasma viewing along the axes of
the plasma tube. Little attempt was made to
optimize these parameters for the reported demon-
stration experiments. While the reported detection
sensitivities reported by Rice et al. (6) are
excellent for high-purity, atmospheric-pressure
helium plasma, significant degradation of the
plasma excitation properties is expected when the
concentration of the air in the helium plasma
exceeds 1%. However, the 45-/xL/min flow rate of
air through the capillary tube into the plasma is
significantly less than 1%; therefore, it appears
that additional air could be metered into the
plasma without significantly degrading its excita-
tion potential. This increased air flow could fur-
ther decrease detection limits for chlorine.
The primary use of the RFIHP detector on the Han-
ford Site would be to measure the concentration of
carbon tetrachloride in the vadose zone air or for
continuous, long-term carbon tetrachloride moni-
toring in well head space. However, indirectly
estimating the concentration of carbon tetra-
chloride in the groundwater by using the Henry's
law constant is another application of the RFIHP
detector. This estimate, however, assumes the
concentration of carbon tetrachloride in air dir-
ectly above the water (within a few inches) is in
equilibrium with the concentration in the water.
In Hack and Shiu (9), the Henry's law constants for
carbon tetrachloride estimate a partitioning ratio
between air and water at equilibrium to be approxi-
mately 0.95 to 1.0. This ratio estimates that for
each microgram per liter of carbon tetrachloride in
the water the air concentration would be approxi-
mately 150 ppbv. Thus, by applying this relation-
ship, the RFIHP sensor could estimate the concen-
tration of carbon tetrachloride in groundwater from
concentration values determined from air just above
the water/air interface. However, it is uncertain
at this point whether perturbations introduced by
fiber-optic probes would cause deviations from
Henry's law behavior.
Other industrial applications of the RFIHP sensor
include monitoring for concentrations of selected
species in air, such as mercury, fluorine, or phos-
phorus. The species could exist in the gas phase
or be associated with an aerosol. The monitoring
could be related to hazardous vapor detectors and
alarms or process control monitoring.
Spark Discharge Probe
Representative spectra from 3000 6000 A for the
air spark appear in Figures 9A-C, in which the
presence of a strong continuum can be seen through-
out the spectral region. In addition, the spectral
region between 3000 and 4600 A contains several
peaks, most likely attributed to NH and N£ mole-
cular bands. The region above 4600 A is considera-
bly less complicated, particularly between 4600 and
4900 A where numerous Cl(II) atomic lines are
located. Figure 10 is a scan of the spectral re-
gion between approximately 4600 and 4950 A without
the presence of a chlorine-containing compound
(lower trace) and with the presence of carbon
tetrachloride (upper trace). The most intense line
in the spectra is at 4779.54 A because of the
Cl(II) emission line. Because of the poor detec-
tion limit (-100,000 ppm) no attempt was made to
produce a calibration curve.
From the results of the initial experiment it is
clear that the SD probe can cause molecular dis-
sociation and elemental electronic excitation.
However, it clearly is not the detector of choice
for measuring the concentration of chlorine-
containing compounds in air. The real potential
for this probe may be for measuring the concen-
trations of selected metals in groundwater. The SD
probe has more possible applications than the RFIHP
probe. It has the potential to be used to excite
vapor samples, liquids, or aerosols to determine
concentrations of selected elements (metals). How-
ever, it cannot be used in a explosive environment
because of its excitation mechanism.
CONCLUSIONS
Two concepts for fiber-optic spectrochemical emis-
sion probes based on emission spectroscopy have
been demonstrated. Further work is required to
optimize operating parameters for the RFIHP probe
to determine the ultimate lower detection limit for
chlorine. The SD probe will be evaluated in aque-
ous solutions during 1989. The target detection
limit for chlorine using the RFIHP probe is 1 ppm,
and the detection limit for selected metals using
the SD probe will have to be below the drinking
water limits of those selected metals (5-50 ppb).
Both the RFIHP and the SD probe have great poten-
tial for environmental monitoring applications
where real-time, in situ measurements are required.
The use of fiber-optic spectrochemical emission
probes will be an advantage in cases where chemical
interferences to other types of sensors [surface
acoustic wave (SAW) and fluorescent] are antici-
pated and where chemical species of concern have no
highly specific reaction. Finally, the probes do
not use any expendable materials; their lifetimes
should greatly exceed that of fiber-optic probes
based on colorimetric or fluorimetric chemical
reactions. Consequently, they are well suited for
continuous environmental monitoring over extended
periods of time.
ACKNOWLEDGMENT
This work was supported by the U.S. Department of
Energy under Contract DE-AC06-76RLO 1830.
119
-------
REFERENCES
1. Stenner, R. D., Cramer, K. S., Higley, K. A.,
Jette, S. J., Lamar, D. A., Mclaughlin, T. J.,
Sherwood, D. R., and Van Houten, N. C.,
"Hazard Ranking System Evaluation of CERCLA
Inactive Waste Sites at Hanford Volumes I,
II, and III", PNL-6456, Pacific Northwest
Laboratory.
2. McCormack, A. J., long, S. C., and Cooke, W.
D., "Sensitive Selective Gas Chromatography
Detector Based on Emission Spectrometry of
Organic Compounds", Anal. Chem.. 37, 1965,
pp. 1470-1476.
3. Slatkavitz, K., Uden, P., Hoey, L., and
Barnes, R., "Atmospheric-Pressure Microwave-
Induced Helium Plasma Spectroscopy for Simul-
taneous Multielement Gas Chromatographic
Detection", J. of Chromatographv. 302, 1984,
pp. 277-287.
4. Slatkavitz, K., Uden, P., Hoey, L., and
Barnes, R., "Element-Specific Detection of
Organosilicon Compounds by Gas Chromatography/
Atmospheric Pressure Microwave Induced Helium
Plasma Spectrometry", Anal. Chem.. 57, 1985,
pp. 1846-1853.
5. Ester, S., Uden, P., and Barnes, R.,
"Microwave-Excited Atmospheric Pressure Helium
Plasma Emission Detection Characteristics in
Fused Silica Capillary Gas Chromatography",
Anal. Chem.. 53, 1981, pp. 1829-1837.
6. Rice, G. W., D'Silva, A. P., and Fassel, V.
A., "A New He Discharge-Afterglow and its
Application as a Gas Chromatographic
Detector", Spectrochimica Acta. Vol. 40B,
No. 10-12, 1985, pp. 1573-1584.
7. Cremers, D., Radziemski, L., and Loree, T.,
"Spectrochemical Analysis of Liquids Using the
Laser Spark", Applied Spectroscopy. 38, 1984,
pp. 721-729.
8. Bauer, H., Christian, G., and O'Reilly, J.,
eds., Instrumental Analysis. Allyn and Bacon,
Inc., Boston, Massachusetts, 1978.
9. Mackay, D. M., and Shiu, W. Y. "Henry's Law
Constants for Organic Compounds", J. Phvs.
Chem. Ref. Data. Vol. 10, No. 4, 1981.
Table I. Experimental Operating Parameters for the
RFIHP Probe System
Plasma Operations
Pressure: 200 torr
Helium Flow Rate: 50 mL/min
Air Sampling Flow Rate: 45 juL/min
Excitation Frequency: 278 kHz
Excitation Voltage: 9.1 kV p-p
Power: 52 W (load), 70 W (forward)
Electro-Optical System
Chopper: 140 Hz
Blocking Filter: ORIEL RG780
Slit Width: 500 pm
Scanning Rate: 3A/sec
Lock-In Sensitivity: 10-100 mV
Lock-In Time Constant: 0.08 sec
Table II. Experimental Operating Parameters for the
SD Probe System
Box Car Averager
Gate Width: 6
Sensitivity: 100 mV, IV, 2V
Input: DC/IMn
Averaging: 10 pulses
Rep. Rate: 20 pps
Photomultiplier
Tube: Hamamatsu R1828.01
(WA 1324)84.11
Voltage: 1.2 kVDC
Spectrometer
Slit Width: 20 n/
Filter: WG335,
GG475 (ORIEL)
Scan Rate: lA/sec
120
-------
10-^m Bore
Capillary 38
(25 mm Extt
/ I
Silica
mm Long
irnal)
Graphit
Ferrule
B
Ic
To Vacuum Pump ^
Viton O-Ring
Seals
Tee
1 — h r
I-
:rom Ma
Contr
I Electrode ,
\ I :
\
Fused Silica Capillary Tube
1.2-mm Bore, 6-mm OD,
8 cm Exposed Length
J L
Tee
Operating Pressure: 200 Torr
is Flow He Flow Rate: 50 mL/min
slier Air Flow Rate: 45/jL/min
Power: 52 W
To Fiber
Fused Silica Window
25-mm Diameter
2.4 mm Thick
Figure 1, Schematic Diagram of RFIHP Prototype Probe
EG&G/PAR 1450
Power Design
TW5005
ENI Model HPG-2
Power Supply
1453 Head
Jarrel Ash
1 m Czerny Turner
Scanning Spectrometer
1200 Lines/mm,
17°27'
ITHACO 397EO
Superguide Fused
Silica Fiber 1000-^m
Core (17db/km @
0.63 Ljm)
Matheson Dyna
Blender Model
SP- 760 Mass Flow
Controller
Tektronix 464
Oscilloscope (Signal
Monitor)
Figure 2. Schematic Diagram of Experimental RFIHP System
121
-------
Fiber Input Optics
Window
Cell <""\
Lens 25 mm D, FL 25 mm
Fiber
Potted and Polished
Aluminum Ferrule
Spectrometer Input Optics
Fiber
-67 mm
4
Long-Pass |
Filter
Spectrometer
Entrance Slit
(300 fjm)
Chopper Housing
-165 mm >
Figure 3. Diagram of Input/Output Optics for RFIHP
Fiber
_J
Helium
Inlet
/
Helium >• | . I ^
Grin Lens i i
|
To RF Supply] I I*
i_q^-
•" — To Vacuum
Pump
L_
2 Jjl '///,
— -Jv Silica Capillary Tube
\
RF
Electrode
Ground i
, Electrode !
/i Ambient
n ^r
I xi
Capillary Tube or I
Critical Orifice |
J
Figure 4. Conceptual Design of RFIHP Field Probe
Sharpened 2% Thoriated
ngsten Electrodes
ap ~2 mm)
16 in. Dia.)
Stranded Cable
Silicones Inc.
AWM Style 3239 -^^
25 kV DC 150°C VW-1
m y%
w&^&fiZ
^
t
^
JL
f
yx
'/
/.
\
y
'i
V
/
4
\
^
yr
y,
^
y
y/
y
\
t
'y
y
'/,
^
I
To High Voltage F
Power Supply and -«-^
t
Y/1
y-,
^
6
y
/,
1
y
/,
^
y
lj
OL|
/
ft
y],
y,
^y
y.
"y
4,
^
y,
^
•^
^
y
Body Fabricated
^_^- from Delrin ®
(1.25 in. Diameter,
6.5 in. Length)
Fused Silica
Optical Fiber
1000-nm Core
(Silicone Cladding)
;/ Umbilical
n
Spectrometer
Figure 5. Schematic Diagram of SD Probe
122
-------
Trigger Detector
Trigger Fiber
Jarrel-Ash
1-m Czerny Turner
Scanning Spectrometer
1200 lines/mm, 5000
Probe Fiber (Butt Mounted)
Tektronix 7514
Oscilloscope
Power
Design
TW 5005
DC Supply
+15V
Data Precision
Model 5740
Superguide
Fused Silica Fiber
1000-^m Core
Spark
Probe
EG>M-11A
Trigger
Module
High Voltage
Cables
To Spectrometer
m)
Signal
Input
Simpson 461-2
Digital Multimeter
Figure 6. Schematic Diagram of Experimental SD System
Wavelength
Figure 7. Evolution of Chlorine Emission
for Decreasing TCA Concentration
Figure 8. Comparison Response of RFIHP
to Three Chlorine-Containing
Compounds Versus Air
123
-------
4000 A
5000 A
Air (IV)
Baseline (IV)
5000 A
6000 A
Figure 9. Spark Probe Air Spectra from 3000 to 60004
(a) 3000 - 40004, (6) 4000 - 50004,
(c) 5000 - 60004
124
-------
4781.82)
4781.32 •
4768.68 Cl (H)
4766.62 C (I)
4794.54 Cl (E)
4809.05 \ c. (TT)
481 0.06 /C'(ID
4817.33 C (I)
4818.55 Cl (II)
481 9.46 ,c|(m
4896.77 Cl (E)
14904.76CKH)
-^A.
Air
Figure 10. Resppnse of SD Probe to Carbon Tetrachloride
Relative to Air
DISCUSSION
EDWARD HEITHMAR: What was the radio frequency?
KHRIS OLSEN: 276 kilohertz. That's a 9.1 kilowatt power peak to peak.
EDWARD HEITHMAR: Is that self igniting?
KHRIS OLSEN: Essentially, yes. The system that we have has a pulse that
ignites it. When there is a high analyte concentration in there, ignition has to be
done with a coil.
EDWARD HEITHMAR: So you have a problem with quenching of the
plasma with a high analyte, and then having to re-ignite.
KHRIS OLSEN: With the slow leak system, we do not run into that problem,
since the amount of liquid or vapor that was present in the gas did not cause any
problems. When we saturated the system with air, we had most of our problems.
It doesn't like air very much. In other words, if it's saturated with air, at the 10%
level, it is slow in igniting. In a helium atmosphere, it goes very quickly.
HANK WOHLTJEN: Is there any reason you didn't go to just a simple
platinum wire ionization type sensor to do halogens?
KHRIS OLSEN: No.
125
-------
POROUS GLASS FIBER OPTIC SENSORS
FOR FIELD SCREENING OF HAZARDOUS WASTE SITES
by
S. M. Finger, P. B. Macedo, Aa. Barkatt, H. Hojaji,
N. Laberge, R. Mohr, M. Penafiel
Catholic University of America
Vitreous State Laboratory
Washington, DC 20064
ABSTRACT
Rugged, continuous porous glass fiber
optic sensors have been developed and
successfully demonstrated for pH and
temperature measurements. Porous glass
fiber optic sensors are made by selective
leaching of a jacketed borosilicate glass
fiber. The degree of leaching can be
controlled to provide a monolithic
structure with a pre-determined pore size
which can be varied to allow these
sensors to be used for measurements in
liquids, gases, or mixed matrices, e.g.
sludges. The monolithic structure also
maintains the strong attachment between
the sensor portion and the rest of the
fiber, which acts as a light pipe. Since
the sensor is an integral part of the
fiber, losses between the sensor and
light pipe regions are minimized. The
end of the sensor is coated with a thin,
porous- layer of gold to reflect the
incident and response radiation back into
the light pipe for analysis.
Porous glass fiber optic sensors can be
designed for a wide variety of analytes
by changing the active species (or
combination of active species for multi-
analyte measurements) bonded to the large
internal surface area of the sensor. The
pH and temperature sensors which were
produced used dyes as the active species.
However, other active species, such as
enzymes and other biochemicals, could be
attached to the internal glass surface.
Sensitivity can be controlled by varying
the length of the porous sensor region
and by varying the concentration of the
bonded active species. It should also
be noted that the light makes two passes
across the porous sensing region since it
is reflected at the end of the sensor.
A mathematical analysis of a fluorescence
sensor was performed. It was found that
the sensitivity of a 2 cm long porous
senaor would be over 200 times as great
as the sensitivity of a two-fiber
fluorescence sensor because the porous
sensor is able to capture essentially all
the fluorescence given off within the
fiber. In the two-fiber system (the two
fiber tips being tangent and at an angle
of 22° ), only a fraction of the
fluorescence is captured in the receiving
fiber; the rest is lost in the open
region between the fibers.
The ruggedness, wide applicability and
inherent sensitivity of porous glass
fiber optic sensors offer significant
advantages for the development of low-
cost, portable, real-time field screening
methods which are critical for effective
hazardous waste site evaluation and
surveillance monitoring.
BACKGROUND
The demand by the Environmental
Protection Agency (EPA) for
environmental monitoring data has
increased explosively over the last ten
years. There are several factors for
this increased demand. First, the Agency
has moved to expand monitoring coverage
from a focus on conventional pollutants
to a strong focus on toxic pollutants.
Second, the number of matrices has been
enlarged from a focus on air and fresh
surface water to monitoring coverage of
soil, sediment, groundwater, marine
water, estuarine water, drinking water,
and fish tissue. Third, the number of
EPA Program Offices with their own major
monitoring programs has also increased
dramatically: from air and surface water
only, to the establishment of major
monitoring efforts for Superfund,
Resource Conservation and Recovery Act
(RCRA), Toxic Substances Control Act
(TSCA), and Safe Drinking Water Act
(SDWA) .
127
-------
The result has been an exponential surge
in demand by the EPA Program Offices for
new analytical methods to support these
monitoring programs and needs. In
addition, there is a critical need to
reduce the cost of monitoring. This very
real problem must have a high priority to
ensure that programs addressing critical
environmental problems are not delayed or
limited due to the high cost of
monitoring. These critical and urgent
problems will be exacerbated in the
future as the Agency highlights new
initiatives and intensifies current areas
of activity.
To overcome these problems, a new
generation of field monitoring methods is
required. These methods must be (a) easy
to use, (b) inexpensive to operate, (c)
real-time, (d) in situ, (e) capable of
screening a wide variety of analytes, (f)
reliable, (g) rugged, (h) portable, (i)
have long shelf lives, and (j) be
capable of unmanned operationl. Fiber
optic monitors have the capabilities to
meet these requirements. Fiber optics
have been used as a light pipe in
spectrometric systems2-4 and as an
integral part of chemical sensors5-7.
Fiber optic chemical sensors typically
have reactive species on the external
surfaces of the optical fiber, such as on
the tip or outer wall of the fiber, or
the fiber is immersed in a reservoir of
reagentl. The subject of this paper is a
fiber optic sensor with an integral
porous tip so the optical interaction
takes place throughout the volume of the
sensing element.
DESCRIPTION OF THE SENSOR
The preparation of the porous glass fiber
optic sensor has been described by
Macedo, et a!8 and is summarized here.
Borosilicate glass optical fiber is made
porous by having the fiber go through a
phase separation furnace while it is
being pulled. This causes the
borosilicate glass to separate into a
silica-rich and a silica-poor phase. The
soft, silica-poor phase is then removed
by acid leaching at 90-95 C, which is
followed by a long water rinse to remove
substances such as silica gel, which
precipitate within the pores upon
leaching. This process produces glass
fibers with pore sizes as small as 40
microns. The active species is then
attached by silanization. We have used
this procedure to attach a variety of
dyes. The dye is covalently bonded to
the glass surface through a silane
coupling agent which prevents loss of the
dye (or other active reagent). Last,
sputtering was used to coat the end of
the sensor tip with a thin, porous layer
of gold. The gold acts as a mirror so
that a single optical fiber can be used
for both incident and response light
transmission. Thus, for example, this
design allows for two-pass absorbance
measurements. Also, fluorescence
measurements can be performed with only
one fiber, rather than two, since both
the incident and emitted light are
transmitted within a single fiber.
The optical fiber with the porous sensor
tip, shown in figure 1, is attached via a
standard connector to a directional
coupler which in turn, is connected to an
appropriate source and detector, as shown
in figure 2.
Porous glass fiber optic sensors have
several advantages over other fiber optic
sensors. These include: (a) ruggedness
(the sensor element is integrally
attached to the fiber wave guide and
because the active species is bonded to
the interior of the sensor), (b) minimal
reflection at the light pipe-sensor
interface (the porous tip is integrally
attached to the light pipe), (c)
effective control of sensitivity (the
length of the porous sensor tip can be
varied as can the concentration of
active reagent on the glass surface
within the sensor), and (d) ability to
perform multiple simultaneous analyses
(the system can provide incident
radiation and measure absorbance,
fluorescence or raman scattering at
several wavelengths at the same time).
These features make porous glass fiber
optic sensors well-suited for in situ.
real-time, remote measurements by
personnel with minimal training.
EXAMPLES OF APPLICATIONS
A. Absorbance Measurements
pH and temperature porous glass fiber
optic sensors have been successfully
prepared. pH sensors used phenol red or
cresol red as the indicator dye. Optical
absorbance measurements at 575 nm were
found to provide a good response to pH
changes. Figure 3 shows this response
from pH 5 to 8. Figure 4 shows the
response of a porous glass fiber optic
temperature sensor from 28 to 70 C.
This sensor used pinacyanol chloride as
the indicator dye; absorbance
measurements were made at 625 nm.
Absorbance measurements using dyes would
also be effective for monitoring the
concentration of heavy metals and
chlorinated organics. Beer's Law can be
used to estimate detectability limits:
A = eLc
(1)
where A is the absorbance, e is the molar
extinction coefficient, L is the length
128
-------
of the porous sensor tip, and c is the
concentration of the analyte. Assuming a
minimum absorbance measurement of 0.02, a
molar extinction coefficient of 2x104
liters/cm-mol, and a porous sensor length
of 4 cm, the minimum detectable
concentration of analyte would be:
Cm = 2.5x10-7 mol/liter
If the molecular weight of the analyte is
assumed to be 100, the minimum detectable
concentration could be expressed as:
Cm = (2.5x10-7 mol/L)(lOOg/mol) =
2.5x10-5 g/L
25 ppb
Thus, absorbance measurements in the low
ppb range are achievable with porous
glass fiber optic sensors. Since there
are many dyes which can be used for such
sensors, these monitoring devices can be
adapted to a wide range of analytes.
Also, since the spectral response of the
dyes is in the visible range, losses in
the light pipe portion of the sensor
would be minimal.
The response of the fiber optic
absorbance sensor could be optimized for
maximum response for a desired
concentration range. The intensity of
the light decreases in accordance with
Beer's Law. Stated in terms of intensity
rather than absorbance, the relationship
I = lo exp (-eLc]
(2)
where I is the intensity of the light as
it passes through the sensor region and
lo is the intensity of the incident
light. Differentiating twice, with
respect to concentration and sensor
length, gives:
d2I/dLdc = lo(eLc-l) exp(-eLc)
13)
Setting equation 3 equal to zero to
determine the relationship for maximum
response, we obtain:
eLc = 1
(4)
Equation 4 can be used for maximum
response of the sensor for a given
concentration of analyte. While it would
not be possible get maximum response near
detectability limits, this relationship
ia useful to optimize the sensor at
higher analyte concentrations.
B. Fluorescence Measurements
As noted earlier in this paper,
fluorescence measurements could be
performed using a single porous fiber
rather than a two-fiber system as is
currently required. In addition to being
simpler in design, porous fiber
fluorescence systems provide a
significant enhancement in signal
intensity. This is due to the analyte
molecules and the excitation and
fluorescence radiation all being
contained within the porous sensor. In a
two-fiber fluorescence system, the
analyte molecules fall in the open region
between the two fibers so that only the
excitation and emitted radiation within
the numerical aperture of the fibers is
measured; the rest is lost to the
surroundings. The comparison is
exemplified by the two situations
depicted in figure 5. The angle between
the two fibers in Situation I is chosen
as 22o and the diameter is chosen as 0.6
mm (as is used in the fluorescence sensor
described in reference 2 ) . The
fluorescence intensity of the two-fiber
system is given by the following
equation:
FI = Hloo/rTDI'] [£AC] l(DI cos 9)
(DI sin 0) (2DI) (7f/4 f~ j
= Ioof,/vcDI(TT/2 ) cos6 sin6
(5)
where loo is the intensity of the
incident radiation, D is the fiber
diameter, f is the fluorescence
efficiency, v
-------
fiber fluorescence system! This 2.
increased intensity could be used to
perform measurements at lower
concentration levels, or it could allow
the use of an incandescent sottrce 3.
insteadof a more complex but powerful
laser source. Either way, the porous
fiber sensor would offer distinct 4.
advantages over the two-fiber system.
CONCLUSIONS 5.
The ruggedness, wide applicability and
inherent sensitivity of porous glass 6.
fiber optic sensors offer significant
advantages for the development of low-
cost, portable, real-time field 7.
screening methods which are critical for
effective hazardous waste site evaluation
and surveillance monitoring. 8.
REFERENCES
1. L. A. Eccles, S. J. Simon, and S. M.
Klainer, "In Situ Monitoring at
Superfund Sites with Fiber Optics,"
U. S. EPA Environmental Monitoring
Systems Laboratory, Las Vegas,
Nevada.
J. E. Kenney, G. B. Jarvis, W. A.
Chudyk, and K. O. Pholig, Anal.
Instr. . 1*5(4), 423 (1987).
D. A. Van Dyke and H. Cheng, Anal.
Chem.. 60. 1256 (1988).
O. S. Wolfbeis and A. Sharma, Anal.
Chim. Acta. 208. 53 (1988).
L. A. Saarl and W. R. Seitz, Anal.
Chem.. 55. 667 (1983).
W. R. Seitz, Anal. Chem. . 5J5_i 16A
(1984).
R. E. Schirmer and A. G. Gargus,
Amer. Lab.. 30 (Oct. 1986).
P. B. Macedo, Aa. Barkatt, X. Feng,
S. M. Finger, H. Hojaji, N. Laberge.
R. Mohr, M. Penafiel, and E. Saad,
SPIE Proc.. 986. in print (1988).
SOURCE
Figure 1. Porous Glass Fiber Optic Sensor
DIRECTIONAL
COUPLER
CONNECTOR
POROUS
SENSOR
DETECTOR
Figure 2. Arrangement of optical components
130
-------
11 .
10
£ 7
~ 6
m
rT~r-™i—p-p-
400
oO
500 550
Wavelength, mn
650
700
Figure 3. pH porous glass fiber optic sensor respcr.se
12
10
x
oc
400
Temperature
500
600
Wave1en g th, nm
700
800
Figure A. Temperature porous glass fiber oprice sensor response
131
-------
X..
w
>/s.
/'*\
Mirrored
surface
Single fiber porous glass system
2-fiber system
Figure 5. Comparison of the single fiber sensor vs.
a two-fiber fluorescence system
DISCUSSION
DELYLE EASTWOOD: Could you elaborate about the pore sizes, and how
you would vary those for various reagents and possible diffusion problems for
those smaller pore sizes?
STANLEY FINGER: Pore sizes are quite controllable. We have been able to
get down to pores on the order of about 40 to 50 angstroms. Obviously, for the
gaseous systems, you can work with the smaller size pores. For liquid systems,
you're going to need larger pores, on the order of several hundred angstroms.
The pore size is controlled through the heat treatment itself, and by varying the
temperature and residence time of the heat treatment. You can very easily
change the pore size. In addition, there are a number of different temperature
and time combinations that will give you the same range of pore sizes.
While there is one optimum for any given pore size distribution, you still have
a wide range of conditions in which you can get that pore size distribution.
The other factor that comes into the choice of pore size, of course, is the
response time. By enlarging the pore size, you improve the response time,
because of the smaller diffusion path that's required and the larger diameter for
that diffusion path.
So enlarging the pore size is helpful in terms of response times, although we
haven't had real problems with that. These sensors tend to come to equilibrium
within a minute or two -a fairly quick response. We don't see the need for these
sensors to get down to response times on the order of a second or so, although
that could potentially be done. We haven't looked at that.
STEVEN GOHEEN: You fused the porous and the nonporous regions of the
fiber optic. Did you ever quantitate the difference between fused and unfused?
In other words, it could be cumbersome in the field to have to change the entire
fiber optic every time you were measuring a different component, or using a
different sensor. Maybe a sensor would wear out, and you want to replace the
tip. Did you actually quantitate the difference?
STANLEY FINGER: These sensor tips aren't really fused. They are actually
an integral part of the sensor, which is leaching out a portion of that sensor tip.
The term fuse implies something with connecting the tip on, and that's not quite
correct.
The problem of looking at multi-analytes, relates to whether we can look at a
number of different spectral locations at the same time, or use several different
dyes at the same time, and the answer is yes.
There is no reason why you couldn't be looking at different portions of the
spectrum for different analytes simultaneously, or have more than one dye, or
other type of measurement being made at the same time.
The problem of changing these is one of going back to the connector and
changing it at that point. The connector doesn't have to be yards, or any great
distance away from the tip. You can have a connector fairly close to the tip, if
you want to periodically change the sensor tip.
We don't see this as being a problem. In fact, we see it as being one of the
advantages of the porous sensor, and periodically changing these as they wear
out can be done fairly easily.
HANK WOHLTJEN: The porous glass sensor is a very elegant approach, but
there is a fundamental problem with any optical sensor that's based on a
colorimetric change of an immobilized reagent. The band shift that occurs is
caused by changes in the activity of the medium, rather than the concentration.
So you can't really make a pH sensor. You make a sensor that's sensitive to
hydrogen ion activity, and not necessarily hydrogen ion concentration. If you
have a pure solution of known pH, you can get an accurate measure of it, but
if you have an unknown in there that is going to change the ionic strength of the
solution, then you can't get an accurate measure. You'll get good data, but what
will it mean?
STANLEY FINGER: That is typical with all optical systems, especially dye-
type systems, and I don't know that it can be solved completely, except through
extensive calibrations for any particular analyte.
You would have to look for a situation where the expected changes in the
environment would not significantly change the signal, so that what you are
measuring can be related pretty closely to concentration.
Our idea for these sensors is to look at hazardous waste sites in terms of
monitoring. The need is primarily to determine whether there is a significant
amount of a particular hazardous component. The accuracy of our measure-
ment is less important than perhaps in the standard laboratory test, where you
have to take that data and go into court, and fine somebody based upon it. What
we're looking at here is do we have a significant amount of concentration - is
there a problem here, or not?
HANK WOHLTJEN: By the same token, fluorescence measurements may
not have that kind of complication. In your scheme to do fluorescence
measurements, how do you strip off the excitation light? Do you use a filter and
is it adequate?
STANLEY FINGER: Yes, it's done through the coupler, and the filters, and
it is adequate from what we have been able to determine.
132
-------
INSTRUMENTATION AND METHODOLOGY FOR MULTICOMPONENT
ANALYSIS USING IN SITU LASER-INDUCED FLUORESCENCE
Jonathan E. Kenny, George B. Jarvis, and Hong Xu
Department of Chemistry
Tufts University
Medford, MA 02155
ABSTRACT
In order to facilitate remote, in situ monitoring of multiple
contaminants in ground-water, we have improved our
prototype laser/fiber optic instrument by the addition of a
Raman shifter and a diode-array spectrometer detector. In
this paper, we discuss the performance of the Raman shifter
in producing multiple excitation wavelengths, the correction
of observed spectra for source fluctuations, and related topics.
KEY WORDS: fluorescence, fiber optics, in situ, Raman
shifter
INTRODUCTION
The sensitivity of the laser-induced fluorescence technique,
along with the many advantages of fiber optics for remote
sensing, can be combined to produce powerful
instrumentation for in situ monitoring of groundwater (1).
Although the ingenious design of molecule-specific or
functional-group-specific optrodes (2) makes it possible to
detect even non-fluorescent analytes using fluorescence-based
techniques, our work has concentrated on the monitoring of
naturally fluorescent species. These include about half of the
substances on the EPA's list of 119 priority pollutants.
Recently, we developed a simple, sensitive, and portable
instrument which uses the ultraviolet output of a frequency-
quadrupled Nd:YAG laser to induce fluorescence in various
aromatic molecules in aqueous solution (1). We demonstrated
the sensitivity of this instrument to be in the parts-per-billion
regime for substances like phenol and unleaded gasoline,
when the light was carried to and from the remote sample by
a pair of 20- to 25-meter long, 600-/J diameter fused-silica
optical fibers (3).
One limitation of this first-generation instrument was its lack
of specificity. Since fluorescence detection offers the
possibility of considerable specificity, we have attempted to
work towards molecule-specific detection in a remote-sensing
instrument.
Warner and coworkers (4) have developed and reviewed
luminescence-based techniques for identifying specific
components in mixtures. Of course, the principle is simple:
measure enough parameters of the sample, and you will be
able to effect a mathematical "separation" in the computer
instead of a physical separation in the laboratory. The most
fundamental aspect of molecular luminescence is its
dependence on wavelength: the absorption and emission
spectra of a molecule provide two characteristic "signatures"
by which that molecule may be identified. Our focus will be
on the collection and analysis of Excitation-Emission Matrices
(EEM's) of pure compounds and mixtures, which contain the
encoded spectral signatures of each fluorescent component.
Now we briefly explain EEM's. For a single fluorescing
component, the measured fluorescence intensity, M, as a
function of excitation wavelength, Ax, and emission
wavelength, \n, is given by
= k
(1)
where k is a scale factor, I^) is intensity of excitation light
delivered by the laser and excitation fiber, e^) and ^(A,,,)
are the components of the absorption and emission spectra of
the compound, D(Am) represents the efficiency with which
the emitted light is transmitted through the detection fiber(s)
and detected by the sensing instrument, and c is the quantity
of interest, i.e., the concentration of the substance being
detected. For a dilute mixture, the total signal is just a sum
of such terms, one for each fluorescing component. The
EEM of a single component or mixture is a matrix of M
values, with \^ and ^ being the row and column indices.
Since D(Am) is constant in time for a given instrument, only
variations in source intensity, IC^), must be divided out from
the experimental measurement to allow the EEM to be
analyzed. For example, the least-squares method can be used
to determine the concentration of each fluorophore present in
the sample. Other data filtering and reduction schemes have
been developed, including autocorrelation techniques (5)
which can be useful for fingerprinting of complex mixtures.
Laboratory-based instruments for EEM analysis generally
feature lamp-monochromator sources with good temporal
stability which may be easily tuned across the desired range
of Ax's with any desired interval size. When performing
remote analyses using uv light and fiber optics, lasers become
the only practical sources; unfortunately, they are more
difficult to tune and are subject to greater intensity variation,
even over short periods of time. The work to be discussed
herein relates to a convenient, multiple-wavelength excitation
source based on stimulated Raman scattering, and a
methodology for monitoring laser power fluctuations.
INSTRUMENTATION
A schematic of the instrumentation used for remote
multidimensional fluorescence measurements is shown in
Figure 1. A Nd:YAG laser (Quanta-Ray OCR-11) with
second-, third-, and fourth-harmonic generating crystals is
the primary excitation source; its output at 1064, 532, 355, or
266 nm may be used directly, or Raman shifted to produce
various other wavelengths. The desired wavelength is chosen
133
-------
by rotating the Pellin-Broca prism at the Raman shifter
output; this directs the beam through a variable attenuator
and a focusing lens onto the tip of the excitation fiber, which
delivers it to the sample. A beamsplitter before the
excitation fiber directs a fraction of the light onto a
monitoring detector. Fluorescence from the sample is carried
to the diode-array spectrometer (Aries Monospec-27) by one
or more detection fibers. The EG&G blue-enhanced OMA
detector is interfaced to a Compaq 386-20 computer via an
IEEE-488 bus.
GENERATION OF MULTIPLE EXCITATION
WAVELENGTHS
To overcome fiber transmission losses in the ultraviolet,
where many fluorescent compounds must be excited, laser
sources must be used. For single-wavelength excitation, we
have found a frequency-quadrupled Nd:YAG laser to be
convenient its good power, simple and troublefree operation,
and air cooling make it a viable source in a portable
instrument. To generate EEM's, multiple A/s are needed in
the range ~250-500 nm. In general, the amount of
vibrational structure in the absorption spectra of the .target
molecules dictates the spacing of these wavelengths; intervals
of 2-5 nm are typical in laboratory work done in nonpolar
organic solvents, while larger intervals should suffice in
aqueous solution. The only practical laser sources for
generating the required excitation wavelengths in the uv are
frequency-doubled dye lasers and Raman shifters (6).
Dye lasers offer considerable advantages: high power, good
beam quality, and continuous tunability over a wide spectral
range. However, they are fairly complicated and expensive,
and, to cover a wide spectral range, several dye solutions are
required. Furthermore, to access the shortest wavelengths
needed, e.g., 250-350 nm, the output of the dye laser itself
must be converted by frequency-doubling.
Raman shifters, on the other hand, are extremely simple,
having only one moving part (a tuning prism) and one
working fluid. As dye lasers work on the principle of
stimulated fluorescence, Raman shifters work on stimulated
Raman scattering. Thus, unlike dye lasers, whose output is
continuously tunable over some part of the fluorescence
spectrum of the dye, Raman shifters produce output only at
fixed frequency shifts from that of the pumping laser.
Output frequencies occur at intervals of ±1, ±2, ±3, etc.,
times a fundamental vibrational frequency of the fluid in the
Raman shifter. If a N4YAG laser with its various harmonics
is used as a pump laser, the shifted outputs from each of the
inputs provide reasonable spectral coverage, even if only a
single working fluid, like H2 gas, is used. Additional output
wavelengths may be obtained by using a second gas; both
may be present at once, so no fluid changes are required
during tuning.
The output power in each of the available excitation
wavelengths from such a system can vary greatly; however,
for a reasonably compact NdYAG laser pump, and using
only H2 in the Raman shifter, we have generated 21 different
wavelengths in the region from 220 to 532 nm having
sufficient intensity for remote sensing work, as shown in
Figure 2. We are currently studying the effect of using a
mixture of H2 and methane; limited results using this mixture
were reported elsewhere (7), and the system looks promising.
MONITORING SOURCE INTENSITY CHANGES
As shown by Eq. 1, the intensity of the observed signal is
directly proportional to the intensity of excitation light, l()±),
incident on the sample. The removal of this unwanted
dependence in quantitative work is often called power
normalization; I(\) is measured and both sides of Eq. 1 are
divided by the reading to produce a quantity proportional to
concentration, c. In conventional fluorimetry, a beamsplitter
normally splits off part of the excitation light beam at a
suitable point and directs it towards a detector. For in situ
measurements, there are a number of complications to this
simple procedure. First, the focal spot of the laser beam is
likely to be highly nonuniform in intensity distribution.
Since it is being focused onto a small fiber, care must be
taken that an exactly corresponding portion of the split-off
focused beam hits the power normalizing detector.
Obviously, the second, related problem is to eliminate any
vibration or other motion that might cause misalignment,
once it is achieved. Third, the attenuation of the excitation
fiber varies greatly with wavelength, and this variation must
be properly accounted for. Manufacturers' data or actual
attenuation measurements could provide this information;
however, the transmission properties of fibers which carry
large amounts of uv light are known to change considerably
over the period of exposure (so-called "solarization").
We have found that these problems may be addressed by
monitoring scattered excitation light returning through the
detection fiber(s), in addition to the measurement of
excitation intensity at the source, as described above. The
actual experimental work was done using our undispersed-
fluorescence detection system, but the principle is equally
applicable (and probably easier to implement) in the diode-
array spectrometer system used to collect EEM's.
For these power normalization experiments, two changes were
made to the block diagram of Figure 1. A Laser Precision
joulemeter was put behind a transparent sample holder (a
fused silica cuvette filled with distilled water) to monitor the
true source intensity reaching the sample. In place of the
diode array spectrometer, a detection module (1) was used
which could separately monitor total fluorescence and 266-nm
Rayleigh scattering. The responses of the pre-launch laser
power monitor and the detection module's Rayleigh scattering
monitor were measured against the true source intensity at
the sample. The results showed good linearity over a suitable
range of excitation energies, as depicted in Figure 3. Further
details of this experiment and related experiments are given
elsewhere (8).
Having demonstrated the viability of using either monitoring
scheme for power normalization, we now discuss the rationale
for using both. The pre-launch reading indicates directly the
performance of the laser-Raman shifter system. The
Rayleigh monitor directly indicates the performance of the
entire instrument. Proper readings on both indicate a
functioning system even in the absence of any fluorescence
signal (should such a pristine aquifer be located in field
work). This is a factor of major importance in a field
instrument which may very well be operated in a mode
wherein permanently installed fiber optics probes, inaccessible
to the operator, are periodically serviced by a mobile
instrument. Before a probe is first installed, the initial
relationship between the two power normalization channels
should be established with the probe immersed in distilled
134
-------
water. Under normal conditions, the Rayleigh probe's
reading should be used for actual fluorescence intensity
corrections, since it more accurately measures power incident
upon the sample as the excitation fiber ages. A comparison
of the two readings, in fact, allows solarization of the
excitation fiber to be monitored. Thus, the Rayleigh probe
ensures that proper source intensity corrections are made
throughout the useful life of the excitation fiber, as well as
indicating when that useful lifetime has been reached.
Of course, either near-zero or excessively large readings on
the Rayleigh monitor indicate conditions under which other
types of corrective action must be taken (8).
SUMMARY
The Raman shifter appears to be a promising source for the
production of multiple excitation wavelengths, although a
two-gas mixture may have to be used as the working fluid to
produce a sufficiently fine grid. The proposed power-
normalizing scheme, which utilizes two separate
measurements, has the additional advantage of providing a
means of instrument self-check and monitoring of the
solarization of fibers which deliver high-power uv light to
the remote sample.
ACKNOWLEDGMENT
We wish to acknowledge the financial support of the U.S.
Environmental Protection Agency through a grant to the
Center for Environmental Management at Tufts University.
REFERENCES
(1) Kenny, J.E., Jarvis, G.B., Chudyk, W.A., and Pohlig,
K.O., "Remote Laser-Induced Fluorescence Monitoring
of Groundwater Contaminants: Prototype Field
Instrument," Analytical Instrumentation 16, no. 4, 1988,
pp. 423-445.
(2) Seitz, W.R., "Chemical Sensors Based on Fiber Optics,"
Anal. Chem. 56, no. 1, 1984, pp. 16A-34A.
(3) Chudyk, W.A., Carrabba, M.M., Jarvis, G.B., and
Kenny, J.E., "Prototype Laser Fluorescence-Fiber
Optics Groundwater Contaminant DetectOT,"Specialty
Conf. on Environ. Engin., EE Div., ASCE-Boston, MA,
July 1-5, 1985, pp. 98-103.
(4) Warner, I.M., Patonay, G., and Thomas, M.P.,
"Multidimensional Luminescence Measurements," Anal.
Chem. 57, no. 3, 1985, pp. 463A-481A.
(5) Neal, S.L., Patonay, G., Thomas, M.P., and Warner,
I.M., "Data Analysis in Multidimensional
Luminescence Spectroscopy," Spectroscopy 1, no. 3,
1986, pp. 22-28.
(6) Mollenauer, L.F. and White, J.C., eds., Tunable
Lasers, vol. 59 in Topics in Applied Physics, Springer-
Verlag, New York, 1987.
(7) Duardo, J.A., Nugent, LJ., and Johnson, F.M.,
"Combination Lines in Stimulated Raman Emission
from Gas Mixtures," J. Chem. Phys. 46, no. 9, 1967,
pp. 3585-3591.
(8) Jarvis, G.B. and Kenny, J.E., "Considerations for
Power Normalization of Remote Laser-Induced
Fluorescence Measurements," in preparation.
135
-------
DIODE
ARRAY
INTERFACE
COMPUTER
FIGURE 1, BLOCK DIAGRAM OF MULTIDIMENSIONAL LUMINESCENCE SPECTROMETER FOR
IN SITU MONITORING. HG=HARMONIC GENERATOR, PB=PELLIN-BROCA PRISM, I=IRIS,
BS=BEAMSPLITTER, QC=QUANTUM COUNTER LASER MONITOR, L=LENS, S=SAMPLE AND
GS=GRATING SPECTROMETER.
-------
100 -i
10 -
CO
0)
D 1 '
O
-H
rH
-H
E
.01 -
.001 -
20
1
oo
1
STIMULATED RAMAN SCATTERING
I
3OOO 4OOO 5OOO 6OOO 7OOO BOOO
Wavelength (Angstroms)
FIGURE 2. OUTPUT OF 2OO psig HYDROGEN-FILLED RAMAN SHIFTER
PUMPED BY HARMONICS OF YAG LASER
-------
600
CO
300 ••
LLJ
50
100
150
200
250
300
JOULEMETER (mV)
FIGURE 3. RESPONSE CURVES FOR THE TWO LASER POWER MONITORS VERSUS JOULEMETER READING AT THE SAMPLE
DIAMONDS REPRESENT THE PRE-LAUNCH (QUANTUM COUNTER) MONITOR: SQUARES, THE RAYLEIGH MONITOR.
-------
DISCUSSION
DELYLE EASTWOOD: The excitation emission matrix approach which, as
you pointed out, largely developed by Christian and Momer, is really more
applicable to diagnostic purposes than to monitoring.
When I was with the Coast Guard, we found that we had so much extra data it
was confusing.
Isiah Warner and Steve Fuh do have a fiber optical multi-channel analyzer,
capable of excitation emission matrices already. I don't remember what laser
they used.
UNIDENTIFIED PARTICIPANT: How does the excitation matrix approach
account for energy transfer interferences among the competing fluorophores?
JONATHAN KENNY: It really doesn't. The limitation of this approach is that
it will work in dilute solutions, and if the solutions are dilute enough, the
chromopores don't interact. So it's predicated on the assumption of Beers law
for linearity and additivity.
MIKE CARRABBA: We saw earlier this morning a slide by Professor
Chudyk, in which he compared the GC/MS data to his data. There were
instances where there were two to three order of magnitude differences.
Do you think that that could be accounted for in the power normalization
scheme that he is using, versus the one that you're using?
JONATHAN KENNY: The slide he showed of the detection unit looked just
like the slide 1 showed. I think that what might happen is if you're in a nondilute
environment, the absorption of some of the scattered light before it gets back
into the RLIF probe could be hurting you. It's hard to speculate. I really don't
know.
139
-------
INFLUENCE OF NATURALLY OCCURRING VOLATILE COMPOUNDS
ON SOIL GAS RESULTS
by
Royal J. Nadeau
U.S. Environmental Protection Agency
Environmental Response Team
Edison, New Jersey
Joseph Tomaszewicz
ERT Technical Assistance Team
Edison, New Jersey
ABSTRACT
Naturally occurring volatile
compounds are present in the vadose
zone particularly in soils of high
organic content. These compounds
are the product of living plants or
degradation of plant parts. Some
of these compounds have the same
retention times as some common
hydrocarbon pollutants, producing
false postive results.
Knowledge of the types and
levels of these naturally occurring
compounds is important to the soil
gas analyst to make accurate
analytical interpretation of
chromatographs particularly from
portable gas chromatographs.
INTRODUCTION
Soil gas surveys are being
widely used to determine the
presence of contaminants in the
groundwater and overburden
materials. Many times these
surveys will require collecting
soil gas samples in heavily
vegatated areas that are believed
to be out of the contaminated zone.
Often the samples collected from
these areas are considered free
from volatile contaminants but yet
contain many volatile compounds
that are present in the soils from
natural origins.
Some of these naturally derived
volatile compounds possess similiar
physical chemical characteristics
to certain volatile contaminant
compounds. Under these conditions,
the analyst using a field detection
instrument eg. hand-held total
volatile organic detector or
portable gas chromatograph could
make a false positive determination
for that sample.
We have encountered a number of
naturally volatile compounds in a
number of the soil gas surveys that
were performed at sites across the
country. This paper will dwell
upon these compounds and the
problems that were presented to the
analyst and data interpreter.
METHODOLOGY:
The method utilized for
obtaining soil gas samples consists
of creating a 1.5cm diameter
vertical hole in the ground using a
slam bar followed by inserting a
.75cm diameter O.D. stainless steel
probe into the hole, sealed at the
top with modeling clay. The hole is
evacuated for one minute to remove
atmospheric infiltration using an
air sampling pump. A total organic
volatile detector (HNu Model PI101)
was then attached to the stainless
steel tube to provide a gross level
of the total volatile organics
present. Following this
determination, a sample of the soil
gas was collected into a 1.0 L
Tedlar bag.
The bagged samples were then
analyzed within 24-36 hours from
time of collection using a portable
gas chromatograph (Photovac Model
141
-------
10A10 or 10S5-0) equipped with a
packed column and a photoionization
detector (10.7 eV)(GC/PID).
Compound identification was made by
matching retention times and area
responses with external standard
gas mixtures composed of selected
aromatic and chlorinated
hydrocarbons. Syringe blanks were
run between each analysis to ensure
against carryover contamination.
After the Photovac analyses,
selected samples were then adsorbed
on Tenax/Carbon Molecular Sieve
(CMS) tubes for compound
identification by GC/MS analysis.
One of the criteria for selecting
samples for confirmation was the
presence of compounds that were not
readily identified by the Photovac.
These sample tubes were analyzed by
thermal desorption onto a cryogenic
trap, then analyzed by a
Hewlett-Packard 5993 GC/MS using
"Compendium of Methods for the
Determination of Toxic Organic
Compounds in Ambient Air" (EPA
600/4-84-041, Apr. 1984).
CASE STUDY #1. (Methane Site)
Explosive levels of a
combustible gas determined through
the use of a Combustible Gas
Indicator Meter (CGI) and Organic
Vapor Analyzer (OVA) had been found
in the basements in several homes
in a new housing development. Some
of the homes had been constructed
adjacent to subsurface trenches
filled with surface vegetation eg.
trees, understory shrubs cleared
for development (Figure 1.)
Soil gas samples were collected
at several locations in the
backyards of the homes that were
closest to the trenches. Additional
soil samples were collected in back
of homes that had been constructed
in another part of the development.
Off-site samples were collected in
the backyards of several residences
that were not part of the
development to determine background
levels for the survey.
The field survey data indicated
the presence of high concentrations
of organic volatiles in the soils
with methane comprising most of the
organics in the soils behind the
homes closest to the trenches. The
backyards of those homes used as
background samples contained very
low or non-detectable levels of
organic volatiles.
A Century OVA - 128 was used in
the gas chromatographic mode for
the methane analysis. The OVA in
the gas chromatographic mode
produced results that were
consistent with the field data; the
locations that had high
concentrations as measured by the
field instruments had high
concentrations of methane in the
bag samples.
Analytical results from the
various analytical tiers used at
the Methane Site are presented in
TABLE 1. The confirmatory
analysis from the GC/MS revealed
the presence of low molecular
weight hydrocarbons at many
locations throughout the site.
These low molecular weight
hydrocarbons along with olefinic
hydrocarbons have been detected in
soils by previous investigators and
are thought to be the result of
microbial activity (Francis et.
al.,1975). Terpene isomers were
identified at several locations,
however the highest concentration
was observed at the location with
the highest concentration of
methane.
CASE STUDY #2.
(Chlorinated
Solvent Site)
Previous sampling efforts had
determined the presence of various
chlorinated hydrocarbon compounds
in private residential wells (e.g.
1,1-dichloroethane,
1,1,l-trichloroethane and
dichloroethylene). The area of
concern contained two aquifers, a
shallow overburden and a deeper
bedrock aquifer. The residences in
this area were situated along a
north-south roadway with a riverbed
located approximately 100 feet to
the east. Most houses were on the
east side of the roadway with
several businesses located west and
south of the area. A soil gas
survey was conducted to better
define the contaminated areas and
potential sources.
Analysis of the GC/PID field
data did not detect the presence of
the expected chlorinated
hydrocarbons. TABLE 2. contains
the results of both the field PID
142
-------
TABLE 1.
LOCATION
NATURALLY OCCURRING VOLATILE ORGANICS
AT
METHANE SITE — SEPTEMBER 1987
trip EAST OF
blank No. 22 No. 1 No. 11 No. 8 No. 10 TRENCH
chemical name
***********
acetaldehyde (1) ND .013 .008 .Oil
terpene isomer (1) ND .037 ND ND
terpene isomer (1) ND .031 ND ND
methane (2) ND 207 ND 631
Total Volatile
Organics (3) NR 11 0.5 4.0
(1) = GC/MS (in ppm)
(2) = OVA (in ppm)
(3) = HNu Portable Photoionization Detector
background as benzene equivalents)
ND = not detected
NR = no reading
bmdl= present but not quantifiable
*******************
bmdl
.086
.018
>112000
. 4
.037 bmdl
.140 ND
.053 ND
361 >40000
105
NR
( in ppm above
TABLE 2.
NATURALLY OCCURRING VOLATILE ORGANICS
AT
CHLORINATED SOLVENT SITE
SEPTEMBER 1986
G7E1
******
ppb
t***4
1100
ND
67
760
ND
63
140
LOCATION
EW-5
H-16
*************************************
*
chemical name
scan
************************
terpene isomer
terpene isomer
terpene isomer
terpene isomer
terpene isomer
terpene isomer
terpene isomer
319
350
382
394
414
421
458
ppb
******
2000
ND
780
ND
ND
ND
ND
PPb
******
3200
550
2500
ND
160
ND
ND
***********
The GC/MS analysis indicated
the presence of aromatic
hydrocarbons as observed by the
field GC as well as various C6 -
C10 alkanes and C% alkyl benzenes.
The appearance of this mixture of
compounds can be indicative of the
presence of petroleum distillates
and was possible considering the
proximity of one transect (G) to a
defunct gasoline station and the
other transect (R) along the
railroad line. The sample points
that indicated significant HNu PID
readings, but did not show any of
the target compounds by field GC
did contain four major unknowns.
These non-target compounds were
identified in the GC/MS analysis as
several terpene isomers. The
terpene compounds could be expected
in areas where coniferous
vegetation is present. In fact, a
reference was made in a field log
indicating the presence of pine
trees along a portion of the EW
transect. No attempt was made to
further identify the specific
terpene isomers.
143
-------
DISCUSSION
CONCLUSIONS
It is commonly accepted that
naturally occurring volatile
organic compounds can influence
soil gas measurements (see Table
3.) Direct readout field instru-
ments that measure total volatile
organics are likely to include cer-
tain naturally occurring compounds
along with the anthropogenic com-
pounds of interest.
Some portable instruments with
photoionization detectors can be
equipped with specific lamps that
are not sensitive to the presence
of some naturally occurring
volatiles but will include others
if present. For example, an
instrument used at both these sites
was equipped with a 10.2 eV lamp
that does not detect low molecular
weight paraffins (eg. methane,
ethane, propane) but is sensitive
to certain aliphatic aldehydes and
ketones (eg. acetaldehyde,
propionaldehyde and acetone) and
certain olefins (eg. propylene,
ethylene). Some investigators have
found concentrations of ethylene at
greater than 20 ppm in soil gas in
saturated soils under anaerobic
conditions5. A perched water table
was present at the Methane Site
(Figure 2.) at those locations
where the highest concentrations of
methane was observed.
Although the presence of
ethylene was not confirmed by the
GC/MS, soil conditions were optimum
for its production in addition to
the methane.
The presence of terpene isomers
in the soil gas at the Chlorinated
Solvent site in conjunction with
the conifers substantiates the
observations of previous
investigators. More terpene
isomers were observed at this site
than at the Methane site. This can
attributed to the difference in
arboreal species (hardwords at the
Methane site versus conifers at the
Chlorinated Solvent site). Higher
concentrations of terpenes were
observed at the Chlorinated Solvent
site also
The presence of
naturally-occurring volatile
compounds can influence the results
of soil gas surveys. Although each
naturally-occurring compound may be
present in small amounts, when
composited, these compounds can
influence the total volatile
organic level. Caution should be
exercised in using total volatile
organics as an indicator of
pollutant levels present especially
at those sites where naturally
occurring volatile compounds are
abundant. Using total volatile
organic level as a measure of
pollutants in soils should be done
only in those situations where
naturally-occurring volatile
compounds do not influence the
results.
It is especially prudent in
those surveys designed to delineate
groundwater contamination to
incorporate instrumentation that
can differentiate between the
contaminants and naturally
occurring volatile compounds
present.
Determining background levels
of naturally-occurring volatile
compounds is extremely important
for characterizing their potential
influence on the utility of soil
gas results.
ACKNOWLEDGEMENT
The authors express their
appreciation for the technical sup-
port provided by Alan M. Humphrey,
Project Leader for the Chlorinated
Solvent site study and to TAT mem-
bers, Brian McGeorge and Carl Arm-
buster for their analytical sup-
port. They also thank the peer
reviewers; Dr. Thomas Spittler,
EPA-Region I and Dr. Joseph Lafor-
nara, ERT Team Leader for their
helpful comments.
144
-------
TABLE 3. Naturally Occurring Volatiles released, from biological
sources.
SOURCE: Stotzky and Schenck, 1976.
COMPOUND
•:********)!
monoterpenes
(eg. a-pinene,
b-pinene,
limonene
myrcene
camphene
SOURCE
t******>i
conifers (pines)
aromatic shrubs
Artemisia
VAPOR PRESSURE
************
3.85
2. 81
1. 49
1.99
sesquiterpenes
conifers and
deciduous trees
alcohols
(eg. 1-butanol)
2-butanol
isopropanol
t-butanol
ethanol
naphthalenes
formaldehyde
acetaldehyde
propionaldehyde
acetone
ethylene
propylene
formic acid
terpenes
aldehydes ( <9 carbons)
alkanes ( <11 carbons)
(eg. undecane)
alcohols
ethylene
aliphatic aldehydes
and alcohols
methylated heavy metals
eg. mercury, selenium,
tellurium and arsenic
hydrogen cyanide
* vapor pressure expressed in mm Hg as extrapolated from several
sources, (see literature citations).
corn and sunflowers
(anaerobic
conditions)
potato
germinating seeds of
a variety of plant
types
dead woody plants
anaerobic/aerobic
metabolism
fungi (mostly)
some bacteria
5.0
12. 6
26.8
29. 9
38. 6
0.06
3366.0
748. 3
257. 5
185. 6
*******
7600.
31. 8
. 4
145
-------
CI'llD
(1) BOUBLIK, Ttroas, FRIED, Vojtech
and HALA, Eduard. 1984.
Tire Vapor Pressures of Pure
Substances. Elsevier Publishing
Company, New York.
(2) ETMOCIS, A. J.; J.M. DUXBURY and
M. ALEXANDER. 1975.
J. Soil Biology and
Biochemistry.
\fol. 7 pgs. 51-56.
(3) (SEW, Don W. 1984. Perry's
Chemical Engineers' Handbook.
McGraw-Hill Publishing Co. New
York, New York.
(4) WEAST, Robert C. (ed) 1975.
Handbook of Chemistry and Physics.
Tne Chemical Rubber Company. Cleve-
land, Olio.
(5) aUTH, K.A. and S.W.F. RESTALL.
1971. Ihe Occurrence of Ethylene
in anaerobic soil. J. Soil Science
22(4): 430-443.
(6) STOTZKY, Guenther and SCHHMCK,
Susan. 1976. Volatile Organic
Compounds and Microorganisms.
Critical Riviews in Microbiology.
Hie Chemical Rubber Company.
Cleveland, Ohio.
.*.<.' V A'^
MK.C Tt •» i-TL. . X
Figure I Burled Vegetation at Methane Site
146
-------
figure 2 Juiist/rto
Production
DISCUSSION
JONATHAN NYQUIST: Have you noticed when working with high meth-
ane, a small negative peak at the beginning of your Photovac chromatogram?
We've spotted this, and learned that although methane cannot be detected by
a P1D detector, methane is a UV absorber, and it will reduce the response to the
detector while it's going through.
ROYAL NADEAU: In relation to that, the HNU sometimes goes negative
when you are working in areas where methane is likely to be present. This is
similar to your situation, because the same type of detector system is present
in the HNU as is in the Photovac.
JONATHAN NYQUIST: We found that it can actually lower the response of
the HNU, or the Photovac tip to other compounds, because methane becomes
part of the carrier gas and knocks down the instrument response.
ROYAL NADEAU: That's a very good point, and one I didn't emphasize, but
the matter of coelution is very real with these. It's not just a masking that could
take place, if you have a lot of natural compounds present.
JACK McLAUGHLIN: Regarding the picture of performing soil gas in the
snow, is this really practical, because of the cold weather, and the lack of
volatilization of some of the volatile organics?
ROYAL NADEAU: Obviously, working in a snow-covered field, where you
have a frost cover, is not ideal for doing soil gas. It beats the mosquitoes, but
it has its drawbacks.
We have been taking temperatures throughout a lot of our soil gas surveys, and
once you get three or four feet below the frost zone, you will find there is
amazing consistency, that the soil is warmer than you would think.
For instance, I have done soil gas in Colorado, in December and January, and
found it was 54'. So if something is there, you will find some volatilization. Of
course if your probe isn't long enough to get through the frost zone, you will
not be able to collect much of these volatiles in the sample, which will influence
your results, especially if you're trying to track a groundwater flow.
JACK McLAUGHLIN: Are you aware any work done on some kind of
peripheral device that would heat the zone, so that you could actually take a
measurement in the particular zone, if you were down at three or four feet.
ROYAL NADEAU: We have been doing some work in Edison, using
electroprobes to heat up a zone. It takes so much electrical energy since the soils
are such poor conductors, that you'll end up fusing the soils to form silicabefore
you will get much of a zone to sample. I have heard of some people trying to
use steam stripping - injecting steam into the soils with the idea of then trying
to drive out the volatiles — so that we're not limited by the time of the year for
soil gas surveys.
TOM SPITTLER: When the ground is frozen, you actually have conditions
more favorable to doing a soil gas survey, if you keep in mind what the principle
focus of the soil gas survey is - to find out whether the vapor in the soil is higher
at one place than in another, since it reflects the volatile contamination
underground - the underlying material - whether it's a plume or a spill.
With the frozen ground, you basically place a cap on the loss of vapor through
the soil, and as a result, you end up concentrating the vapor in the vadose zone,
so that you can get a more sensitive reading of what's down there on the water
table.
The temperature differences are negligible, because once you're below the
frozen surface, you're not experiencing winter time temperatures, you're
experiencing normal soil temperatures, which don't fluctuate more than five
degrees throughout the year.
So in a sense, it almost enhances a soil gas survey, because you don't get that
constant loss through the porous surface of the soil.
ROYAL NADEAU: In addition, groundwater is usually fairly consistent in
temperature, so if there are volatiles, they will be coming out of the ground-
water.
147
-------
AN IN SITU TECHNIQUE FOR MEASURING SOIL-GAS DIFFUSIVITY
P. M. Kearl1, T. A. Cronk2, and N. E. Korte1
Oak Ridge National Laboratory*
ABSTRACT
Measuring the diffusivity of a gas in the
unsaturated zones has a direct application
to soil-gas surveys and to the prediction of
contaminant movement via vapor phase
transport. This paper proposes an in situ
technique to measure the diffusivity of
selected gases. The method involves an
uncased borehole in which an interval is
isolated using pneumatic packers. A gas is
injected into the interval and allowed to
diffuse into the surrounding rock. Changes
in gas concentration in the borehole are
measured by a downhole nondestructive
method. The equations for describing
diffusion out of a borehole are analogous to
those for heat flow theory. The diffusion
equation with the initial value and boundary
conditions appropriate for this problem are
solved. The resulting mathematical
description of diffusion and the design for
the downhole instrument are presented in
this paper.
INTRODUCTION
Attention has recently been focused on the
role of the unsaturated zone in the storage
and transport of hazardous waste. At sites
where hazardous waste is used, stored, or
disposed of, it is common to find
unsaturated zone monitoring as part of an
early warning system to detect groundwater
contamination. Subsequently, soil-gas
surveys are being used to define the extent
of a groundwater plume at sites where
groundwater contamination has already
occurred.
Vapor transport is an important mechanism
for the transport of contaminants in the
unsaturated zone. Quantification of vapor
transport rates, however, is still limited.
By measuring vapor transport, it is possible
to predict the rate of movement of
contaminants in the unsaturated zone and
select the proper sampling depths and
locations for soil-gas surveys.
Under isothermal and isobaric conditions
gaseous diffusion is expected to be the
major mechanism for vapor transport. This
is because vapor-phase diffusion
coefficients greatly exceed those for the
aqueous phase. Organic vapors in the
gaseous headspace of unsaturated aquifers
are a significant aspect of volatile
pollutant transport (5).
Several methods are available to measure the
diffusion coefficient. Stallman (8) relates
the diffusion of a gas in air to a porous
media using an empirical coefficient related
to tortuosity and the air-filled porosity.
Laboratory measurements of diffusion
coefficients of oxygen in porous media are
discussed in papers by Taylor (9), and
Shearer and others (7). In papers by Raney
(6) and Lai and others (3), measuring oxygen
diffusion in a porous media is extended to
in situ field sites. Raney (6) describes a
diffusion chamber which consists of a probe
inserted into the soil to a depth of 12
inches. A chamber on the probe is filled
with nitrogen, a small port is manually
opened, and diffusion of oxygen from the
soil into the chamber allowed to take place.
At 10-minute intervals a valve is closed and
the concentration of oxygen in the chamber
is measured. Determination of the diffusion
coefficient for flow into a spherical
boundary is described by Crank (2).
Lai and others (3) describe a field method
in which needles are inserted into the soil
and oxygen is injected. The change in
oxygen concentration is determined by gas
chromotography through the same needle after
increasing periods of time. The resulting
spherical boundary value problem is also
solved by Crank (2).
This paper presents a method to measure the
diffusion coefficient of selected organic
gases in an uncased borehole. Unlike
earlier methods which remove a portion of
the gas to analyze the concentration, this
1 Environmental Sciences Division
2Health and Safety Research Division
* Operated by Martin Marietta Energy
Systems, Inc. under DOE contract No. DE-
AC05-840R21400.
149
-------
method uses a non-destructive technique to
measure the in situ concentration of
selected gases. Heat flow equations
presented by Carslaw and Jaeger (1) are
adapted to describe the boundary value
problem of gaseous diffusion from a cylinder
of radius, a, at an initial concentration C0
into the surrounding porous media initially
at zero concentration.
GOVERNING EQUATION
The mathematical representation of the
diffusion of a gas through a porous medium
is analogous to the mathematical
representation of the flow of heat by
conduction through a solid. Both diffusion
and heat flow are described by the same
differential equation. When no sources or
sinks are present, the equation describing
transport is,
conductivity of the medium outside the
cylinder.
y8V2F(x,y,z;t)
where F(x,y,z,t) -
dF(x.v.z:t1
(1)
temperature, in the case of heat flow
concentration, in the case of diffusion
expressed as a function of position and
time,
and, /3 — a proportionality constant
characteristic of the medium.
For heat flow, {> - n, (the thermal
diffusivity of the medium).
And for vapor diffusion, /3-D (the
effective diffusivity of the gas in the
porous medium when the medium is considered
macroscoplcally homogeneous).
It is assumed in this paper that /3 is
constant and is independent of the function
F.
Equation (1), has been solved for a wider
variety of initial value and boundary
conditions in the context of heat flow than
for diffusion. For our application,
solution for heat flow which is analogous to
the diffusion problem is discussed in
Carslaw and Jaeger (1).
For the infinite cylinder of radius a,
consider the region T > a initially at zero
temperature. At r = a, the region is in
contact with a perfect conductor of specific
heat c and mass density p , initially at
temperature V0. The specific heat and mass
density of the medium, r > a, are c and p
respectively. There is no contact
resistance at r — a.
The temperature V(t), in the region T < a is
given as
4aV
-«tu2/a2 du
o uA(u)'
(2)
2c p
m m
-
, . , . .
a. parameter which is twice the
r
s s
ratio of the heat capacities of an
equivalent volume of the medium to that of
the cylinder.
and,
A(u) =
(3)
where J (u) and Y (n) are the ordinary
Bessel Functions.
The solution of the diffusion equation with
appropriate initial values and boundary
conditions for the case presented in this
paper has a similar form and is given as,
C =
-DtuVa2 du
lo uA(u)' (4)
where C = the concentration in the cylinder
as a function of time, D = the effective
diffusivity of a selected gas in the porous
media, a = the radius of the borehole, C0 is
the initial concentration in the borehole,
A(u) is given in Eq. (4), and, a = is a
parameter which is equal to twice the
porosity of the media.
The diffusivity of a selected gas in the
porous media, D, is a function of the
gaseous diffusivity in free air, the
tortuosity of the media, and any transient
storage capacity of the media.
To calculate the diffusivity from test data,
relative concentration values, (C/C ), of
gas in the borehole are plotted as a
function of time on a semi-logarithmic
scale. Curve matching techniques are then
used to match the test data to the type of
curves presented in Figure 1.
All practical precautions should be taken to
ensure that the experimental apparatus
designed to apply this theory to site
conditions upholds the initial assumptions
as closely as possible.
These assumptions are:
• An infinite, isotropic, macroscopically
homogeneous medium.
The diffusivity of the medium is a
constant independent of concentration and
position.
The appropriate initial concentrations.
Diffusion of particles takes place only
through the voids in the medium, r > a.
The solids in the region r > a are
impermeable to diffusion.
The only influence on particle transport
is from the concentration gradient. Air
movement and pressure gradients are
either absent or negligible.
where V = the temperature in the cylinder as
a function of time t, and K — the thermal
150
-------
FIELD EQUIPMENT
It is apparent that borehole diffusion can
be described mathematically and that the
measurement of the parameters used in this
description would be useful for modeling the
transport of volatile contaminants in the
vadose zone. Unfortunately, such
measurements are not presently made in the
field because potentially useful analytical
techniques would cause too much disturbance
to the system. Ideally, the measurement
should be made in a packed-off section of a
borehole as shown in Figure 2. Most
commonly used analytical methods require
pulling a vacuum to collect the sample.
Thus, even if the amount of sample removed
or recirculated in the borehole is small,
the diffusion measurement will be affected.
Commonly available field instruments such as
HNUs, TIPs or non-dispersive infra red
detectors cannot be used without making some
correction for the effect of the measurement
process on the measurement itself.
There are, however, some possible
alternatives. One promising technique is
the use of fiber optic optrodes. These
devices employ a chemical substance at their
tip which is sensitive to the contaminant of
interest. This substance undergoes a
chemical change that can be related to the
concentration of the contaminant.
Unfortunately, the development of such a
tool is in its infancy and the device is not
easily used for chemicals such as
trichloroethene a prime candidate for
these measurements.
Recently, however, Oak Ridge National
Laboratory has developed and licensed a
device called an I-GAS chip (4). This
device consists of a chemical "goop"
sensitive to a particular contaminant. The
heat conducting properties of the "goop" are
affected by the presence of the contaminant.
The change in thermal conductivity is
directly proportional to changes in
contaminant concentration. This system, in
theory, is ideal for performing borehole
diffusion measurements. The technique is
non-destructive, it pulls no vacuum, nor is
air circulated in any way. Only a small
amount of heat is applied to the chip to
make the measurement. We hope to report
actual in-situ borehole diffusion
measurements within a short time. We
believe that these measurements will greatly
aid the development of modeling capabilities
for volatile contaminants in the vadose
zone.
CONCLUSIONS
The diffusion coefficient is important in
predicting isothermal/lsobaric transport of
organic vapors in the unsaturated zone. An
in-situ technique is described which
directly measures the diffusion coefficient
of selected organic gases. This technique
has direct implication for predicting vapor
transport in the unsaturated zone and
assisting in the design of soil-gas
surveys. Future work should incorporate
skin affects, the dependence of the
diffusivity on gas concentration, the
effects of soil water content on the
transport of vapor, and the influence of
different densities on gas movement in the
borehole. In addition, the parameters
influencing the effective diffusion
coefficient should be mathematically stated.
By independently measuring these parameters,
selected unknowns can be evaluated based on
effective diffusion measurements.
REFERENCES
(1) Carslaw, H.S., and Jaeger, J.C.,
Conduction of Heat in Solids. 2d
ed.(Clarendon Press, Oxford, 1986), p.
342.
(2) Crank, J., The Mathematics of
Diffusion. (Clarendon Press, Oxford,
1956), p. 84.
(3) Lai, S. H., J. M. Tiedje, and A. E.
Erlckson, 1976, In-situ measurement of
gas diffusion coefficient in soils.
Soil Sci. Soc. Am. Journal, vol. 40, pp
3-6.
(4) Lauf, R. J., B. S., Hoffheins, and C.
A. Walls, 1987, An intelligent thick-
film gas sensor: development and
preliminary tests. ORNL TM 10402.
(5) Peterson, M. S., L. W. Lion, and C. A.
Shoemaker, 1988, Influence of vapor-
phase sorption and diffusion on the
fate of trichloroethylene in an
unsaturated aquifer system. Env. Sci.
and Tech.. Vol. 22, pp. 571-78.
(6) Raney, W. A., 1949, Field measurements
of oxygen diffusion through soil. Soil
Sci. Soc. of Am. Journal, vol. 14, pp.
61-65.
(7) Sheraer, R. C., R. J. Millington, and
J. P. Quirk, 1966, Oxygen diffusion
through sands in relation to capillary
hysteresis. Soil Sci.. Vol. 101, No.
6, pp. 432-36.
(8) Stallman, R. W. , 1964, Multiphase
fluids in porous media - a review of
theories pertinent to hydrologic
studies. U.S. Geologic Survey. Prof.
Paper 411-E, Washington, D.C.
(9) Taylor, S. A., 1949, Oxygen diffusion
in porous media as a measure of soil
aeration. Soil Sci. Am. Proc. Vol. 14,
pp. 55-61.
151
-------
1.0
0.75
0.50
0.25
-1.0
0.0
1.0
'0910(Dt/a2:)
FIGURE 1 CONCENTRATION IN A CYLINDER OF A PERFECT CONDUCTOR, INITIALLY
AT A CONCENTRATION C IN AN INFINITE MEDIUM INITIALLY AT ZERO
CONCENTRATION. PLOTS ARE §HOWN FOR VARIOUS VALUES OF THE PARAMETER =< .
©
(7) Cond-u.it Pipe, 1 3/8"
^^^ Diameter
(.2) Tigre Tierra Pnuematic
Packers
(3) Perforated Iniection
^ Conduit
@ I-GAS Chip
\5_) Solenoid valve
(J3J Liquid Product
m-
•®
©
©
©
FIGURE 2 ILLUSTRATION OF EQUIPMENT CONFIGURATION FOR IN-SITU
BOREHOLE DIFFUSION MEASUREMENTS
152
-------
DISCUSSION
JOHN EVANS: Can you expand on your argument that sampling points some
distance from the borehole destroys the soil properties in some way. There are
large volumes of soil in between, which is the main diffusivity medium, and
conversely, if that's true, you've also disrupted the soil with the boreholes.
TOM CRONK: Regarding installing sensors in the soil and disrupting the soil
properties, you mentioned that a large distance of soil would exist between the
measurement sensors. That would, in effect, give you a realistic interpretation
of diffusivity. That would be true, but you would have to introduce enough gas
to diffuse through those large volumes you are talking about. In effect, that
wouldn't be a good policy, we feel, because we would be introducing quite a
bit of gas into the soil for a measurement.
Also, if the distance between the measuring devices was indeed large, the time
factor required for the gas to diffuse through that distance would tend to make
the measurements rather long, maybe weeks.
JOHN EVANS: You have disrupted the column by a fairly large diameter hole.
How representative are the walls of the borehole, for example, relative to the
undisturbed soil?
Also, what gas are you using?
TOM CRONK: We're going to test a device with toluene or benzene. Your
point is well taken regarding the borehole. Those are questions that need to be
addressed with further research with this device.
The rate at which soil passes from the borehole into the soil medium will be
affected by the borehole. Perhaps there will be other effects and phenomena.
We will need to make corrections, if we find that the error is substantial. Much
will depend on the different types of soil we are using.
TOM SPITTLER: One of the problems I think you're going to encounter is
the rate and the amount at which soil will adsorb, or absorb -1' m not sure which
process-chemicals like benzene and toluene. How are you going to distinguish
diffusion through the soil from adsorption to the soil?
TOM CRONK: That point has also been considered, and we're going to see
how much error is introduced through actual measurements with our device.
Hopefully, it will be a small effect, maybe even short term, but we can test by
taking measurements first in fresh soil, where the gaseous concentration was
zero. Then, after the soil has been saturated, or has adsorbed as much as it can,
we'll repeat the measurement, and see what different values we get.
TOM SPITTLER: I think you're going to find it's a fairly substantial factor
for things like the aromatic hydrocarbons. You might have to put in a casing if
these boreholes are not stable. Could your theory be modified to include a
screen or casing, so that you would have a two-layer diffusivity equation?
TOM CRONK: Yes. The modification would take place in the porosity
constant, and in effect, would physically change the porosity of the soil at the
boundary, where the diffusion takes place. As long as we knew the grid spacing
of the casing, I think we could incorporate that easily.
JOHN EVANS: There is a lot of literature on bacterial effects wiping out
aromatics in particular. Isn't this a problem over short distances? I have seen
numbers that say you have to be within a couple of feet of the water table to see
any soil gas profile at all for those species.
TOM CRONK: That could be a problem, and we do need to address that with
further research.
JOHN EVANS: Maybe you would be better off using something other than
benzene and toluene.
RICHARD GLANZMAN: How are you differentiating transmissivity due to
a pressure differential from the diffusivity?
TOM CRONK: We're assuming that there are no pressure gradients in the soil.
Your question refers to how we intend to first establish the initial concentration
in the borehole, without causing a pressure gradient.
The first technique utilizes a vial of highly volatile liquid, opened via a solenoid
valve, followed by volatilization in the borehole. We feel that the pressure
gradient introduced from this method will be negligible.
The second method we are thinking about is to inject the gas from a cylinder
and vent the borehole at the same time, in effect holding the pressure constant.
The borehole is full when we get the same concentration out the vent. Both are
then shut at the same time.
PHILIPDURGIN: That was an excellent problem to work on. We're dealing
with that kind of question with underground storage tanks. We're asking, are
the backfilled materials sufficiently porous and permeable to allow gases to
diffuse, so you can have vapor monitoring? It's a good tool to be used in the
future, if it works out.
153
-------
A Field Method for Determination of Volatile Organics in Soil Samples
Thomas M. Spittler and Mary Jane Cuzzupe
USEPA Regional Laboratory
60 Westview St.
Lexington, MA 02173
J. Tyler Griffith
Goldberg, Zoino & Assoc.
27 Naek Road
Vernon, CT 06066
Inc.
With the growing concern for leakage from
underground storage tanks and the realiza-
tion that such leaks can be a serious
threat to water supplies, there has arisen
a need for better and more rapid methods
to assess such problems. The use of soil
gas analysis coupled with various field
monitors, notably portable gas chromato-
graphs, has given investigators a power-
ful tool to assist in locating and pin-
pointing leaks. Once a leaking tank has
been identified, there is the question
of how much contamination is residual
in the surrounding soil. Several at-
tempts have been made to develop a method
for determining soil contamination by vo-
latile solvents. At the present time,
none is truly satisfactory as a field
method, though several show some promise.
Following a suggestion from our lab, Po-
jasek and Scott developed a surrogate
screening method in 1981 (1). They took
an aliquot from collected cores and placed
it in a tared VGA vial, added 10 ul of 1%
HgC12 to retard biodegradation and filled
with organic free water from a buret.
The volume of water was recorded for dilu-
tion calculations. Samples were stored
?t 4 C till analyzed. Before analysis,
the sample was homogenized ultrasonically
and ten ml of aqueous solution was with-
drawn into a 40 ml VOA vial. After 30
minutes for equilibration, headspace an-
alysis was performed.
In 1983 we suggested a rapid screening
procedure (2) using a similar but more
field suitable technique. Pretared vials
were spiked with 20 ul of 2% HgC12 and
filled with 20 ml of organic free water.
Into these labeled bottles approximately
10 g of soil was placed while collecting
samples in the field. Where immediate
results are needed, the sample can be
shaken for a few minutes and headspace
analysis performed on the spot. After
postweighing to determine the weight of
soil, calculations can be made on concen-
tration in the original soil.
Based on some results from the Ada, OK
Laboratory, shown the author by Dr- John
Wilson (3), the technique was modified so
that only about one gram of soil was used
in 30 ml of water. This method was dis-
cussed with Tyler Griffith who tested the
technique using typical aromatics from
hydrocarbon fuels (Benzene, Toluene, Eth-
ylbenzene and Xylenes). This investiga-
tion was conducted as Mr. Griffith's MS
thesis at the University of Connecticut
(4).
At about the same time, John Fitzgerald
was developing a field screening method
for gasoline-contaminated media for his MS
thesis at the University of Lowell (5).
Most of this work was done using a port-
able total organic analyser (Hnu PI-101)
and measuring headspace vapor above soils.
He found fair correlation between samples
and discovered that the larger the samp-
ling jar, the better the results. This
was owing to the rather high flow rate of
the Hnu sampling pump, and the consequent
dilution of headspace in small jars.
EPA's Superfund program has published a
method for analysis of volatiles in soil
(SW846, 5030). This method involves col-
lection of the sample in a VOA vial and
return to the laboratory where an aliquot
is extracted with methanol. The methanol
extract is then diluted in water and sub-
mitted for standard purge and trap GC/MS
analysis. We found that this technique
resulted in severe loss of volatiles with
time. For example, TCE levels dropped 80%
after only 1 day holding time and were 90%
lower after 14 day holding at 4°C.
All of the above methods have strengths
and weaknesses. No in-depth study of the
methods has been conducted to date. How-
ever, it is possible to point out the
problems and some possible solutions based
on one or more of the techniques described.
The presentation will discuss in greater
detail the method investigated by Griffith
and show some data, on the problems with
155
-------
5030. The limitations of Fitzgerald's
method are obvious because there is no
possiblity of determining the exact com-
position of a volatile sample using a
total analyser like the Hnu PID monitor.
Pojasek and Scott's technique has some
obvious advantages when a large number of
samples must be screened without the ad-
vantage of Purge and Trap GC/MS equip-
ment .
It is clear from this discussion that there
is still room for method development in
this important field. Rapid and accurate
analysis of field samples is frequently
needed to determine the extent of contami-
nation in order to make decisions on
removal, air stripping or vacuum extrac-
tion cleanup. It is especially necessary
to know when a site is clean enough to stop
remediation and begin site closure. Un-
fortunately, many regulations have already
been put in place to determine "how clean
is clean", but few, if any, can specify
with confidence the methodology by which
field samplers, chemists and site managers
are to make that determination.
REFERENCES
(1) Pojasek, R. B. and Scott, M. F.,
"Surrogate Screening for Volatile Organlcs
in Contaminated Media," Hazardous Solid
Waste Testing: First Conference, ASTM STP
760, R. A. Conway and B. C. Malloy, Eds.,
American Society for Testing and Materi-
als, 1981, pp. 217-224.
(2) Clark, A. E., Lataille, M. and Tay-
lor, E. L., "The Use of a Portable PID
Gas Chromatograph for Rapid Screening of
Samples for Purgeable Organic Compounds
in the Field and in the Lab," SOP for
USEPA Regional Lab, Lexington, MA., June
29, 1983.
(3) Private communication from John 0.
Wilson, Robert S. Kerr Environmental Re-
search Laboratory, Ada, OK.
(4) Griffith, J. T., "A New Method for
Field Headspace Analysis of Soils Con-
taminated with Aromatic Hydrocarbons," MS
Thesis, Univ. of CT, Storrs, 1988.
(5) Fitzgerald, J. J., "Analytical Scre-
ening of Gasoline-Contaminated Media," MS
Thesis, University of Lowell, Lowell, MA,
1987.
DISCUSSION
JOHN EVANS: In some cases, sensitivity is a problem, if you insist on using
GC/MS rather than gas chromatography.
THOMAS SPITTLER: I don't insist on using GC/MS. A GC/MS is nothing
but an expensive gas chromatograph that does a very nice job if you have to get
absolute, incontrovertible identification. But it hasn't got sensitivity. Work is
now being done to get it down to the part per billion level, or below, where we
can get down to the part per trillion level in the field, with an instrument that
costs about five percent as much.
JOHN EVANS: How about other ways of improving the exchange, such as
adding methanol to the water? What is your opinion on direct purge-and-trap
methods for soil, without any addition of solvents, provided you have a mobile
laboratory available.
THOMAS SPITTLER: That's one of the SW846 options, that you can take
the sample, add it to water, and purge it directly. The second suggestion is
excellent - maybe a small methanol concentration in the water solution to im-
prove the speed with which you do the extraction.
Now for the aromatics, let me caution that these are not real-world samples.
These are spikes, and as any good chemist will tell you, a spike is like an
expressway ramp: easy on, easy off. But when you take a sample that's been in
the real world for a year, or a decade, or a half a century, it's a totally different
matter to get it out of the soil quantitatively.
This is a very important point. Do you have to get it out quantitatively? If
rainwater flowing through it for 30 years can't wash it out of the soil, why
should we dig up that soil and extract it with methanol? It isn't going to go
anyplace.
DON FLORY: Have you compared the values that you would get by the purge-
and-trap technique, which is what's called for in SW846 and the CLP for low
levels?
THOMAS SPITTLER: Yes Mr. Griffith did use the standard method. He had
it done at the State Laboratory in Connecticut, and they compared pretty
favorably with his answers, until the samples were held more than three or four
days. And then the state lab results were very, very low, compared to what he
had in those original soil samples. There was a volatile loss.
TOM STOLZENBERG: To address the question about direct purge and trap,
we had a soil contaminated with tetrachloroethylene. We took splits of that
sample, analyzed it by direct purge and trap, in water, and also by TCLP. We got
seven times more tetrachloroethylene by the TCLP leaching method than we
did get by direct purge and trap. In the TCLP leaching method, the soil is
subjected to a much longer period of extraction, so to speak, which we reasoned
had a lot to do with the rate at which the perchloroethylene is emanating from
the soil particle itself.
This leads to the question, did you spike those lab soil samples then look at grain
size distributions and the effects?
THOMAS SPITTLER: The ones with the aromatic hydrocarbons were
spiked. The samples with the TCE were real-world samples. That was the first
set of work that I showed you.
TOM STOLZENBERG: We feel that there are a big differences in the rates
of extraction, depending on whether the soil has been directly contacted with
the PCE itself, or has been exposed via vapor. We feel that it takes a lot more
to extract the perchloroethylene out of soil that has been in direct contact with
the pure product.
THOMAS SPITTLER: You're absolutely right. That points out the need for
some good, sound research on chlorinated solvents. Here's a very simple
experiment you can try, if you get back to your lab, and you have a GC. Take
a drop of trichloroethylene, put it into a bowl or vial full of water, and then take
another bottle and put the same size drop of benzene or toluene or any aromatic
hydrocarbon, and watch them.
If you do the trichloroethylene first, and then go to the benzene, you'll be able
to watch them both. If you do the benzene first, and then the trichloroethylene,
the benzene will be dissolved before you can get back to look at it.
The trichloroethylene will sit there for about three or four hours. A little drop,
far below the solubility limit of any of the chlorinated solvents, goes into
156
-------
solution. You can shake it up, and you may think you put it into solution, but
what you've done is created a suspension of very tiny microdrops. They will
be in true solution by tomorrow morning, but it takes a long time.
The kinetics of solubility pose very interesting problems with chlorinated
solvents. Therefore, when you're trying to extract them from the soil, you're
relying on those kineticss which complicate the issue.
We recognize the problem. What we have to do now is devise a method for
increasing the rate at which this happens. One of the suggestions I heard
yesterday at the meeting was to take a little sonicater with you, with a very low
power requirement. You can plug it into a small inverter or a small power
supply. Sonicate the sample, when you're looking at the chlorinated solvents
in soil. I have a feeling it would work.
BERNIE BERNARD: In the late '60s and '70s, almost identical work was
being done by a group of people in geochemical prospecting for oil. In those
studies, holes were poked in sediments offshore, and soil gases were done
onshore. The same kind of data found here were generated with a couple of
minor differences.
First, the volatile gases would partition into the gas phase on equilibrium, in
relation to their solubility coefficient, so that their actual concentrations as
measured from head space would have to be normalized by the partition
coefficient, namely for benzene. For instance, the gases would partition to the
extent of only 60% to 70% in the headspace, under the conditions you gave.
So the answer you obtained would have to be divided by say 0.6, or 0.65 to get
a true answer. If you strip the headspace off and reequilibrate, you get another
60%, etc.
THOMAS SPITTLER: The headspace principle is to prepare a standard with
the known concentration in the aqueous phase, and then equilibrate the same
way you do the unknown. Then you use the headspace measurement as a
surrogate for the liquid phase concentration. This is a very accurate and precise
quantitative analysis of dissolved solvents, and we've got data to show that this
holds down into the part per billion range for aqueous solvents.
But you've got to take into account the in-rock constant if you're going to use
the concentration in the headspace as the final measure. It's just a surrogate, and
you compare it with the concentration of known standards to calculate what's
in the aqueous phase.
BERNIE BERNARD: We found a particle size effect, that shales and very fine
clays caused the equilibrium not to push toward the headspace nearly as quickly
as sandy material. So in the real world, there may be some effect of particle size.
THOMAS SPITTLER: I'm sure there is that effect, in that the evidence of the
effect is here. But this is not real-world sampling. This is thesis, and it's got to
be tested in the laboratory, where we've got some control.
BERNIE BERNARD: You were commenting on the ultimate sensitivity being
in the mid parts per trillion. Our experience with purge and trap is that we are
trying to achieve several parts per billion level with a five-cc sample, and you're
talking about achieving an order or two orders of magnitude lower than that,
with a 200 mL sample.
THOMAS SPITTLER: Three orders of magnitude.
BERNIE BERNARD: With the same types of detection?
THOMAS SPITTLER: No, this is the photoionization detector, not a mass
spectrometer.
BERNIE BERNARD: I'm talking about photoionization detection in labora-
tories.
THOMAS SPITTLER: You're not talking about Photovac PID, because it's
about two orders of magnitude more sensitive than the conventional PID's that
have been on GC's for years.
BERNIE BERNARD: No, I'm not. I guess I'm referring to the fact that most
labs around the country are having trouble achieving the method detection
limits for method 502.2, which are on the order of 0.01 to 0.05 ppb. They are
having trouble with that. Yet, you're saying you can easily achieve sensitivities
using a hundred-fold less sample size with this particular detection.
THOMAS SPITTLER: Ask a few of the people here with those Photovacs
who are GC practitioners. It's unbelievable what you can do with a good trained
chemist.
UNIDENTIFIED PARTICIPANT: Have you tried thermal desorption?
THOMAS SPITTLER: Yes, it can be done, but it's still going to give you just
a relative measure, and what we're looking for now is some way to get a
quantitative measurement, especially when you're talking about a multi-
million dollar decision about whether to dig it up, vacuum extract it, turn it into
concrete, or any of the other things being done at some Superfund sites, before
we know what the real problems is. That's a serious issue, and a lot of money
goes down the drain on very poor, ill planned technology.
157
-------
SOIL-GAS SAMPLING AT A SITE WITH DEEP CONTAMINATION BY FUELS
H. B. Kerfoot, S. R. Schroedl
Special Projects Office
Lockheed Engineering and Sciences
Company, Las Vegas, Nevada
J.J. D'Lugosz
Advanced Monitoring Division
U.S. EPA Environmental Monitoring
Systems Laboratory
Las Vegas, Nevada
ABSTRACT
Soil-gas sampling and analysis is a
technology finding widespread acceptance
as a preliminary screening method for
delineation of subsurface contamination.
In an arid climate, a complex of
4-21" outside diameter underground
storage tanks were investigated using
soil-gas sampling, EPA method laboratory
analysis of soil samples, and GC/MS
analysis of groundwater samples.
Because of the relatively thick vadose
zone (85' - 90' to the water table) and
low volatility of the fuel contaminants,
there was a significant chance that the
technology would not detect the
contamination. However, the results of
the soil-gas survey reflect steep
hydrocarbon gradients near the tanks and
more gradual gradients some distance
from the tanks. These results, as
paralleled by the soil analysis and
ground water analysis, indicate shallow
soil contamination near the tanks that
leans to deeper ground water
contamination at a distance from the
tanks.
INTRODUCTION
Soil-gas surveying is a technology
finding widespread acceptance as a
preliminary screening method for
delineating of subsurface contamination
(1) and underground storage tank leakage
(2). The technology originated in the
1920's for use in petroleum exploration
(3). More recently, soil-gas sampling
and analysis has been used for
investigation of vadose zone properties
(4).
Environmental application of the
technology was first done in Europe and
more sophisticated sampling and analysis
methods were later applied in studies of
landfill gases and detection and
delineation of ground-water
contamination by volatile organic
compounds (VOCs). The U.S.
Environmental Protection Agency has
funded evaluations of techniques (1, 5,
6, 7) and EPA (8) and private (9, 10)
workers have applied forms of the
technology. Other studies have applied
the technology to determine vadose-zone
transport properties (11). Bulk grab
sampling (1), sorbent grab-sampling
(10), and passive-sampling (6)
techniques have successfully been used
for measurement of VOCs in soil gases.
On-site analysis and remote analysis of
samples have both been performed, and a
wide variety of instrumentation,
including detector tubes, portable
organic vapor detectors, and gas
chromatography with mass spectrometry
(GC/MS) have been used.
Much progress has been made in the
understanding of the factors that
determine soil-gas concentrations in the
area of the subsurface contamination .
Chemical and physical properties of the
contaminants, physical transport of VOCs
from the contamination to the sampling
location, and subsurface fate of the
target VOCs during transport influence
soil-gas VOC concentrations at a given
sampling location (12).
The physical properties of the
contaminant that most affect soil-gas
concentrations are vapor pressure,
diffusion coefficient, and Henry's Law
constant. The vapor pressure and
Henry's Law constant are important for
estimating the magnitude of the vapor
concentration in contact with
contamination from soil or ground water
contamination, respectively. Lower
soil-gas concentrations in contact with
contamination can result in lower
shallow soil-gas concentrations thereby
reducing the sensitivity of the
technology. Lower volatility fuels,
159
-------
such as diesel fuel, can be a problem
for soil-gas surveying technology for
this reason (13). The diffusion
coefficient of the compound directly
influences the vadose-zone transport
rate (see below) and is typically
acceptable for vapor pressures and
Henry's Law constants. The thicker the
vadose zone is at a site, the more
important this physical property is,
because of the increased influence of
transport in shallow soil-gas VOC
concentrations.
Because diffusion is the primary
mechanism for gas transport in soils
(14, 15), vadose zone properties that
affect diffusion are a major factor in
determining vadose-zone transport rates.
Modified forms of Fick's Law have been
derived to describe diffusion in porous
media, and all show that diffusion is
negligible when the connected air-filled
porosity of the medium falls below
approximately 10 percent. This means
that continuous layers of clay and
perched water bodies can serve as
barriers to diffusive transport. As in
all mass transport, VOC concentrations
downstream of the slowest step of the
process will be depressed. For that
reason, such situations can create a
situation where the technology is not
applicable.
In this paper we describe a soil-gas
survey for jet fuel with monitoring-well
and soil-boring confirmation of the
results at a site with 85 to 90 feet to
the water table. The contamination at
the site is due to leaky waste-fuel
tanks that contained jet fuels that are
dominated by low-volatility
hydrocarbons. Because of the relatively
thick vadose zone and low volatility of
the fuel contaminants, there was a
significant chance that the technology
would not detect the contamination. We
discuss our findings and what level of
contamination the technology was able to
detect.
SITE DESCRIPTION
The site used in this study contains
four 21-foot diameter concrete
underground storage tanks (numbers 22,
23, 24, and 25) arranged linearly and
separated by 3 feet. Between 1946 and
1974, the site had been operated as a
heating oil storage and pumping station
and since 1974 as a solid waste and
waste fuel storage area. The tanks, as
well as waste drums and crates are
enclosed within a 100-foot-wide by 200-
foot-long, 5-foot-high chain-link fence.
Near the fenced area, three monitoring
wells have been drilled to the water
table (85 - 90 feet). Figure 1 provides
a detailed map of the site indicating
concrete tanks and monitoring wells and
Table 1 provides a description of soil
types encountered in the soil borings.
METHOD
The data compared in this study was
obtained using three different
techniques, including soil-gas grab
sampling with on-site analysis,
laboratory analysis of soil samples
using EPA approved methods, and review
of monitoring well data. Each technique
is described in more detail below.
Soil-Gas Sampling and Analysis
Soil-gas measurements consisted of 66
sampling points in and around the
compound (see Figure 2), determined as
the survey developed. Soil-gas samples
were taken from a depth of 7 feet using
0.25-inch OD/0.125-inch ID stainless
steel probes approximately 7.5 feet
long. Probe emplacement involved
hammering a 1-inch driving bar into the
ground to a depth of 7 feet, removing
the bar, inserting the probe in the
void, and backfilling with native fine-
grained material. With the probe in
place, a manifold with septum was
attached as was a 100-cc MSA Samplair
vacuum pump. Samples were taken through
the septum using a clean 1-cc Hamilton
Gastight syringe after 200 cc (two pump
volumes) was purged through the probe.
The sample was taken immediately to the
field trailer for analysis. Analysis
was by either Shimadzu GC-3 gas
chromatography/flame ionization detector
(GC/FID) for total hydrocarbons or AID
Model 511 has chromatography/electron
capture detector (GC/ECD) for the
chlorinated compounds 1,1,1-
trichloroethane (TCA), trichloroethene
(TCE), and tetrachloroethane
(perchloroethylene or PCE). The
Shimadzu was operated with a 6' x 1/8"
SS column packed with 10% SP-100
chromosorb at a temperature of 65°C and
a 200°C injector and detector
temperature. The AID operated with 6" x
1/8" SS 0.1% AT-1000 Graphpac column.
Column and detector temperatures were
both 105°C and the injector temperature
was 112°C.
160
-------
Soil boring and sampling were done at 7
locations in and around the area of
investigation (locations A through G in
Figure 3). A conventional auger was
used to drill each boring to a depth of
20 feet. The locations of these borings
were chosen to investigate the areal
extent of shallow (<20 feet) soil
contamination at the site. Borings A -
C were intended to confirm the
indication, based on the soil-gas data,
of shallow soil contamination near the
tanks. Borings D, E, F, and G were
intended to confirm the boundaries of
shallow horizontal migration as
indicated by the soil-gas results.
Samples were taken at 5 feet, 10 feet,
15 feet, and 20 feet using a spit-spoon
sampler containing 12 1-inch long brass
sample rings. Each split-spoon sample
was divided into four equal portions,
one for soil identification and three
for laboratory analysis. Each of the
splits was wrapped in aluminum foil and
duct tape to contain volatiles.
Monitoring Well Data
Monitoring wells were drilled by James
M. Montgomery Engineers to a depth of
110 feet. The number of wells drilled
was predetermined independent of any
subsurface contamination data and well
locations of one up-gradient and two
down-gradient from a potential
contamination source were chosen on the
basis of past experience.
RESULTS AND DISCUSSION
Soil Gas
The data gathered during the soil-gas
survey included the characterization
samples as well as duplicate samples,
serial samples, and temporal variability
samples used for comparative analysis.
Table 2 provides the temporal
variability data as an example of the
type of reproducibility encountered.
The results of the comparative analyses
between these types of samples indicated
an order-of-magnitude delineation of
soil-gas concentrations was a
representative indication of soil and
groundwater contamination by
hydrocarbons. Figure 4 shows this
order-of-magnitude interpretation as
isoconcentrations of the soil-gas total
hydrocarbon results. Further analysis
of this interpretation shows horizontal
concentration gradients close to the
tanks are large, consistent with shallow
soil contamination as the source of the
VOC vapors. The lower gradient pattern
of soil-gas total hydrocarbon
contamination that extends generally to
the east/east-northeast is in the
general direction of the local ground-
water gradient and is consistent with
contamination located much deeper.
There is also an indication of a small
lobe of elevated soil-gas concentrations
to the south of tanks 22 and 23,
suggesting some southerly horizontal
migration of fuels there.
Figures 5, 6, and 7 show chlorinated-
hydrocarbon soil-gas isoconcentration
contours for tetrachloroethene (PCE),
1,1,1-trichloroethane (TCA), and
trichloroethene (TCE). Both the TCA
and PCE spatial distributions are very
similar to the total hydrocarbon spatial
distributions. However, for TCE, the
pattern of soil-gas concentrations is
quite different, showing the major
amount of detectable TCE concentrations
near and south and east of tanks 22 and
23 .
Soils
Tables 1 and 3 list the results of soil
analyses for selected volatile organic
compounds and total hydrocarbons.
Analysis of these results show two
distinct trends; i) contamination, when
present in the 20 foot boreholes, is
found only in the boreholes nearest the
storage tanks (boreholes A, B, and C),
and ii) the contamination profiles for
the boreholes show both increasing and
decreasing contamination with depth.
These trends are consistent with the
soil-gas trends in showing shallow near-
tank soil contamination and deeper
(beyond 20 feet) contamination away from
the tanks. However, the added
information of the contamination profile
increasing with depth indicates the
contamination is not merely surface
spill oriented but also involves a
source at some depth beyond 10 feet
below the surface.
Monitoring Wells
Well log data taken from field notes
show the water table to be at
approximately 80 feet with an 8 foot
non-aqueous petroleum layer (NAPL) on
the water table in well 12. HNU
readings were low (0.2 to 2.2) and non
varying in wells 11 and 13 and increased
without interruption in well 12 from 1.4
at the surface to 140 at 70 feet.
161
-------
Table 4 shows the result of analyses of
ground-water samples form wells 11, 12,
and 13. GC/MS analysis of ground water
from wells 11 and 13 indicates low to
non-detect concentrations of VOCs.
GC/MS analyses of the ground water and
of the NAPL showed the presence of C-7
to C-12 compounds in the NAPL and
benzene, ethylbenzene, toluene, and
xylene in the water. Field GC/FID
analysis of the ground water and NAPL
indicates a strong component of pre-
benzene hydrocarbons in both.
Characterization of Contamination at the
Site
As stated earlier, the soil-gas sampling
locations were chosen as the survey
progressed. Although an objective of
the sampling plan was to have less than
1 order of magnitude change of total
hydrocarbon concentration between any
adjacent pair of sampling locations,
this was not feasible near the tanks
because of the very high horizontal
concentration gradients there. In
contrast to the high total hydrocarbon
horizontal concentration gradient near
the tanks as seen in Figure 4, gradual
soil-gas concentrations gradients can be
seen to the east of tank 22 and to the
south of tanks 22 and 23. Evaluation of
the distance-dependence of Fick's Law as
well as past soil-gas investigations
(16) indicate that at sites having a
significant depth to ground water, such
as this site, shallow contamination (ca,
10% of the depth to ground water)
produces a much higher soil-gas
concentrations gradient than is seen
from a ground-water VOC source
characterized by gradual isopleth
gradients.
Using the high vs. gradual isopleth
gradient comparison, initial
interpretation of the soil-gas data lead
to the conclusion that the site contains
shallow soil contamination near the
tanks and a minimum ground water
contamination plume indicated by the
100 mg/m contour in Figure 4. Further
site investigation of monitoring wells,
soil borings, and transport theory
supported this interpretation by 1)
determining the presence of free product
in well 12 which is within the 100 mg/m3
contour and the absence of product in
wells 11 and 13 which are outside of the
100 mg/m3 contour; 2) an uninterrupted
increasing HNU readings from 1.4 to 140
at 70 feet thereby indicating the
groundwater as the only source of
hydrocarbons; 3) the absence of soil
contamination in certain boreholes to at
least 10 feet as a result of the soil
boring analyses indicating a subsurface
source as well as surface spill
contamination; and 4) providing a
mechanism for horizontal migration of
hydrocarbons regardless of the ground-
water gradient through experiments in
test chambers of hydrocarbons in contact
with groundwater (17) thereby explaining
the cross-gradient lobe to the south of
tanks 22 and 23.
Soil-gas surveying for TCA, TCE, and PCE
was done to investigate the possibility
of surface spills of solvents within the
site. The coincidence of the TCA, TCE,
and PCE spatial distributions with at
least part of the petroleum hydrocarbon
distribution, the parallel tendency of
increasing chlorinated hydrocarbon and
hydrocarbon soil contamination with
depth, and the low volatility/high
immobility of chlorinated hydrocarbons
indicating the inability to move on
their own, leads to the possibility they
were present as impurities in the fuel
that leaked.
CONCLUSION
In this field study, several noteworthy
results were obtained. It was shown
that soil-gas hydrocarbon VOC
concentrations can serve as an indicator
of non-aqueous petroleum liquids on a
deep (80 feet) water table. The fact
that steep horizontal contamination
gradients are indicative of shallow soil
concentration above a deep aquifer was
also borne out by results.
In addition, the data indicate that
temporal variability of soil-gas
concentrations over a period of 24 days
did not create significant bias or data-
comparability problems with soil-gas
data at this site. As with any field
method, the data should be considered
preliminary and requiring additional
supporting data from additional
monitoring wells and analyses of soil
borings from more than 15 feet.
ACKNOWLE DGEMENTS
The authors wish to acknowledge the
support and assistance of James M.
Montgomery Engineers.
162
-------
NOTICE: Although this research was
funded in part by the U.S.EPA through
Contract 68-03-3249 to Lockheed
Engineering and Science Company,
Incorporated, it has not undergone
Agency review and does not necessarily
reflect Agency policy.
1. Marrin, D. L. and Thompson, G. M.,
Groundwater. 25(1). 21-27 (1987).
2. Santa Clara Valley Water District,
"Groundwater Monitoring Guidelines",
Santa Clara Valley Water District,
Pub. No. 101 R5358tp, August, 1985.
3. Horvitz, L. , Science. 229 (4716),
821 - 827 (1985) .
4. Weeks, E.P.,Earp, D.E., Thompson,
G.M., Water Resources Research.18(5)
1365-1378 (1982).
5. Kerfoot, H.B., and Barrows, L.J.,
"Soil Gas Measurement for Detection
of Subsurface Organic Contamination",
1986, U.S. EPA, Las Vegas, NV.
6. Kerfoot, H.B., and C.L. Mayer, The
Use of Industrial Hygiene Samplers
for Soil-Gas Measurement,
Groundwater Monitoring and Review.
VI(4), 74-78, 1986.
7. Kerfoot, H.B., Mayer, C.L., Durgin,
P.B., D'Lugosz, J.J., Ground
Water Monitoring and Review. 6,
74-78 (1986).
8. Spittler, T.M., Clifford, W.S.,
Fitch, L.G., In Proceedings of
the Sixth National Conference on
Management of Uncontrolled
Hazardous Waste Sites, Hazardous
Materials Control Research Inst,
Silver Springs, MD, 420-433, 1985.
9. Jowise, P.P., Villnow, J.D.,
Gorelik, L.I., Ryding, J.M.,
In Proceedings of
the Sixth National Conference on
Management of Uncontrolled
Hazardous Waste Sites, Hazardous
Materials Control Research Inst,
Silver Springs, MD, 193-199, 1985.
10. Zdeb, T. In Proceedings of Organic
Chemicals and Petroleum Hydrocarbons
In Ground Water: Prevention.
Detection, and Restoration. National
Water Well Association, Worthington,
OH,
11. Kraemer, O.K., Weeks,E.P., Thompson,
G.M., Water Resources Research,
24(3). 331-341.
12. Marrin, D.L. and Kerfoot, H.B.,
Environmental Science and
Technology, 22 (7) . 740-744 (1988) .
13. Evans, O.D. and Thompson, G.M. In
Proceedings of Petroleum
Hydrocarbons and Organic Chemicals
in Ground Water: Prevention,
Detection, and Restoration, National
Water Well Association, Dublin, OH
(1986).
14. Bruell, C.J. and Hoag, G.E., In
Proceedings of Petroleum
Hydrocarbons and Organic Chemicals
in Ground Water: Prevention,
Detection, and Restoration, National
Water Well Association, Dublin, OH
(1986).
15. Mattes, G., The Properties of
Groundwater. Whitey-Interscience,
New York, 116-117, 1982.
16. Marrin, D.L., In Proceedings of
22nd Symposium on Engineering
Geology and Soils Engineering,
1986
17. Schwille, F., Dense Chlorinated
Solvents in Porous Media; Model
Experiments, Lewis, 1988.
163
-------
56. 50.
. 5(11)
. 25 26
. 18 .51
,3-$,
.57 .59
Soil-Gas Sampling Location
Monitoring Well
Monitoring Well
Figure 1. Site Map
Figure 2. Soil-Gas
Sampling Locations
164
-------
12-rV
H Soil Boring Location
V Monitoring Well
Figure 3. Site Map with
Soil Boring Locations
Figure 4. Total Volatile
Hydrocarbon Soil-Gas Concen-
trations (mg/m3)
165
-------
Figure 5. PCE Soil-Gas
Concentrations (mg/m3)
Figure 6. TCA Soil-Gas
Concentrations (mg/m3)
166
-------
Figure 7. TCE Soil-Gas Concentrations (mg/m )
167
-------
Table 1. Selected Volatile Organic Compounds (EPA Method 8240)
Concentrations (ug/kg)
Sample Soil Type
A10
A15
A20
BIO
B15
C5
CIS
D5
D10
E5
E15
F5
F15
G5
G15
CL w/gravel
CL
ML w/sand & gravel
CL w/sand
SC-SM w/gravel
ML w/gravel
CL
SC w/gravel
CL w/gravel
SC w/gravel
CL w/gravel
CH
CL
CL
CH w/gravel
Toluene Ethyl- Total TCA
benzene Xylenes
160
4600
400
130
10
<10
120
2
<2
<2
<2
<2
<2
<2
<2
<10
7300
240
34
<10
<10
<10
<2
<2
<2
<2
<2
<2
<2
<2
<10
43000
5400
170
63
26
13
<2
<2
<2
<2
<2
<2
<2
<2
<10
<100
<100
130
<10
<10
62
<2
<2
<2
<2
<2
<2
<2
<2
PCE
33
<110
<100
71
20
120
23
<2
<2
<2
<2
<2
<2
<2
<2
Table 2. RESULTS OF TEMPORAL EFFECTS ON SAMPLING PROBES
(Concentrations in ng/cc)
Probe
Location
4
16
17
21
Date
Sampled
3/22/88
4/15/88
4/14/88
4/19/88
4/14/88
4/21/88
4/15/88
4/21/88
Total
Petr. Hydrocs
51,000
80,000
63
13
35
40
50
71
TCA
-
0.58
0.62
0.37
0.26
1.30
1.24
TCE
-
3.13
1.61
0.39
0.12
0.67
1.14
PCE
-
0.1
0.1
0.0
0.0
0. 1
0.1
168
-------
Table 3. Total Hydrocarbons in Soil Samples
Borehole
Depth
Soil Type
Headspace Cone
(ng/cc)
Lab Cone*
(ng/g)
5
10
15
20
5
10
15
20
5
10
15
20
5
10
15
20
5
10
15
20
5
10
15
20
CL 11.8
CL w/gravel
CL 19,500
ML w/gravel 11,400
CL w/gravel 226
CL w/sand 1317
SC-SM w/gravel 17.3
24.8
ML w/gravel 20.8
CL 123
CL 204
SC-SM w/gravel 15.1
SC w/gravel 10.2
CL w/gravel 24.4
CL 146
99.4
SC w/gravel 42.7
SC-SM w/gravel 23.2
CL w/gravel
ML 17.8
CH 20.8
CL w/gravel <20
CL 20.8
CL 17.8
270
480
<100
<100
<100
<100
190
<100
<100
100
270
<100
160
1000
120
<100
<100
1000
G
5
10
15
20
CL
CL w/gravel
CH w/gravel
CL-ML
10.6
8.4
13.6
32.1
320
170
100
100
* Lab analysis used EPA Method 418.2
Table 4. Ground Water and Non-Agueous Petroleum Layer
Analysis Data (values in mg/1)
Compound Tested
Well 11 Well 12 NAPL Well 13
Benzene
Ethylbenzene
Toluene
m,p-Xylenes
o-Xylene
Bromoform
Chloroform
Dichlorobromome thane
0.3
ND
ND
ND
ND
0.3
0.2
0.2
6800
440
9100
770
740
ND
ND
ND
520
820
4000
1200
900
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
169
-------
SOIL GAS ANALYSES TO DELINEATE A PLUME OF VOLATILE ORGANIC COMPOUNDS
FROM A HAZARDOUS WASTE SITE IN WILLIAMSON COUNTY, TENNESSEE
Roger W. Lee
U.S. Geological Survey
A-413 Federal Building
Nashville, TN 37203
Mario Fernandez
U.S. Geological Survey
4710 Eisenhower Blvd. Suite B-5
Tampa, FL 3363A
ABSTRACT
Volatile organic compounds, such as toluene and
chloroethylenes, are known to migrate from under-
ground disposal sites by advection in ground-water
flow and to volatilize into the unsaturated zone
above the water table. The dynamic characteristics
of these compounds permit their detection by proper
sampling and analysis of soil gases in the unsat-
urated zone. The disposal site, where as much as
44,000 gallons of industrial wastes were buried in
pits in 1978, consists primarily of disturbed clay-
rich regolith overlying Ordovician carbonate rock.
In sampling soil gases at the site, a hollow steel
shaft, about 6 feet in length and 3/4 inch in
diameter, fitted with a drill head was used to bore
to various depths into the unsaturated zone. Soil
gases were drawn through ports located behind the
drill head into capillary tubing inside the shaft,
through a septum-sealed port at the top of the
shaft, and into a gas syringe. Where contaminants
were repeatedly encountered, hollow soil probes made
of copper tubing were thoroughly decontaminated and
used to collect the soil gases. Prior to obtaining
samples, blank samples were collected through the
copper tubes and analyzed to insure that the tubes
were free of volatile organic compounds. By use of
the drill to bore a pilot hole, the tubes were
inserted into the soil for gas sample collection.
Detection and analyses for volatile organic com-
pounds were performed by using a portable field gas
chiomatographic system, with photoionization
detection.
Results of the sampling indicated that as many as 20
different volatile organic compounds could be
detected in soil gases above disposal pits on the
site. Plots of locations of volatile organic com-
pounds in soil gas indicated two small plumes that
extend up to about 50 feet beyond the site to the
west and southwest, respectively. No detectable
levels of the compounds were found in water or soil
gas toward the south, in the direction of ground-
water flow in the shallow water-table aquifer.
Key words: soil gas methods, volatile organic
compounds, hazardous wastes
INTRODUCTION
Analyses for volatile organic compounds (VOC) in
soil gases have been described by several re-
searchers (1; 2; 3; 4; 5), and methods developed
have been applied successfully to many case studies
designed to delineate the extent of ground-water
contaminations (3; 4; 6). Thompson and Marrin (6)
have described in detail the field methods and
sampling protocols to properly delineate contaminant
plumes. The purpose of this paper is to describe
the application of some of these field methods and
results of soil gas sampling at a U.S. Environmental
Protection Agency Superfund site in carbonate
terrain in Middle Tennessee (Figure 1).
In 1978, approximately 44,000 gallons of industrial
wastes were disposed in pits on a farm in Williamson
County, Tennessee. The waste products, consisted
primarily of semi-solid adhesive process waste with
some containing solvents, hexane, toluene, chloro-
ethylenes, organic fillers, and water soluble
adhesives (7). These were poured into an open pit
from a former phosphate strip mine, and four
excavated trenches.
Preliminary investigations in 1985 by the State of
Tennessee determined the presence of many of these
organic compounds in soil and shallow ground water
below the site (7). Investigations of the geology
in the area (D.B. Withington, U.S. Geological
Survey, 1988, written commun.) and hydrogeology
(P. Tucci and others, U.S. Geological Survey, 1988,
written commun.), and a remedial action investi-
gation (7) have been completed. These investiga-
tions have indicated hydrocarbon movement of less
than 100 feet in soil and water on the site and in
the immediate area surrounding it.
The purpose for studying VOC in soil gas was to
locate contaminants and delineate plumes emanating
from the disposal site.
HYDROGEOLOGIC SETTING
The surficial geology of the area is composed of 3
to 15 feet of clay-rich regolith, consisting of soil
171
-------
and weathered rock. The underlying correlative
Bigby and Cannon Limestones, locally referred to as
the Bigby-Cannon Limestone, are underlain by the
Hermitage Formation, all of Ordovician age (Figure
2). Below the Hermitage Formation is the Ordovician
Carters Limestone, which is underlain by the Lebanon
Limestone, also of Ordovician age. The Bigby-Cannon
Limestone is exposed on the upper hillslopes, but is
not present nor considered a significant aquifer
near the disposal site. The principal water-bearing
zone affected by the contaminants is the Hermitage
aquifer. The lower part of the Hermitage Formation
functions as a confining unit to ground-water flow
to deeper rocks; thus, the contaminants are confined
to the saturated and unsaturated zones of the rego-
lith Hermitage aquifers (P. Tucci and others, U.S.
Geological Survey, 1988, written commun.). Hydro-
logic and chemical data from observation wells
indicate that contaminants have not moved appre-
ciably in the regolith Hermitage aquifer neither
laterally nor vertically downward to underlying
aquifers.
METHODS
Qualitative analyses of VOC from soil gases were
performed using a portable gas chromatograph
equipped with a photoionization detection system.
Separation of the organic compounds was achieved in
1/16-inch columns packed with SE-30 substrate.
Organic compounds in ground water from the site were
previously identified by laboratory analyses using
gas chromatography coupled with mass spectrometry
(7). Peaks from the field system were tentatively
identified using pure standard samples and comparing
retention times of pure compounds with soil gas
samples. Individual analyses were conducted at
ambient air temperatures (columns were not heated),
which varied from 20 °C to 35 °C. Because retention
times decrease with increasing ambient temperatures,
a benzene standard was periodically injected. The
retention time of benzene was used to calculate
relative retention times of other standard com-
pounds. Relative retention times were reproducible
to within +5 percent error. Most commonly identi-
fied compounds were hexane, trichloroethylene,
trichloroethane, toluene, perchloroethylene, and cis
and trans dichloroethylene. An example chromatogram
is shown in Figure 3.
Sampling of gases from the unsaturated zone was
accomplished using two field procedures. The first
method employed a 6-foot long hollow steel shaft 3/4
inch in diameter fitted with a carbide twist drill
bit. The probe was driven to sampling depth using
an electric drill. After flushing the system with
at least 10 times its volume of soil gas, the gas
was withdrawn through ports located behind the drill
head, into stainless steel tubing 1/32-inch diameter
inside the shaft and into a gas syringe. Samples
were injected into the chromatograph within a few
minutes of collection.
This tool was effective during exploration for VOC
in soil gas, but proved difficult and time-consuming
to decontaminate where large concentrations of VOC
were present in the unsaturated zone. In areas
where VOC concentrations were greatest, hollow soil
probes made of copper tubing were used. The 5/16-
inch diameter tubes were 1- to 3-foot long and
fashioned with a chisel drive point and four perfor-
ations just behind the tip. A pilot hole was bored
to within 6 to 12 inches of the sampling depth
(usually 3 feet). The copper tube was inserted in
the pilot hole and driven into the clay-rich rego-
lith to the proper sampling depth. The flexible
tubing from a peristaltic pump was fitted over the
top of the copper tube and gas was withdrawn through
the length of the copper tube. At least 10 but no
more than 100 tubing volumes of soil gas were drawn
through this system, and the pump was stopped just
prior to insertion of the needle tip of the gas
syringe. The needle tip was inserted near the
connection of the flexible tubing and the copper
tube, and the needle tip pushed past the connection
into the top of the copper tubing. The sample was
withdrawn and injected into the gas chromatograph
within a few minutes of collection.
Because of the persistence of VOC in the soil gas
sampling apparatus, it was necessary to decontami-
nate each part of the sampling system following each
positive encounter with VOC in the soil gas. The
drill probe and copper tubes were decontaminated
with the following procedure—alconox wash,
distilled-deionized water rinse, methanol rinse,
alconox wash, and distilled-deionized water rinse.
The drill probe was further dried by drafting
ambient air through the line using a portable
peristaltic pump. The tubes were washed in the lab
and oven-dried overnight at 150 CC. In the field,
both the drill probe and copper tubes were certified
free of VOC by drafting ambient air through each and
testing each one for VOC contamination, using the
gas chromatograph. Gas syringes were similarly
washed, rinsed, dried, and certified clean prior to
filling with actual soil gas sample. These pre-
cautions were essential to the substantiation of the
presence or absence of VOC at any sampling location,
although some additional field time (about 15 per-
cent) was required by this part of the operation.
This procedure was very successful in decontam-
inating equipment.
RESULTS AND DISCUSSION
Compounds detected in and around the disposal pits
are shown on the chromatogram in Figure 3. The most
commonly identified compounds showed some vari-
ability in relative concentrations, based on mea-
sured peak areas, however the overall pattern of
compounds or "fingerprint" of the contaminants was
consistent in the contaminated areas. On the basis
of data collected at about 60 locations, sampled in
June and August 1987, and data collected from other
site activities (7), the contaminants were found in
the area shown in Figure 4. In general, the highest
levels of contaminants, as determined from relative
responses of soil gas samples in the gas
chromatograph-photoionization system, were above the
disposal pits, and have been reported to exceed
1,000 parts per million as total volatile organic
compounds (7), near the central part of the plume.
Ground water has been shown to be contaminated at
the site in excess of 100 parts per million total
volatile organics, also near the center of the plume
(7) (Figure 5).
172
-------
(1)
(3)
The shape of a VOC plume in soil gas is determined REFERENCES
by many factors including geology, hydrology,
topography, chemical nature of the VOC, quantity of
VOC present, time since disposal, disposal prac-
tices, and the location, size, and shape of the
disposal pits. Two lobes of the contaminant plume
from the disposal pits extend beyond the fence line
to the west. The smaller lobe is in a shallow
depression a few inches deep, which extends due west
from the disposal pits. The larger lobe extends to
the southwest, following the surface drainage
pattern. At the lower end of the plume, a drainage
channel up to 4 feet deep drains into the south
field. At the fence crossing, the south field is
about 5 feet lower in altitude than the site itself,
and the surface drainage is southwesterly. About 20
soil gas samples from the south field (Figure 5)
indicated no VOC present in the unsaturated zone.
Samples from ground-water monitor wells in this area
showed no detectable concentrations of VOC (7).
Movement of the contaminants has been limited to
date, but some movement has occurred, principally
along surface features of the site. These features
may replicate subsurface conditions of fractures or
solution channels in the carbonate terrain, which
may control ground-water flow and transport of the
VOC. In addition to volatilization of VOC into soil
gas from contaminated ground water, an alternate
contaminant pathway is possible. The presence of
the contaminants along the surface drainage channel
may be due to runoff washing contaminants from the
pit areas into the drainage channel, concentrating
them in the soil and soil gas. Furthermore, this
indicates the possibility of transport of the
contaminants from the site by rainfall saturation of
the contaminated soil in the pit areas and overland
transport of VOC through the south field and into
the Little Harpeth River or possibly into other (7)
parts of the shallow ground-water system in the
area. Further work is needed to determine the
relative importance of these two transport mechan-
isms. Future investigation of the environmental
effects of this site will incorporate these aspects
of contamination.
Clark, Arthur E., Moira, Lataille, and Taylor,
Edward L., "The Use of a Portable PID Gas
Chromatograph for Rapid Screening of Samples
for Purgeable Organic Compounds in the Field
and the Lab," Methods for Organic Chemical
Analysis of Municipal and Industrial Waste-
water, EPA-600/4-82-057, 1982, pp. 1-12.
(2) Clay, Paul F., and Spittler, Thomas M., "The
Use of Portable Instruments in Hazardous Waste
Site Characterizations," Proceedings of
Management of Uncontrolled Hazardous Waste
Sites, 1986.
Kerfoot, Henry B., "Soil-Gas Measurement for
Detection of Groundwater Contamination by
Volatile Organic Compounds," Environ. Sci.
Technol., Vol. 21, No. 10, 1987, pp. 1022-1024.
(4) Marrin, Donn L., and Kerfoot, Henry B., "Soil-
Gas Surveying Techniques," Environ. Sci.
Technol., Vol. 22, No. 7, 1988, pp. 740-745.
(5)
Spittler, Thomas M., Fitch, Lester G., and
Clifford, W. Scott, "A New Method for Detection
of Organic Vapors in the Vadose Zone," Proc.
Conf. Characterization and Monitoring of the
Vadose Zone, National Water Well Association,
1985.
(6) Thompson, Glenn M., and Marrin, Donn L., "Soil
Gas Contaminant Investigations: A Dynamic
Approach," Ground-Water Monitoring Review, Vol.
7, No. 3, 1987, pp. 88-93.
Geraghty & Miller, Inc., "Hazard Evaluation and
Remedial Alternatives Study for the Kennon
Site, Brentwood, Tennessee, Volume-1-Hazard
Evaluation," Geraghty & Miller, Inc., Ground-
water Consultants, Oak Ridge, Tennessee, 1987.
173
-------
6°47'22"
86°45'
35°58'
35°57'30'
Location Map
Nashville
Williamson' Study
County area
DISPOSAL SITE
—A' LINE OF SECTION
3A^ OBSERVATION WELL AND NUMBER
39» DOMESTIC WELL AND NUMBER
O HACKETT'S SPRING
23»
(4 7 ,47 A
Base Irom U.S.
Geological Survey
1:24,000, Franklin.
1981
0 0.6 KILOM E T E H
CONTOUR INTERVAL 100 FEET
DATUM IS SEA LEVEL
Figure 1.—Location of study area, disposal
site, and observation and domestic wells.
174
-------
sw
NE
600
2,000 FEET
VERTICAL EXAGGERATION X5
EXPLANATION
[ji-ijgsl REGOL1TH
[—i—|] BIGBY-CANNON LIMESTONE
E-q-3 HERMITAGE FORMATION
fa=id CARTERS LIMESTONE
I LEBANON LIMESTONE
, ARROWS SHOWING APPROXIMATE
'' GROUND-WATER FLOW
DIRECTIONS AND RELATIVE
MAGNITUDE
RAINFALL RECHARGE
Figure 2.—Generalized geohydrologic section of study area.
Cis-Dichloroethylene
Trans-Dichloroethylene
Hexane
1,1,1-Trichloroethane
Trichloroethylene
Toluene
Perchloroethylene
UJ
in
5678
TIME, IN MINUTES
1011
12
13
Figure 3.--Chromatogram of volatile organic compounds from soil gas
at the hazardous-waste site near Williamson County, Tennessee.
175
-------
35°57'30"
APPROXIMATE BOUNDARY
OF VOLATILE ORGANIC
COMPOUNDS IN SOIL GAS
LOCATION OF SOIL-GAS
SAMPLE POINT
DIRECTION OF SURFACE-
WATER RUNOFF
INTERMITTENT STREAM
— »— FENCE LINE
CONTOUR INTERVAL 25 FEET
DATUM IS SEA LEVEL
Figure 4.--Location of soil-gas sampling points and approximate location
of volatile organic compounds in soil gas at the disposal site.
176
-------
35°57'30'
ORGANIC COMPOUNDS
— 10- LINE OF EQUAL CONCEN-
TRATION. IN PARTS PEF
MILLION - Interval
10X
—\ — FENCE LINE
— - INTERMITTENT STREAM
0 300 FEET
0 W02_°./Wo2-1
100 METERS
CONTOUR INTERVAL 25 FEET
DATUM IS SEA LEVEL
Modified from Geraghty and Miller, Inc., 1986
Figure 5.—Total volatile organics in ground water at the disposal site, April, 1986.
177
-------
SOIL-GAS SCREENING: ITS THEORY AND APPLICATIONS TO
HAZARDOUS WASTE SITE INVESTIGATIONS
Lvnne M. Preslo, R. Pavlick and Waller M. Leis
Roy F. Weston, Inc., 1001 Galaxy Way, Suite 107
Concord, California 94520
During the last three to four years, soil-gas sampling has grown from a
virtually unknown, seldomly used, technique, to become one of the mainstays
and an essential tool in the geosciences field for site investigations. Soil-gas
screening, if conducted properly, is a very effective and comparatively inexpen-
sive technique for the following applications:
• Source-area identification: Source areas of volatile chemicals within the
vadose, or unsaturated, zone can be identified using soil-gas techniques.
• On-site vs. off-site sources: Soil-gas can assist in the delineation between on-
and off-site sources.
• Plume-tracking: Soil-gas screening can be used to track plumes of chemicals
within the groundwater, depending upon site conditions.
• Migration of landfill gases: Soil-gas screening can also be used to identify the
type of chemicals present in and the migration patterns of landfill gases (eg.
Calderon Bill Requirements in California).
• Optimize subsequent monitoring points. Soil-gas screening is used to opti-
mally locate, and therefore reduce the total number of, the more expensive and
more intrusive monitoring points (eg., soil borings and groundwater monitor-
ing wells).
The presentation includes the theory behind soil-gas screening, its applica-
tions at various sites and its limitations depending on site conditions.
DISCUSSION
The Preslo et al. paper was not presented at the symposium. The work was
summarized by Thomas Spittler, USEPA Region I. The following discussion
was then held regarding on-going and future efforts in the soil gas area.
HALSTUBER: A point that ought to be emphasized is that we use a Henry's
law constant for the distribution of an organic compound over aqueous
solution, whereas there is no such constant over a soil. You can actually have
orders of magnitude differences between what you would expect, and what you
would get if you spiked a soil today, say with ethylene dibromide, compared to
a soil that has been around for 20 years, where the volatilization of the ethylene
dibromide is extremely low. You can be misled in that area.
1 agree with the value of GC for any kind of specific information. About ten
years ago, I worked for the Geological Survey and dealt with a lot of synthetic
fuel wastewaters. In particular, the natural groundwaters that I dealt with were
about 50 parts per million of organic carbon, with 5% by weight organic
material. They are black in color, yet they have virtually no volatiles whatso-
ever. The GC is blank, almost. GC is incredibly valuable. Anybody doing
specific compound plume following should use it, but I would like to see more
back-up of that kind of work- the total organic carbon -because there are a lot
of nonvolatile things that we might be missing.
THOMAS SPITTLER: Yes. there are a lot of sites where the nonvolatiles far
outweigh the volatile constituents. It's also true that at many of these sites, the
nonvolatiles have very little inclination or capacity to migrate, and the real
problem is the material that has the capacity to migrate that causes problems.
For example, you can dump your oil from your car in your backyard from now
until doomsday, and if your well is 50 feet away, it probably will never get there.
Bui you put two or three ounces of gasoline into the soil, it will be there in the
length of time it takes for rainwater and groundwater to bring it there, and it will
be there for a long time to come.
Soil is a question of the relative impact. We can have sites contaminated with
heavy oil. We can have sites contaminated with PCB's. If it isn't going
anyplace, and if a kid isn't sitting there eating it, it has very little health impact.
But a half a gallon of gasoline in that same site could wipe out all of your
neighbors'drinking water wells in a matter of a year, or two. or three, depending
on the rate of flow.
So the emphasis on the volatiles is fairly well placed. It doesn't mean we are
ignoring ihe others, but in terms of prioritizing and putting our effort where the
biggest payoff is, we've got to get a handle on controlling the volatile
contamination situations, or we're not going to have enough drinking water
around to talk about it.
HAL STUBER: 1 agree thai most of the concern is with the volatiles. the
chemicals that we have been emphasizing the most. They are the most toxic and
have Ihe greatest mobility.
Just a caution - Ihere definitely are types of compounds, polar, very water
soluble organic compounds, that can move in groundwater.
THOMAS SPITTLER: Yes, there are lots of other problems out Ihere. The
problem we face, I guess, is a classical problem of limited resources and
enormous problems. There is a tendency to throw a lot of money al the problem
that's the closest at hand.
HARRY McCARTY: While the oil industry is responsible, in pan, for a lot of
the problems, particularly with underground storage tanks, a lot of work has
been done on soil gas on sniffers, and other techniques for petroleum explora-
tion, even related to oceanographic fields, in terms of off-shore exploration thai
could benefit to this program. Go beyond some of the classical environmental
literature and look at some of the petroleum literature.
Although there is a lot more modern technology, you could apply the same sorts
of techniques using more modem instrumentation. You get a lot more informa-
tion out of it.
The only caveat is that the petroleum literature is notoriously slow in coming
out, so the techniques you read about this month in the AAPG bulletin were
submitted two years ago, and the work was done five years ago. But there is
much that could be picked up, and obviously your people have looked at a lot
of the problems. Some of the solutions may be found in some of Ihe existing
literature as well.
THOMAS SPITTLER: 1 second that. In fact. I've got a package of articles
stacked up on one of my file cabinets thai I've senl out to people all over the
country. The third article in the package is a historical article thai appeared in
Science about four or five years ago. It's basically a review of the whole halo
approach lhat oil companies used for decades to locale pockets of oil and
natural gas great depths in the ground, by simply profiling on the surface, doing
vapor monitoring, and finding that there were patterns that were quasi-circular
patterns (halo effects) with increasing concentration as you moved into the
center. When they got to Ihe center, they drilled right into pockets of oil.
That technology we owe to the oil company, and that's basically where soil gas
got its start.
Indeed, we should have a better exchange of newer and more sophisticated
technologies when they come up, but the same problem exists with all scientific
literature. What's done today will show up three years from now, and what we
need is to know what people are doing today, by more verbal exchange, instead
of waiting around to read about it as if it were medieval history.
JIM BERYA: In some of our soil gas study areas vapors seem to move laterally
above the contamination plume and then collect in pockets. We are driving the
probe further down into the ground, plugging the end up with about four, five
inches of soil, pulling it out, and analyzing thai soil by thin-layer chromalog-
raphy to look for semi-volatiles and nonvolatile compounds. Thai confirms the
soil gas approach, especially when we are working with JP5 and Jet A samples.
THOMAS SPITTLER: Good observation. There are many geological forma-
tions-clay lenses, sand lenses over caliche formations-thai are very hard and
almost impermeable, so that the vapor, when il migrates, sometimes lakes
unusual paths.
179
-------
You'll find all of these phenomena observed one way or another, directly or
indirectly, and people noting that it isn't at all like working out in the middle
of a desert, with a totally homogeneous zone. We're dealing with very nonho-
mogeneous media and problems, and we encounter it at every site. Every site
is unique.
TRACIE BILLINGTON: I have been working on an underground tank from
a gas station, which happens to be 250 feet away from a drinking water supply
well, in an area where the only source of drinking water is groundwater. It's a
lot bigger problem than many people perceive. The State of California does
have an underground tank program, but site investigations are probably bigger
than we can handle.
THOMAS SPITTLER: I spent two hours with a planning commission in
southern Rhode Island, which includes six or eight major towns. The whole
meeting was centered around what these towns can do to protect their own
water supplies.
I started out the session by saying, don't wait for us (EPA). The federal
government will not protect your water supplies. The state government will not
protect your water supplies. We have, in every state in New England, thousands
of towns, each with their own private water supply andmillions of private home
owners with their own water supply. There is no way we can provide ground-
water protection.
We can, however, encourage a strategy that a couple of towns have imple-
mented. Ashton, Massachusetts six or seven years ago bought their own
portable gas chromatograph. They hired a Ph.D. chemist (who was one of the
early proponents of soil gas monitoring,) on a part-time basis to use that gas
chromatograph to do soil gas studies, to study the contamination sources in the
town, to monitor around all the gas stations, underground tanks and the
principal problem in the town, a big chemical company that had contaminated
40% of their town water supplies. This town has been doing that work for six
years. This town knows more about their water supplies, more about where
their future problems are going to come from, and more about what they have
to do to head off those problems than any place else in the country, and they have
done it all on their own. They are the ones who had the water supply problem.
And who is a more logical group to do something about it?
These people can't afford a $100,000 van and a GC/MS, but they could afford
a $10,000 GC and a part-time chemist, and a space for him to work, the water
district headquarters.
It's a perfect example of what can be done. If you want to protect your water
supplies, you'd better get together and start doing something about it yourself.
We can give you the help, we can give you the advice, we can give you the
technology that's available, we can tell you how to do it, and how not to do it,
but we cannot come out and do it for you.
And on the other hand, why should we, if the tools are there, and the willingness
is there, and the capability? Take simple gas chromatography. The simpler the
tool, the more certain it is you're going to come back with an answer. As soon
as the tools get unnecessarily complex (with digitizing, and temperature
programming, and computer interfaces, and a telephone link to send the data
back) it becomes more difficult to get the sample into the instrument and the
chromatogram and standards out.
That's a plea for simplicity, not cheapness. We're talking about good instru-
mentation with a good track record in the field, but we're not talking about
something you can't afford.
In fact, I don't think there is a town in the country that can afford not to do
something about the potential contamination of their own water supplies. This
is not EPA policy, this is what I think we should be doing and encouraging.
JAMES DELAVIN: How do you educate people to do that? How do you get
that kind of information out to people? Can you even consider attending all the
town meetings?
THOMAS SPITTLER: I don't know what it takes, really. But I can tell you
what I do. I talk about water supply problems to dozens of towns in New
England. I have done it in my own town, that has five gas stations sitting only
1,000 yards down gradient. If EPA statistics are correct, that means five leaking
underground tanks.
They have been thinking about doing for a year. I think they are ready to act.
PHILIPDURGIN: The people in EPAregions say they need quality assurance
guidelines, and our response has been totryandsetupaguidancedocumentthat
deals with quality assurance and quality control. We're going to try to keep it
simple, but it will provide some guidance for people who want to have some
quality assurance and want to use it in court cases.
THOMAS SPITTLER: Chemists have been doing quality assurance ever
since they have been analyzing samples. If you have a good chemist, and he
knows what he's doing with the gas chromatograph, he will prepare required
standards, and practice quality assurance.
This is not difficult technology. This is relatively simple gas chromatography
that requires careful work, good quality control, and proper standardization.
Those are not difficult things to achieve. A good guidance document would be
most welcome.
180
-------
ATMOSPHERIC ANALYSIS BY OPEN PATH INFRARED SPECTROSCOPY
Philip L. Hanst
Infrared Analysis, Inc.,
1424 North Central Park Avenue
Anaheim, CA 92802
ABSTRACT
The infrared absorption spectrum of the atmosphere
has been recorded with open air paths of 90 meters
and 720 meters. A 23 meter long multiple-pass
cell was used with a Digilab FTS-40 spectrometer.
An MCT detector was used, and spectral resolution
was 0.5 cm"-'-. It is shown that even in the
presence of the strong absorption patterns of
water and C02, pollutant gases may be measured
at mixing ratios as low as 10"° (1 PPB).
Pollutant gases distort the absorption spectrum
of clean air in small but recognizable ways.
These distortions are calibrated for quantitative
analysis by means of digitized reference spectra.
Examples are given for carbon dioxide, methane,
nitrous oxide, carbon monoxide, ammonia,
difluorodichloromethane, benzene, butane, ethylene,
formic acid, formaldehyde, isoprene, methanol,
nitrogen dioxide, nitric oxide, nitric acid,
sulfur dioxide, ozone and acetone.
INTRODUCTION
The present work is a continuation of long path
infrared studies of the atmosphere that have been
carried on intermittently for many years.
Previous work using long folded optical paths is
described in references 1 and 2. These references
contain additional citations of previous work.
In earlier work, an enclosed optical path was
nearly always used. An exception was the outdoor
long path studies by Herget, who used a single
long path between large 30-inch transmitting
and receiving telescopes. (Reference 3) We now
report on open path studies using a three-mirror
multiple-pass optical system (White cell), with
total optical paths up to one kilometer. At a
chosen pathlength, a multiple-pass cell can
transmit and re-focus the same amount of energy
as a large pair of telescopes, even though the
cell mirrors are much smaller than the telescope
mirrors. This is the principal advantage of the
three mirror cell, as can be inferred from the
title of White's original paper in 1942:
Long Optical Paths of Large Aperture. (Reference 4)
MEASUREMENT TECHNIQUE
A Digilab FTS-40 spectrometer was used, working
with a resolution of 0.5 cm~^. The radiation
from the scanning interferometer was projected
out the side port of the spectrometer into a
3-mirror multiple-pass cell with a 22.5 meter
base path. The two collecting mirrors were
semi-circular in shape (D-mirrors). They were
cut from a single round mirror of 10 inch
diameter. These mirrors were mounted on a
pedestal placed in the open air. A wooden cover
over the mirrors shaded them from the sun.
The field mirror was of 12 inch width and was
situated close up against the spectrometer.
The cell mirrors were coated with silver,
protected with a ceramic over-coating. This
type of coating has reflectivity higher than
99% throughout the infrared region. The coating
is resistant to tarnish and other types of
corrosion. It is superior to gold both in
reflectivity and durability.
The arrangement of optical components is
diagrammed in Figure 1. The helium-neon laser
radiation that originates in the spectrometer
and is centered in the infrared beam was found
to be bright enough to be seen on the silver
mirrors, even in daylight. This red light was
used for alignment and pathlength verification.
After the infrared radiation had been passed
through the long path cell, it was captured
and directed to the detector by a 3-mirror
transfer optics system mounted on a base plate
in the sample compartment. The spectrometer
could be returned to normal use merely by removing
the transfer optics plate and moving aside the
plane mirror that coupled the infrared beam out
the side-port. Following are some of the matters
considered in the choice of operating conditions.
a. Open Path. If the path is open, there are
no wall effects. Reactive compounds are properly
measured. Photochemical equilibria are not
disturbed. If the air is moving, the trace gas
measurements are averaged over all the air mass
that moves through the optical path during the
181
-------
time of scanning. Previously, the present author
made an effort to surround his light path with
an enclosure—a pipe, or some kind of tunnel.
One reason for this was to allow recording of a
background spectrum when the light path was
evacuated or flushed with nitrogen. Another
reason was to stabilize the air sample so that
refraction effects due to air turbulence would
not cause excessive noise in the spectrum.
A third reason for the enclosure was to permit
filling the light path with water vapor and pure
tank air so that a water reference spectrum
could be obtained. It now appears that, if
necessary, these reasons for enclosing the light
path can be ignored.
b. Lack of Noise from Air Turbulence. When there
are many cell traversals in an open path, air
turbulence causes the red laser beam to move
about randomly. The laser beam image at the cell
exit shows jitter and pulsations. Probably the
infrared image has similar variations. The
spectrum, however, does not seem to show noise
from this image movement. Probably, this discrim-
ination against the "seeing noise" is due to the
high frequencies at which the scan modulates
the infrared signal. These modulation frequencies
are in the kilocyle range; while the "seeing
noise" frequencies are probably ten to one
hundred times lower. Low frequency modulations
would appear as noise only if one were working
in the far infrared.
c. Water Vapor. The concentration of water
vapor in the air will be 1CP to- 10^ times higher
than the concentrations of the trace gases being
measured. Absorption by water vapor therefore
dominates the infrared spectrum. It is customary
for spectroscopists either to remove the water
from their optical path or to prepare a background
spectrum with the same amount of water absorption
as in the sample spectrum. Unfortunately,
working at kilometer pathlengths complicates
preparation of background spectra of water vapor.
With the other infrared-absorbing gases, including
CC>2, one can fill a relatively small absorption
cell with a high pressure of the compound and
match the absorption in the kilometer path of
air. This is not possible with water, because
of its limited vapor pressure. To make a water
reference spectrum for a kilometer path, one
needs to fill the whole path with water vapor
mixed with pure air or nitrogen. This was
attempted in previous studies, but not here.
Subtracting the water lines does not add any
information to the spectrum; it just makes it
easier to read the information that is there.
With the kilometer path subtraction being so
cumbersome, we have decided to dispense with it.
The main thrust of the present work is to show
that one can read the spectrum directly for trace
gases, without removing any water lines.
d. Resolution. At atmospheric pressure, the
width of spectral lines is about 0.2 cm~ .
To see all the detail in the air spectrum
therefore requires resolving power on the order
of 0.1 cm 1. Since the average laboratory FT-IR
system does not do that well, a lower resolution
must be accepted. Lower resolution is also
advantageous in requiring less computer memory
and yielding shorter computation times. In the
present work, we have used resolution of 0.5 cm"-'-
Wg have been able to utilize most of the detail
in the gas phase spectra while working with our
modest-priced instrumentation.
e. Choice of Pathlength. When working in a
spectral region where water and COo do not absorb
strongly—like the region 1200 cm"* to 800 cm"-*-—
lengthening the optical path increases the
measurement sensitivity. When working in a
region of strong water and (X>2 absorption,
however, the path may need to be shortened to
allow transmission of enough energy for a
measurement. For some molecules there is a
choice between using a strong absorption band
that falls in a region of heavy interference,
or a weak band that is in the clear. In the case
of S02, for example, our choice is to use the
weak spectral features at 1130 cm~l with a
maximum pathlength. In the case of N02, our
choice is to use the strong band at 1600 cm"-'-,
but to shorten the path to give about 30%
transmittance at the measurement frequency.
f. Throughput Advantage of the Multiple-Pass
Optical System. Radiation projected by an
optical instrument spreads as it goes out.
The intensity incident on a distant receiver
decreases with the square of the distance from
the source. The three-mirror multiple-pass
cell brings the collecting mirror close to the
source, even though the path is long. This
gives the long path of large aperture, with
high energy throughput.
g. Convenience_pf the Multiple-Pass Optical
System. The use of a multiple-pass cell allows
the transmitter and receiver to be together,
as in the present work, where they were part
of a commercial spectrometer. The field mirror
is mounted with the spectrometer. The objective
mirrors are set up as a separate portable unit.
A kilometer path can be set up in a room—or
on a roof—or in a parking lot. If the
spectrometer unit is in a van and the objective
mirrors are on a tripod, the whole system can
easily be transported from place to place.
h. Choice of Detectors. Any gain in detector
sensitivity is equivalent to an increase in
pathlength. In spectral regions of strong
interference from water or carbon dioide, a
gain in detector sentivity is better than an
increase in pathlength. A spectrometer system
for trace gas analysis should be equipped with
a nitrogen-cooled photo-detector of the highest
available sensitivity. In the present work,
a "wide-band" mercury-cadmium-telluride detector
was used. For compounds whose analytical bands
fall in the high frequency region—like HF,
HC1, and I^CO—an indium antimonide photo-
detector should be used.
182
-------
i. The Background Spectrum. Computer manipulations
of the data require that the spectra be in
absorbance form. To make an absorbance plot, one
needs a background spectrum. In gas studies, the
spectrum of the empty cell usually serves as
background. Since the open path cell cannot be
emptied, we had to resort to "shorting out" the
optical system to provide a background spectrum.
A pair of plane mirrors did this. One mirror at
the entrance to the cell sent the infrared beam
across to the other mirror at the cell exit.
This second plane mirror sent the infrared beam
to the detector. When these background spectra
were used with the sample spectra obtained through
the multiple-pass cell, the resultant absorbance
plots were quite flat across the whole spectrum.
This proves that the silver mirrors with their
ceramic over-coating do not have any dips in
reflectivity in the spectral region used.
j. The Use of Digitized Quantitative Reference
Spectra. Iri recent years the major improvements
in infrared technique have come through the
computer software. Thirty years ago when the
author was engaged in infrared studies of the
Los Angeles smog, a transmittance spectrum was
obtained from point-to-point hand measurements
of the empty cell spectrum and the sample spectrum.
The measuring and re-plotting for one trans-
mittance spectrum would take all afternoon. The
computer now does this in a second, or less.
The software capabilities now include automated
quantitative analysis. For this, one needs
digitized quantitative reference spectra, which
have not generally been available. Currently
available collections of infrared spectra are
mainly designed for identification, not quant-
itation. For quantitative analysis of gases
one needs to allow properly for the effects of
narrow line widths which lead to deviations
from the logarithmic absorption law. In order
to avoid those deviations when working at modest
resolution, the reference and sample spectra
must be used only in the low absorbance region
(0.1 or less, for compounds whose spectra have
single lines not fully resolved). For the
present work a collection of quantitative
reference spectra was prepared in digital form
using the same instrument that was used for the
long path studies. When these spectra are used
at low absorbance, the logarithmic absorption
law is always obeyed.
THE ATMOSPHERIC TRANSMISSION
Shown in Figures 2 through 5 are the transmittance
spectra obtained at a 90 meter path (4 traversals)
and a 720 meter path (32 traversals). These
spectra were recorded on July 3, 1988, which was
a warm but only moderately humid day. These
spectra show which spectral regions are available
for measurements and which are not.
In Figure 2 we see that even for the shorter
path, the region between 3900 cm"-'- and 3550 cm"1
does not transmit enough energy to be of any use.
This is the region of the OH bands. Thus alcohols
and acids must be detected by bands other than
those involving the OH stretch. Likewise, Figure 4
shows that the carbonyl region between 1800 cm"1
and 1610 cm"1 offers practically no energy, even
at the 90 meter path. The detection of carbonyl
bands in open air studies is therefore out of the
question. This is especially unfortunate, because
many strong and characteristic molecular bands
fall in the carbonyl region. The region 1580 cm"1
to 1400 cm"1 is also practically useless, but there
are not very many important bands here, so this
is not a serious loss.
The plots show that the regions 3200 cm"1 to 1800
cm"1 and 1400 cm"1 to 700 cm-l are the prime
spectral regions for open path gas measurements.
It will be shown that these open regions of the
spectrum reveal bands and lines for almost every
polyatomic and hetero-nuclear diatomic molecule
in the air. When the revealed band is a strong
one, the detection capability extends down to the
level of a few parts-per-billion, or lower. When
the revealed bands are weaker, the detection limits
are correspondingly higher.
EXAMPLES OF TRACE GAS MEASUREMENT
Carbon dioxide is a minor constituent of the
atmosphere, but its normal mixing ratio of 340
parts C02 per million parts air (PPM) puts its
concentration some 200 times higher than the
concentration of the next trace gas. Next is
methane, at about 1.6 PPM. After methane comes
nitrous oxide at 0.3 PPM and carbon monoxide
at about 0.16 PPM. Figures 6 and 7 show some of
the C02, CH4, N20 and CO lines that may be used
for direct measurements in an open atmospheric
path. The lines are of course interspersed with
water lines; but some are in the clear andean be
used for quantitative measurements. These
recommended lines are marked by arrows.
In order to verify our 720 meter pathlength and
also to test the validity of our quantitative
reference spectra, we have calculated from the
spectrum the concentrations of the four naturally
occuiring molecules mentioned above. To do this
we used the Digilab software subtraction routine.
To measure a compound, the sample spectrum and the
reference spectrum were displayed, and the
interactive subtraction was carried out. When the
band being displayed was removed from the sample
spectrum, the amount was readily calculated from
this formula:
cone.-path
product for
reference spec
subtraction concentration
X factor on = of compound
screen in sample
720 meters
The measured values are tabulated below.
Compound
C02
CH4
N?0
CO
Normal Background Measured
(Parts-per-million) Amount
340 350
1.6 1.9
0.30 0.29
0.16 0.27
183
-------
The above measured values appear to be an
adequate verification of our measurement method.
It is especially important to obtain a nearly
correct value for N20, which is not a pollutant.
Previous studies have shown that the N20 concen-
tration never deviates measurably from its
background value.
The low measured value of carbon monoxide indicates
that the air mass under study was not polluted by
auto exhaust in any significant degree. In an
urban area, the CO concentration will usually be
about ten times higher than the amount measured
here. It appears that on July 3rd, a Sunday,
Ossining was not in a plume of pollution moving
North from New York City. Instead, it is presumed
that the town was immersed in a clean air mass
that had moved in from the West with the passage
of a weather front during the previous day.
To further illustrate the detection of pollutant
gases, we present Figures 8 through 17. In each
figure the bottom plot is a portion of the atmos-
pheric spectrum for July 3, 1988. Above is a
reference spectrum of the constituent under
consideration. The third spectrum presented in
each figure is a synthetic spectrum obtained by
adding together the atmospheric spectrum and the
reference spectrum weighted to correspond to the
presence of the indicated number of parts-per-
billion (PPB) of the compound. These synthetic
spectra therefore show the distortions of the
clean air spectrum that are indicative of the
presence of pollutants. Particular lines or
bands that may be used for quantitative analysis
are marked by arrows. These examples cover a
number of important molecules, but the set is
far from complete. In the future we expect to
expand the set of examples to include halogen
acids, nitrates, peroxides, nitriles, nitros-
amines, alkenes, alkynes, hydrides, organo-
metallics, and other groups of compounds.
Discussion of individual cases follows.
Ammonia. Figure 8-left shows that ammonia has
several lines and bands favorably located for
detection. The two features marked in the figure
are probably the best choices for quantitative
analysis. Measurement sensitivity is high.
The July 3rd spectrum does not show any absorption
attributable to ammonia. Comparing the synthetic
spectrum with the real spectrum puts the detection
limit at 1 or 2 PPB.
Dichlorodifluoromethane. Figure 8-right shows
that there is a very strong spectral feature for
CF2C12 at approximately 1161 cm"-'-. There are
some weak N20 lines in this region, which are
barely resolved. We cannot see the CF2C12 that
is present in clean air at about 0.4 PPB. The
synthetic spectrum shows that 3 PPA of the
compound could easily have been detected.
Probably, high spectral resolution would be help-
ful in detecting the compound at less that 1 PPB.
Benzene. The strongest feature in the benzene
spectrum falls at 674 cm"-'-. This is a region of
strong C02 absorption and therefore one might
consider using a weaker benzene band that falls
in a region with less interference. There is such
a band centered at 1037 cm" , but it turns out
that this band gives maximum absorbance only 2%
as great as the maximum absorbance in the band
at 674 cm . This is a case where it is best to
choose the stronger band and minimize interference
by backing off on path length. Figure 9 shows the
case for 90 meters of air. There is enough
transmission between the C02 lines to allow
detection of the benzene. The lower spectrum
shows that the three absorbance minima centered
around 674 cm"1 line up nicely in the absence of
benzene. In the synthetic spectrum we see that
200 PPB of benzene clearly distorts the pattern
in a way that can be used for quantitative
measurement.
Butane. Methyl and methylene groups in organic
molecules absorb in the region 3000 cm"-'- to 1850
cm"-'-—the C-H stretch region. At a 720 meter
path the clean air spectrum shows a weak C-H band
due to the organic matter. Figure 10 shows how
the absorption would be increased if 200 PPB
of butane were added to the air. From this
synthetic spectrum we determine that the amount
of C-H absorption in the July 3rd atmospheric
spectrum was equivalent to the absorption by
about 50 PPB of butane.
Ethylene. The strongest feature in the ethylene
spectrum falls at 950 cm"-'-, about one cnT^ to the
side of a water line. At a resolution of 0.5 cm"l
a few PPB of ethylene will reveal their presence
as a "shoulder" on the water line, as seen in
Figure 11-left. In the July 3rd clean air spectrum
there is a small shoulder at the bottom of the
water line, but this is probably not due entirely
to ethylene. There are also weak C02 lines in
this region, one of which probably contributes
to this shoulder. Higher resolution would separate
the lines better.
Formic Acid. Formic acid reveals itself in a
shoulder on a water line near 1105 cm~l (Figure
11-right). This molecule is a product of the
atmospheric photochemistry. The 1105 cm"! band
is seen clearly in the spectra of the Los Angeles
smog. It has also been observed in spectra
recorded through the stratosphere.
Formaldehyde. The C-H stretch band of formaldehyde
is rich in lines and falls on the low frequency
side of the C-H bands of most other molecules.
At a 720 meter path many weak water lines overlap
the formaldehyde spectrum, as shown in Figure 12.
At least five formaldehyde lines fall between
water lines and may be used for measurement. These
are marked by arrows. The use of higher resolution
and the use of an InSb detector would increase the
measurement sensitivity for formaldehyde.
Isoprene. The diolefin isoprene (CjHo) shows two
strong spectral features near 900 cm"", one of
which is in the clear—Figure 13-left. It is
important to be able to measure isoprene in the
open air because this compound is released into
the atmosphere in large quantities by living and
decomposing vegetation.
184
-------
Methanol. Methanol, Figure 13-right, is measured
with good sensitivity by its sharp spectral feature
at 932 cm"1. Because of the use of methanol in
gasoline, the compound is now regularly detected
in the urban air.
Nitrogen Dioxide. NC>2 measurement is a case where
it is necessary to back off on pathlength so that
its major band may be seen through the water
interference. This is shown in Figure 14-left.
At 90 meters path there is a small "window" at
1600 cm"1 that shows some of the structure due to
the NC>2 band. There is structure in the spectrum
of the atmosphere on July 3rd that corresponds
to perhaps 30 PPB of NO^. The synthetic spectrum
shows the deepening of the structure when 100
PPB of N02 are added. The use of higher spectral
resolution would probably make the detection
of N02 easier.
Nitric Oxide. NO, like N02> must been seen
through water interference. With NO, the inter-
ference is less, so the 720 meter path has been
used. The absorption coefficient for NO is much
smaller than the absorption coefficient for
N02> so that the detectability levels of the
two molecules turn out to be about the same.
Figure 14-right shows the appearance of an NO
"line" at 1900 cm-1. This line is in fact an
unresolved doublet, so here again is a case
where higher resolution would benefit the
detection.
Nitric Acid. For nitric acid measurement we
choose spectral features of moderate strength
that fall in the region 900 cm"1 to 880 cm"1
(Figure 15-left). There are stronger features
in the nitric acid spectrum, but they are not
in the clear. The arrows indicate two features
suitable for quantitative analysis. After
seeing the distortion due to 20 PPB of HN03,
one can look back at the "clean air" spectrum
of July 3rd and conclude that it contains
absorption due to approximately 5 PPB of HN03.
Sulfur Dioxide.
Unfortunately, the strongest
centered at 1360 cm"1, is well into
band. In measuring S02 we must
S02 band,
the main water
therefore rely on the relatively weak spectral
features between 1200 cm"1 and 1100 cm"1.
Bands useful for quantitative analysis are
marked in Figure 15-right. The amount of S02
used for the example was relatively high—
500 PPB. Increases in sensitivity are available
from using a long scanning time, a longer path
and a more sensitive "narrow band" MCT detector.
Ozone. Ozone detection is straightforward,
as shown in Figure 16. The clean air spectrum
of July 3rd showed absorption due to about
30 PPB of ozone.
Acetone. The last example—for acetone—
Figure 17. shows that compounds with broad
bands are detectable, but generally with
less sensitivity than compounds with sharp
features. Acetone is a poor case for infrared
open path work because its very strong carbonyl
band is hidden by water.
ACKNOWEDGEMENT
The assistance of Mr. Jeffrey D. Bernson in
the experimental portion of this work is
gratefully acknowledged.
REFERENCES
1. Hanst, Philip L., Wong, Ngai Woon, and
Bragin, Joseph, "A Long Path Infrared-red Study
of Los Angeles Smog", Atmospheric Environment",
Vol. 16, No. 5, 1982, pp. 696-981.
2. Tuazon, E. C., Graham, R. A., Winer, A. M.,
Easton, R. R. , Pitts Jr., J. N., and Hanst, P. L.,
A Kilometer Pathlength Fourier-Transform
Infrared System for the study of Trace Pollutants
in Ambient and Synthetic Atmospheres",
Atmospheric Environment", Vol. 12, No. 4,
1978, pp. 865-875.
3. Herget, W. F., "Air Pollution: Ground-based
Sensing of Source Emissions", in Ferraro, J.,
Basile, L. (Eds), Fourier Transform Spectroscopy,
Volume 2, Academic Press, New York, 1979,
pp. 111-127.
4. White, J. U., "Long Optical Paths of Large
Aperture", J. Opt. Soc. Am., Vol. 32, 1942,
pp. 285-288.
185
-------
FTS-40
Optical Bench
MCT
detector
V
Flip mirror
Side port
Field mirror
Removable base plate
with transfer optics
Objective
mirrors on
portable
pedestal
Figure 1. Optical system for open air long path infrared spectroscopy.
36OB.e 3550.1
aim.I
UAVDUCERS
Figure 2. Infrared transmittance of air at Ossining, New York - July 3, 1988.
Digilab FTS-40 spectrometer - MCT detector - 0.5 cm~ resolution - 30 minute scan.
186
-------
270B.0 26S0.I
265H .0 2600.
2400.0
WAVEMJMBERS
Figure 3. Infrared transmittance of air at Ossining, New York - July 3, 1988.
Digilab FTS-40 spectrometer - MCT detector - 0.5 cm resolution - 30 minute scan.
Figure 4. Infrared transmittance of air at Ossining, New York - July 3, 1988.
Digilab FTS-40 spectrometer - MCT detector - 0.5 cm resolution - 30 minute scan.
187
-------
850.0 800.
500.0 450.0
WAVENUHBER3
Figure 5. Infrared transmittance of air. at Ossining, New York - July 3, 1988.
Digilab FTS-40 spectrometer - MCT detector - 0.5 cm resolution - 30 minute scan.
SUN JUL 03 13107147 19BB
3020.0 3000.0 2960.0 2860.0 2940.0
WAVENUMBERS
Figure 6.IDENTIFYING METHANE LINES - 720 METERS AIR - METHANE REFERENCE
CO Reference
SUN JUL 03 13!07:47 1988
SAMP = 720MJU3S
RES = 0.S
SCANS = 1024
2150.0
WAVENIHBER5
2840.0
Figure 7. IDENTIFYING N20, CD AND C02 LINES - 720 METERS OUTSIDE AIR
188
-------
1.0 950.0
VAVEHJMBERS
SAMP = 720MJUL3 RES = 0.5 SCANS = 1024
ADD 20 PFB AMMONIA TO 720 M AIR
SUN JUL 03 13:07:47 19B8
920.0 1160.0
1160.0
UAVENJHBERS
SAMP = 720MJUL3 RES = 0.5 SCANS
ADD 3 PPB CF2CL2 TO 720 H Alfl
SUN JUL 03 13:07147 198B
1140.1
1024
Figure 8. Ammonia and Difluorodichloromethane.
Benzene reference
73B.B 720.0
SAMP = BBMJUL3
700.0 680.0
WAVENUHBERS
RES = 0.5
660.0 650.1
SCANS = 5B0
90 M AIR - 90 M AIR WITH 200 PPB BENZENE - BENZENE REFERENCE
SUN JUL 03 15:00:26 198B
Figure 9. Benzene
189
-------
3100.0 3000.0 290B.0 2803.0
WAVENUHBERS
SAW = 72BHJU3S RES = 0.S
720 H AIR - 720 H AIR WITH 200 PPB BUTANE - BUTANE REFERENCE
SUN JUL 03 13107147 1B88
2700.0
SCANS = 1024
Figure 10. Butane.
\J
t
V
/
V
i
IT
Formic acid reference
970. B 96B.B 940.0 K
UAVEM>eCRS
SAW = 720HJUL3 RES = 0.5 SCANS = 1024
ADO 20 PPB ETHKLEHE TO 720 H AIR
SUN JUL 03 13:07147 1988
w
V
V
V
u
U
•B 1140.B 1100.0 101
WAVENUHBERS
SAMP = 720MJUL3 RES = 0.5 SCANS = 1024
ADD 40 PPB FORMIC ACID TO 720 M AIR
SUN JUL 03 13107147 19ee
Figure 11. Ethylene and Formic Acid
190
-------
BUT = 72BMJUL3
WAVENUHBERS
RES = 0.5
ADD 100 PPB FORMALDEHITDE TO 720 H AIR
SUN JUL B3 13107147 1968
2750.0 2720.1
SCANS = 1024
"Figure 12. Formaldehyde.
VAVDUCER3
SAW r 7204TUL3 RES = 0.S SCANS a 11124
ADO 40 PPB ISCPRENE TO 720 M AW
SUN JUL 03 13:07147 1966
VAVENJCERS
SAH» = 72BMJUL3 RES = 0.5 SCANS = 1024
ADO 30 PPB HETHANOL TO 720 H AIR
SUN JUL 03 13:07 = 47 1968
Figure 13. Isoprene and Methanol.
191
-------
ttf
Nitrii oxide reference
\J
0.8-
0)
u
Absorbar
y
V
u
II!
11
U
1.0-
u
c
o
w
II 11
V
If
u
\J
1620.0 1600.0 1580.0 \92B.B 1900.0 IBB
VAVEMJHBERS VAVEMJH3ERS
SAMP = 90MJUL3 RES = 0.5 SCANS = 500 s/af = THJHJUJJ RES = a.s SCANS = ,024
Am 102 PPB N02 TO 90 M AIR
SUN JUL 03 15:00:26 1988
ADD 150 PPB NO TO 720 H AIR
SUN JUL 03 I3I07M7 1988
Figure 14. Nitrogen Dioxide and Nitric Oxide.
Nitric acid reference
JU
1
1
1
I
1
V
1
1
1 —
V
V
OJ
u
c
2
u Jt
, -0.01
V
1
U
U
W
l/ii
'V
B8B.0
VAVENUfBERS
> = 72BHJUL3 RES = 0.5 SCANS = 1024
ADD 20 PPB NITRIC ACID TO 720 H AIR
SUN JUL 03 13:07147 1988
860.B 1Z00.0 1150.0 Ml
WAVENUtBERS
SAHP = 720MJUL3 RES = 0.5 SCANS = 1024
ADD 500 PPB SULFUR DIOXIDE TO 7EB M AIR
SUN JUL 03 13:07:47 1988
Figure 15. Nitric Acid and Sulfur Dioxide
192
-------
1100.0
1000.0 900.0 a
WAVENUHBERS
SAhP = 720MJUL3S RES = B.S SCANS = 1024
720 H AIR - 720 M AIR WITH 100 PPB OZONE - OZONE REFERENCE
SUN JUL 03 13107147 1968
Figure 16. Ozone
1300.0 1200.0 1100.0 1060.1
WAVENUOERS
SAMP = 720HJUL3 RES = 0.S SCANS = 1024
720 H AIR - 720 M AIR WITH 500 PPB ACETONE - ACETCJC REFERENCE
SUN JUL 03 13:07147 1988
Figure 17. Acetone
193
-------
DISCUSSION
JOSEPH SOROKA: How careful do you have to be in subtracting your
reference or your clean air spectra from your actual spectra. In work that I'm
familiar with, subtraction has been a very sticky problem especially determin-
ing when you've reached a point where you've subtracted enough and not too
much.
PHILIP HANST: With the interactive subtraction, you have computer control.
You subtract very slowly, and the subtraction factor appears on the screen
continuously, so when you have finished subtraction, you get that number.
Divide it by your path length, and there's the answer in parts per million - that
is, if you have a reliable reference spectrum proper for quantitative work.
There are collections of thousands of infrared spectra, and they're made only
for a qualitative analysis. They're not designed for quantitative work. Always
use the strongest band for your measurement, because that's where you get
sensitivity. If you look at published spectra, the strong bands usually bottom
out. They're too intense, and they deviate from Beer's law for various reasons,
which prohibits using those spectra for quantitative analysis. So you have to
make your own reference spectra and keep the absorbance very low. You use
the MCT detector, so you have a good signal to noise ratio. You don't let the
absorbance of any bands go above say, 0.1. Then you have a quantitative
reference spectrum that you can believe in, which follows Beer's law. You read
the subtraction factor off your screen, and you have your answer.
I have a Digital Lab spectrometer. We're working together, preparing a library
of quantitative reference spectra for gas analysis that are digitized, and
available on floppy disk. You can get them and do the same thing that I've done.
JOE SOROKA: Would you estimate that a different operator on the same
instrument, for example, or on a different instrument, would be able to do the
kind of analysis you are doing relatively easily?
PHILIP HANST: Yes. You wouldn't have to know spectroscopy. I think the
software has made it so easy that as soon as you learn what to look for, anyone
can do it. The quantitative correct answer can be obtained down to a part per
billion level.
JOE SOROKA: Do you feel confident that you've hit most of the known
possible components in the air, that you're not going to get any interferences?
PHILIP HANST: From experience, there are half a dozen important pollut-
ants. Usually, the strong lines of bands are at different places in the spectrum.
Overlap and interference between the trace gases hardly ever comes up. It's the
interference between the water and the trace gases that you're always fighting.
194
-------
DEVELOPMENT OF THE MINITMASS, A MOBILE TANDEM MASS SPECTROMETER
FOR MONITORING VAPORS AND PARTICULATE MATTER IN AIR
Henk L.C. Meuzelaar, William H. McClennen, Neil S. Arnold
Tim K. Reynolds, Wallace Maswadeh, Patrick R. Jones and Dale T. Urban
Center for Micro-Analysis and Reaction Chemistry
University of Utah, EMRL, Room 214
Salt Lake City, Utah 84112
ABSTRACT
A fieldable, miniaturized Ion Trap Mass
Spectrometer (MINITMASS), based on a Finnigan MAT
ITMS system, was developed and tested. In
addition to regular electron ionization, the
MINITMASS is capable of operating in chemical
ionization as well as collision-induced
dissociation modes and features selective mass
storage and axial modulation options.
A specially designed air sampling inlet allows
direct analysis of permanent gases and
condensable vapors. A limited amount of
chromatographic pre-separation can be obtained by
means of "transfer line chromatography".
Furthermore, a novel electrostatic aerosol
sampling inlet has been developed which deposits
air particulate matter on Curie-point pyrolysis
filaments thereby enabling subsequent desorption
of adsorbed volatiles and/or pyroylsis of
nonvolatile organic matter.
The MINITMASS weighs less than 115 kg, uses
approx 1000 W and fits into a single electronic
rack. A 8x7x7 ft mobile laboratory module, which
can be transported with a regular 3/4 ton pick-up
truck, provides access to remote test sites and
enables operation under demanding environmental
conditions. Moreover, the MINITMASS system can
be remotely controlled from distances up to
several miles, thus facilitating operation under
hazardous conditions.
Preliminary test results with three model
compounds, namely a permanent gas (sulfur
hexafluoride), a condensable vapor (diethyl
malonate) and a biological aerosol (bovine serum
albumen), illustrate the special capabilities of
the MINITMASS system.
KEYWORDS: fieldable mass spectrometer; tandem
mass spectrometry; chemical ionization; direct
air sampling; transfer line chromatography;
aerosol characterization; mobile laboratory
module; remote control; Curie-point pyrolysis.
INTRODUCTION
Highly desirable characteristics of fieldable
mass spectrometry (MS) systems for hazardous
waste site investigations include high
sensitivity and specificity, in addition to
real-time analysis capability and
user-friendliness. Furthermore, the system
should be versatile enough to handle a broad
range of different sample types, i.e., gases,
liquids and solids. Preferably, all this should
be embodied in a small, lightweight, ruggedized
and affordable instrument package.
No system reported combines all of the above
desirable characteristics. Instead, common
trade-offs involve: specificity vs. size, weight
and cost (viz. in tandem MS systems (1));
specificity vs. real-time analysis (viz. in GC/MS
systems (2)); or sensitivity vs. real-time
analysis (e.g., in purge-and-trap systems (3)).
Recent advances in Ion Trap MS technology (4) now
open up the possibility of designing powerful
tandem MS systems with chemical ionization
capabilities without compromising size and weight
requirements. Moreover, a novel direct air
sampling technique developed in our laboratory
and based on the so-called "transfer line gas
chromatography" (TLGC) approach can provide sub
ppm detection levels in combination with enhanced
specificity while keeping response times low.
Finally, a unique on-line aerosol sampling inlet
has been developed which enables characterization
of nonvolatile organic components in air
particulate matter or of adsorbed volatile
constituents.
Preliminary test results obtained with the
MINITMASS, a miniaturized Ion Trap mass
spectrometer equipped with specially designed
inlets for direct air sampling as well as aerosol
collection, will be reported here.
INSTRUMENTATION
The MINITMASS is an Ion Trap MS based instrument
capable of operating in Chemical Ionization (CI)
(5), Collision Induced Dissociation (MSn) (6)
and Selective Mass Storage (SMS) (7) modes.
Further, it has the advanced capabilities of
195
-------
Automatic Gain Control (AGC) (8), Automatic
Reaction Control CI (ARC-CI) (9) and Axial
Modulation (AM) (7).
In normal scanning electron ionization mode,
Finnigan MAT Ion Trap Detectors are specified to
measure 1 ng of naphthalene with a 30:1 S/N ratio
(10). Use of the above described special
operational modes can substantially improve
minimum detectable quantity levels (complete mass
spectra have been reported from subpicogram
sample quantities (11) using CI), as well as
dynamic range (AGC and ARC-CI) and specificity
(MSn and CI). The new SMS and AM modes are
especially powerful for trace analysis in complex
sample matrices while also enhancing overall
sensitivity (12).
The complete MS system with vacuum pumps and
COMPAQ 386/20 microcomputer workstation weighs
less than 115 kg, uses approx. 1000 W of
electrical power and can be installed in a single
1.5 m high electronic rack (see Figure 1). This
rack also accomodates all gas cylinders (carrier
gas + CI reagent gases) as well as the power
supplies and control electronics for the vapor
and aerosol sampling inlets.
Vapor and Aerosol Sampling
Figure 2 shows the vapor and aerosol sampling
inlets, both of which have been specially
designed at the University of Utah. Since some
technological innovations incorporated in each
inlet are currently being patented, only a
general description follows here. The vapor
sampling inlet (Figure 2a) can be heated to 523 K
to eliminate condensation losses of low volatile
atmospheric components and connects directly to a
standard 1 m capillary transfer line available on
all Finnigan MAT ITD systems. During sampling a
short 1-4 s "pulse" of ambient air is admitted
into the capillary transfer line at 15-30 s
intervals. Between sampling the inlet is sealed
off dynamically by a curtain of inert gas.
The aerosol sampling inlet shown in Figure 2b
consists of a pump drawing 0.1 1/s of air through
an electrostatic charging section into an
electrostatic precipitation zone where charged
aerosol particles are deposited on the tip of an
0.5 mm dia. ferromagnetic filament. After a
predetermined collection period, the filament is
mechanically drawn into a sealed-off reactor
where it is rapidly heated by means of a 500 kHz
rf field. This so-called Curie-point pyrolysis
technique has been described in detail elsewhere
(13). Evaporation and/or pyrolysis products
formed are sucked into the heated capillary
transfer line and transported into the vacuum
chamber of the MS system.
Automation and Data Processing
The basic ion trap MS system is controlled by
standard Finnigan MAT software (ITMS, revision B)
to which we have added several routines for
automation of sampling inlet functions and CI gas
control. After initial set-up, the complete
system can be remotely operated from a distance
of up to several kilometers using twisted pair
wire connections, high speed (60 kbaud) line
drivers and commercially available Carbon Copy
Plus (Meridian Technologies) software, and a
second PC-AT compatible computer work station.
Figure 3 shows a block diagram of the various
control systems. Besides enabling operation of
the MINITMASS in hostile environments, one of the
main benefits of operation under Carbon Copy
Plus is that this allows transfer of data to
remote computer stations while new data are still
being acquired by the local work station.
Final data processing of complex signals, e.g.,
obtained in the presence of chemical interferents
and/or background signals, can be performed by
means of SIGMA-PC, a special software package for
multivariate statistical analysis of
spectroscopic data developed at the University of
Utah. SIGMA-PC is a new PC compatible version
of the SIGMA program originally written for IBM
9000 workstations and described elsewhere (14).
Major subroutines of SIGMA-PC include:
normalization, univariate analysis, factor
analysis, discriminant analysis, canonical
correlation analysis, variance diagram, K nearest
neighbor classification and various graphic
routines for visualizing spectroscopic data.
Mobile Laboratory Module
In order to enable field-operation of the
MINITMASS while providing the basic amenities of
an instrumentation laboratory, we have designed
and constructed a 10 m^ lab module which fits
the bed of a regular 3/4 ton pick-up truck (see
Figure 4). An aluminum/polystyrene sandwich
construction minimizes total weight while
providing excellent insulating properties. The
mobile lab module has several functions. First
of all, it provides physical support for the
MINITMASS system and associated sampling inlets.
Moreover, it is equipped with a battery (12 V,
500 Amp hrs) powered 110 V ac power supply
capable of supporting MINITMASS operation for up
to 4 hrs. Further, a 3.5 kW (generator powered)
air conditioning system enables operation in hot
environments (up to 309 K tested).
Also, dual propane tanks allow up to 1 week of
freezer operation (sample and solvent storage),
as well as for gas heater and cooking stove use.
Finally, the mobile lab module provides adequate
working and living space for 2 technicians and,
when necessary, rudimentary sleeping quarters.
EXPERIMENTAL CONDITIONS
Vapor sampling data reported here were obtained
under the following conditions. SFg vapors
were diluted into a six gallon bucket inverted
over the vapor sampling inlet. The dilution
produced a sample concentraton of approx. 40 ppm
which was analyzed with an El scan function
utilizing the SMS capability of the system. The
m/z 127 ion was isolated and the resulting ions
scanned from 60 to 150 u. Ionization time was
196
-------
2000 msec and the scan rate was 4 s~l. Vapors
were sampled repetitively for 2 s each while the
volumetric flow into the instrument was 0.025 ml
s~l. The vapor inlet and transfer line were
operated at ambient temperature (approx. 300 K).
Diethyl malonate (DEM) data were obtained by a
^O-CI-MS/MS method. The sample was ionized
for 3 ms and allowed to react with the HgO
reagent gas for 100 ms before isolating the m/z
161 parent ion. After 5 ms of collision-induced
dissociation the daughter fragments were scanned
from 60-200 u. Sampling times were approximately
4 s each with the helium flow into the instrument
set at 0.02 ml s~l The transfer line and
inlet temperatures were set at 373 K.
Bovine serum albumin (BSA) aerosol sampling data
were obtained from aerosols produced in a
laboratory nebulizer from a solution of 1 mg/ml
BSA in water. The aerosols were collected
electrostatically onto a 1040 K Curie-point
pyrolysis wire for 11 minutes and then pyrolyzed
for 1.5 sec. The inlet was heated to 473 K and
the transfer line was operated at 433 K. A 2:1
split of pyrolysis products entered the transfer
line. Data were obtained using a standard A6C
scan function. The scan rate was 4 s~l from
50-200 u.
PRELIMINARY TEST RESULTS
Sulfur hexafluoride (SF5) is a frequently used
leak detection and environmental tracer gas.
Figure 5 shows three repetitive samples of 40
parts per million (ppm) levels of SF5 in air
For this sequence the samples were taken every 15
s although the very short elution time on the
capillary in the TL6C would allow sampling
intervals as short as 5 s. The SF5 is detected
easily with electron ionization (El) due to a
very strong SF5+ ion at m/z 127 which carries
over 80% of the ion intensity. However, since
this positive ion is very stable, there is little
or no specificity gain when operating in the
MSn mode.
Diethyl malonate (DEM) was chosen as a test
chemical for field tests simulating compounds
with relatively high boiling points and low vapor
pressures. This organic compound with a boiling
point of 471 K fragmented readily and produced
the positive ion spectra shown in Figure 6 using
El, H20-CI, SMS and MS/MS, respectively. The
El spectrum in Figure 6a shows characteristic
fragments at m/z 133, 115, 105 and 87 with a very
small amount of the molecular ion at m/z 160
except when self-CI takes place at high
concentrations. The F^O-CI (Figure 6b) gives
an abundant (M+H)+ ion at m/z 161, but still
shows some of the characteristic fragments. The
SMS function (Figure 6c) enhances sensitivity by
allowing all ions outside a predefined narrow
mass range of interest to be dumped from an
overfilled trap prior to the analytical ion
scan. MS/MS on the ion signal of interest
(Figure 6d) then regains the selectivity and
specificity of the fragment ions.
Transfer line chromatograms from the repetitive
analysis of DEM vapors are shown in Figure 7
using H20-CI-MS/MS (as in Figure 6d). For
these samples, a 1 ml/min stream of saturated DEM
vapor in air at ambient temperature was dilated
in a larger air stream (72ml/min) to give a final
concentration of approximately 6 ppm. The
chromatogram traces of m/z 115 and 133 show the
strong specific ions from the DEM each with a
signal to noise greater than 100.
An example of biological aerosol analysis is
demonstrated in Figure 8. In Figure 8 the BSA
solution has been aerosolized, collected by
electrostatic precipitation on the Curie-point
wire, and then analyzed by Py-TLGC. This data
represents approximately 150 ng of BSA collected
over 8 min and analyzed by a TLGC run of 30-40 s
duration. The ion chromatogram profiles in
Figure 8 are believed to represent amino acid
dimers similar to the structure shown. Further
interpretation of these ion signals, currently
underway in our laboratory, will be greatly aided
by MS/MS daughter ion spectra. Moreover, several
critical parameters for aerosol collection
require further optimization to improve on what
is presently estimated to be only a 2-3 % aerosol
collection efficiency. However, these results
clearly demonstrate the wealth of sample specific
information produced by our TLGC technique on a
very short time scale when analyzing complex
vapor or aerosol samples.
CONCLUSIONS
The data presented here demonstrate the
feasibility of constructing a fieldable MS
system, with MSn, CI and other advanced
capabilities while meeting size, weight and power
requirements compatible with transportability to
remote or otherwise less accessible test sites.
Specially designed vapor and aerosol sampling
inlets can be directly connected to the MS system
by means of a heated capillary transfer line
which provides a definite amount of
chromatographic separation, thereby increasing
specificity while maintaining short response
times. Moreover, our novel direct vapor inlet
technique has been shown to produce sharp air
"injection" peaks, thereby enabling optimum use
of the TLGC approach, whereas the aerosol inlet
has been demonstrated to produce characteristic
pyrolysis fragments of complex organic aerosol
components.
A relatively high degree of automation has been
achieved which enables operation of the entire
system from remote computer workstations up to
several miles away, e.g., when performing vapor
and aerosol analyses in hostile environments.
Also, a 10 m^ mobile lab module was constructed
which has proven to be a suitable operating base
for the MINITMASS under demanding environmental
conditions while providing accessibility to
off-road test sites using a regular 3/4 ton
pick-up truck vehicle.
197
-------
Present shortcomings of the MINITMASS include:
(a) inability to move the system during operation
(due to the presence of a turbomolecular vacuum
pump) and (b) low collection efficiency of the
aerosol inlet (probably due to incomplete
optimization).
ACKNOWLEDGEMENTS
Credit is given to Finnigan MAT Corporation (San
Jose, CA), Geocenters, Inc. (contract
#GC-1728-88-002) and to the Advanced Combustion
Engineering Research Center (funds for this
Center are received from the National Science
Foundation, the State of Utah, 23 industrial
participants, and the U.S. Department of Energy)
for sponsoring this research.
The invaluable help and support of Drs. Michael
S. Story and Michael Weber-Grabau (Finnigan MAT,
San Jose, CA) in designing and assembling the
MINITMASS system is gratefully acknowledged. Dr.
A. Peter Snyder (U.S. Army CRDEC, Aberdeen, MD)
is thanked for his continued advice and
participation in various phases of the project.
REFERENCES
1. Pritchett, T.H., Mickunas, D., Bernick, M.
and Weston, R.F., "TAGA 6000 E QA/QC Analysis
Procedures and the Associated Data Gathered
during Four, Two Week Blocks of Indoor Air
Analyses at the Love Canal Emergency
Declaration Area (EDA)," Proc. 36th ASMS
Conf. Mass Spectrom. All. Topics, June 5-10,
1988, San Francisco, CA, pp. 1342-1343.
2. Trainor, T.M. and Lankein, F.H., "Field
Screening of Soil and Water for Volatile
Organics by Mobile Gas Chromatography/Mass
Spectrometry," Proc. 1st Intnl. Symp. Field
Screening Methods for Hazardous Waste Site
Investigations, October 11-13, 1988, Las
Vegas, NV, in press.
3. Robbat, A., Jr., and Xyratas, G., "Evaluation
of Field Purge and Trap Gas Chromatography
Mass Spectrometry," Proc. 1st Intnl. Symp.
Field Screening Methods for Hazardous Waste
Site Investigations, October 11-13, 1988, Las
Vegas, NV, in press.
4. Weber-Grabau, M., Kelley, P.E., Syka, J.E.P.,
Bradshaw, S.C. and Brodbelt, J.S., "Improved
Ion Trap Performance with New CI and MS/MS
Scan Functions," Proc. 35th ASMS Conf. Mass
Spectrom. All. Topics May 24-29, 1987,
Denver, CO, 1114-1115.
5. Brodbelt, J.S., Louris, J.N. and Cooks, R.G.,
"Chemical lonization in an Ion Trap Mass
Spectrometer," Anal. Chem. Vol. 59, 1987, pp.
1278-1285.
6. Louris, J.N., Cooks, R.G., Syka, J.E.P.,
Kelley, P.E., Stafford, G.C. and Todd, J.F.J,
"Instrumentation, Applications, and Energy
Deposition in Quadrupole Ion-Trap Tandem Mass
Spectrometry," Anal. Chem. Vol. 59, 1987, pp.
1677-1685.
7. Weber-Grabau, M., Kelley, P.E., Bradshaw,
S.C. and Hoekman, D.J., "Advances in MS/MS
Analysis with the Ion Trap Mass
Spectrometer," Proc. 36th ASMS Conf. Mass
Spectrom. All. Topics, June 5-10, 1988, San
Francisco, CA, pp. 1106-1107.
8. Yost, R.A., McClennen, W.H. and Meuzelaar,
H.L.C., "Enhanced Full Scan Sensitivity and
Dynamic Range in the Finnigan MAT Ion Trap
Detector and the New Automatic Gain Control
Software," Finnigan Mat Application Note,
Number 209, 1987.
9. Keller, P.R., Harvey, G.J. and Foltz, D.J.
"Analysis of Fragrance Materials Using
Automatic Reaction Control Chemical
lonization on the Ion Trap Detector," Proc.
36th ASMS Conf. Mass Spectrom. All. Topics,
June 5-10,1988, San Francisco, CA, pp.
643-644.
10. "Ion Trap Detector Operation Manual,"
Finnigan MAT Manual Section 1, P/N
94011-98025, Revision E, June 1987, p. 1.
11. Lim, H.K., Sakashita, C.O. and Foltz, R.L.,
"The Application of Chemical lonization to
Drug Analysis," Spectra, Vol. 11, No. 2,
Spring 1988, pp. 10-14.
12. Tucker, D.B., Hameister, C.H., Bradshaw,
S.C., Hoekman, D.J. and Weber-Grabau, M.,
"The Application of Novel Ion Trap Scan Modes
for High Sensitivity GC/MS," Proc. 36th ASMS
Conf. Mass Spectrom. All. Topics, June 5-10,
1988, San Francisco, CA, pp. 628-629.
13. Richards, J.M., McClennenr W.H. , Bunger,
J.A., Meuzelaar, H.L.C., "Pyrolysis
Short-Column GC/MS Using the Ion Trap
Detector (ITD) and Ion Trap Mass Spectrometer
(ITMS)," Finnigan Mat Application Note,
Number 214, 1988.
14. Windig, W., and Meuzelaar, H.L.C., "Numerical
Extraction of Components from Mixture Spectra
by Multivariate Data Analysis," Chapter 4 in:
"Computer-Enhanced Analytical Spectroscopy,"
Plenum Publishing Co., Amsterdam, The
Netherlands, 1988, pp, 67-102.
198
-------
MINITMASS
Inlet Power Supplies
and Control Equipment
Uninterruptible
Power Supplies
Helium Bottles
Vacuum and Air
Sampling Pumps
Figure 1. The MINITMASS system and associated equipment.
Glass-Lined Air Intake
(continuous flow)
— Aluminum
Block Heater
Roof Line
Aerosol Sampling
Control Hardware
Electrostatic
Precipitation
and Collection
Region
Air Sampling
Region
Heated Transfer Line to
Ion Trap
Pyrolyzar
•nd Sample
Inlet
Figure 2. a) Highly schematized line drawing of the vapor
sampling inlet; b) of the aerosol pyrolysis inlet.
-------
MINITMASS
INLET SUPPLIES & CONTROLS
RF FILAMENT MULTIPLIER
CONTROL CONTROL CONTROL
FREQUENCY
SYNTHESIZER
Figure 3. Block diagram of MINITMASS control system.
BATTERY
POWERED
no v ac
SUPPLY
FREEZER MINITMASS
RACK
(DESK-TOP
WORK SPACE)
HIGH
STORAGE
CABINET
S
PROPANE
TANKS
(BENCH-
TOP
WORK
SPACE)
SINK
STOVE
REAR DOOR
Figure 4. Mobile laboratory module
200
-------
1001
a, TOTAL ION PROFILE
' •X-./v-.-V'——j-jv-f
2 80 b, SFj* ION PROFILE
SAMPLING POINTS
12
25
TIHE (s)
37
Figure 5. Repetitive vapor sampling or approH. 40ppm SF6 in air showing a) the total ion
chromatogram and b) the SF5 + Ion chromatogram. Samples were taken for 2 s at 16 s Intervals with
EI-MS analgsis for that main fragment ion SF5+ at m/z 127.
10-
5-
10-
5-
10-
5-
in-i
5-
6
a. EI-NS
i
68
87
61
T875
I
!!
b. H20-CI-HS
35 .il
L5
1,
,128
[
1
161
, 142 1
it
i:
115
88 185
,64 7B 76 81
1 97
c. Cl-HS with
78 78 85
94188
d. CI-NS-HS
. 122
3
142 1
Lt
sns
112 123 136 148
1.
185
87
77 83
i 88
94 101
1 ' '• "l ' '
iae
.1, .
175
1
175
1
1 172 188
5
i:
122
3
162
, ,14i 158 ,
, 178
i.l . ,
i "| 'i | i - , • i i • |
128 148 168 188 H/Z
Figure 6. Diethyl malonate (DEM) mass spectra using a) electron ionization (El), b) water
chemical ionization (H2O-CI, c) CI-MS with Selective Mass Storage (SMS), and d) Cl tandem
MS (MS/MS). The protonated molecular ion (M+H)+ occurs at m/z 161 and is isolated using
SMS prior to collision-induced dissociation to maximize both sensitivity and selectivity.
201
-------
35
o
2
< 33
o
i-
^>
a, TOIflL
SWUNG POINTS
V I
V \
I I
f
1
I
1
b, H/Z 115
1. I
1
[ 1
t\
I
v_
c, M/Z 133
15
31
TIME (s)
47
I
62
Figure 7. Repetitive analysis of DEM vapors in air using H2O-CI-MS/MS (see Figure 6d). Data
is represented by chromatograms for a) the total of all ions from m/z 50 to 200, b) DEM
fragment ion at m/z 115, and c) DEM fragment ion at m/z 133. The 3 s samples were taken at
20 s intervals at the points indicated.
Figure 8. Pyrolysis TLGC-CI-MS analysis of bovine serum albumin (BSA) aerosol collected by
electrostatic precipitation. Ion profile chromatograms are shown for the total of ions
between m/z 150 to 250 (a) and single ion intensities at m/z 168 (b), m/z 182 (c), m/z 196
(d), and m/z 210 (e) tentatively identified as amino acid dimer ions of the general form shown.
202
-------
DISCUSSION
TOM PRITCHETT: You said that basically you can do any type of chemical
ionization rather than having to just rely on ambient air?
HENK MEUZELAAR: Yes, you could do ammonia, isobutane, water chemi-
cal ionization. Others have played with other reagents.
TOM PRITCHETT: What about the inlet system?
HENK MEUZELAAR: The inlet does not limit the chemical ionization.
MARC WISE: We're also working with an ion trap mass spectrometer for
environmental applications. Preliminary tests on samples in water and soil
matrices indicate that we have very good detection limits at sub-part per billion
levels, with very good linearity and quantitative reproducibility.
In real-life samples, do you have any problems with really unwanted ion
molecule reactions at long chemical ionization reaction times?
HENK MEUZELAAR: Yes, when the sample amounts become high, then.
especially in the electron ionization mode, you're likely to start seeing some
chemical ionization and other types of high-pressure, long-residence time ef-
fects.
However, when you choose the chemical ionization mode, and the agent gas is
the dominant reactant, then you can avoid those in most cases. It does require
balancing, sometimes, of reaction times and total residence times. There is
some experimental software that allows you to almost Macintosh-style pro-
gram how the trap will scan, how long you will react. The problems that you
mention exist, but I think they can be overcome.
MARC WISE: For real-life samples where you're looking at a complex
mixture, it's desirable, of course, to have an automated software program for
parent ion scanning, particularly for tandem mass spectrometry. Have you de-
veloped any type of software for that? I think it can be done probably fairly
easily.
HENK MEUZELAAR: No we haven't.
203
-------
THE PREPARATION, CERTIFICATION, AND USE OF SUMMA
CANISTER EXTERNAL PERFORMANCE EVALUATION
SAMPLES IN SUPPORT OF THE TAGA 6000E INDOOR AIR
ANALYSES DURING THE LOVE CANAL EMERGENCY
DECLARATION AREA HABITABILITY STUDY
K. J Cavislon. Richard E. Means. Rita M. Harrel. BJ. Carpenter
Northrop Services. Inc.
David Mickunas & Mark Bernick
Roy F. Weston, Inc (REAC)
Thomas H. Prilchert
U.S. EPA Environmental Response Team
Because of the past problems encountered with the TAGA 6000E in quan-
titative air analyses and because of the high scrutiny that the Love Canal
Emergency Declaration Area (EDA) air data would ultimately undergo, it
became necessary to develop the procedures to provide external audit and
performance evaluation (PE) samples to the TAGA during the study. Northrop
Services, through the Quality Assurance Branch of EMSL/RTP, was requested
by U.S. EPA Region II to provide this support. Northrop Services worked in
conjunction with the Environmental Response Team (ERT) and its contractor to
develop this support capability. The ERT and REAC, its contractor, first defined
the procedures to be used to analyze the samples and then Northrop Services
developed the procedures to prepare and certify the PE samples at the applicable
concentrations. Both 16 Liter and 6 Liter Summa canisters were utilized during
the study. Problems were initially encountered with sample stability in dry
balance gas and with the certification analyses. These problems were solved.
The 16 Liter canisters were successively used during all four phases of the EDA
study and the 6 Liter canisters were successively utilized during the last two
phases. These results will be summarized.
DISCUSSION
TOM PRITCHETT: One of the problems with doing any type of air or soil
gas analysis, is that it is very difficult to get performance evaluation samples/
standards that you can spike into bags or air samples. We have taken this
procedure, and we are now - in conjunction with Northrop Services within the
EPA-utilizing it for just about all our air analysis and soil gas analyses.
As a matter of fact, there is a directive coming out from the Environmental
Response Division of OSWER that will require these types of performance
evaluation samples to be used in soil gas and air sampling.
Basically, you can make up a mixture in a Summa canister, take it out in the
field, spike it into your bag, send it back to the lab; or you can spike it onto a
tube and send the tube back. This is just now coming on board, and it has a lot
of potential beside the TAGA application described here.
205
-------
UNAMBIGUOUS IDENTIFICATION AND RAPID QUANTITATION IN FIELD
AIR MONITORING USING A FULLY MOBILE MASS SPECTROMETER
Frank H. Laukien, Ph. D. and Thomas M. Trainor, Ph. D.
Bruker Instruments, Inc.
Manning Park
Billerica, Massachusetts 01821
ABSTRACT
A fully mobile field mass spectrometer, based on
electron-impact ionization and medium-resolution
quadrupole mass analysis, has been found to be
useful for the detection and quantitation of
volatile organics associated with hazardous waste
site emissions or toxic emergencies. Ambient air
monitoring may take place on a continuous basis
in real-time and with excellent sensitivity,
accuracy, and reproducibility, using an air
sampler interfaced to the ion source of the mass
spectrometer. Alternatively, time-integrated
measurements may be made by concentrating whole
air on a suitable sorbent tube for subsequent
thermal desorption/high resolution gas
chromatography-mass spectrometry analysis in the
field.
INTRODUCTION
The detection of organic compounds in air remains
a challenging analytical problem for the environ-
mental community today. The frequency and type of
operations surrounding hazardous waste site inves-
tigation and remediation activities, coupled with
the variable wind velocity and direction normally
encountered, invariably lead to unpredictable
and intermittent excursions above background
concentration levels for toxic compounds. During
toxic emergencies, such as storage tank leaks or
chemical transport accidents, immediate results
are mandated. Many sampling and analysis
approaches have been proposed for air monitoring
of toxic organics associated with hazardous waste
sites, with the currently most popular including
direct (whole air) field-based measurements with
gas chromatographs (1) or off-site laboratory
analysis of samples collected in specially treated
steel canisters (2) , Tedlar bags (3) , or pre-
concentrated on suitable sorbent tubes (4) .
However, each of these approaches has inherent
limitations which act to preclude universal
applicability to investigations surrounding site
emissions. For instance, the limited specificity
afforded by field GC's has been a concern in many
programs. Besides the data turnaround and logis-
tic problems, researchers are increasingly noting
the technical problems involved in off-site
analysis, which include sample quality degradation
caused by the combined sampling/transport/storage
series of steps. The use of polished stainless
steel canisters has been found to minimize these
problems, but is recognized as introducing new
concerns, including the cost of individual
samplers (ca. $ 500 each) and problems concerning
normal sample levels of water in subsequent
cryotrapping GC and/or GC-MS analysis (2).
The use of field instrumentation is recognized as
invaluable in adequately addressing the variable
nature of airborne emissions. Atmospheric pressure
ionization mass spectrometry has been suggested
(API-MS) (5) as a means of conducting both the
sampling and analysis directly in the field.
However, deploying API-MS, particularly tandem
GC/MS/MS, in the field represents the expense of
an initial capital investment and highly trained
staff of mass spectrometrists that few organiza-
tions can justify. Moreover, the complex, and
often irreproducible (6) , ionization behavior
inherent in API-MS is in practice an impediment to
routine problem-solving.
One technique that has been found to obviate the
majority of these limitations is based on the
development of a fieldable GC-MS which employs
traditional electron-impact (El) ionization. The
Bruker Mobile Environmental Monitor has been
designed to sample, identify, and quantitate
target organic compounds in ambient air. The
advantages realized by utilizing this mobile
instrument include :
1. Direct sampling and analysis of whole air on a
truly continuous real-time basis.
2. Generation of unambiguous electron-impact
data, with the capability of both full-scan
or selected ion monitoring (SIM) data
acquisition.
3. Single keystroke operation, designed for
non-expert operators.
4. Capable of operation with 24 volt rechargeable
battery power under extremes of temperature,
humidity, mechanical and electrical shock,
without a need for compressed gas cylinders.
This presentation will cover the present range of
applications and limitations of the Bruker MEM in
air monitoring programs. Examples of both
continuous and time-averaged sampling approaches
will be discussed and actual results presented.
INSTRUMENTATION
The results reported here were carried out with
standard Bruker MEM mobile mass spectrometers
equipped with both the direct air/surface sampler
and the optional thermal desorption-capillary gas
chromatograph option. MEM control and data
acquisition was carried out by either the internal
MEM controller or the optional external data
system (Compaq Portable III). The MEM, based on an
electron-impact source and quadrupole analyzer, is
described in detail in a separate paper at this
symposium which should be consulted for general
207
-------
information and specifications (7) . The data
gathered here was obtained with MEM units mounted
in a variety of vehicles, including a 4-wheel
drive Chevy Blazer truck. Instrument power was
provided by four 6-volt storage batteries con-
nected in series. Gas standards were prepared
using a Thermo Electron model 360 standards gener-
ator, or alternatively, by spiking known amounts
of materials into Tedlar bags filled with UHP
nitrogen. Tenax tubes were obtained from SKC and
used as received. The fused-silica capillary GC
columns employed included a 30 m x 0.32 mm DB-624
for the auxiliary capillary GC oven, and a 3.5 m x
0.32 mm SE-54 for the direct air/surface sampler.
CONTINUOUS AIR MONITORING BY SELECTED ION
MONITORING GC-MS
The simplest approach to continuous air monitoring
with the MEM is by means of sampling with the
air/surface sampler (7) while acquiring data using
the program AIR MONITOR. AIR MONITOR allows for
the continuous acquisition and display of selected
ion monitoring data for up to a maximum of 22
compounds simultaneously, using from 1 to 4
selected ions per compound. An individual METHOD
file, of which a maximum of 15 may be
permanently stored in the MEM EE-prom memory, may
contain a preprogrammed library of up to 60 target
compounds, thus allowing for a total of 900 unique
compounds stored in memory. An example of a
SUBSTANCE DATA page from a typical method is shown
in Figure 1. Here, four characteristic ions with
the corresponding per cent relative intensities
are entered, along with the compound name and
numerical values for five additional parameters.
During actual air monitoring, a real time display
of MEM response, on a logarithmic basis, is shown
on the video screen. Figure 2. Target compound ion
groups are labeled A L along the x-axis, while
the corresponding ion abundances are plotted along
the y-axis. A constant measurement of the high
vacuum region's pressure is also displayed along
the right-hand edge of the screen. Prior to
initiating an AIR MONITOR experiment, the MEM
conducts a background measurement by scanning each
ion five times which permits the calculation of a
value known as the Minimum Detection Amount (MDA)
for each ion. This MDA is defined as 3 times the
square root of the mean ion response observed
during the background measurement. This MDA is
graphically represented throughout the subsequent
AIR MONITORING run by the unshaded area under the
curve drawn for each designated compound (Figure
2) .
The MEM has been programmed to determine in real-
time both the qualitative identification and quan-
titation of target compounds based on the SIM
results. During each acquisition cycle, the first
ion of each compound is scanned and the observed
background corrected signal is compared to the
WARNING LEVEL value programmed for the correspond-
ing compound (Figure 1). If the intensity of the
first ion (II) is greater than this WARNING LEVEL,
then the three secondary ion signals are recorded
(12, 13, and 14). Figure 3 displays a flow chart
illustrating the decision pathway that is followed
by the MEM software in carrying out an identifica-
tion based on the actual measured ion currents.
The parameter RELIABILITY is a numerical value
which governs the allowable variance between the
library ion abundance ratios and the observed
ratios (spectra matching criteria). INTERFERENCE
is a parameter that functions to suppress false
positive identifications as a result of signals
from interfering compounds present in much larger
concentrations relative to the target compound.
During the AIR MONITOR acquisition, the video
screen is constantly updated in terms of the ion
response information. Ion current significantly
above background levels is depicted as the shaded
areas directly above the MDA unshaded areas. As a
further visualization aid to the operator, prior
to the display update the observed signals of the
secondary ions (12, 13, 14) are ratioed to the
expected relative abundance per cent values con-
tained in the method library, and this normalized
value is actually plotted. By carrying out this
transformation, the end result is that for good
library matches the ion peak heights will be near
equal, leading to readily discernible "flat-
topped" signal envelopes (Figure 3). Clearly, in
instances where unequal ion peak heights above
background are frequently observed, the operator
immediately knows that non-target compounds are
present, which may suggest taking a full mass
spectrum to characterize the air components, and
perhaps, making changes to the current list of
compounds analyzed by AIR MONITOR.
Once an identification has been made in AIR
MONITOR and a compound signal exceeds the
preselected ALARM LEVEL, the compound name and
maximum signal amount is displayed on the screen.
If the instrument has been previously calibrated
for the compound using gas standards of known con-
centrations, then a direct conversion to actual
air concentrations (ppb or ppm) is made. Alterna-
tively, for those compounds in which no instrument
response factor is currently in memory, the raw
log signal value is displayed. Figure 4 contains
a typical calibration curve exhibited by the MEM
for a series of benzene standards in which
accurate readings in the range of 1 to 10 ppm were
of interest for a particular application.
An extensive and thorough evaluation of the MEM
for direct air sampling was recently reported by
the US EPA Region II Environmental Response Team
(ERT) (8). Under actual field conditions, using a
variety of common volatile organic pollutants
delivered as mixtures from certified gas
cylinders, detection limits in the range of 10 to
100 ppb were found to be readily achievable.
Linearity data was reported over the range of 10
ppb to 10 ppm for several compounds using the
direct air/surface sampler. The instrument
stability over time was such that quantitation
error levels fell in the range of +2.2% to -39.1%
for a series of test analytes introduced at the
100-600 ppb range.
AIR MONITORING BY THERMAL DESORPTION GC-MS
Frequently, for highly complex air matrices in
which many interferences exist at the ions
monitored for the actual target compounds of
interest and/or instances where the detection
limits achievable with the direct air/surface
sampler are not adequate, we have found the
optional MEM capillary GC to be of benefit. The
improved chromatographic resolving power of the 30
m capillary column is particularly suited to
multi-component mixtures. The thermal-desorption
injection system incorporated in the GC allows for
the direct analysis of air samples concentrated on
glass tubes containing suitable sorbent materials.
An example of this approach was an analysis
developed on the MEM for the suspected carcinogen
bis-dichloromethyl ether (BCME) in air associated
with a manufacturing facility. Due to the high
levels of interferences from other volatile
organics present at the site, and the need for
sub-ppb detection limits, initial concentration on
Tenax followed by thermal-desorption capillary
GC-MS was utilized. Typical results are shown here
for a calibration curve (Figure 5) and a 1.0 ppb
BCME standard run (Figure 6) achieved using Tedlar
bags for the preparation and delivery of suitable
calibration standards to the Tenax traps. Under
the AIR MONITORING data acquisition conditions,
this level of BCME resulted in a signal
approximately lOx the background level.
208
-------
For many programs the generation of full-scan
electron-impact data on capillary GC separations
is necessary. With the optional MEM external data
system (Compaq Portable III) this high-speed,
memory intensive activity is possible in the
field. An application of this is depicted by the
MEM chromatogram shown in Figure 7. Here,
sampling via Tenax of a site suspected of
hydrocarbon contamination was of interest. By
careful examination of several samples taken both
upwind and downwind of the site, during remedia-
tion activities carried out under highly variable
wind conditions, potential off-site impacts were
determined and appropriate precautions taken
immediately.
CONCLUSIONS
A gas chromatograph-mass spectrometer uniquely
designed for operation in the field has been shown
to provide solutions to analytical problems
associated with the ambient air monitoring of
organic compounds.
ACKNOWLEDGMENTS
Assistance provided by Dr. Jochen Franzen and
Dr. Alex Loudon, of Bruker-Franzen Analytik Gmbh,
is greatly appreciated.
REFERENCES
1. Jerpe. J.; Davis, A. ; "Ambient Capillary
Chromatography of Volatile Organics With a
Portable Gas Chromatograph"; J. Chrom. Sci.,
1987, vol 25, 154-157.
2. EPA Method TO-14; "The Determination of
Volatile Organic Compounds in Ambient Air
Using Summa Passivated Canister Sampling and
Gas Chromatographic Analysis"; US EPA
Environmental Monitoring Systems Laboratory,
Research Triangle Park, NC 27711.
3. Levaggi, D. A.; Siu, W.; Oyung, W.; Zerrudo,
R. V.; "The Use of Tedlar Bags for Integrated
Gaseous Toxic Sampling: The San Francisco Bay
Area Experience"; presented at the 1988
EPA/APCA Symposium on Measurement of Toxic
and Related Air Pollutants; Raleigh, North
Carolina, May 2-4, 1988.
4. Brown, R. H.; Purnell, C. J.; "Collection and
Analysis of Trace Organic Vapour Pollutants
in Ambient Atmospheres. The Performance of a
Tenax-GC Adsorbent Tube"; J. Chromatogr.,
1979, vol 178, 79-90.
5. Shushan, B. I.; DeBrou, G.; Mo, S. H.;
Webster, W.; "Mobile Tandem Mass Spectrometry
for t.fca Characterization of Toxic Air
Pollutant (TAP) Sources"; presented at the
1987 EPA/APCA Symposium on Measurement of
Toxic and Related Air Pollutants; Research
Triangle Park, North Carolina, May 3-6, 1987.
6. Pritchett, T. H.; Hague, R. E.; Willingham,
T. ; "Results from the Environmental Response
Team's Evaluation of the TAGA 6000E Direct
Air Sampling Mass Spectrometer/Mass
Spectrometer"; presented at the 1988 EPA/APCA
Symposium on Measurement of Toxic and Related
Air Pollutants; Raleigh, North Carolina,
May 2-4, 1988.
7. Trainor, T. M.; Laukien, F. H.; "Design and
Performance of a Mobile Mass Spectrometer
Developed for Environmental Field
Investigations"; presented at the First
International Symposium: Field Screening
Methods for Hazardous Waste Site
Investigations, Las Vegas, NV, October 11-13,
1988.
8. Hague, R. E.; Pritchett, T. H.; Cho, K.;
Shapiro, B.; "Result's from the Environmental
Response Team's Preliminary Evaluation of a
Direct Air Sampling Mass Spectrometer, the
Bruker MM-1."; presented at the 1988 EPA/APCA
Symposium on Measurement of Toxic and Related
Air Pollutants; Raleigh, North Carolina,
May 2-4, 1988.
209
-------
SUBSTANCE DATA 01:04
BTXH AIR
* SUBSTANCE NO 23
NAME: TOLUENE
MASS 91.0 u
REL. INTENSITY 100.0%
MASS 92.0 u
REL. INTENSITY 60.1 %
MASS 65. Ou
REL. INTENSITY 20.1 %
MASS 63.0 u
REL. INTENSITY 10.4%
WARNING LEVEL 0.5
ALARM LEVEL 1.0
INTERFERENCE 3.0
RELIABILITY 5.0
MONITOR CODE 11
0=0 D=13J=19 P=25V=31+=51
... E=14K=20Q=26W=32 ,=52
9=9 F=15L=21 R=27X=33-=53
A=10 G=16M=22S=28Y=34.=54
B=11 H=17 N=23 T=29 Z=35 7=55
C=12 1=180=24U=30 =40: =56
FIGURE 1, EXAMPLE SUBSTANCE DATA PAGE FOR A TARGET COMPOUND
(TOLUENE) STORED IN AN MEM AIR MONITOR METHOD FILE.
210
-------
AIR MONITOR
V B/T/X SURFACE
8-
7-
6-
5-
4-
3-
2-
J XYLENE
D BENZENE
F TOLUENE
01:29
A 5.7
A 5.4
A 5.2 42
ABC DEFGHI JK L
FIGURE 2, EXAMPLE OF THE MEM AIR MONITOR REAL-TIME VIDEO
DISPLAY.
211
-------
MEM AIR MONITOR
Measure 11
Subtract background
Measure 15, 13, S 14]
Reliability
Criteria
Met?
Interfering
Compound'^
Issue Warning
Issue Alarm
FIGURE 3. FLOW CHART DESCRIBING THE MEM AIR MONITOR PROGRAM.
212
-------
.
o
w
en
LL)
50000
45000 - -
40000T
35000 -:
30000-
25000-
20000 - :
15000 --
10000-
5000-
BENZENE CALIBRATION CURVE
MEM AIR MONITOR MODE
00 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0
BENZENE CONCENTRATION IN AIR, ppm
FIGURE 4. CALIBRATION CURVE FOR THE DETERMINATION OF BENZENE
USING AIR MONITOR,
213
-------
Ul
en
z
o
D.
CO
UJ
cr
UJ
s:
CD
O
6.0
5.5--
5.0
4.5
4.0 --
3.5
3.0
2.5--
2.0
0.5
1.0
BCME CALIBRATION CURVE
TENAX THERMAL DESORPTION
1.5
2.0
2.5
3.0
LOG BCME AMOUNT. NG
3.5
4.0
FIGURE 5. CALIBRATION CURVE FOR BCME ISOLATED ON TENAX.
214
-------
AIR MONITOR
S GC-BCME
B BCME
8-1
7-
6-
5-
4-
3-
2-
0710
T 3.4
113
2
5
ABCDEFGHI JKL
FIGURE 6. AIR MONITOR RESPONSE FOR 1,0 PPB BCME (10 1
AIR SAMPLE),
215
-------
10708n
9843-1
8577 H
7512 H
6447 H
5381
4318
3251
2185
1120
55
f. Benzene
2. Toluene
3,4. Xy/enes
0.0 2.0 4.0 6.0 8.0
TIME, minutes
10.0 12.0
FIGURE 7, TOTAL ION CURRENT PROFILE FROM THE THERMAL
DESORPTION GC-MS ANALYSIS OF A TE-NAX FIELD
SAMPLE,
216
-------
DISCUSSION
TOM PRITCHETT: What absorbents have you looked at on your GC? Have
you evaluated different absorbents?
FRANK LAUKIEN: We have not really done very scientific studies. We have
looked at 10 or 15 different absorbents. We use either Tenax, or most of the time
for the VOCs, we use two-thirds charcoal and one-third Tenax to prevent break-
through of the very volatile chemicals. We have relied on others' experiences.
DON FLORY: I've seen a lot of advantages listed here for taking instruments
out into the field, especially in terms of turnaround time for results. But we
should note that we're dedicating these instruments to this project when we take
them out into the field. And if we dedicate the lab instruments to the project in
the same manner, we'd get about the same turnaround time.
FRANK LAUKIEN: In the field applications, getting the final report takes
just as long as it does to get a final report out of lab. But when you're in the field
making a decision the next morning, there is no way that you can get any results
from any laboratory in less than 48 hours. I have been out in the field with field
analytical equipment where samples came in at 3:00 in the afternoon and by
7:00 the next morning they needed to know what dirt to move, or someones
begins wasting about $6,000 a day. There's just no way you can meet that
schedule with afixed lab. When a site gets out of control, something unexpected
has happened. That's the advantage of a field analytical lab over a fixed lab.
217
-------
THE PREPARATION OF BUMMA CANISTER PERFORMANCE SAMPLES
AND THEIR SUBSEQUENT ANALYSIS BY THE TA6A 6OOOE MS/MS
Rita M. Harrell, Richard E. Means, and Kenneth J. Caviston
NSI Technology Service* Corporation
P.O. Box 12313
Research Triangle Park, NC 277O9
Mark Bemick and David Mickunas
Roy F. Memton, Inc. (REAC)
88A Raritan Depot
Edicon, NJ O8837
Thomas H. Pritchett
US EPA Environmental Response Team
GSA Raritan Depot
Edison, NJ O8837
William J. Mitchell, PhD
U.S. EPA Environmental Monitoring Systems Laboratory
Research Triangle Park, NC 27711
ABSTRACT
For the indoor air portion of the
Lova Canal Emergency Declaration Area
Habitability Study, summa canister
performance evaluation samples were
prepared and certified by NSI
according to specifications defined
by the U.S. EPA Environmental Response
Team. These samples were then analyzed
in the field by the TA3A 6OOOE MS/MS
using procedures developed by the ERT
and Roy F. Weston. Preparation and
analysis procedures are discussed
along with various problems which had
to be overcome initially. TAGA results
for these canisters are then presented
along with deviations fro* expected
results. Finally, these results are
compared to the data quality
objectives of the study.
INTRODUCTION
directed*
NSI was directed* by the Quality
Assurance Division of EPA/EMSL/RTP to
provide QA support to EPA's Love Canal
Emergency Declaration Area Habitability
Study at Niagara Falls, NY.
A major portion of this support was
the preparation of blind 6L and l&L
summa canisters containing the two
selected Love Canal Indicator
Chemicals (LCIC's) chlorobenzene and
chlorotoluene. These canisters,
prepared at the 2O-5O PPM level, were
used to check the performance of the
TA3A 6000E MS/MS instrument being used
in the study.
This report describes the procedures
used to clean the canisters prior to
sample preparation, preparation, NSI
analysis by BC/FID, the TAGA analysis
results, and comparison of the TA8A
results with the QA objectives
outlined in the study's Quality
Assurance Project Plan (QAPP).
EXPERIMENTAL
Canister Cleaning
Figure 1 shows a schematic drawing of
the canister cleaning apparatus used.
It consists of a vacuum pump, a 1/4 "
coiled copper tubing trap, a Dewar
flask for liquid N2, a thermocouple
vacuum gauge with a range of 1-10OO
219
-------
>um of Hg, a bubble flowmeter, an oven
suitable for heating 2 6I_ canisters
simultaneously, 1 3-way valve, 3
2-way valves, and copper connective
tubing.
Prior to canister cleaning, the copper
coil is cleaned to remove
contaiminants and H20. With valves B
and D open, the coil is purged for
about 10 min with high purity N2 while
being heated with a heat gun. After
the coil has cooled, valves B and D
are closed, the Dewar is replaced and
carefully filled. The vacuum pump is
then turned on. Valve A is opened and
after a few minutes valve C is opened.
When the vacuum gauge indicates a
vaccum of 10>um, the 3-way valve, D,
is opened to the canisters. The
canister valves remain closed until
the gauge again reads 10 /urn One of the
canister valves is then opened and
that canister evacuated to 5OO/Jim
It's valve is then closed while the
other canister is evacuated to 500 Aim.
Both canisters are then opened to the
vacuum system and while being heated
at 75-85 °C evacuated to 1O-25 ,um. The
evacuated canisters are then
repressurized with humidified zero air
and the cleaning process repeated.
In order to determine the efficiency
of cleaning, the canisters are again
pressurized with humidified zero air
and 0.250-0.5OOL samples are
cryogenically collected and analyzed
by GC/FID under the same conditions
used for analyzing standards. Under
these conditions <_ 0.5 PPB are
detectable for most VOC's. For the
Love Canal Habitability Study 25 7. of
the canisters were checked in this
manner and reevacuated for use.
Canister Preparation
A volumetric procedure was used to
prepare the two component mixtures
used in the Love Canal Study. This
procedure was selected because of it's
simplicity and the time constraints
involved for preparation, NSI
analysis, and shipment to the field.
After calculating the amounts of neat
chlorobenzene and chlorotoluene
required for preparing 6L and/or 16L
canisters in the desired concentration
range, aliquots of pure compounds
were transferred via Class S syringes
to a small septum vial, where they
were mixed thoroughly to form a
master solution. Prior to introducing
the master solution, the absolute
pressure of each canister was checked
using a absolute pressure gauge to
insure that it was still under vacuum
(see Figure 2). The canister valve
was then closed and later reopened
simultaneously with a fine metering
valve in line between the canister
and a cylinder of zero air. The flow
rate of air was adjusted so as to
maintain an absolute pressure of
about 15 PSIA. Using a class 5 gas/
liquid tight syringe, an aliquot of
the master solution was removed from
the vial and the liquid column drawn
up into the syringe barrel for
measurement. It was then injected into
the flowing air stream at the septum
fitting. After injection, any liquid
remaining in the syringe was drawn
into the barrel and measured, so that
the actual volume of liquid injected
was known. The absolute pressure at
the time of injection was also noted.
Mild heat was also applied in the
injection area to facilitate sample
evaporation. After 5—1O min the
canister valve was opened fully and
the desired final pressure reached in
2O—30 min. The valve was then closed
and the canister allowed to sit for
about Ihr before taking a final
pressure reading. This pressure was
also noted and the theoretical PPM
of chlorobenzene and chlorotoluene
in the canister determined.
Calibration
Calibration curves were prepared
using diffusion tubes (see Figure 3)
containing chlorobenzene and
p-chlorotoluene as the functioning
elements of a diffusion chamber
calibration system. Diffusion of
material through the tube neck was
predictable and was measured very
precisely. By measuring weight loss
over an extended period of time,
during which the tube was being held
at a constant temperature with a
constant flow of N2 across the tube
(see Figure 4), an accurately known
quantity of the diffused material
in the gas stream was obtained. This
in turn developed a very precise
primary standard which was traceable
to NBS. Using basic diffusion theory
the tube parameters, weight loss,
and a programmable calculator the
diffusion rate for each component
was determined in ng/min.
The 3-5 point calibration curves
were prepared by trapping chamber
stream samples cryogenically while
holding the flow rate to the trap
constant and varying the trapping
times, then analyzing them usina a
220
-------
250 °C
Tracor 550 GC/FID. An OV-17 glass
capillary column 5Om x 0.5mm ID with
a 2. 5/am film thickness Mas used.
Addtional 6C parameters were as
Tol lows i
Detector Temp.
Column Flow 5cc/min
Make Up Gas Flow 30cc/min
Cryofocusing manually Imin
before end of
trapping
Oven Program ti=30 °C
hold Imin
manually close
oven door
5 °C/min
tf=130 °C
After all analyses had been completed,
ng were determined and in this case
converted to jug before plotting
versus the corresponding peak areas.
For the period of interest,
chlorobenzene gave a linear response
for^Ajig versus area but chlorotoluene
showed some scattering at higher
concentration levels. However, on a
curve to curve basis, it was found
that 24 of 25 correlation coefficients
for chlorobenzene were between 0.99
and 1.01 with 8 values of exactly
1.00. For p-chlorotoluene 15 of 16
values were between 0.99 and 1.O1 and
again 8 values were exactly 1.00.
The curve that did not meet these
criteria was excluded.
Canister Analysis
Using the same GC/FID conditions as
for chamber analyses, 3-5 replicate
7.5 mL samples from each canister
were analyzed and the jug of
chlorobenzene and cnlorotoluene
determined for each sample. PPM were
were calculated for the individual
samples and the means of the
acceptable values used for reporting
overall canister concentration.
Figure 5 illustrates a typical
chromatogram with the appropriate ,ug
and PPM.
RESULTS AND DISCUSSION
The QAPP for the Love Canal
Habitability Study lists eight QA
objectives two of which were TAGA
6000E accuracy and precision.
TAGA accuracy was to have been
determined in all 4 phases of the
study using blind canisters supplied
by NSI. However, during the early
stages of the study the canisters
were not used as true performance
evaluation samples because NSI
analysis procedures were still in the
development stage and a standard had
not been set for acceptance or
rejection of canisters. A major
problem that NSI had to deal with in
analyzing the canisters was the
concentration level that was
necessary so that the TASA could
efficiently analyze the field
diluted samples. A gas dilution
system was unavailable and in order
to avoid overloading the GC column
and detector a relatively small 7.5mL
(5 cc/min, 1.5 min collection time)
sample from the canister was analyzed.
There were also problems with build
up and/or hold up of the compounds in
the lines and valves between the
canisters and the cryogenic trapping
system for the GC. These problems
were compensated for by allowing the
sample to purge through the entire
system for a few minutes before
sample collection and purging the
system between canisters with zero
air or helium for 2 hours with the
GC oven at 2OO °C. The purge stream
was also analyzed before proceeding
to another canister- Also, after each
day's analyses, the system was purged
overnight with the GC oven at 200 ° C.
Thus, during this development stage,
the TAGA instrument accuracy was
primarily measured using Scott
standard cylinders of the type used
for TAGA calibration. The analyzed
cylinder was never the same cylinder
which was used for the applicable
calibration. Starting in phase 2 the
16L were used to determine the
overall accuracy of the TAGA and in
phases 3 and 4 both 6L and 16L
canisters were used.
The accuracy QA objective for the
TAGA warn that the magnitude of error
in its' analysis values be 25 7. or
less from the reference values.
Relative accuracy as measured by the
Scott cylinders never exceeded an
absolute value of 25 '/. on any day
during the four mobilizations. The
largest magnitudes of the relative
errors measured by this analysis were
23.1 "/. and 23.0 '/. for chlorobenzene
and chlorotoluene, respectively.
During phases 3 and 4, the relative
error measured by the 6L canisters
never exceeded an absolute value of
23 '/. for either compound. Overall,
for all the 6L canisters analyzed by
both NSI and the TAGA the 23 7. limit
was exceeded 3 times for chlorobenzene
and 4 times for chlorotoluene.
Investigation showed that the first
two of these analyses had differences
221
-------
exceeding 25 7. because of a problem
in the TASA delivery system which was
later corrected. Table I summarizes
NSI theoretical PPM chlorobenzene and
chlorotoluene, NSI analysis values,
TABA analysis values, and the 7.
differences from the theoretical
values. The magnitude of the relative
error for the compounds as measured
by the 16L canister analyses only
exceeded the 25 '/. criteria once for
chlorotoluene during phases 2-4.
However, on that day the magnitude of
the relative errors for chlorotoluene
as measured by the 6L canisters and
the Scott cylinder were 8.8 "X. and
8.0 7., respectively. Figure 6
illustrates graphically the TABA 7.
differences from the theoretical
concentrations when plotted in the
order in which the canisters were
analyzed. In general, chlorobenzene
and chlorotoluene follow the same
pattern showing a positive bias. Mean
values of +1.7 for chlorobenzene and
+2.5 for chlorotoluene confirm this.
Figure 7 shows a similar plot for NSI
analyses. For the most part NSI
chlorobenzene results show a + bias
and chlorotoluene compliments it with
a - bias that is almost a mirror image.
Using this data NSI means (8.8 for
chlorobenzene and —8.4 for
chlorotoluene) were determined and
2 x SD used to prepare a control
chart for determining data
acceptability. These values ware 29.2
for chlorobenzene and 24.4 for
chlorotoluene. Out of the SO or BO
canisters prepared for the project
only 3 did not meet this requirement
and they were not used.
TABA precision was determined by
periodic cylinder/canister analyses.
Again the criteria to be met was 25%.
Day-to-day precision of the TABA was
actually determined from 2 sets of
replicate analyses. First, as per the
QAPP, the precision was determined
from the daily analysis of the 16L
summa polished canisters. The
maximum relative standard deviations
measured in these analyses throughout
the study for chlorobenzene and
chlorotoluene were 15.6 7. and 17.9 7.,
respectively. Because of the larger
number of analyses for each sample,
the relative standard deviations were
also calculated for Scott standard
cylinder analyses. The maximum
relative standard deviations
measured in those analyses for
chlorotoluene and chlorobenzene were
5.4 and 5.5, respectively. Both sets
of data demonstrated that the TASA
analysis met the required data quality
objective of 25 7..
In summary, it is clear that as a
whole the TABA 6OOOE MS/MS analyses
met the Quality Assurance objectives
for precision and accuracy. In those
instances where this did not occur the
problem was quickly located and
corrected.
ACKNOWLEDGEMENTS
The authors would like to thank B.J.
Carpenter, Shirley Henry, Annette
King, and Karen Oliver at NSI
Technology Services Corporation for
their contributions to the study.
REFERENCES
(1). This support was provided under
contract number 68-O2-4444 with
the U.S. EPA.
(2). Altshuller and Cohen, Anal. Chem..
8O2, I960.
222
-------
Schematic of Canister Cleaning Apparatus
Cross-section area
of diffusion path
Fine Metering Valve
-Zero Air
Balance Gas
\
Pennwolt Absolute
Pressure Gouge
«— Injection Septum
-Canister
z. Apparatus for Canister Preparation
Using A Master Solution
Diffusion Tube
Length of
diffusion path
— 4"-
FISURE 3. DIFFUSION TUBE
Carrier Gas Out
O-Ring
Water In
Clamp-
Permeation Tube
-Water Out
Isolation Tubes
Screen
Stainless Steel Cylinder
for Spacer
Carrier Gas In
FIGURE ». DIFFUSION CHAMBER CALIBRATION SYSTEM
Diffusion Tube
223
-------
o
i-i 01
9 ci
CHT SPD 0.40
Zero 10.0
Attn2T5
Aux Sgnl A
Sip Sens ).35
Area Reject 300
0.00 VLV/EXT -3
1.50 VLV/EXT 3
1.50 VLV/EXT 2
1.50 VLV/EXT 1
2.50 VLV/EXT-2
ij- -v
..._J!jdiL
Component Ret. Time Area UG PPM
Chlorobenzene 9.01 64.5110 1.2725 36.5
Chlorotoluene 12.20 69.2200 1.3048 33.3
Sample Chromatogram for Chlorobenzene
and Chlorotoluene on the Tracer 550 GC
10 20 30
Analysis Order of Canisters
40
TAGA Percent Difference by Analysis Order
224
-------
3 10 20 30 40 50 60
Analysis Order of Canisters
NSI Percent Dillerence by Analysis Order
TABLE I
CANISTER
ID
LC-0027
LC-0037
LC-0031
PEB-2
LC-OOO7
LC-0070
LC-O04O
LC-0009
LC-0012
LC-OOO3
LC-0063
LC-O047
LC-0039
PEB-1A
PEB-2A
PEB-1B
LC-0013
LC-0025
LC-OO26
LC-0052
LC-0069
LC-0072
LC-OO66
LC-OO13
LC-O03O
LC-0033
LC-O049
PEB-1B
PEB-2B
LC-0040
LC-OO03
LC-0056
LC-O012
LC-OO20
LC-0039
NSI THEO CONC
CBENZ CTOLU
38.60
32.23
26.37
34. 4O
23.45
32.52
47.80
47.60
31.44
28.78
42.13
28.27
28.81
36.60
37.55
4O.63
37.33
34.67
41. 07
33. 6O
37.33
31.47
37.33
37.33
37.33
37.88
30.40
40.63
37.10
37. 3O
28.24
35.17
33.57
43.71
38.70
PEB-1C 37.00
33.2O
27.74
22.66
29.60
20. 26
27.94
40.77
36.64
27.02
24.75
36.52
24.50
24.96
31.49
32.28
34.89
32.36
29.78
35.27
29.12
32.06
27.03
32.66
32.36
32.36
32.88
26.35
34.89
31.87
32.36
24.27
30.23
28.86
37.88
33.26
31.78
1 NSI RESULTS
! CBENZ (Dif) CTQLU(Dif)
1
[39.
!37.
125.
!4O.
1
|33.
131.
!43.
!47.
136.
i36.
',40.
!34.
126.
147.
141.
1
!42.
139.
!35.
!32.
138.
!39.
!34.
1
!34.
139.
!34.
!39.
132.
142.
!37.
141.
134.
!41.
!36.
!4O.
136.
138.
2(+1.6)
0(-H4.8)
6(-3.O)
3(+17.2)
9 ( +44 . 3 )
5 ( -3 . 1 )
5(-9.0)
4 ( -O . 4 )
7(-t-16.9)
0(+25.O)
4(-4.O)
0 ( +20 . 1 )
2(-9.O)
0 ( +28 . 4 )
2 (+9. 6)
1 ( +3 . 6 )
4(+5.5O
7 ( +3 . 0 )
3 (-21. 4)
2(+13.7)
K+4.7)
2(+8.7)
5(-7.6)
4(+5.5)
4(-7.4)
2(+3.4)
0(+5.3)
1 ( +3 . 6 )
1(0)
7(+11.8)
5(+22.1)
5(+18.0)
1 ( +7 . 5 )
6(-7.1)
5(-5.7)
4(+3.8)
32
18
21
25
15
27
33
23
26
24
25
23
25
33
30
34
26
28
32
31
33
26
30
26
29
27
23
34
30
28
23
29
29
33
33
30
.0(-3.
.4(-33
.8(-4.
.5(-13
.9(-21
.9(0)
.9(-16
6)
.6)
0)
.9)
.3)
.9)
.8(-35.0)
.5(-l.
.6(-0.
. 4 ( -3O
.l(-5.
. 5 ( +2 .
.3(+5.
.6(-5.
.3(-l.
.8(-17
. 9 ( -3 .
. 3 ( -8 .
. 3 ( +7 .
.4(+4.
.8(-0.
. 1 ( -7 .
.B(-17
.0(-10
.3(-17
.1(-12
.3(-l.
.2(-5.
.7(-ll
. 4 ( -3 .
.7(-l.
.6(+2.
-8(-10
. 3 ( +O .
. 4 ( -4 .
9)
8)
.4)
7)
0)
4)
3)
7)
.2)
0)
4)
5)
2)
9)
8)
.2)
.4)
.0)
.30
7)
2)
.3)
6)
8)
6)
.8)
1)
3)
TABA RESULTS
CBENZ (Dif) CTOLU (Dif)
26
22
25
31
27
32
48
43
33
39
4O
21
31
37
36
46
42
34
42
42
41
32
41
37
37
36
32
4O
40
42
30
39
37
35
30
38
.K-32.1)
. 4 ( -30 . 5 )
.6(-3.0)
.7(-7.8)
.4(+16.8)
.2(-1.0)
.9(+2.3)
.8(-7.9)
.3(+5.9)
. 6 ( +37 . 6 )
.6(-3.6)
.6(-23.6)
.6(+9.7)
.9(+3.6)
.3(-3.3)
.K+13.5)
.5(+13.B)
.3(-l.l)
. 7 ( +4 . 0 )
.0(+25.0)
. 0 ( +9 . 8 )
.5(+3.3)
•0(+9.8)
. 4 ( +0 . 2 )
.9(+1.5)
.9(-2.6)
.6(+7.2)
.2(-1.0)
.0(+8.0)
.9(+15.0)
. 0 ( +6 . 2 )
.9(+13.8)
.5(+11.7)
.0(-19.9)
. 2 ( -22 . 0 )
. 2 ( +3 . 4 )
19
14
21
36
23
28
34
39
29
35
36
13
3O
30
31
40
31
29
38
36
36
29
36
27
34
32
29
36
37
37
27
35
34
30
27
34
.5(-41.3)
.K-49.2)
.8(-4.0)
. 9 ( +24 . 7 )
.4(+16.1)
.7(+2.7)
.8(-14.6)
. 1 ( +6 . 7 )
. 6 ( +9 . 5 )
.2(+42.2)
. 6 ( +0 . 2 )
.4(-45.3)
. 9 ( +23 . 8 )
.8(-2.1)
.7C-1.8)
.7(+16.6)
. 0 ( -4 . 2 )
. 5 ( -0 . 9 )
.B(+10.O)
.2(+24.3)
.6(+14.2)
.4(+8.8)
.9(+15.1)
.5(-15.0)
.K+5.4)
•2(-2.1)
.K+1O.4)
. 2 ( +3 . 8 )
.8(+18.4)
.9(+17.1)
.K+11.7)
.9(+18.8)
.0(+17.8)
.9(-17.B)
.7(-16.7)
. 4 ( +8 . 2 )
16L canisters arm indicated by PEB in canister ID listings.
225
-------
DISCUSSION
AVRAHAM TEITZ: What do you use for confirmation?
TOM PRITCHETT: We sample the air stream until we see the signal go back
up for the ions we're looking for. And then, depending on the type of analysis
we 're doing, we' 11 pop either a Summa canister or use some other confirmation
analysis. If we're doing time-weighted average sampling in conjunction, we
use the time-weighted average results for confirmation.
But typically, the compound is an unknown. We can't really do too much,
except a Summa canister off to the side, at the same time we're sampling. We
go out with anywhere from two to four evacuated Summa canisters for doing
these confirmation analyses.
AVRAHAM TEITZ: Would that be analyzed the standard way, by GC/MS?
TOM PRITCHETT: Yes. One of the things we're considering is having an on-
board portable GC, so at least we have the confirmation by retention time.
UNIDENTIFIED PARTICIPANT: Does the sensitivity change by the same
factor for different compounds?
TOM PRITCHETT: We haven't looked at all the compounds. The compound
I described ionizes by charge exchange. The reason the sensitivity went down
is that the source chemistries were in competition. As we increased the amount
of water present, it actually scavenged the regent ions with greater ionization
potentials.
I've monitored multiple parent ions for the same compound and seen the
relative intensity of those parents, particularly oxygenated compounds, change
through the day. In some cases, one parent, the M+ ion, will increase in
sensitivity while another, the M-l, decreases. This work was with parent long
only and was independent of any variation in fragmentation.
UNIDENTIFIED PARTICIPANT: Have you looked at using some internal
standards, since you know what compounds you are looking for?
TOM PRITCHETT: When we know what we're looking for, we have
considered using isotopic internal standards. There is so much unknown going
on in source chemistry, we don't know how compounds track with each other.
We haven't been able to do these comparisons under controlled conditions.
We are seriously considering using an isotope in surrogate standards to be
continually doped into the mixture. If we see the intensities of those surrogates
changing, we could do a series of spikes in that matrix in order to gather a data
base on changes in relation to response, which is needed to start using actual
internal standards.
UNIDENTIFIED PARTICIPANT: Do you have any experience with the
atmospheric pressure source of the TAGA? Does it have similar problems?
TOM PRITCHETT: The group leader in our contractor staff came out of a
consulting company that did an extensive amount of atmospheric pressure
chemical ionization (APCI). They ran into some problems with ammonia and
amines.. Also, Brian Stewart of the National Research Council has done some
controlled humidity studies with an APCI source.
He started off with very dry air, and there was almost no sensitivity. As water
was added, the sensitivity rose drastically and then started tailing back off.
Initially there is no water chemistry, because there's no water present. The
sensitivity tailed off once he started getting too much hydration. All you're
doing is pushing the hydration up to higher levels. You actually decrease the
acidity of your reagent ions, because as you hydrate a hydronium ion, its pH
actually drops a little.
MR. PRITCHETT: What we're talking about is actually humidity effects
before you do any declustering. It's the cluster itself that protonates the
compound in the source, not the hydronium ion. For example, the 37 ion or the
55 ion is actually doing the protonation. As you increase the 57 in the source,
versus the 37, you actually decrease the effective acidity of your reagent gas.
MARC WISE: Obviously, the way to get around these problems is to
continuously inject an internal standard into the gas stream. Do you have a
viable means of doing that at this time?
TOM PRITCHETT: As a matter of fact, we do. We're using gas standards to
dope into our sample measures. We do hour calibrations in ambient air, so it's
just a matter of putting a very low level, let's say to 10 ppb, of a continuous feed
into the sample matrix. You want to make sure you pick some surrogate
compounds that aren't going to cross interfere with some of your other
compounds.
The main reason we haven't gone to internal standards is that we don't know
enough about the charge exchange source chemistry to predict which com-
pound tracks which. In other words, if I see this response go down for this
compound, what does that mean for the response of compound Y?
Something else we found with these chemical ionization chemistries is that if
you change the sample matrix, you drastically change your calibration.
I used to go out with this instrument, calibrate in ambient air, and then stick it
in an incinerator, monitoring the stack gas. The CO2 concentration and the
absolute humidity concentrations are not comparable. We look for that now. If
we change the reagent gas, or if the matrix changes, we do some type of QA/
QC to verify that we haven't changed the response factors.
MARC WISE: So your response factors that you determined twice during the
day and used as intermediates are in the presence of the compounds you're
analyzing? You determine those with the compounds?
TOM PRITCHETT: Yes. We measure the response factors in the matrix
before and at the end of the day. We do the spikes into the matrix - two-point
spike. And it's always in the sample matrix that we're analyzing.
MIKE FINNIGAN: It seems that ionization is an exceedingly complex
problem. You've mentioned there are three methods of chemical ionization -
what compounds are ionized by what mechanism, what mechanism is affected
by the presence of CO2 or water vapor, and so on.
Why not add a short stage of chromatography in front of this to space out these
compounds - to allow the chemistry in the source to be dominated either by
chromatography, or some other technique, as you do in an ion trap, so that the
chemistry in the source is a fixed one, and not just dependent upon the water
that's coming in with the sample or the CO2?
TOM PRITCHETT: That is an advantage of your system versus what we're
dealing with. You've got your source chemistry stable. We haven't gone into
depth in doing that, primarily because this problem was known for a long time.
If I could figure out how to do it, (add a short GC and retain the calibration
source chemistry) and make that transition quickly, I probably would shift that
way, because I would like to be able to use some chromatography on the front
end. It would really give me some real strong advantages being able to
distinguish between compounds.
226
-------
RESULTS FROM THE ENVIRONMENTAL RESPONSE TEAM'S PRELIMINARY
EVALUATION OF A DIRECT AIR SAMPLING MASS SPECTROMETER
Robert E. Hague
Department of Environmental Science
Cook College, Rutgers University
New Brunswick, New Jersey 08903
Thomas H. Prichett
U.S. EPA Environmental Response Team
GSA Raritan Depot
Edison, New Jersey 08837
Kwong Cho
Roy F. Weston, Inc. (REAC)
GSA Raritan Depot
Edison, New Jersey 08837
Ben Shapiro, formerly of
Enviresponse, Inc. (EERU)
GSA Raritan Depot
Edison, New Jersey 08837
ABSTRACT
During the summer of 1987 an investigation of the
Bruker MM-1 mobile mass spectrometer was performed
under the auspices of the EPA Environmental
Response Team (ERT) in order to evaluate this
instrument's applicability for emergency response
and Superfund-related tasks. The unit was tested
with a series of ambient air samples spiked with
mixtures of organic compounds over the low-ppb
to low-ppm range. An evaluation of the
instrument's suitability was made on the basis of
sensitivity, specificity, ruggedness and overall
usefulness under field conditions.
INTRODUCTION
The instrument evaluated in this paper is the
Bruker mobile mass spectrometer model MM-1
manufactured by Bruker Instruments, Inc. The
MM-1 is a quadropole mass analyzer and electron
impact ion source mass spectrometer which has been
designed specifically for use under field
conditions. The instrument and its ancilliary data
handling package is self-contained, and is designed
to be ransportable and operable while mounted in a
four wheel drive vehicle. During mobile operations,
power is supplied using a rechargeable 24 volt
DC battery.
Operation on a real-time analysis format. The air
sample is drawn into the instrument at a constant
flow rate in ambient air and analyzed instantane-
ously and continuously. There is no separation
of organic compounds or sample preparation. All
compounds are analyzed simultaneously. Analytical
results, instrument parameters, and listings of
identified compounds are displayed on a video
screen and continuously updated. During routine
air sampling, sets of four ion peak intensities
representative of a given compounds are displayed
as histograms on the screen on a real time basis.
Histogram height varies as the base 10 logarithm
of the ion's intensity. The intensity scale range
allows measurement over 7 orders of magnitude. All
results are normalized, so that direct comparison
of measurements from all analyses can be made.
Additions to the spectrum library may be made by
either defining a compound in terms of ion intensity
ratios entered from the keyboard, or by exposing
the sampling head to the vapor of that compound,
and selecting ion mass frequencies from the
resulting spectrum. Printed copies of analyses,
library contents and system status may be displayed
at any time on the CRT or downloaded to a small
panel-mounted printer.
Sampling is performed using a sampling line
directly coupled to the mass spectrometer,
both of which are operated by a simple control
keyboard. (Figure 1) Samples are introduced
into the system via heated sample head located at
the end of the sample lilne. The sample head is
a nickle gauze coated with a semipermeable
silicone membrane. Samples are pulled through
this membrane by a sampling pump at a rate of 1 to
3 cc/nimute. The membrane serves to protect the
sample line itself, which consists of a 3.5 meter
0.32 mm quartz capillary column coated with an SE-
54 Phase surrounded by an insulated, heated jacket.
At the end of this line, the sample is drawn
through a second silicone membrane and into the
mass spectrometer. In the event of line
contamination, there is a backflush system. The
samplilng line is capable of temperature ramping
and some primitive separation of compounds can be
achieved. Once inside the mass spectrometer, the
compounds are ionized utilizing electron
bombardment by an electron source.
In general, gas chromatographic/mass spectrometer
systems are designed to present individual
molecular species to the mass spectrometer follow-
ing their separation from the sample matrix on the
chromatographic column. In real time analysis,
such as is the case with the MM-1, the compounds
are not separated and are presented to the detector
simultaneously. The ions from different compounds
are combined and the resulting ion mass ratio is
the s>im of the ions from all of the compounds
present in the sample. This leads to the previo-
usly mentioned difficulties in compound
identification and quantification. In most
environmental analyses of compounds, the three
most prevalent characteristic ion masses of each
compound are utilized in identification. In an
effort to circumvent the problems encountered in
227
-------
real-time analysis with identification, the MM-1
utilizes four characteristic ion masses and their
abundance ratios. Even so, with complex mixtures
of compounds, the detector has difficulties with
identification, and false positive and false
negatives are common.
EXPERIMENTAL
The MM-1 was tested over a range of concentrations
representative of both chronic and acute emission
levels of organic compounds in air. A dual sample
dilution manifold was constructed to provide the
capability of diluting cylinder gas standards
in ambient air over the ppb to ppm range.
The system consisted of two parallel sample dilut-
ion lines of differing diameters with side-tapped
glass and Teflon sampling ports. Both manifolds
use individually sized mass flow controllers
at their discharge ends and share common diluant
sources, (ambient air and zero air) common organic
vapor sources and a common low pressure multi-
stage blower. Use of ambient air as the diluant
allowed evaluation of the instrument under actual
daily variations in temperature, humidity, and
background concentration. The manifold lines
were sized so as to provide usable concentrations
over a feasable sampling interval. Two lines
were used. A 7/8 inch ID Teflon line for the high
dilution line provided metered flow values of 30-
150 liters/miln. for dilutions to between 5 to
100 ppb. A 3/8 inch ID stainless steel line
providing metered flow rated of 2-7.5 liters/min.
for vapor concentrations over the range from 1.0
to 10.0 ppm. Both lines were heat traced and
maintained at 40 degrees C. Concentrations at
each dilution were verified by gas chromatography
before being presented to the MM-1.
The shared vapor sources were of two types: (1)
Standardized NBS traceable multi-componant
compressed gas mixtures that were metered
through a mass flow controller into the sample
lines for ambient air dilution to produce the
desired concentration and, (2) a heated
vaporization source for compounds which are
unsuitable as pressurized cylinder standards.
The vapor source was pairs of midget blass
impingers containing the organic compound of
interest kept in a thermostatted water bath.
A controlled air fow was passed through the
impinger train, saturated with vapor, and diluted
to its final desired concentration with ambient
air.
The evaluation of the MM-1 was divided up into
four phases. Phase 1 evaluated the instrument's
lilnearity and sensitivity by providing
progressively higher dilutions of standard gas
mixtures cylinder by cylinder over the test
concentration range. A "detectability limit'
for those compounds in that mixture determined.
Phase 2 presented the instrument with unknown
concentrations of known compounds within the list
of compounds presented to in Phase 1. Phase 3
presented the MM-1 with mixtures of compounds
from a list of non-cylinder-stable compounds,
including aniline, pyridine and m-creasol. The
analyst was not provided with information on
compounds to be expected or their concentration.
A list of the compounds present in each cylinder
is given in Table 1.
RESULTS
Phase 1 testing indicated that sensitivity was
chiefly limited by interferences from the other
compounds present in the mixture. Typical high-
sensitivity compounds were detectable at 10 ppb.
(Table 2). The detectability of a given compound
was a function both of the other compounds pre-
sent and the ambient background concentration of
organic materials. Complex mixtures create inter-
ferences which raise the detection limits consid-
erably. Common-ion effects apparently prevented
the MM-1 from recognizing some compounds even at
the 10 ppm level. However, it should also be
stated that the linearity of response for many
of the compounds was quite good, with correla-
tion coefficients exceeding 0.95 over three or-
ders of magnitude. Typical response curves are
shown in Figures 2 to 5.
Phase 2 the results of tests performed are given
in Tables 3 & 4. in complex mixtures, both
false positives and negatives were common. For
those compounds whcih were correctly identified,
quantitation error ranged from +2.2% to -39.1%
and depended on the specific compound. Variation
in response for replicate samples was less than 8%.
Phase 3 presented a series of mixtures to the in-
strument whose concentration and composition were
unknown to the analyst. An effort was made to
select compounds which are of interest at Super-
fund sites, but which were not available as cy-
linder gas standards. Calibration curves were
not prepared for these compounds and the emphasis
was strictly on qualitative identification . The
results of Phase 3 are shown in Table 5.
OVERALL COMMENTS
•SENSITIVITY AND ACCURACY
When single compounds with no common ions are be-
ing analyzed, the sensitivity of the MM-1 under
typical ambient conditions ranged from 10 to 25
ppb. The stability of the instrument is such
that repeat readings over a period of time agree
within a fraction of a response unit. Linearity
of response was found to be good over the range
of 10 ppb to 10 ppm for many compounds. As
stated previously, mixtures of compounds sharing
one or more common ions drastically reduces both
the sensitivity and accuracy of the instrument.
Accuracy was also found to be low where compounds
not included in the target library were present
in the mixture.
"PORTABILITY
Subsequent to the Phase 3 testing, the MM-1 was
mounted in a four wheel drive vehicle and field
tested. It was found that the durability of the
unit was truly remarkable. After off the road
use, the instrument still retained its calibra-
228
-------
tion and full internal' vacuum. The battery pack
was found to generate sufficient power for a 6
to 8 hour operating schedule before recharging
was necessary.
Setup time is minimal at about 15 minutes with
an additional 30 minutes for calibration and
sample line purging. The analysis time is approx-
imately 15 seconds, with triplicate analyses in
less than one minute.
•LEVEL OF OPERATOR TRAINING REQUIRED
The Bruker MM-1 is designed with extremely simple
operating procedures based on a series of menus
which are all accessable from a simple keyboard.
This allows an operator to analyze ambient air
samples with a minimum of training. It should
be noted however, that although the instrument
is simple to operate, the data have limitations
and should be assessed by an experienced mass
epectroscopost who is aware of potential interfer-
erences and the limitations inherent in real-time
analysis.
The experienced operator may in some cases be able
to recognize false positives and false negatives
by reviewing the ion mass list and abundances of
those ions. In complex mixtures of organic com-
pounds, useful data may be limited. With this
understanding, an operator will be better pre-
pared to recognize any problems which may occur.
"RELIABILITY
The reliability of the instrument from a mechani-
cal standpoint was exceptional. Over the course
of three months, the only breakdowns were a leak-
ing calibration gas valve and an electronics over-
load caused by a power failure and subsequent
voltage surge. The manufacturer responded prompt-
ly in both cases, and had the instrument function-
ing perfectly within one day.
In order to evaluate the unit under true field
conditions, during the last part of the study,
the instrument was mounted in a four wheel drive
vehicle and used as a mobile unit at a Superfund
landfill site. It experienced no mechanical pro-
blems even while being operated under conditions
of high temperature, high humidity and while be-
ing driven off-road. The data obtained however,
again were limited in their usefulness by high
concentration in the soil gas samples of a wide
variety of miscellaneous hydrocarbons, which,
given this instrument's lack of chromatographic
capability allowed only a few compounds to be
tentatively identified.
•OBSERVATIONS AND RECOMMENDATIONS
In the configuration used in this study, the in-
strument is not applicable to most site assess-
ment work, unless it is known that a relatively
small number of compounds known to be present,
or if the compounds of interest are present at
concentrations significantly higher than potent-
ial interferants. The instrument is designed for
real-time air monitoring and performs well in
this capacity. The complex mixtures of compounds
which can be found in waste sites are likely to
create identification problems with false negat-
ives and positives occuring.
In the case of a chemical spill emergency, given
knowledge of the compound or compounds released,
the instrument would be the method of choice in
delivering quick and precise data. Its trans-
portability and ruggedness would allow a plume
of vapor to be readily traced and ouantitated.
A new gas chromatographic attachment is now avail-
able. Although it was not evaluated in this
study, manufacturer's information indicates that
many of the problems found in this study should
be resolved by this modification.
REFERENCES
(1) Bruker-Franzen, "Analytic GmbH,' The MM-1
Mass Spectrometer User Manual. Bruker-
Franzen Analytik GmbH, Bremen, West Germany,
1986, pp. 2-28.
(2) Bruker-Franzen, 'Analytic GmbH,' The MM-1
Mass Spectrometer User Manual. Bruker-
Franzen Analytik GmbH, Bremen, West Germany,
1986, pp. 2-32.
TABLE 1.
CYLINDER GAS STANDARDS—PHASES 1,2
TANK A
Toluene
1,1,1-Trichloroethane
1,4-Dioxane
Acetone
1,2-Dichloroethane
TANK B
Vinyl Chloride
Benzene
Methylene Chloride
1,1-Dichloroethylene
Trichloroethylene
TANK C
Methyl Ethyl Ketone
n-Hexane
Methyl Isobutyl Ketone
Tetrachloroethylene
1,4-Dioxane
TANK D
Cyclopentane
Ethyl Acetate
1,1-Dichloroethane
1,1,2-Trichloroethane
Carbon Tetrachloride
TANK E
Chlorobenzene
o-Chlorotoluene
TANK F
Isopropanol
Ethyl Ether
3-Chloropropane
Styrene
Ethyl Benzene
Freon 11
229
-------
Table 2. MM-1 Sensitivity and Selectivity
of Compounds in Phase 1 Mixtures
Compound Limit of Detection
(ppbv)
Table 3. Phase 2 Mixtures
Acetone
Vinyl Chloride
Cyclopentane
Benzene
Methylene Chloride
Hexane
Ethyl Acetate
1 , 4-Dioxane
Toluene
1 , 1-Dichloroethylene
1 , 1-Dichloroethane
Ethyl Benzene
Chlorobenzene
o-Chlorotoluene
Trichloroethylene
1,1, 1-Trichloroethan
1,1, 2-Trichloroethan
Tetrachloroethylene
Styrene
Tr ichlor of luorome thane
Ethyl Ether
Methyl Isobutyl Ketone
1, 2-Dichloroethane
Methyl Ethyl Ketone
Isopropanol
3-Chloropropene
Carbon Tetrachloride
100
1000
10
25
10
25
10
1000
25
25
100
5
10
10
25
100
25
25
100
ND
1000
10000
100
ND
ND
ND
ND
ND—Not Detected Due to Common Ion Effects
Compound i
Acetone
1,4-Otonnt
1,2 DIcMonxethine
1,1,1 Trlchlorwthine
Toluene
Chi oro benzene
o-Chlorotoluene
Acetonltrlle
DIchloroKthine
Acittldehyde
(pebj Identified H*?.
icc.s >
423.2 >
47S.7 X
474. S *
467.7 X
522. ( x
527.2 I
Filie
rot.
x
X
I
W-l Response: .Unknown II
Compounds
Benzene
Vinyl Oilorlde
Hethylene Oilorlde
1,1-Olchlorwthylene
Trlchlonxthylene
Ethyl Benzene
Ethyl Ether
Freon-11
Styrene
Acetonltrlle
Allyl Oilorlde
Cotipoundj
Hethyl Ethyl Ketone
Hexine
1,4-Dloxine
1,1 Olchloroethine
1,1,2 Trkhloroethme
Cyclopent4ne
Ethyl Acetite
Cirbon Tetrjchlorlde
Tetrjchloroethylene
Acetonltrlle
Methyl Isobutyl Ketone
Cone. F«He
(ppb) IdenttflKl Keg.
133.7 X
133.7 >
134.4 x
125.3
142.4
264.7
249.3
255.3
279.2 X
249.7 >
W
-------
Compound
Taiis 4. Accuracy of Pia.sa 2 Responses
Generated Measured Percent
Concentration (ppb) Concentration (ppb) Deviation
CMorobenjene
Toluene
o-Chlorotoluene
522
463
527
321
507
540
-38.5
+ 7.6
+ 2.5
Compound
Accuracy of W-l Response: Unknown 12
Generated Measured Percent
Concentration (ppb] Concentration (ppb} Deviation
Benzene
Dlchlororoe thane
Styrene
134
134
279
137
US
170
+ 2.2
+15.5
•39.1
FIG. 2 METHYLENE CHLORIDE
5 -
106
Compound
Accuracy of HM-1 Response: Unknown 13
Generated Measured Percent
Concentration (ppb) Concentration (ppb) Deviation
FIG. 3 O-CHLOROTOLUENE
1,1,2 Trlchloroethane
Ethyl Acetate
1,1 Dlchloroethane
Methyl Isobutyl Ketone
258
252
247
131
NQ - detected,
155
219
229
NQ
not quantified
-39.9
-13.1
- 7.3
Accuracy of MH-1 Response: Unknown 14
Compound
Vinyl Chloride
Cyclopentane
1,1-Olchloroethjne
Trkhloroethene
1,1,2,-Trtchloroethane
Generated
Concentration
251
564
569
263
593
Measured
Concentration
170
607
555
165
532
Percent
Deviation
-32.2
+ 7.6
- 2.5
-38.4
-10.3
Table 5. Phase 3 Unknowns
Unknown (tl
Concn.
(ppm)
False False
Identified Positive Negativ
Aniline 1.34
m-Cresol 0.5
Pyridine 45.0
Ethoxyethanol 7.0
Acetaldehyde
Unknown #2
Concn.
(ppm)
False False
Identified Positive Negativ
ro-Cresol
o-Xylene
Aniline
Acetaldehyde
Ethyl Benzene
Unknown #3
0.5
79.0
6.7
Concn.
(ppm)
False False
Identified Positive Negative
Chloroform 7.8
Butyraldehyde 4^6
n-Hexane 6 0
n-Butanol
Dichloromethane --
I *
FIG. 4 1.1,2-TRICHLOROETHANE
FIG. 5 CYCLOPENTANE
t
OJ
0
tfo
ite
231
-------
DISCUSSION
MIKE FINNIGAN: The samples that you picked and the volatiles and the
artificial standards, consisted of four to six compounds. You got up to two of
them, three of them, and so on. What is the real world when you go out to a site?
How many compounds are you going to wind up seeing, and what's your
prognosis about that?
TOM PRITCHETT: It depends. We typically run with four compounds per
cylinder. We run two cylinders of mixtures. I have been tasked to go out and
monitor for as many as 12 to 13 compounds simultaneously at a waste site. I
have analyzed headspace samples (leachates) where I had the world come at
me. You name it, it was there.
At other places, you're only looking for one or two compounds. And that's all
you're really going to see. We gave it a worst-case scenario. I'd say a good 60%
to 70% of the sites don't ever run into that mishmash. But because it worked
well at one site, doesn't necessarily mean it's going to work well at another,
particularly once you start throwing in the typical hydrocarbon soup you see
around multiple-use landfills.
We said the compound mix shown was a worst-case scenario. Actually I've
seen compound mixes that are much, much worse from a landfill. Three
quarters of the compounds in that chromatogram were hydrocarbons (a hydro-
carbon soup). You'll see the oxygenated hydrocarbons in there also, and some
chlorinated compounds.
If PCE and TCE are in a landfill, and it starts biodegrading, you can start
walking down the degrees of chlorination until you hit vinyl chloride and
chloroethane. You'll see all levels of chlorination if you've got those two
compounds, and you've got a biologically active landfill.
You'll also see all levels of oxygenation there - aldehydes, ketones. That is a
real landfill.
You can have a very clean site at a spill, or it could be a multi-use landfill where
you have the world thrown at you.
ARTHUR BOYER: I've wondered if you've developed any methods or rules
of thumb when to use what instrument? For example, should I use GC/MS?
Should I use MS or should I use MS/MS?
TOM PRITCHETT: I don't think anyone who does this work has really
locked on what to do when. As you learn from experience, you find certain
things work at different sites. New technologies come out. I do things now for
soil gas analyses that I wouldn't even have tried two or three years or six months
ago. I'm always finding new tools to use and I'm learning more about the tools
that I've got. I'm also realizing maybe this one tool wasn't as good as I thought
it was. I can look back at what I did a year ago, or two years ago, I realize now
what was occurring then that I just didn't see then.
JOE SOROKA: When you do some GC/MS, you try to match the peaks you
get off the GC with the library TIC and use your own intuition and knowledge
to look at the spectra to define what that particular component might be.
Are there other techniques with these kinds of direct-sampling air spectrome-
ters that don't have the separation advantages that you get with a GC at the
forefront? Or can you feel confident that you've possibly got most of the stuff?
What would happen if you have a sample which may have things such as
tetraethylene lead compounds, which probably are not in your library, and were
totally unexpected? We have found them in certain samplings.
TOM PRITCHETT: If Ijusthad one spectrum to look at, there are certain ions
that are fairly universal and there are others that you just can't make any sense
of. Some compound classes will stick out, and you can pull them out fairly easy
from a real mishmash. Others you may very well end up missing compounds
because you can account for all the ions for the compounds that you are seeing.
But there may be something else hidden, because unless you do really quanti-
tative spectral subtracting, the compounds may be lost in the differences of
what you're looking at. It depends on how complicated the spectra are. That's
why if you're doing unknown work, you should use GC on the front end. By
the way, this is a problem that occurs with any direct-air sampling instrument.
JOE SOROKA: On the Bruker, for example, is there an algorithm within their
software that can identify or account for all of the peaks that are in the spectra,
except for particular ones, and highlight them so that you can take a look at
them, and try and figure out where they come from?
FRANK LAUKIEN: When you come to a site, and you're not sure what
you're looking for, we recommend starting in the single-ion monitoring mode.
In many cases, the compounds you're monitoring for will be identified and r
quantitated without any additional interference.
Then there are cases where you see there are some identifications, and there's
a lot of unidentified ion activity on the screen. The next step is to switch from
single-ion monitoring to full-scan mode to see whether there are some com-
pounds which you weren 't monitoring for in single-ion monitoring mode. That
may result another range of problems.
However, you will run into those problems where there is a lot of interference
- even in the full spectral mode, there is too much interference for identifica-
tion. Or as you've described, there might be some additional ions which simply
don't fit into any picture. In this case, you switch over to the chromatography.
232
-------
Introduction to the
Session on Immunochemical Methods
Jeanette Van Emon, Chairman
I would like to begin the session by explaining the Agency's efforts in
immunochemistry. EMSL-Las Vegas is the lead laboratory within the Agency
for immunochemical methods, and our program consists of two facets
development of immunoassays and the evaluation of these techniques.
Our in-house efforts are partially conducted through Lockheed, ESCO, and
we've just had the opportunity to hire Richard White, who comes to us from
industry with the knowledge of development and evaluations of immunoassay
kits. We are also have efforts through the University of Nevada- Las Vegas, En-
vironmental Research Center by Kaz Lindley an analytical chemist who is
eagerly embracing immunoassay techniques as well as helping us out svith con-
ventional wet chemistry. Development efforts are also supported by our
contractor, Acurex, who is busy preparing work for us now. We have students
from UNLV, particularly Julie Alamo, who assists me in just about even' detail
of my program.
Our in-house efforts are supported by the U.S. Department of Agriculture
(USDA). We have an inter-agency agreement with them through the Food,
Safely and Inspection Services. We are also networking among other govern-
mental agencies.
We have cooperative agreements with the University of California at Davis
with Drs. Bruce Hammock and Mark Kurth. Those of you familiar with
immunoassay for environmental contaminants are already quite familiar with
the name of Bruce Hammock. We're very fortunate to have these laboratories
intimately involved in our program. Right now they are developing methods tor
analyzing nitro aromatic compounds.
We also have a cooperative agreement with the University of California at
Berkley, the hybridomas facility headed by Dr. Alex Karu. This is a very fine
monoclonal laboratory. We're also fortunate to work with this facility.
We had a cooperative agreement just recently awarded to Westinghouse Bio-
Analytic Systems for developing immunoassays for some of the aromatic
hydrocarbons.
We are also working with the private sector in coordinating evaluation studies
- again, with Westinghouse Bio-Analytic Systems. We just finished an evalu-
ation study for their immunoassay on pentachlorophenol. and now w e are ready
lo go to our next step, the field demonstration.
We are \\ orking on evaluations with other governmental agencies — the USDA
evaluations criteria committee and with the Association of Official Analytical
Chemists lAOAC). We have proposed Agency guidelines on evaluation
studies. A general referee for immunochemical methods has recommended that
these guidelines be incorporated further into AOAC's methods committee's
guidelines on evaluation of immunoassays. And we are beginning to work with
the FDA, which probably stems from the efforts of the Office of Technology
Assessment and their recent assessment on immunoassay.
Regarding the developmental side of our program, there are many reasons the
Agency needs to develop or wants to develop immunoassay techniques.
Hrst, the technique is easily amenable to analyzing human body fluids and
biomarkers, so exposure assessment studies become very real. Immunoassay s
have a high sample capacity so you can get data in real time. And you can studs1
animal populations around Superfund sites, or even human populations.
The technique is rapid. You can perform analyses on site, or you can outfit a
mobile van and use laboratory-type based immunoassays. ForSupert'und sites.
remedial actions and designated hot spots, can be monitored.
Immunoassay is capable of analyzing some products of biotechnology. In fact.
for analyzing products of genetically engineered microorganisms, immunoas-
say may really be the only method of choice.
Our evaluation process works like this. For immunoassays submitted to the
EMSL-Las Vegas, there are certain developmental criteria we would like to see
fulfilled before we undertake an in-depth evaluation study.
One of these is that the assay is well characterized, or mature, has a standard
operating procedure, accompanying QA/AC. data quality objectives, and so
on. Depending upon the fulfillment of these developmental criteria, the next
step would follow.
This would be either a laboratory confirmation, or a laboratory evaluation. It
the assay is not well characterized, then the Agency would do a very minor type
of evaluation study, or a preliminary evaluation.
Assays that are well characterized undergo a more detailed evaluation. After
fulfilling a successful single-laboratory or multi-laboratory evaluation, the
assay would go out on site. We have several demonstrations planned for the up-
coming year - first, the pentachlorophenol, w hich could de\ clop into a rather
large study for providing some useful data on pentachlorophenol degradation,
as well.
The important part of immunoassays. or any new technology, is implementa-
tion. So how do we do this? The Agency is working on this, as well.
First is the validation or evaluation of immunoassays, for w hich the Agency has
an ongoing program.
Next is to introduce immunoassay s into anal) tical laboratories. We also have
efforts in this area. We recently participated in an FDA-sponsored workshop on
immunoassay1 technology. We assisted the Congressional Office of Technology
Assessment (OTA) in their assessment of immunoassays w hich they term, an
emerging technology. 1 was fortunate to present ,i seminar to the OTA in
Washington. D.C. a couple of months ago.
The Agency's immunoassay program was chosen as a subject for a computer-
animated graphics program package. We are hoping that this can be used as a
first-level training tool for persons who .ire interested in immunoassay. to be
followed up with more detailed training programs.
The selection ot initial compounds is important, as we want to make sure that
we're giving the technology a lair assessment. The problems with new
technologies are thai you give them compounds that are impossible to analyze
by other means. Immunoassay is amenable to a w ide range of compounds,
probably more so than any other analytical technique. How ever, it is important
to choose the compounds for assay development that w ill also fulfill Agency
and environmental monitoring needs.
We have run announcements in Commerce Business Dailv aimed more at the
business community, and in Seie/iee. aimed at academicians, to find out w hat
has already been developed and who is working on this type of technology. so
w e can make more effectix e use of our resources. In fact, these announcements
resulted in several unsoticitated proposals.
We're compiling a list ot assay systems that have already been de\ eloped, and
also a list ot compounds ot high priority tor immunoassay development for the
Agencv.
233
-------
DELIVERY SYSTEM FOR RAPID, SEMI-QUANTITATIVE ANALYSIS OF
LOW MOLECULAR WEIGHT CONTAMINANTS AND RESIDUES
Philip L. McMahon
Director, Research and Development
IDEXX Corporation
Portland, Maine 04101
Robert Suva
Research Scientist
IDEXX Corporation
Portland, Maine 04101
Chris Brooks
Project Engineer
IDEXX Corporation
Portland, Maine 04101
ABSTRACT
Current testing protocols and assay
systems for rapid screening of samples
for the presence of a contaminant or
residue often require auxiliary equipment
and/or reagents. This precludes their
use in routine, field surveillance
programs. In addition, functional
characteristics often require critical
timing and volume measuring steps.
Here we report on a self contained test
system which offers features of self-
measuring of reagents, in addition to
internal controls for validating
performance of the assay. The device
itself can be modified to allow semi-
quantitative as well as yes/no results.
Descriptions and performance
characteristics of the test system will
be discussed.
INTRODUCTION
The intent of field testing is to rapidly
screen samples for the presence of a
compound of interest. When presumptively
identified, the sample can be sent to a
laboratory for confirmation and further-
identification. From this premise, field
tests should be specific to correctly
identify the compound, sensitive to the
appropriate level and easy to perform to
allow screening of a large number of
samples.
In field testing situations, instrument
and assay reliability are of paramount
importance. In addition to these
concerns, any additional technique
sensitive procedures will yield
inconsistent results. Environmental
contaminants lend themselves to analysis
by immunoassays. However, immunoassays,
while often being easy to perform in
field situations, have required auxiliary
equipment such as measuring pipets,
washing equipment and instruments for
readout.
Recently, delivery systems for
immunoassays have appeared which offer
some advantages in ease of use. However,
they still suffer limitations of
additional equipment and reagent
requirements. The optimum system would
be a zero technique imrounoassay which
integrates reagents and the delivery
system and which is self-contained and
disposable.
DELIVERY SYSTEM DESCRIPTION
The system consists of three general
components. The device is a disposable,
self-contained unit capable of measuring
sample and retaining any excess reagents
within the body. The tray contains
appropriate wells for all reagents
necessary to perform the test. The final
component is the reagent set, premeasured
and sealed in the tray.
The device consists of two, initially
separated, absorbent materials. These
are contained in a plastic body with a
test head for sample/reagent entry and a
depressable plunger for activating the
system. The first absorbent takes up a
premeasured volume, independent of
starting sample volume. The sample is
absorbed through ports in the test head
which pulls all the volume through the
bioreactive zones. The head therefore
can vary in design depending on the
application. If two ports are available,
one can be for assay validation and
calibration, the second serves to detect
the analyte. In this configuration, if
color develops in the calibration port,
the reagents are functional and steps
have been followed correctly. Color
develops in the sample port proportional
to the analyte concentration. By
adjusting the calibration level to equal
the cutoff of the test, then the presence
or absence of analyte at that level is
made by direct visual comparison of the
color in the sample port to the
calibration port.
235
-------
Because the calibration port is subjected
to the same environmental conditions as
the sample, including the sample matrix,
conditions will be equivalent in each
test zone, removing any temperature or
matrix effects.
Multiple ports in the test head are
another possible configuration. The
ports can serve to simultaneously measure
separate analytes or serve as multiple
calibration levels. Thus, calibration can
be set at an upper and lower limit of
concentration of the analyte being
measured.
A prefilter can be attached to the test
head which will remove any particulate
foreign matter from the solution prior to
entry into the device. The tray can be
designed to remove the prefilter after
the sample incubation step.
The second absorbent serves as a
reservoir for waste. Any sample or
reagent that is pulled through the ports
will be retained, removing the need for
auxiliary washing equipment or waste
receptical.
The tray itself consists of separate
wells into which premeasured volumes of
reagents have been dispensed. The assay
procedure calls for removing the sealing
film, adding the sample, and sequentially
moving across the tray, subjecting the
test head to the various reagents. By
this means, the test head can be exposed
to sample, conjugate, wash solution,
substrate solution (if an immunoassay)
and stop solution (if required to
stabilize the endpoint of the reaction).
The reagents themselves consist of
antibodies or binding proteins of
characterized activity, conjugate to
detect the analyte, appropriate wash
solution to remove any unbound materials,
substrate solution to visualize the bound
enzyme, and, if appropriate, sample
diluent. The sample diluent can be used
to dilute any interfering substances.
This increases the utility of the assay
system by allowing sample extraction to
occur under relatively harsh conditions
to insure dissolution of the analyte.
These chemicals can then be diluted to a
point where they no longer interfere with
the biological activity of the assay.
REDUCTION TO PRACTICE
The utility of the test system has been
assessed by applying the technology to a
variety of systems, each with separate
constraints on sample type and
sensitivity requirements.
One assay developed required part per
billion level sensitivity and organic
solvent extraction to solubilize the
analyte. The analyte was Aflatoxin Bl, a
product of fungal contamination in feed
comodities. The extraction procedure
developed consisted of grinding the
sample in methanol:water (70:30). The
reagents included antibody to Aflatoxin
Bl which was raised in rabbits and
passively adsorbed onto the solid
support. The conjugate was Aflatoxin Bl
conjugated to alkaline phosphatase,
indoxyl phosphate was used as the
substrate. Calibration intensity was
adjusted to represent a signal level
equivalent to 20 ppb Aflatoxin Bl in the
sample. The final test protocol was as
follows.
1) Sample (25 g) is ground in 100ml
methanol:water (70:30) for 5
minutes.
2) Sample extract was then placed in
the first well in the tray and
the device was inserted.
3) Sample and device were allowed to
incubate for 3 minutes.
4) The device was removed from the
sample well, the secondary
absorbent was then pressed into
contact with the primary
absorbant and the device was
placed in the conjugate well
(prefilled with 300 ul of
conjugate) and subsequently
incubated for 1 minute.
5) The device was removed from the
conjugate well and placed in a
wash well containing approxi-
mately 500 ul of wash solution.
Following absorption of wash
reagent the device was removed.
6) The device was placed in a
substrate well for 15 seconds,
removed and allowed to develop
for 1 minute.
7) The reaction was stopped with
water or stop solution.
8) The color intensity of the
calibrator port was compared to
that of the sample port.
The assay, as described is a competitive
immunoassay. The presence of analyte in
the sample will therefore compete with
the conjugate and decrease the color
intensity of the sample spot. The
calibration spot develops color
independent of analyte concentration and
sets a reference level of color. If the
sample has less than 20 ppb Aflatoxin Bl,
the color intensity of the sample spot
will be greater than the calibration
spot. If the sample spot color intensity
is less than or equal to the calibration
spot, the sample is presumptively
positive for the toxin.
The assay works well in the presence of
high organic solvent (70% of methanol)
and has shown no sample matrix effects
236
-------
from corn, peanuts or peanut butter.
Current procedures for Aflatoxin Bl
include thin layer chromatography or high
pressure liquid chromatography. This
test system for Aflatoxin Bl correlates
well with these procedures.
Another application of the technology was
developed to detect the antibiotic
penicillin in milk samples. In this
case, a prefilter was used to remove fats
from the milk. The test consisted of
incubating the milk sample in the
presence of the biological, then allowing
the mixture to enter into the device.
The entire protocol is as follows:
1) Add 7 drops (0.25 ml) of milk
with disposal plastic pipet to
the first well contianing the
conjugate and allow to incubate
for 5 minutes.
2) Remove device leaving the
prefilter in the first well and
insert device into the second
well and incubate for 3 minutes.
3) Insert device into wash well
until all liquid is absorbed.
4) Insert device into substrate for
15 seconds, remove and allow to
develop color for 3 minutes.
5) Stop reaction and read results.
The principles of the test are to allow
any analyte present in the sample to bind
to the conjugate and therefore prevent
the subsequent binding of conjugate to
penicillin covalently coupled to the test
device, washing excess reagents through
the test head and developing color. If
penicillin is present in the sample being
tested, then the sample port color
intensity will decrease in proportion to
the analyte concentration. By visual
comparison of the color intensity of the
calibration port to the color in the
sample port, the sample can be classified
as negative or presumptively positive for
penicillin. The calibration level was
set at the action level of penicillin
(5ppb).
Existing methodologies for detection of
penicillin include culturing with the
organism Bacillus stearothermophilus or a
binding assay using radioactively
labelled penicillin. The correlation
between all methodologies was excellent.
CONCLUSION
The availability of immunological
reagents specific for contaminants and
residues coupled with a convenient
delivery system allows assays to be
developed which can accurately detect
analytes in a variety of sample types.
Prefiltration by the device itself and
subsequent containment of the solutions
provides ease of use features which
should make environmental monitoring more
practical. Organic solvents to some
extent can be tolerated by the
biologicals; if not, then dilution
of the sample to reduce the solvent
concentration can be accomplished. The
net result should be tests which allow
expanded surveillance for contaminants by
individuals not as skilled as trained
laboratory personnel.
DISCUSSION
PETER DUQUETTE: How do you distinguish between the closed-ring
penicillin and the open-ring penicillin?
ROBERT SUVA: The open-ring penicillins don't bind to the binding protein
that we're using.
237
-------
"Fieldable Enzyme Immunoassay Kits for Pesticides"
Peter H. Duquette
Patrick E. Guire
Melvin J. Swanson
Bio-Metric Systems, Inc.
R & D Division
9924 West Seventy-Fourth Street
Eden Prairie, Minnesota 55344
Abstract
Receptor proteins (especially antibodies and
enzymes) have been demonstrated useful for
rapid, convenient detection and semiquantitative
analysis of pesticides. BSI has developed a
rapid, disposable, self-contained, sensitive EIA
device designed to allow untrained personnel to
test for pesticides or other specific
environmental contaminants in water or soil.
The analyte in the test sample competes with an
enzyme-analyte conjugate for a limited number of
immobilized antibody sites. This Pinch Test
format can detect paraoxon at one micromolar in
water, with positive results indicated by
clearly visible color development within ten
minutes. It is operational in salt, brackish,
and fresh raw water. This format is designed to
have all dry components and to have an ambient
shelf life of greater than one year. The format
is readily adaptable for use with other
environmental contaminants.
Introduction
the analyses of pesticides has
mainly on conventional
HPLC) and
development of
such as
it possible to
pesticides by
Until recently,
been based
chromatographic (i.e., GC,
colorimetric techniques. The
immunochemical procedures
radioimmunoassay (RIA) have made
analyze low concentrations of
combining immunospecificity with the sensitivity
of radiochemistry (1-7). However, because of
the inherent disadvantages involved with using
radioactive materials (i.e., disposal of
radioactive waste, possible health hazard, and
use of expensive counters), there has been an
interest in developing related nonisotopic
immunoassays. An excellent alternative is the
use of an enzyme as a nonisotopic label and
recently solid-phase enzyme-linked immunosorbent
assays (ELISA's) have been described for
pesticide residue analyses (8-12). These assays
circumvent the hazard and waste disposal problem
associated with the use of radioisotopes, yet
are comparable in sensitivity to RIA. An
additional advantage to the use of antibodies
versus conventional chromatographic techniques
is that of high specificity of the antibody for
the analyte, thus often avoiding extensive
initial extraction of the lipophilic pesticides
from the sample (13). The enzyme immunoassay
procedures have many advantages of the RIA (i.e.,
specificity, sensitivity), but require only
inexpensive equipment. Possibly the greatest
advantage of enzyme immunoassays is that they can
be adapted to either automated or fieldable
methods (14).
Experimental Approach
Bio-Metric Systems, Inc. (BSI) has developed and
is in the process of demonstrating a rapid enzyme
immunoassay (i.e., Pinch Test) for the detection
of organophosphate pesticides. The Pinch Test
could also be easily adapted to the measurement
of other environmental pollutants. The tasks
required to develop the test for detecting
organophosphate pesticides include: 1) preparing
suitable derivatives of organophosphates that can
be coupled to proteins in such a way that
specific antibody can bind to them with high
affinity; 2) preparing conjugates of the
organophosphate derivatives with an appropriate
enzyme derivative. The resulting conjugate must
have suitable properties that allow rapid binding
to immobilized antibody (a function of the number
of haptens coupled and specific methods of
coupling) and good recovery and stability of
enzyme activity; 3) immobilizing antibody in a
suitable form to a solid support such as
cellulose; and 4) optimizing the components of
the test (e.g., signal-generating enzyme).
Previous work at BSI has indicated that a more
stable analog of paraoxon (i.e., diethyl 4-
aminobenzylphosphorate) (DABP) (Figure 1, 3}
exhibits excellent cross-reactivity with paraoxon
antisera. Analogs of DABP were synthesized for
use in preparing a suitable enzyme-hapten
conjugate. The DABP was reacted with succinic
anhydride to attach a carboxyl group and a spacer
to the 4-aminophenyl group (Figure 1, 4). This
compound (Figure 1, 4) was then reacted with N-
hydroxysuccinimide (NHS) and dicyclohexyl-
carbodiimide (DCC) to form the activated N-
oxysuccinimide ester (NOS) which can be easily
coupled to various biomolecules.
Antibody specific for paraoxon was purified from
antiserum that had been prepared by immunizing
rabbits with a bovine serum albumin-paraoxon
239
-------
conjugate. The immunogen had been prepared by
coupling diethyl-p-aminophenylphosphate (Figure
1, 2) to BSA by diazotization as previously
described (9). The antiserum was fractionated
by precipitation with 40% saturated ammonium
sulfate and further purified by affinity
chromatography on Sepharose-DABP. The crude IgG
preparation was passed through a column of DABP-
Sepharose in 0.01M 2-(N-morpholino)ethane-
sulfonic acid (MES) buffer at pH 5.0. After
washing through the unbound protein, the
specific antibody was eluted with 0.1M acetic
acid and collected in tubes containing 0.1 ml of
l.OM NaHC03 at pH 9.0 to neutralize the antibody
solution. The antibody was evaluated using the
following ELISA method. DABP coupled to human
serum albumin was adsorbed onto polystyrene
microtiter plates. The antibody preparation was
added and plates were incubated for one hour.
After incubating, peroxidase-labeled anti-rabbit
IgG was added and plates were incubated for one
hour a second time. The plates were then washed
with 0.05% Nonidet P-40 in PBS followed by
addition of ^2 and 2,2'-azino-bis(3-ethylbenz-
thiazolinesulfonic acid (ABTS). After twenty
minutes the plates were read at 405nm with a
microtiter plate reader.
Because of the instability of the paraoxon, the
more stable DABP derivatives were used to
prepare enzyme-hapten conjugates. These
conjugates were determined to have excellent
immunological and enzymatic activity. Recently,
we have prepared spacer-modified glucose oxidase
conjugates with various haptens which has
resulted in greater stability and faster binding
of the enzyme-hapten to the immobilized antibody
compared to conjugates of native GO. The DABP
or carboxylic acid derivative (40 can then be
coupled to GO or modified GO by any one of the
following methods: a) direct attachment of the
DABP analog (4) (free carboxylic acid moiety of
the succinic acid spacer) to the amines of the
protein by use of l-ethyl-3-(3-
dimethylaminopropyl)carbodiimide (EDC), a water
soluble activated ester (i.e., sulfo-NOS), or
coupling of the N-hydroxysuccinimide ester
(i.e., NOS) of this hapten by established
methods; and b) by direct attachment of the DABP
by diazotization to the enzyme or modified
enzyme.
Assay Development
Following antibody evaluation and synthesis of
the enzyme-labeled hapten, the components are
being tested and analyzed using an "enzyme-
receptor Pinch Test." The "Pinch Test" was
developed at BSI and is currently being
developed into assay kits for drugs and
naturally occurring low molecular weight toxins
(Figure 2). The Pinch Test is a non-
instrumented apparatus for qualitative or semi-
quantitative determination of an analyte from
biological fluids or environmental samples. In
general terms, the format consists of two or
more reaction zones which contain the
appropriate reagents in liquid permeable solid
media. These reaction zones are placed in such
physical arrangement as to allow control over
the sequence and time of exposure with test
samples. The EIA has been miniaturized to
maximize speed, portability, and ease of use. It
is capable of detecting concentrations as low as
1-10 ng/ml of one or more analytes from
biological or environmental samples on a single
strip simultaneously.
The test format (Figure 2) consists of four
parts: 1) antibody disks (A); 2) read-out disks
(B); 3) absorbent blotting reservoir (C); and 4)
a tube containing lyophilized enzyme-hapten
conjugate (D). The antibody disks (A) contain
immobilized antigen-specific antibody, while the
read-out disks (B) contain: an immobilized
enzyme; the dye; and a substrate for the enzyme-
hapten conjugate. A reservoir (C), containing an
absorbent pad is located beneath the antibody
disks. An enzyme-hapten conjugate of the analyte
is lyophilized and contained in a small tube (D).
The user simply reconstitutes the lyophilized
conjugate (D) to a specified volume with water,
allowing the conjugate enough time for complete
dissolution (ten seconds). Next, five drops of
sample are applied to the antibody disks and
incubated. The disks are washed with five drops
of PBS or water to remove any extraneous
material. One drop of the reconstituted
conjugate is next added to the antibody disks (A)
and incubated for one minute. The user then
folds the top plate (Segment 1) containing the
read-out disks (B) over the bottom plate (Segment
2) containing antibody disks (A), and pinches for
approximately three seconds. The results are
read in five minutes with a positive result being
indicated by color formation. This assay format
has distinct advantages over many other assays.
For example, the sample size is not limiting
because the antibody disk can be exposed to a
relatively large volume of sample (up to 1 ml).
This allows the antibody to bind any analyte
residue which may be in the sample, thus
increasing the sensitivity of the assay.
Secondly, the wash step with PBS allows any
possible interfering substances still present in
the sample to be washed out of the antibody disk,
thus reducing the possibility of non-specific
color development. Additionally, the enzyme-
hapten conjugate can be added after the sample,
thereby increasing sensitivity. Finally, this
format allows one to use built-in controls and
can be adapted as a multi-toxin assay.
Since antibody and enzyme activities are affected
by environmental factors (e.g., temperature,
humidity), it is difficult to prepare a rapid
enzyme immunoassay suitable for field use.
However, we have demonstrated that antibody and
enzyme activities can be maintained at 80%
efficiency after storage at 70°C for 14 days.
One method to achieve kit fieldability is to
incorporate an internal reference which responds
to environmental factors but is independent of
the presence of analyte. Such a reference can be
incorporated into our assay format by the
inclusion of a control anti-enzyme binding region
with the immobilized antibodies against the label
enzyme.
240
-------
Summary and Conclusions
Data from the above developmental experiments
indicates BSI can produce a sensitive and
convenient screening test for organophosphate
insecticides. BSI has already developed similar
EIA's for various drugs of abuse (i.e., cocaine,
morphine, phenobarbital) with sensitivities of
1-10 ng/ml. It is believed that the assay
format could easily be adapted to measure low
levels of environmental chemical hazards such as
polychlorinated biphenyls, pentachlorophenol,
and other compounds found on the EPA Priority
Chemical List.
Acknowledgements
The authors wish to acknowledge the following
support: USEPA SBIR Contract No. 68D80035.
References
1. Langone, J.J. and Van Vanakis, H., 1975,
"Radioimmunoassay for Dieldrin and Aldrin," Res.
Commun. Chem. Pathol. Pharmacol H):163.
2. Albro, P.M., Luster, M.I., Chae, K.,
Chaudhary, S.K., Clark, G., Lawson, L.D.,
Corbett, J.T., and McKinney, J.D., 1979, "A
Radioimmunoassay for Chlorinated Dibenzo-p-
Dioxins," Toxinol. Appl. Pharmacol. 50:137.
3. Fatori, D. and Hunter, W.M., 1980,
"Radioimmunoassay for Serum Paraquat," C1in.
Chem. Acta 100:81.
4. Luster, M.I., Albro, P.W., Chae, K.,
Lawson, L.D., Corbett, J.T., and McKinney, J.D.,
1980, "Radioimmunoassay for Quantisation of
2,3,7,8-Tetrachlorodibenzofuran," Anal. Chem.
52:1497.
5. Ercegovich, C.D., Vallejo, R.P., Gettig,
R.R., Woods, L., Bogus, E.R., and Mumma, R.O.,
1981, "Development of a Radioimmunoassay for
Parathion," J. Agric. Food Chem. 2^:559.
6. Rinder, D.F., and Fleeker, J.R., 1981,
"A Radioimmunoassay to Screen for 2,4-
Dichlorophenoxyacetic Acid and 2,4,5-
Trichlorophenoxyacetic Acid in Surface Water,"
Bull. Environ. Contam. Toxicol. 26:375.
7. Wie, S.I., Sylvester, A.P, Wing, K.D.,
and Hammock, B.D., 1982, "Synthesis of Haptens
and Potential Radio!igands and Development of
Antibodies to
Diflubenzuron and
Chem. 30:943.
Wie, S.I.
of
Insect Growth Regulators
BAY SIR 8514," J. Agric. Food
8.
"Development
and Hammock, B.D., 1982,
Enzyme-Linked Immunosorbent
Assays for Residue Analysis of Diflubenzuron and
BAY SIR 8514," J. Agric. Food Chem. 30:949.
9. Hunter, K.W. Jr. and Lenz, D.E., 1982,
"Determination and Quantification of the
Organophosphate Insecticide Paraoxon by
Competitive Inhibition Enzyme Immunoassay," Life
Sciences 30:355.
10. Schwalbe, M., Dorn, E., and Beyermann,
K., 1984, "Enzyme Immunoassay and
Fluoroimmunoassay for the Herbicide Diclofop-
Methyl," J. Agric. Food Chem. 32:734.
11. Kelley, M.M., Zahnow, E.W., Peterson,
W.C., and Toy, S.T., 1985, "Chlorsulfuron
Determination in Soil Extracts by Enzyme
Immunoassay," J. Agric. Food Chem. 33:962.
12. Huber, S.J. and Hock, B., 1985, "A
Solid-Phase Enzyme Immunoassay for Quantitative
Determination of the Herbicide Terbutryn," Z_._
Pflanzeckr. Pflanzenschutz 9_2:147.
13. Newsome, W.H., 1986, "Potential and
Advantages of Immunochemical Methods for Avalysis
of Foods," J. Assoc. Off. Anal. Chem. 69:919.
14. Hammock, B.D. and Mumma, R.O., 1980,
"Potential of Immunochemical Technology for
Pesticide Analysis," In:Pesticide Analytical
Methodology, ACS Symposium Series No. 136, J.
Harvey Jr. and G. Zweig (Eds.), 321-352.
C.H.Oa
C,H,0'P
-NH.
Paraoxon (1)
Aminoparaoxon (2)
DABP (3)
DABP-Succinic acid (4)
Figure 1. Paraoxon Analogs
241
-------
—^ Segment 1
Pares: A. Antibody Disk
B. Read-out Disk
C. Blotting Reservoir
D. Lyophlllzed Conjugate
Procedure:
1. Resuspend Conjugate:
- Open conjugate tube making sure pellet Is on
the bottom of tube.
- Add 12 drops PBS with disposable plpet.
- Replace cap and mix thoroughly.
Sample, Wash, Conjugate Application:
- Take a disposable plpet and add 5 drops
sample to test well.
- Add 5 drops PBS to wash the test well.
- Take disposable pipet and add one drop of
resuspended conjugate to test well. (Add
conjugate carefully and slowly. The addition
of two or more drops will make the test Invalid.)
Wait one minute.
3. Pinch:
- Pinch the test by folding read-out piece into
the test well while holding upright. Pinch
all the way until plastic is touching plastic,
hold for 2-3 seconds, and release.
4. Read Result!
- Read result 5 minutes after the pinch by comparing
to color chart.
0,1 • negative
2 or darker • positive
Figure 2. The "Pinch Test
DISCUSSION
HANK WALSH: How much time and effort is involved with coming up with
a new assay? Suppose I wanted to detect PCB's instead of organophosphates?
PETER DUQUETTE: It takes a while. You have to make the antigen and
conjugate. Then you have to get your tilers up. The conjugate must have good
immunological activity. That's been a problem. We have some "spacer technol-
ogy" where we couple the analyte to the active enzyme and splice them. That
allows us to have good immunological activity. The time that it takes to develop
these is quite significant. How long it's going to take to do the whole assay will
vary. The read-out disk is the same in all the assays, so those basically are stable.
You still are going to have to go through and check out your conjugate and your
antibody.
HANK WALSH: What would it be - half a man-year to five man-years?
PETER DUQUETTE: We have a development contract with » private
company now, and we proposed for five different analytes that would take a
one-person-level effort for six months.
RICHARD WHITE: Regarding the data on the stability of the enzyme, was
that an enzyme that was in association with the solid phase or entrapped in the
solid phase?
PETER DUQUETTE: That enzyme is immobilized, and there are some
stabilizers added. It's a combination of polymers and proteins, and that's about
all I can tell you.
RICHARD WHITE: In the pinch test that you talked about, the conjugate
would be not really immobilized in the solid phase, right?
PETER DUQUETTE: No.
242
-------
INTEGRATED IMMUNOCHEMICAL SYSTEMS FOR ENVIRONMENTAL MONITORING
J. M. Bolts, S. E. Diamond, J. F. Kolc, S. H. Lin and F. J. Regina
Allied-Signal Corporate Technology, P. 0. Box 1021R, Morristown, NJ 07960
and
P. G. Koga, G. C. Misener and J. C. Schmidt
Bendix Environmental Systems Division, P. 0. Box 9840, Baltimore, MD 21284
ABSTRACT
The physiological functions of antibody molecules
include the specific recognition of foreign
materials, and the potentiation of the host's
immunological defenses. The extraordinary
specificity of the natural immune response can be
exploited and adapted for environmental monitoring
by designing integrated analytical systems which
combine immunochemical elements with advanced
sensor instrumentation.
In this paper, our prototype immunodetection
systems, suitable for environmental field use, are
described. Two different integrated immunochemical
systems are discussed, one based upon a
chromatographic approach and the other utilizing an
optical fiber immunosensor design.
Using these technologies, a wide range of analytes
have been successfully assayed in both air and
water sample streams, including a variety of toxic
chemicals and hazardous microorganisms. The
specific example of bacterial detection is used to
illustrate the broad applications of these
technologies, and the options available in the
design of customized immunoreactor/sensor component
systems.
Key Words: Immunochemical sensor, optical fiber,
size-exclusion chromatography.
INTRODUCTION
Integrated immunochemical sensing systems represent
a most attractive route to the development of
versatile and sensitive instrumentation for
environmental monitoring applications. In our
laboratories, two prototype immunodetection
systems, suitable for environmental field use, are
under development. One design features separate
immunochemical and detection modules. In this
system, air or liquid samples are first analyzed in
an antibody reactor, and then the products of the
immunoreaction are conveyed to a sensor module for
quantisation and display. In the alternate design,
the immunoreaction occurs directly on the surface
of a transducer device.
This paper will describe a chromatographic sensor
system based on the first design, and a fiber optic
system exemplifying the second design. The
performance capabilities of both immunodetection
systems will be illustrated with reference to the
immunochemical detection of low levels of bacteria
in environmental samples.
SYSTEM DESCRIPTION
Our integrated immunodetection systems are modular
in design, and can accomomdate a variety of
interchangeable sensor modules. A common "chasis"
module is comprised of a sample acquisition system,
air and liquid transport systems, and a host
electronics and signal processing apparatus.
Battery packs for power in the field, and
reservoirs of buffers and immunoreagents, are also
aboard the chasis module. In contrast, the
individual sensor modules contain only components
which are unique to their respective operating
requirements. This modular system design allows
the user maximum flexibility to select the sensor
option best suited to his particular needs. In
addition, the modular design helps to ensure that
future advances in sensor technology can be readily
adapted for use with immunodetection units already
in the field.
The overall system is designed to be portable,
approximately one cubic foot in volume, and capable
of unattended field operation over a twenty-four
hour period. In addition, the system is designed
to respond within minutes to specified threshold
levels of targeted molecules and organisms.
CHROMATOGRAPHIC SENSOR MODULE
A technique designated Size Exclusion
Chromatography (SEC) has been developed in our
laboratories for application to the rapid
immunochemical detection of microorganisms and
other large analytes. The basic principle of the
SEC method is depicted schematically in Figure 1.
Note that the SEC technique distinguishes between
large and small analytes, and that large analytes
elute rapidly from the column while smaller
analytes are delayed within the column and elute
243
-------
later. When used in conjunction with an
immunochemical pre-incubation, an SEC column can
rapidly separate the larger antibody-antigen
complexes from the smaller, unconjugated antibody
molecules, as long as there is a sufficient size
difference between the immune complexes and the
immunoglobulin molecules (molecular weight 150,000
Daltons).
We have exploited the SEC technique to design a
sensor module for the rapid detection of bacteria
and other large analytes. Figure 2 illustrates the
operation of the module in a specific application
involving the detection of pathogenic Salmonella
bacteria in a water sample. In this assay, a
sample aliquot is first briefly incubated with a
concentrated solution of fluorescein-labelled
anti-Salmonella antibodies (FITC-Ab). During this
reaction, Salmonel1 a bacteria in the sample form
immune complexes with the labelled antibodies, and
the extent of this reaction reflects the
concentration of Salmonella in the sample. The
large antigen-antibody complexes are then separated
from unbound labelled antibodies by flowing the
mixture through an SEC column. As shown in Figure
2, curve B, the bacteria bound to labelled
antibodies elute from the column after
approximately 8 minutes, while the remaining,
unreacted labelled antibodies do not elute until
nearly 16 minutes have elapsed. A negative control
run, curve A, confirms the 16 minute elution time
for fluorescein-labelled antibodies in the absence
of bacterial analyte. Positive control runs, not
shown in Figure 2, give no fluorescent signal above
background at any elution time; this is due to the
fact that the bacteria themselves, in the absence
of fluorescein-labelled antibodies, are not
fluorescent at the detection wavelengths.
The SEC technique is a rapid and convenient means
of assaying the results of an immunochemical
reaction. It is highly resistant to interferences
and false positives, because the early-eluting
signal peak appears only when the sample satisfies
two criteria simultaneously: the sample must not
only react with the specific labelled antibodies,
but must also fall within a particular size-
dependent window of elution times. Immunologically
cross-reacting materials of incorrect size, as well
as other bacteria or particulates which are not
recognized by the labelled antibodies, do not
produce a fluorescent signal peak.
Futhermore, the SEC technique is virtually ideal
for repetitive sampling and monitoring
applications, because SEC columns are self-
clearing. Unlike traditional immunoaffinity
columns, where antigens accumulate on an affinity
matrix, an SEC column does not saturate with use
and does not require replacement or regeneration as
some maximum binding capacity is approached.
Rather, with an SEC system, both antibody-bound and
unbound materials freely flow through and out of
the analytical column, and the column may then be
reused as soon as the unreacted labelled-
antibody peak from the previous assay has cleared
the detector. Indeed, the Salmonella sensing
module used to produce the data shown in Figure 2
has been used for weeks to assay a variety of
positive and negative environmental samples under a
range of different conditions. The only constraint
was that an interval of at least 30 minutes was
required between runs, in order to assure the
complete clearance of the labelled-antibody peak
from the preceeding assay.
In summary, our protoype integrated immunochemical
sensor, incorporating an immunoreaction chamber and
an SEC column in tandem, has proven to be a
versatile and practical instrument for the rapid
detection of large environmental analytes.
OPTICAL FIBER SENSOR MODULE
One alternative to the chromatographic sensor
module discussed above is a module incorporating an
optical fiber immunosensor. Like the
chromatographic sensor, the optical fiber sensor
ultimately measures the fluorescence of
fluorescein-labelled antibodies. However, unlike
the SEC system, the fiber optical system is not
subject to significant size constraints on the
antigens being detected, and can be used to assay
analytes either larger or smaller than an antibody
molecule. Also in contrast to the chromatographic
module, the optical fiber module does not have a
separate immunoreaction chamber. Instead, the
immunochemical reaction occurs directly on the
surface of the optical fiber sensor.
A schematic diagram of the optical system used in
the fiber optic sensing module is shown in Figure
3. In this prototype system designed by ORD, Inc.,
exciting light is focussed onto the cylindrical end
face of an optical fiber. The light propagates
within the fiber by total internal reflection, and
establishes a narrow evnescent wave zone at the
interface between the fiber and the surrounding
liquid medium. Fluorescent materials in the liquid
phase may be excited by this light in the
evanescent wave zone, and the resultant fluorescent
emission may then be captured by the fiber and
carried by total internal reflection back to the
fiber's end face for detection. The excitation and
emission maxima of 485 and 530 nm, respectively,
are selected to optimize the sensor for the
detection of fluorescein. Additional details
concerning the instrument design and its principles
of operation are available in Reference (1).
In order to adapt this optical fiber sensor for the
immunochemical detection of environmental analytes,
a sandwich immunoassay configuration is used.
Unlabelled capture antibodies are covalently
immobilized on the surface of the optical fiber.
Brief incubation of an antigen-containing sample
aliquot with an immunochemically derivatized fiber
results in the selective capture of antigen
directly from the sample onto the fiber surface.
The quantity of surface-bound antigen is then
probed by exposing the fiber to fluorescein-
labelled second antibodies. In the absence of
244
-------
analyte, these labelled antibodies exhibit very low
nonspecific binding on the surface of an optical
fiber bearing only capture antibodies. As a
result, few labelled antibodies are brought into
the narrow evanescent wave zone at the fiber/liquid
interface, and little fluorescent emission is
stimulated by the light propagating within the
fiber. However, if antigen has been captured from
the sample in the previous step, the labelled
second antibodies bind with high affinity to the
antigens on the fiber surface. The result is a
significant fluorescent signal, due to the presence
of fluorescein-labelled antibodies within the
evanescent wave zone. The intensity of the
fluorescent emission correlates with the amount of
antigen captured on the fiber surface and,
indirectly, with the initial concentration of
antigen in the sample.
Using this fiber optic sensor module, a wide
variety of molecular and supramolecular analytes
may be rapidly and accurately assayed. For the
detection of Salmonella typhimurium, results have
been obtained which are roughly comparable in
sensitivity and response time to those discussed
above for the chromatography module. Significantly
better performance has been observed for other
bacterial analytes, and the fiber optic module has
also been successfully employed for the sensitive
detection of much smaller molecular analytes. For
repetitive sampling applications, a mechanism for
fiber replacement is necessary, since efforts to
regenerate antibody-coated fibers following antigen
binding have proven to be unsatisfactory thus far.
However, the flexibility to detect either large or
small analytes with a single sensor module makes
the integrated optical fiber immunosensor system an
attractive choice for many environmental monitoring
applications.
SUMMARY
Two integrated immunochemical systems for
environmental monitoring have been described, one
based upon a chromatographic approach and the other
utilizing an optical fiber immunosensor design.
The performance of these prototype sensor systems
has been illustrated with reference to the
immunochemical detection of low levels of bacteria
in environmental samples. However, a wide range of
analytes, including toxic chemicals and hazardous
microorganisms, may be monitored successfully using
one or another of the sensors being developed for
use with our modular detection systems. Points of
distinction have been highlighted which may serve
as technical bases for selecting among the various
options available in the design of customized
immunoreaction/sensor component systems.
REFERENCES
(1) T. Hirschfeld and M. Block, "Assay Apparatus
and Methods", United States Patent No.
4,558,014 (12/10/85).
Basic Idea Of The SEC Method
Ab'
.Ab* + Ab*
SEC
0)
c
0)
flj
C
O)
(75
(S -Large Agent (e.g., Bacteria)
Ab* = Labeled Antibody
E.Ab*
Elution Time
245
-------
SEC of S. typhimurium
Ab*
50 r
i
« 40
o
c
0)
0
(A
£ 30
o
3
U-
90
(
(
1
•
1
i
i 1
1 I
• 1
Ag.Ab* * /
10 -
r\
A : 0 Cells (FITC-Ab Alone)
B : 1 x 106 Cells
0 2 4 6 8 10 12 14 16 18 20 Minutes
Fiber Optic Waveguide Sensor Optical Schematic
I Light Source
Fiber
(1 x 60 mm)
Aspheric Lens
f = 8.5
Filter 485/20
f = 8.5
Aperture 800JJL
f = 8.5
f = 8.5 Dichroic Filter Shutter f = 8.5 Photodiode
Beam Splitter 530/30 Detector
510 LP
[Redrawn from ORD User's Manual with permission]
246
-------
DISCUSSION
BOB HARRISON: You're talking about using a fiber optic weight guide
sensor. Could you comment on the regeneration or replacement of the sensor,
whether you have disposable sensors and what the implications are for practical
field application?
FRANCIS REGINA: For the sensor, we throw the fiber out after one use. The
fibers are basically 1 mm by 60 mm.
STEVE GOHEEN: When you immobilized antibodies, did you ever see or
experience any problems with winding near the active site with the antibody?
FRANCIS REGINA: No. Generally, we have good activity of our immobi-
lized antibodies.
JEANETTE VAN EMON: How about adapting this to low-molecular-weight
compounds, such as environmental contaminants?
FRANCIS REGINA: The size-exclusion chromatography method would be
very difficult to adapt to low-molecular-weight compounds, because you need
that large analyte that will not be retained in the column to come out with the
void volume. If you try to analyze for a protein analyte, you'll basically be
trying to separate one protein from a complex of two proteins. In the time frame
that we're trying to do it in, using columns that are basically eight or nine centi-
meters long, with total volumes for the whole column of maybe three or four
milliliters, that would be very difficult. The fiber optic wave guide sensor could
be very adaptable to small analytes. Protocols wold have to be developed to do
competition assays, instead of sandwich assays.
247
-------
IMMUNOCHEMICAL QUANTIFICATION OF DIOXINS IN
INDUSTRIAL CHEMICALS AND SOILS.
Martin Vanderlaan, Bruce Watkins, and Larry Stanker
Lawrence Livermore National Laboratory
Livermore, CA 94550
We have previously reported characteristics of several monoclonal antibod-
ies (Mabs) for the immunoassay of polychlorinated dibenzo-p-dioxins and -
furans (PCDD and PCDF; Toxicology 45:229) We have applied these to the
detection of PCDD and PCDF in a wide range of contaminated samples. For 15
different samples, a direct comparison was made between the levels of PCDD
and PCDF contamination determined by conventional gas chromatography-
mass spectroscopy (GC/MS) and determined by immunoassay. Samples in-
cluded fly ash, soil, technical grade chemicals, motor oils, PCB transformer oil,
and still bottom residues. These ranged in contamination from less than 1 ppb
to several thousand ppb of PCDD and PCDF. The Mabs bind preferentially to
tri-, tetra- and penta- PCDD and PCDF containing lateral chlorines, suggesting
that they will be well suited for screening samples for the most toxic congeners.
The GC/MS data included information on total tetra- and penta- PCDD and
PCDF as well as levels of specifically regulated toxic congeners. There was
good correlation between the results of the immunoassay and those of conven-
tional GC/MS analysis in spite of these differences in the exact congeners
detected by the two different techniques. In general, the immunoassay required
substantially less sample clean-up than did GC/MS, thereby offering the
promise of substantially reduced costs and time for sample analysis. Sample
clean-up consisted of two columns: a carbon column followed by an acid-silica
column with dioxin recovery exceeding 70%. Efforts at speeding and automat-
ing the clean-up process will be presented.
Funding for this work was provided by the US EPA through Interagency
Agreement DW-89931433-01-0 and Hoechst AG. Work was performed under
the auspices of the United States Department of Energy by Lawrence Livermore
Laboratory under Contract No. W-7405-ENG-48. Chemical samples and GC/
MS analysis were provided by Hoechst AG, Frankfurt, FRG.
DISCUSSION
BRUCE MOLHOLT: You mentioned some reconstruction experiments with
soil, where you added dioxin and the motor oil back to soil. Did you try that with
fly ash to see if you could get 100% recovery of your immunogenecity? With
the soil experiments, did you try that over a period of time to see if dioxin binds
to soil so that it can't be seen?
MARTIN VANDERLAAN: No. I think we're going to have the same
extraction problem that everybody else has. If you're talking about aged New
Jersey dioxin in soils, then you probably are going to have to beat on it with a
soxlet extractor or something like that to get it out. Immunoassay doesn't solve
that problem. It coextracts anything that causes a problem, and it shortens up
the overall clean-up procedure a bit.
MARK GREENE: What is the affinity of your antibody for the ligand? Did
all of the antibodies that you made for dioxin cross inhibit one another?
MARTIN VANDERLAAN: I haven't done the cross inhibition check. My
assumption is that with a small molecular weight like this, you really can only
put one antibody on a molecule and that once it's bound, nothing else will bind
it. But that's an assumption I haven't tested.
I also haven't measured the exact affinity constant, because that's fairly tricky
for something that isn't water soluble. To get these assays to work, you've got
to use a little detergent. And the dioxin is probably sitting in detergent micelles.
You could make measurements, but I'm not sure what they would mean. I feel
much more comfortable reporting the sensitivity of the assay and the perform-
ance of the assay. That's the measure that is of greatest value. I think that that
is related to antibody affinity, but it isn't a strict quantitative measurement.
JEANETTE VAN EMON: Can you describe in small detail your sample
preparation?
MARTIN VANDERLAAN: We will extract the sample by whatever means
you like. If it's a case of fly ash, it was treated with acid and then soxlet
extracted. If it's the case of soils, we mix it with sodium sulfate and then shake
it in hexane for half an hour or so, with a couple of changes of hexane.
Once the sample is extracted, it's run over a carbon column. There are a couple
of washes of solvents there where other things wash through. Then it's diluted
in toluene. We then mix that with acid-impregnated silica and dry that down
onto the silica, so we're doing a sulfuric acid treatment on a solid-phase silica
matrix. We pour hexane over that, and the dioxin comes off because it didn't
react. That gets dried down into a little bit of detergent and resuspended with
the antibodies.
249
-------
REMOTE, CONTINUOUS, MULTICHANNEL BIOCHEMICAL
SENSORS BASED OF FLUOROIMMUNOASSAY TECHNOLOGIES
j-NLin, P. Kopeckova*, J. Ives, H. Chuang, J. Kopecek*, J. Herron, H-R
Yen, D. Christensen, J.D. Andrade
Department of Bioengineering
University of Utah
Salt Lake City, UT 84112, U.S.A.
*Institute for Macromolecular Chemistry
Prague 6, Czechoslovakia
The advances in fluoroimmunoassay, coupled with the rapid development of
fiber optics, planar waveguides, and integrated optics, make it feasible to
construct and apply remote, continuous multichannel biochemical sensors. The
development of such immunosensors requires: a) preparation, characterization,
selection, and immobilization of the needed antibodies or antibody fragments;
b) a means to deliver remotely the fluor or fluorescently-labelled competing Ag;
c) a means to regulate the Ag-Ab binding properties to allow reasonable
response times or externally controlled "zeroing" of the sensor; d) an inexpen-
sive and reliable means to excite fluorescence with minimal excitation of
general background fluorescence; e) a convenient means of detecting, collect-
ing, and processing the emitted fluorescence so as to obtain a quantitative assay;
f) means to provide many sensing channels to permit the assay of 2 or more
analytes together with the needed reference and blank channels.
We are addressing all of these problems in a coordinated fluorosensor
development program. Progress on each of the component areas will be briefly
discussed using results from prototype sensors for prothrombin and anti-
thrombin III.
DISCUSSION
MEL SWANSON: You stated that for covalent immobilization to occur, an
antibody has to adsorb first. Would that be true only of hydrophobic surfaces,
or would it also be true of hydrophilic surfaces?
JOE ANDRADE: It's got to get there first, and it has to have a residence time
there greater than a thermal fluctuation. And that means basically an adsorption
event. That adsorption event and the covalent immobilization can occur almost
simultaneously, but it's unlikely, because it's coming up to the surface.
There is a repertoire of functional groups on the surface of a protein - the amino
groups that are used to covalently immobilize it (sulfhydryl, etc.) There is also
a whole array of nonpolar groups on most protein surfaces - so-called hydro-
phobic patches. On a hydrophobic surface, the protein which is the most tightly
adsorbed is certainly adsorbing through the face with the greatest hydrophobic
character.
There's a whole variety of collision events and processes going on. You must
imagine the protein interacting with the surface, colliding with that surface
through all possible orientations. Some of those orientations essentially lead to
no residence time or just diffusion-limited residence time. Other orientations
lead to a substantial residence time. Those orientations are not necessarily the
same orientations required for covalent immobilization. After the abdsorption
event, the thing diffuses around on the surface and partially denatures. And in
that squirming on the surface, which takes on the order of seconds to minutes
to even longer, it finally finds a reactor group. And then the reactive event may
occur. The adsorption event is critical to the immobilization event in most
systems.
251
-------
A Microbial Bioassay Developed for Rapid Field Screening
of Hazardous Waste Sites
I. Cecil Felkner, Ph.D.
Tom Christison, M.T.
Brenda Worthy, M.T.
Christine F. Chaisson, Ph.D.
ABSTRACT
A laser/microbial bioassay technology that
is cost effective, accurate, and
quantitative for various classes of
chemical toxicants present in the
environment has been developed. This
system is capable of identifying and
quantifying toxicants in various matrices
which are likely to contain the
environmental pollutants on the U.S.
Environmental Protection Agency's priority
list, even when these substances are
present at low levels. The system has the
significant advantage of being field
portable, rapid, and sensitive while still
incorporating a high level of specificity
for numerous toxic chemicals and chemical
classes. Furthermore, the system is
capable of distinguishing between those
substances that possess cytotoxicity only
and those which have genotoxic properties;
hence the potential for identifying those
which might produce acute toxic symptoms
but possibly chronic disease, e.g.,
cancer, as a consequence of chronic
exposure at low levels. The bioassay
system consists of a laser photometer that
makes 1200 measurements/second at lb
unique angles over a 180 degree arc and an
isogenic set of Bacillus subtilis mutants.
The laser detects toxicity to the bacteria
(in a liquid sampling cuvette) by
differential light scattering (DLS) of a
fine beam which is measured quantitatively
in terms of intensity at the various
angles. The parameters recorded are the
number of bacteria, their size, shape and
distribution, and any increases or
decreases in all of these parameters. The
quantitative determination of toxicant
concentration is a function of the dose-
response kinetics by the bacteria in a
given sample; the specificity of response
to a given chemical/chemical class is made
possible by monitoring bacterial mutants
that differ from each other by only one
property, which causes them to be more
sensitive than the other set members based
on the mechanism of toxicity. Therefore,
a "fingerprint" unique to each chemical
can be generated from the differential
response of the 19-member test set of
bacteria. Since the system also includes
a metabolic activation capability in a
unique "solvent" cocktail, carcinogenic
chemicals such as benzo(a)pyrene and
dimethyl hydrazine are easily identified
and quantified. Finally, individual
samples can be read and analyzed by an
integrated computerized system so that 5
to 10 concentrations of an aqueous sample
can be processed in just over one hour.
INTRODUCTION
Instrumentation and methods used in
environmental monitoring have typically
been those from analytical chemistry for
assaying chemical residues. This has been
for a very good reason -- the availability
of numerous, specific chemistry-based
methods that have been validated by well-
established academic, federal, state, and
private institute laboratories who are
frequently called upon to perform analyses
in a rather large variety of matrices,
e.g., water, soil, air, and food. In fact,
the United States Environmental Protection
Agency (USEPA) set forth the "Guidelines
Establishing Test Procedures for the
Analysis of Pollutants," (FR 44:No.233,
1979, under 40 CFR Part 136), in which
methods such as gas chromatography (GC),
high performance liquid chromatography
(HPLC), mass spectroscopy (MS) and
inductively coupled plasma optical
emission spectroscopy (ICP) were listed as
the acceptable analytical procedures for
organic and inorganic analyses. These
methods are highly reliable when linked
with the appropriate extraction methods,
compound-specific reactions, and suitable
detectors. However, these analytical
procedures do not provide, and are not
intended to provide, a direct assessment
of the toxicity of chemical residues.
Also, unique procedures must be developed
for each chemical or chemical class;
performance of tnese conventional
analytical procedures destroys the sample
because the method usually entails such
253
-------
preparatory steps as extraction, chemical
treatments to facilitate separation of the
sought components, and chemical
conversions to generate the detectable
chemical species. By necessity,
analytical laboratories are
geographically fixed, making it necessary
to store and transport samples and thereby
create a holding time. During the holding
time, it is expected that problems for the
analyses will be created, particularly
when the sample materials are volatile,
bind to the container (e.g., the
pesticides paraquat and diquat stick to
glass or borosilicate containers),or can
degrade under the conditions of storage
and transportation. Finally, it is
expensive to use traditional analytical
methods because they require sophisticated
instrumentation, highly purified solvents
and gases, frequent equipment
standardization, and highly trained
technical professionals.
In view of the requirements and attendant
resource expenditures cited above, there
is a serious problem created for truly
adequate monitoring of the environment for
toxicants. Felkner, et al. (1988) have
noted the need for "cheaper, quicker, and
more sensitive analytical methods for
detection of chemicals in the media in
which the chemicals may exist," and have
emphasized that such needs are recognized
by Congress, regulatory agencies, consumer
advocates, scientific societies, private
industry, and other segments of society.
Corporations and/or collaborating federal
agencies involved in toxic site cleanup
are burdened with expensive and time-
consuming analytical methods to determine
the extent to which a cleanup is
effective. Moreover, these procedures do
not benefit from rapid and inexpensive
prescreening technology and must therefore
be used to analyze all of the collected
samples until a statistically
representative sample size has been
reached; hence, there will be numerous
analyses performed on samples lacking the
toxicants being sought (in fact, the vast
majority of samples are negative) at great
expense to those who must finance this
monitoring. This optimizes the chance
that humans will be exposed to harmful
levels of toxicants, and the lack of
adequate monitoring may also prevent the
further development and use of potentially
beneficial chemicals, not due to the
health risk but because methods to
quantify their residues in food, water, or
soil may not exist or may be too costly to
warrant their development.
New technological approaches that offer
inexpensive and efficient ways to screen
for toxicants in the environment are
needed, but any developed system must have
adequate sensitivity and specificity and
should be portable, ideally capable of
collecting and quickly analyzing samples
on site. This capability exists for the
laser/microbial bioassay system which is
described in this paper as a candidate for
further development as an on-site system
for monitoring toxic waste cleanup.
METHODS
The laser/microbe bioassay system
developed by Felkner, et al. (1988) has
been described in detail in the published
literature. Briefly, it is a 66 min.
computerized bioassay that utilizes 19
isogenic strains of Bacillus subtilis to
characterize and quantify the toxicants
present in an aqueous solution. The
response of the bacteria to toxic
substances is monitored by differential
light scattering from a laser beam (632.8
nm), which is received by an array of
detectors and input directly into a
microcomputer that analyzes the toxic
response through developed software
programs. A fingerprint, generated from
the members of the isogenic set of
bacteria, specifies the identity of a
given toxicant based on its mechanism of
action (i.e., a biological detection that
relates the structure of the chemical to
its activity). The entire assay is
accomplished by incubating the test sample
with the bacteria in a small vial that is
heated by a heating block at 37 C;
specifically, a small pellet of the tester
bacteria with lyophyllized media is added
directly to the vial containing distilled
water after which the sample increment is
also added. Details of the assay
performance are given below.
The bioassay is performed with or without
metabolic activation. For the
nonactivated assay, lyophilized bacteria-
media (sufficient Brain-Heart Infusion
[BHI] broth to support the growth of the
bacteria during the assay) are introduced
into the incubation cuvette (scintillation
vial) containing 1.0 ml of deionized water
and mixed, held at 60C for 8 min., further
diluted with 4 ml of water and then
incubated at 37 C until an assay-ready
stage is reached (when this culture can be
diluted 1:20 and yield a concentration of
10 bacteria/ml). A computer program
recognizes this stage, at which time the
assay or control (+/-) sample aliquot (0.1
ml in 10ml of distilled water) may be
added. The negative control culture is
required to have an acceptable generation
time (e.g., approximately 40 min. for the
wild type strain or a doubling time of
27.6 min.) before an assay is considered
valid. Readings are taken at 0, 6, and 66
min., respectively, to determine growth
inhibition and/or any changes in shape or
size of cells (relative to the control
culture) as a measure of toxic response.
254
-------
These measured responses are dose-related
and therefore quantitative for the
toxicant concentration.
For metabolic activation, 0.5 ml of S9 mix
(microsomal fraction at 30 mg protein/ml)
is added to the solubilizer (cocktail
contains emulsifier-dispersanc) at the
time the assay sample is added. All other
conditions of the assay are the same as in
the nonactivated assay. However, it
should be noted that although the cocktail
components have no effect on bacterial
growth, they facilitate delivery of water
insoluble compounds to the detector
bacteria; hence, it can be used for assays
with or without S9 mix.
Specific software programs have been
designed to analyze the assay responses
(Felkner, et al., 1988) to determine how
toxicants affect the various members of
bacterial isogenic sets (isosets). By
exposing the bacterial cultures of these
isosets to varying dilutions of the test
material and includin'g a non-dosed
control, concentration dependent effects
can be assessed. Any differences between
strain responses are scored and recorded,
thus becoming a response profile for the
assayed material. If differential
sensitivity is seen between any members of
the 19 B_._ subtilis strains, the test
material is judged as genotoxic,
otherwise, any toxicity measured is
considered to be cytotoxicity and not
likely to have chronic effects such as the
potential for carcinogenicity . In
addition, the responses can be analyzed to
show the mechanism(s) by which toxicity is
exerted, the structure-function
relationship through which the system is
able to specifically identify chemicals.
All data are also analyzed statistically
and carried through data reduction
procedures leading to data printout for
each assay.
RESULTS
The data which follow are not intended to
be exhaustive, but should serve to
illustrate the systems's capability for
detecting toxicants in the environment.
Hence, an example of a toxicant which does
not require metabolic activation, 4-
nitroquinoline-1-oxide (4NQO) and one
which requires metabolic activation,
benzo(a)pyrene (BAP) are presented. In
the five figures, the profiles of the test
bacteria are generated by a plot of the
relative intensity of scattered light (y
axis) versus the scattering angle (^_axis)
at which the detectors received the light
beam scattered by the bacteria
(particles). Increases in intensity show
increases in the number of bacteria
(growth) and a shift in the profile to the
smaller scattering angles indicates that
the bacteria are increasing in size
(swelling) whereas a shift in profile to
the larger scattering angles indicate a
size decrease (shrinking). From increases
on the y axis, the number of bacteria are
calculated, so that NQ represents the
bacteria present at 0 time and N
represents the bacteria present after 66
min. of incubation for either the negative
control or a sample dosed at a given
concentration. Hence N/NQ represents the
relative increase in bacteria from which
the generation time ( TAU ) can be
calculated, and TAU/TAUC represents the
generation time of a dosed sample relative
to the control. Thus, an increase in
TAU/TAUC represents growth inhibition, a
decrease means growth stimulation, and a
value of 1.0 means there is no effect by
the t reatment .
Figure ifl represents the normal growth of
a tester strain (strain 10 or B_._ subtilis
strain fh2006-7) which has an N/NQ of 5.2
and a TAU of 40.06 mm. Figure 1^ shows
the same strain that has been dosed with
0.15ug/ml of 4NQO for which N/NQ was
reduced to 2.1 and TAU is 90.94 min.
Figure lc is a combination of the data
from Figures lg and
which can be
compared so that TAU/TAUC can be
calculated. In this case the values are
1.0 for the control and 2.27 for the
sample treated with 0.15 ug/ml of 4NQO.
The TAU/TAUC values for 0.15,1.5, and
4.5ug/ml of 4NQO were determined to be
2.27, 3.35, and 11.50, respectively,
showing that the generation time is
increased in a dose-responsive manner and
that TAU/TAUC corresponds directly to a
specific dose for a specific strain.
Figure 2 represents the response of strain
11 (wild type or normal strain) to 7.6
ug/ml BAP. Here, TAU for the control was
46 min. and the TAU/TAUC value was
calculated as 1.14. Figure 3 represents
the response of strain 19, a mutant strain
that specifically cannot repair the type
of damage caused by the BAP metabolite
produced by S9 activation. The TAU for
the control was 48 min. and the TAU/TAUC
was 2.27. This data shows that metabolized
BAP has a greater effect on the mutant
than the wild type strain and that BAP has
an adverse genetic (genotoxic) effect.
The TAU/TAUC value for strains 11 and 19
at BAP concentrations of 15.1 ug/ml in the
presence of S9 activation were 2.15 and
2.59, respectively which showed that at
this increased concentration even the wild
strain was damaged and that the effect on
the mutant was increased further.
SUMMARY AND CONCLUSIONS
The data presented here, though
abbreviated, show that the laser/microbial
bioassay system can detect a toxic
response from chemicals that are directly
toxic (4NQO) or those that require
255
-------
metabolic activation (BAP). The total
assay time is 66 min., and the data are
collected and stored on software
immediately when the samples are assayed
by the laser through differential light
scattering. The response can be directly
correlated with the dose by calculating
the effect of the test chemical on the
rate of growth (generation time or TAU).
The calculations are done automatically by
this computerized system within a matter
of seconds. The system is not dependent
upon a separate assay for each chemical or
chemical class but depends upon response
of the mutants based on the mechanism by
which each chemical exerts its toxicity.
Therefore, the response is specific and
moreover quantitative when related to the
dose response. Finally, the system can be
made field portable and further developed
for on-site toxic waste cleanup
monitoring.
REFERENCES
Felkner, I.e.,
Christison, T.
J. , and Wyatt,
Worthy, B.,
Chaisson,C.
P.J., 1988,
Kurtz,
Laser/Microbe Bioassay System, Proc.
5th Nat. Conf. on Hazardous Wastes
and Hazardous Materials, pp 81-84.
SCATTERING
Figure 1a. Growth of untreated control culture of Bacillus
subtllls strain fh2006-7 (No. 10) monitored by laser
differential light scattering and scored for
generation time (TAU).
256
-------
Growth of Bacillus subtil Is strain fh2006-7 treated
with 0.15 ug/ml of 4-nitroquinollne-1-oxide monitored
by laser differential light scattering and scored for
generation time (TAU).
25
SCOPING FILES FDR -t-N I TPODU INOL I NE-1-OX I DE DATA 'l;,rlb,lc>
snr SETS
I 1 -to
; : 41
I -1 43
*ug/ml
Figure Ic.
TIMEimi N/NO
6CMIM 5.1
66MIN ."J. 1
6E.NIN 1.6
66MIN 1. :
TAU
KI.OG
90.O-I
13-t. 15
463.71
TAU/TAUC
1 . 00
3TP
1'"'
10
10
10
0. 00
0. 15
1 .50
4. 50
Comparison of Bacillus subtllls strain fh2006-7
control culture to culture treated with 0.15 ug/ml of
4-nltroquinollne-l -oxide by scoring the TAU/TAUC
ratio as related to dose. 4NQO concentrations of
1.5 and 4.5 ug/ml are scored but not shown
graphically.
257
-------
STRflIN 19 B(fl)P — 7.6 UG/ML
SCATTBNO ANOU (Jtyifm)
SCOPING FILES TOP BEN7Q i A'> PYPENE DATA i 3 >
SMP SETtt
4 7 10
5 811
f, 912
*Lig/ml
TIME
66MIN
66MIN
66MIN
TAU
-IB. 03
109.03
124.40
TAU/TAUi:
1 . 00
•;•. 27
2.59
STP CONiT*
19 0.00
19 7.CO
n 15.1
Figure 3. Comparison of Bacillus subtllls potA101.hls.met (strain
19) growth In the presence and absence of 7.6 ug/ml
BAP metabollcally activated with a mlcrosomal fraction
(59) using differential light scattering and scoring for
TAU/TAUC. Treatment with 15.1 ug/ml of BAP Is also
scored but not presented graphically.
STRflIN 11 B(fl)P — 7.6 US/ML
SCATTOBNO ANOLf
-------
DISCUSSION
JEANETTE VAN EMON: About how long does it take to screen a panel for What you're doing takes really about an hour of your time. The computer and
mutants for a particular compound? the software programs process this immediately. Then you would have the
„„„„ ™^. .^1^™ i , , i_i i L • • , i ,f profile for quantitative determination of the concentrations.
CECIL FELKNER: It takes probably whatever the time is (a morning, half F M
a morning, or something like that) once the primary screen to determine the We are beginning to build libraries relative to various chemicals. If you're
profile is made. dealing with nice, clean chemicals, there shouldn't be any problem. But what
... happensif you have a real-world situation? One of the early things that we had
Sometimes you re surprised, and the response is so great that everything is ..,..,. , , , , . . ," „
,,,;., ., ,,,,", to do with this bioassay was to be able to detect toxicants in the face ol con-
wiped out. You then dilute until you get an appropriate level and redo the r , , . , ,, ,.„. , r ..... , ,,
r fusants,such thingsas bentomte.pH differences,andsotorth. We had 11 to 15
" ' different types of confusants that were put in the presence of the toxicants for
which we had to make specific identification. This was a successful operation.
259
-------
Monitoring Volatile Organics in Water by a Photovac Portable Gas
Chromatograph with Multiple Headspace Extraction Method
James S.Ho
Paul Hodakievic*
Joseph F. Roesler
Environmental Monitoring and
Support Laboratory
U.S. Environmental Protection Agency
26 W. Martin Luther King Dr.
Cincinnati, Ohio 45268
* Technology Applications, Inc.
c/o Environmental Monitoring &
Support Laboratory
U.S.Environmental Protection Agency
26 W. Martin Luther King Dr.
Cincinnati, Ohio 45268
ABSTRACT
A Photovac 10S50 portable gas
Chromatograph equipped with a capillary
column and a photoionization detector
(10.6 eV ) in conjunction with a head
space sampling device is evaluated for
monitoring volatile organics in water.
A multiple headspace extraction
technique is described for quantitative
determination of volatile organics in
water samples of various matrices. This
method,first proposed by McAuliffe,
consists of a repeated analysis of vapor
over liquid after replacing the analyzed
equilibrated gas by an equal volume of
pure air with its subsequent
equilibrium. The advantage of this
method is that it simultaneously
determines contents in the equilibrium
gas and also measures the partition
coefficients (K) for the liquid sample
under investigation. This method
eliminates the influence of a sample
matrix on the phase equilibrium.
Therefore, it can determine volatile
organics in all types of aqueous samples
without prior standardization and
without concern for a potential matrix
effect from complex constituents found
in samples of hazardous waste sites.
INTRODUCTION
Portable field instruments for
monitoring and screening at hazardous
waste sites are expected to be used
widely for both quantitative and
qualitative determination of volatile
and semivolatile organic contaminants in
air, water and soil. As a screening
tool, portable monitors minimize the
time and expense associated with
transporting uncontaminated samples back
to the laboratory for analysis.
Moreover, the on-site analytical
instrument provides immediate results by
which prompt and appropriate sampling,
off-site analyses and corrective action
may be undertaken.
A 1985 study involving 183 hazardous
waste disposal facilities demonstrated
that volatile organics were detected
more frequently than other types of
priority pollutants in groundwater (1).
This finding suggested that volatile
organic scans might be used as a
screening technique to establish the
extent of organic contamination of
hazardous waste sites. One of the
instruments, the Photovac portable
Photoionization detector (PID) gas
Chromatograph, which has been widely
used as a monitor for toxic organic
vapors in ambient air, may be used for
analyzing the water samples from
hazardous waste sites.
This paper discusses the use of the
Photovac 10S50 portable gas
Chromatograph with a capillary column
and a photoionization detector (10.6
eV.) in conjunction with a multiple
headspace extraction (MHE) technique for
quantitative analysis of volatile
organics in water.
THEORETICAL PRINCIPLE OF THE MULTIPLE
HEADSPACE EXTRACTION (MHE) TECHNIQUE
MHE was first proposed by McAuliffe
(2,3) to determine the solubility of
hydrocarbons in water. Steps of MHE are
depicted schematically in Figure 1.
First, a volume of sample solution is
placed into a calibrated glass syringe
followed by a volume of clean air
(headspace). During the equilibration
period between the liquid sample of
volume VL and the gas that occupies
volume VG, a portion of volatile
compound in the liquid will partition
into the gas phase. The phase
distribution of the analyte is defined
by a mass balance equation:
CLVL =
CLVL
CGVG
(i)
where CL is the initial analyte
concentration in the liquid, and CL and
Cr are the equilibrium concentrations in
261
-------
the liquid and gas phases respectively.
Applying the distribution law, CL =
KG
•G'
where K is the partition coefficient for
the compound in question, the equation
above may be substituted and rearranged
to yield the following relation:
CL =
(2)
After equilibration, the gas
chromatographic determination of CQ is
conducted,and then the equilibrated gas
is completely expelled from the syringe
and replaced with a fresh volume of air.
The remaining quantity of analyte in
solution (VLCL) redistributes between
the two phases. GC analysis is then
performed on the replacement air. This
pattern may be iterated any number of
times.
The formula for the determination of
volatile organics in solution utilizing
multiple headspace extraction under
static conditions and with equal volume
of VG and VLcan be derived from equation
(2):
First extraction
CL = CG1(K+1)
Second extraction
CL = CG2(K+1)2/K
nth extraction
CL =
and
cGn = Kn"1CL/(K+l)n
(3)
(4)
(5)
= (Cj./K)/( 1/K+l) n (6)
LogCGn = Log(CL/K) - nLog(l/K+l) (7)
A semilog plot of CGn VS n is a linear
relationship with the slope being a
function of K only and .the intercept
being a function of initial sample
concentration, CL, and K.
MATERIALS AND METHODS
Gas Chromatograph
The GC used for this analysis was a
portable Photovac Model 10S50 with a PID
detector(10.6 eV.). It was equipped
with a 0.53 millimeter X 10 meter
encapsuled capillary column coated with
100% chemically bonded
dimethylpolysiloxane (Photovac Inc.) .
Synthetic pure air was used as the
carrier gas and supplied at a pressure
of 40 psi. The column flow was set to
4.3 ml/minute. Since the Photovac is not
equipped with an oven, the column
temperature was not controlled. For
this reason the column temperature
fluctuated but usually stayed in the
range of 32-42 degrees Celsius.
Instrument Evaluation
In order to show that the Photovac is a
dependable instrument a gas phase
standard was analyzed five times and the
precision of the analyses calculated.
The results are reported in Table 1.
Figure 2 shows a typical chromatogram
of the six volatile organics. Detailed
evaluations(4'5> have shown the Photovac
to be an effective and handy portable
GC. The six volatile organics chosen for
this study are trichloroethylene,
benzene, 1,1,1-trichloroethane,
chloroform, 1,1-dichloroethane, and
methylene chloride. These chemicals
were chosen because they represent some
of the more common pollutants found at
contaminated ground water sites( '.
Calibration vapor phase standards were
made by injecting a standard mix of the
volatiles in methanol into a sealed
glass flask. The flask was then allowed
to sit for half an hour before it was
analyzed by the GC. The three point
262
-------
calibration curves generated were linear
as evidenced in Figure 3A and 3B.
Multiple Headspace Extraction (MHE)
Technique
Three different sample matrixes, 5%
NaCl, 10% Nacl, and primary waste water
effluent from the City of Cincinnati's
sewage treatment plant were studied.
The salt solutions were chosen to
represent the effect ionic strength had
on the phase equilibrium of the
volatiles. To perform the MHE, the
liquid sample is poured into a 100 ml
gas-tight syringe filling it to the
brim. The plunger is then pushed to
remove all but 50 ml of the sample.
The calibration standard mix is injected
into the syringe and the syringe is
immediately sealed. The syringe is
allowed to sit for 30 seconds after
which 50 ml of room air is drawn into
the syringe. The sample is then mildly
shaken for 30 seconds and allowed to sit
for 10 minutes. After 10 minutes,
equilibration is reached and the sample
is analyzed by the GC. The Photovac
has an internal pump which draws a
headspace sample into a sample loop.
The sample in the loop is flushed into a
precolumn and column. The amount of
sample entering the column is determined
by the carrier gas flow rate and the
time that the valves joining the loop
and columns are left open. During the
time that the sample is being drawn into
the GC a partial vacuum is created in
the 50 ml syringe headspace interfering
with the sample flow. This problem is
avoided by pushing the syringe barrel up
at the same time that the sample is
being taken. Care must be taken to
avoid drawing any water into the system.
After the first sample is taken, the
remaining air in the syringe is pushed
out and 50 ml of new air drawn in. The
sample is then shaken and analyzed as
before. This process is repeated until
the desired amount of extractions are
analyzed.
Calculations
The peak area for each compound is
plotted on semilog paper for each
extraction. Figure 4 shows the results
of six volatiles spiked in reagent water
which underwent five iterative
extractions. The y-axis intercept is
found by drawing the best fit line
between the points and is equal to
(Equation 7). K is equal to the
reciprocal of the slope minus 1. A
sample calculation for one of the six
volatile analytes, 1,1-dichloroethane,
may be summarized as follows:
y-axis intercept = 9.1
based on the three point calibration
curve for 1,1-dichloroethane and a
response of 9.0 the y-axis intercept
concentration, C-jy/K = 352 ug/L
slope = 1.2
K = I/(1.2-1) = 5.0
CL = 352 ug/L * 5.0 = 1762 ug/L
The true value for CL was 1753 ug/L.
The discrepancy between the known
concentration and the measured
concentration using the above procedure
was 0.5%.
Liquid Standard Technique
The more common technique for headspace
analysis is to perform just one
extraction and compare the peak areas to
a liguid standard. Reagent water was
spiked at the same concentration as the
samples and the first extraction peak
areas were used for calculating the
original concentration of the volatiles
in the samples.
RESULTS
Table 2 lists and compares the results
of the MHE and liguid standard
techniques. Figure 5 shows this
263
-------
comparison in graphical form. These
results show that the MHE provided much
greater accuracy than the liquid
standard technique. The discrepancy
range for the MHE method was from -20.1
±11.2 % for dichloromethane in waste
water to +29.7 + 15.5 % for chloroform
in 10 % NaCl solution as compared to
the discrepancy range for the liquid
standard technique which was -64.2
±20.1 % for trichloroethylene in waste
water to 167.3 +49.0 % for
trichloroethylene in 10 % NaCl solution.
Table 3 lists the K values found for the
six compounds in the three matrixes and
reagent water. The difference in K
values between the 5% and 10% NaCl and
reagent water explains the much greater
discrepancy of results when using the
liquid standard technique for these
matrixes. The K values for the waste
water are not greatly different from
those of the reagent water, but the
accuracy still suffered when using the
liquid standard technique in waste
water. This points out another
advantage of the MHE method; several
extractions are performed, and if one of
the extractions is off, that point may
be discarded. In the case of the waste
water, the first extraction peak areas
were low compared to the subsequent
extractions, due most likely to spurious
adsorption effects similar to those
reported by Drozd^6). Using the MHE
technique on the water, the first
extraction point could be discarded and
still sufficient data would remain for
an accurate determination of CL and K.
The liquid standard method, however,
lives and dies on the first extraction;
a definite disadvantage, especially when
the quantity of sample is limited.
CONCLUSIONS
The above examples have shown how
volatile organic compounds in water
samples of various matrices can be
quantitatively analyzed with the MHE
procedure. The major advantages of
using the Photovac 10S50 portable GC
with the multiple headspace extraction
method are:
1. More accurate data can be obtained
by using the MHE procedure in
comparison with the standard single
extraction liquid standard
procedure.
2. Use of this technique for on-the-
spot analysis reduces the cost of
sending unnecessary samples back to
the laboratory for analysis. Quick
decisions can also be made to
redirect investigations of
hazardous waste sites.
3. With the MHE method, the influence
of a sample matrix on the phase
equilibrium is eliminated. It
simultaneously determines the
analyte contents in the equilibrium
gas of a headspace and measures the
partition coefficient (K) for the
liquid sample under investigation.
Therefore, a Photovac portable GC
combined with the MHE technique
provides a useful field ability to
determine volatile organics in all
types of aqueous samples. It
requires no external aqueous
calibration standard and has no
need for concern about matrix
effects which are likely to occur
in complex samples found at
hazardous waste sites.
REFERENCES
1. "Volatile Organics Scan:
Implications for Ground Water
Monitoring" by R.H. Plumb, Jr. and
A.M. Piatchford. Proceedings of
the Petroleum Hydrocarbons and
Organic Chemicals in Ground Water -
Prevention, detection and
Restoration - Conference and
Exposition - The Westin Galleria
Houston Texas, Nov. 13-15, 1985.
264
-------
"GC Determination of Solutes by
Multiple Phase Equilibrium" by C.
McAuliffe, Chem Tech. 46(1971)
"Head-Space Analysis and Related
Methods in Gas Chromatography" by
B. V. looffe and A.G. Vitenberg,
Lenningrad State University,
Lenningrad, USSR, translated by
Ilya A. Mamantov P42-46 A Wiley-
Interscience Publication John Wiley
& Sons, 1984.
"Evaluation of Photovac 10S50
Portable Photoionization Gas
Chromatograph for Analysis of Toxic
Organic Pollutants in Ambient Air"
by R.E. Berkley, U.S.EPA EMSL
Methods Development and Analysis
Division, Research Triangle Park,
N.C. 27711. Report No. EPA/600/4-
86/041.
"Evaluation of Photovac TIP and
Model 10S50 Gas Chromatograph as
Screening Tools" by M.W. Holden,
D.L. Smith, and G.S. Durell,
atelle Columbus Division,
Columbus, OH 43201-2693, Contract
No. 68-02-4127 Work Assignment 26.
"Spurious Adsorption Effects in
Headspace-Gas Determination of
Hydrocarbons in Water" by J. Drozd,
J. Vejrosta, J. Novak, and J.A.
Jonsson, Journal of Chrom.,
245(1982)185-192.
1" EXTRACTION
2nd EXTRACTION
1" SUBSTITUTION
Nth EXTRACTION
(N-l) SUBSTITUTION
V«n
-v,;c
L-'-'ln
FIGURE 1. DIAGRAM OF MUTIPLE HEADSPACE EXTRACTION
265
-------
Figure 2. Typical chromatogram of the six volatile organics.
Peak 1 is the solvent peak. The other peaks are
as follows: 2. methylene chloride 3. 1,1-dichloroethane
4. chloroform 5. 1,1,1-trichloroethane 6.benzene
7. trichloroethylene.
CO
CD
£_
CO
CD
Q_
16
17 -
16-
15-
14 -
13-
12 -
11 -
10 -
9-
8 -
7 -
6-
5-
4 -
3-
2 -
1 -
Qas Phas* Calibration Curvts
Response vs. Concentration
o 20
Trichloroethylene
40
ug/L
60 BO
+ Benzene
100
oMethylene chloride
FIGURE 3A. THREE POINT CALIBRATION CURVES
266
-------
QAS PHASE CALIBRATION CURVES
Response vs Concentration
i i i i i i iii \ir irirr r i
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
(Thousands)
ug/L
• 1, 1, 1-trichloroethane + Chloroform o 1. 1-Dichloroethane
FIGURE 3B. THREE POINT CALIBRATION CURVES
a:
20.0
10.0-
0
1
FIG.4
234
EXTRACTION NUMBER
Semilog plot of Response vs. Extraction Number in Reagent Water
267
-------
c
o
c
>> \
¥ o
c: o
o <
K
<
<
./
\
y
\
i%
/
,
/
W
IV
_K ^
^
<
/
<
N
Id
,^7^
r
X
X
X
x
X
x
<
H
-•
^
-j
5
1 R
7
>
£
£
\
^r
„
_P
(3
R
^
K
X
X
><
X
r;
Trichloroethylene Benzene 1,1,1-Trichloroethane Chloroform 1,1-Dichloroethane Methylene chloride
MHE fgggl MHE VTA MHE F?^l LS ITVl LS P^H LS
FIGURE 5. COMPARISON OF TWO DIFFERENT HEADSPACE METHODS
Table 1. Instrument Precision for Gas Hiase Standards
Compound
Irichloroethylene
Benzene
1,1,1-Tricnloroethane
Chloroform
1,1-Dichloroethane
Methylene Chloride
Test
Cone.
ug/L
16.4
8
154.4
795.8
172.4
49.1
Avg.
Peak
Area
10.6
5.5
3.2
2.8
4.7
2.6
RSD %
(n=5)
8.2
12.2
8.2
5.7
6.1
4.4
268
-------
Table 2. Results of Analysis of Six Volatile organics Using the Multiple Headspaoe
Extraction Technique and the Liquid Standard Technique
5% NaCl Solution
C ug/L
Spite
Trichloroethylene 167.6
Benzene 80.9
l,l,l-JTrichloroeth 1571.1
Chloroform 8094.5
1,1-Dichloroetnane 1753.5
Methylene Chloride 499.5
Multiple Headspaoe Extraction Liquid Standard Technique
C ug/L RSD Discrepancy
Found % ug/L %
166.4
79.3
1423.9
9281.3
1636.6
427.7
3.7
10.3
4
18.6
0.8
8.9
-1.2
-1.6
-147.2
1186.8
-116.9
-0.7
-2
-9.4
14.7
-6.7
C ug/L
Found
BSD
Discrepancy
ug/L %
-71.8 -14.4
273.4 22.3 105.6 63.1
98.8 10.5 17.9 22.2
1771.4 12.8 200.3 12.8
8755.3 10.6 660.8 8.2
1992.6 8.1 239.1 13.6
564.7 11.9 65.2 13.0
10 % NaCl Solution
Trichloroethylene 167.6
Benzene 80.9
1,1,1-Trichloroeth 1571.1
Chloroform 8 094.5
1,1-Dichloroethane 1753.5
Dichloromethane
499.5
Multiple Headspace Extraction Liquid Standard Technique
C ug/L PSD Discrepancy C ug/L RSD Discrepancy
Found % ug/L % Found % ug/L %
178.5 0.5 10.9 6.5 448.0 20.5 280.4 167.3
17.6 101.1 125.0
12.5 1575.0 100.2
6.5 4708.0 58.2
10.1 1700.4 97.0
6.0 347.5 69.6
87.9
1605.9
10498.6
1591.2
590.7
9.7
2.4
15.5
3.7
2.7
9.7
34.8
2404.1
-162.3
91.2
13
2.2
29.7
-9.2
18.3
182.0
3146.1
L2802.5
3453.9
847.0
Primary Waste Water Effluent
Multiple Headspace Extraction Liquid Standard Technique
C ug/L PSD Discrepancy C ug/L RSD Discrepancy
Found % ug/L % Found % ug/L %
Trichloroethylene 167.6
Benzene 80.9
1,1,1-Trichloroeth 1571.1
Chloroform 8094.5
1,1-Dichloroethane 1753.5
Dichloromethane 499.5
172
74
1659
10603
1744
399
.4
.3
.2
.3
.4
.1
13.7
10.6
8
5.8
18.4
11.2
4
-6
88
2508
-9
-100
.8
.6
.1
.8
.1
.4
2
-10
5
-0
-20
.8
.6
.6
31
.5
.1
81.81
50.11
1188
7158
1328
335.8
20.
9.
6.
9.
5.
1
3
6
7
9
8
-85.79
-30.79
-383.1
-936.5
-425.5
-163.7
-64.2
-37.6
-22.9
10.1
-28.5
-34.9
Table 3. Comparison of the Partition Coefficient (K) Found
for each of the Six Volatiles in Reagent Water and Three
Matrixes.
Trichloroethylene
Benzene
1,1,1-Trichloroethane
Chloroform
1,1-Dichloroethane
Methylene Chloride
Reagent
H2O
2.9
4.0
1.6
7.7
5.3
11.1
K values
5%
NaCl
1.9
3.0
1.2
4.0
4.0
7.1
10%
NaCl
1.0
1.4
0.6
2.9
2.1
6.3
Waste
Water
3.2
4.5
1.7
6.3
4.8
7.1
269
-------
DISCUSSION
JACK MURPHY: You say this saves time. If you do four or five extractions BERNIE BERNARD: I found that two extractions are typically needed for the
and run them on the GC, how long does it take to run one sample? compounds that extract easily. What about something like tetrachloroethylene
,,,-__„„ ,. , , . , _ -if where might get into a situation where every extraction gives you basically the
JAMES HO: You really don t need to take four or five extractions. If you get * J 5 ' '
,..,,_ . . . same amount;
two or three good extractions, you can do it quickly. One extraction takes about
ten minutes, and the analysis takes about ten minutes. Twoorthreegoodextrac- JAMES HO: We haven't tried tetrachloroethylene. It probably would be a
tions would be enough. problem.
270
-------
HAZARDOUS WASTE SITE MEASUREMENTS
OF PPB LEVELS OF CHUDRTNATED HYDROCARBONS
USING A PORTABLE GAS CHROMATOGRAPH
AMOS LINENBERG
PRESIDENT
SENTEX SENSING TECHNOLOGY, INC.
RIDGEFIELD, NJ USA
ABSTRACT
Portable Gas Chrcmatography has been established
as an alternative to laboratory analysis,
especially in hazardous waste site situations
where analysis is required to be obtained at
"Real Time".
For reliability and accuracy, however, the
instrument to be used should include most of the
features of a laboratory gas chromatograph,
particularly, the ability to detect all compounds
in question. This paper will describe the
system used, the experiments conducted, and the
results of various measurements of chemicals
found in hazardous waste sites.
INTRODUCTION
The technique of Gas Chromatography offers
considerable advantages when used for the de-
tection of PPB levels of contaminants in air,
water and soil. The features offered by this
technique are as follows:
I. HIGH RESOLUTION.
II. HIGH SENSITIVITY.
III. DETECTABILITY OF A LARGE VARIETY OF
COMPOUNDS.
IV. REPRDDUCABILITY AND ACCURACY IN
IDENTIFICATION AND QUANTIFICATION OF COMPOUNDS.
While those features are easily obtained in the
laboratory, a field operating gas chromato-
graph must include all the necessary operational
capabilities to obtain these features in an
on-site analysis.
I.
HIGH RESOLUTION.
Resolution is determined by the following
parameters:
A. Ability to change columns.
B. Ability to use columns of different length.
C. Capability to use capillary columns or packed
columns depending upon resolution needed.
D. Efficient injection system - a heated on
column injection or efficient adsorption/
desorption system.
E. Stable and evenly controlled column oven.
F. Temperature programming.
G. Low Volume detector cell to accommodate
capillary columns.
II. HIGH SENSITIVITY.
High sensitivity is obtained by using the proper
detector for the compounds to be detected. A
Photoionization Detector, Argon lonization
Detector and Electron Capture Detector can be
used. A Thermal Conductivity Detector does not
have the sensitivity for low level measurement;
a Flame lonization Detector can be used, but it
is cumbersome to operate in the field.
In addition, a preconcentrator could be used to
obtain ultra high sensitivity or to comply
with certain federally-recommended methodologies.
III. DETECTABILITY OF VARIETY OF COMPOUNDS.
Certain detectors will allow the detection of
different compounds. The Photoionization
Detector can detect most organic compounds if a
proper light source is used. The 11.8 E.V.
light source will detect all organic compounds
except those with a higher ionization potential.
Since the lifetime of this light source is
short,' however, the light source which is used
most frequently is the 10.6 E.V. source.
Although'this detector is sensitive to many
organic compounds, it is quite insensitive to
typical hazardous waste site compounds such as:
Dichloroethane, Trichloroethane, Carbon Tetra-
chloride, Chloroform, Methylene Chloride and
other compounds which have ionization potentials
about 10.6 E.V.
The Argon lonization Detector, with an energy of
11.6 E.V. will detect all-(he above compounds
and is also, a rugged and reliable detector
easily operated on site without a flammable gas
source.
271
-------
The Electron Capture Detector is extremely
sensitive and relatively easy to operate but is
selective only to comnonly found volatile
halogenated compounds.
IV. REPRODUCABILITY AND ACCURACY IN
IDENTIFICATION OF COMPOUNDS.
The ability of the gas chromatograph to
reliably identify compounds is a function of its
ability to reproduce results. This ability
requires stable temperature and flow conditions.
More important, the calibration method used for
the identification should be an accurate one
which uses certified standards for each compound
to be detected. The use of computerized
methods, in which one calibrant is used to
analyze a number of compounds is a relatively
inaccurate method, and errors in both identifi-
cation and quantitation of compounds may occur.
EXPERIMENTAL
The purpose of the experiment was to identify and
quantify various chlorinated hydrocarbons of the
type which are commonly present at hazardous
waste sites. The main object was to separate
and detect those hydrocarbons in the presence of
various solvents such as Hexane, MEK, and
gasoline vapors.
The compounds to be analyzed were:
Vinyl Chloride
1,1 Dichloroethane
1,2 Dichloroethane
1,1 Dichloroethene
1,2 Dichloroethene
111 Trichloroethane
112 Trichloroehtane
Trichloroethene
Methylene Dichloride
Tetrachloroethylene
The instrument used was the SCENTOGRAPH Portable
Gas Chromatograph. Both capillary and packed
columns were used for the separation. Tempera-
ture was varied from 40 deg. C to 110 deg. C.
The detector employed was an Argon lonization/
Electron Capture Detector. The Argon lonization
Detector was suitable for the detection of all
compounds. This detector, when operated as an
Electron Capture Detector, detected the
chlorinated compounds without interferences of
other hydrocarbons. PPB level measurements were
obtained by using the preconcentrator with the
Argon lonization Detector or by an injection of
1 cc of air sample to the gas chromatograph
using the Electron Capture Detector.
Temperature programming was used to shorten
analysis time. Because the SCENTOGRAPH is
operated from a built in lap-top computer, all
results and chromatograms were stored on disk
for future reference. Calibrations were stored
independently so as to allow access at future
waste site evaluations, obviating the need for
subsequent preparation of calibration standards.
RESULTS
Using different columns, complete resolution of
all compounds was obtained. Please see Figure 1
and 2 for a typical chromatogram. The main
difficulty appeared in the determination of
vinyl chloride, for two reasons: Due to its
high volatility, its retention time is relatively
short and it was difficult to detect it using the
same column used to separate the heavier com-
pounds. In addition, the Electron Capture
Detector did not respond well to the vinyl
chloride and, therefore, it was more difficult
to detect it from the other solvents.
Standards for calibration obtained from a
specialty gas company were certified to the PPB
levels. Other standards were prepared by mixing
compounds in teflon bags.
CONCLUSION
Accurate on-site analysis of chlorinated
hydrocarbons were carried out by using the
SCENTOGRAPH. All results were recorded on the
computer disks. Complete resolution of all
compounds was obtained down to the concentrations
in the PPB range. The Argon lonization Detector
coupled with a preconcentrator was used for
general analysis; the Electron Capture Detector
was used when interferences from other solvents
were present.
272
-------
METHYLENE CHLORIDE
CHLOROFORM
1,1.1 TRICHLOROETHAHE
CARBON TETRflCHLORIDE
BROnODICHLOROHETHflNE
TRICHLOROETHYLENE
BROttOFORM
TETRrtCHLORETHYLEME
Column used: 10' long x 1/8" o.d. teflon packed with 20% SP2100/.1% Carbowax 1500 on 100/20
Supelcoport
Datector: Argon lonization, Column Pressure: 22 PSI, Column Tenperature: 50 deg. c,
Preconcentrator: Tenax, Sampling Tine: 10 seconds, Concentration levels of all compounds
tested were at 1 ppm + or - 10%
Column used: 6' aluminum 1/8" o.d packed with 20%
SP2100/.1% Carbowax 1500 on 100/120 Supelooport
Detector: Argon lonization
Column
Pressure: 20 PSI
Column
Tenperature:75 deg. c.
Preconcentrator: Tenax
Sanpling Time: 10 seconds
Concentration Levels of the four
components tested were at:
TrichlozDethylene: .5 ppm
1,1,1 Trichloroethane: 2 ppn\
1,1,2 Trichloroethane: 2 ppn
Methylene Dichloride: 2 ppm
1. METHYUENE DICHLORIDE
2. 1.1.1 TRICHLOROETHANE
3- TRICHLOROETHY1-ENE
4. 1.1.2 TRICHLOROETHANE
273
-------
DISCUSSION
AVRAHAMTEITZ: Have you used any of the equipment for air sampling, AVRAHAM TEITZ: Is it possible to get to the sub ppb level, for say 1,1,1-
and have you had any problems with it for ambient air sampling? I have known TCA?
some organizations that have had some problems. »»!,-><-,¥ Tximnnnr.,-, ,, • , j . , , , , , .
6 H AMOS LINENBERG: It's not intended to go below the ppb level. In the field,
What kind of sensitivity can you get using the Sentex® for ambient air Iwouldsafely say thatwecansee5-10ppbeasily without pushing it too much.
sampling? With chlorinated compounds, especially those that are responding to ECD, we
AMOS LINENBERG: You can go to the ppb level. Ca" g° bd°W the Ppb leve1'
274
-------
CORRELATION CHROMATOGRAPHY WITH A
PORTABLE MICROCHIP GAS CHROMATOGRAPH
Edward B. Overton, Robert W. Sherman,
Charles F. Steele and Hettihe P. Dharmasena
Institute for Environmental Studies
Louisiana State University, Baton Rouge, Louisiana 70803
ABSTRACT
The identification of unknown volatile organic
compounds is of great interest for chemical hazard
assessments. The speed in which these toxic
chemicals can be identified and their concentrations
determined is particularly crucial in emergency spill
situations. A portable microchip gas chromatograph
used with correlation chromatographic techniques
allows for rapid on-scene identification of volatile
organic compounds at part-per-million levels. The
Gas Analyzer uses two temperature controlled
capillary microchip gas chromatographs with
different stationary phases for qualitative and
quantitative analysis. It is interfaced with a
Macintosh personal computer which uses special
software to identify unknown compounds and
integrates the peak areas in a chromatogram. Each
sample is simultaneously analyzed on two different
chromatographic columns having different stationary
phases. The resulting retention times are compared
to a library of normalized retention times for a variety
of volatile organic compounds. Correlation
chromatography is used to insure the positive
identification of components in an unknown sample
by requiring that the library-matched identifications
for both gas chromatographic columns must be in
agreement.
INTRODUCTION
Chemical hazard assessments at spill incidents and
Superfund Sites commonly require the compound-
specific qualitative and quantitative analyses of
unknown volatile organic compounds. Further, these
analyses should be rapid. Since traditional
laboratory-based analyses are time consuming and
expensive there is an urgent need for the develop-
ment of rapid field-deployable analytical methods. A
new compact microchip gas chromatograph (GC),
containing two independent temperature controlled
capillary column gas chromatographs, allows for
rapid on-scene identification of volatile organic
compounds at part-per-million (ppm) levels.
Correlation chromatography techniques (1) have
been adopted to aid in the identification of unknown
compounds by comparing the retention indices
obtained from the simultaneous analysis of samples
on two GC columns coated with stationary phases of
different polarity.
EXPERIMENTAL
The Microsensor Gas Analyzer (model 200,
Microsensor Technology Inc., Fremont, California) is
a small, portable instrument which uses two
temperature controlled capillary microchip gas
chromatographs with different stationary phases.
The columns in the Gas Analyzer were coated with
DB-1 or DB-5 and DB-1701 liquid phases by
Microsensor Technology. Gas chromatographic run
times on the microchip GC were on the order of one
minute.
The Microsensor Gas Analyzer was interfaced with a
Macintosh personal computer (Apple Computers,
Cupertino, California) for acquisition and treatment of
chromatographic data. The software allows for the
display of chromatographic peaks, automatic peak
detection and base-line assignment, area
integration, and matches the sample peaks of a
chromatogram with a normalized library of retention
times or retention indices for unknown compounds
run on each liquid phase.
Qualitative information was obtained for an unknown
sample by use of a modified Kovats Retention Index
System (2). The retention indices for compounds on
columns of different polarities were compared.
Positive identification of a given compound was
made only if the retention index for both columns
matches standard values (1) in a library accessed via
the Macintosh computer. Table 1 gives a portion of
the retention index libraries for selected volatile
compounds of interest in environmental analyses.
A series of gaseous standards were injected via a
pressurized sample loop onto each column and the
retention times determined. From this, retention
indices were calculated. This information served as
a reference for the analyses of unknowns. Various
components in a gaseous mixture were analyzed
using correlation chromatography on the
275
-------
Microsensor Gas Analyzer. A gaseous mixture was
introduced into the Gas Analyzer and simultaneously
analyzed on both chromatographic columns.
Figure 1 shows an example of capillary GC analyses
of a standard mixture on DB-5 and DB-1701
columns. The difference in retention times for
various compounds is illustrated. Figure 2 shows
retention indices calculated for a set of compounds
on DB-1 and DB-1701 columns. These data were
normalized using straight chain hydrocarbons (64
through C-| 1) standards, to provide retention indices
that were used for compound identification.
SUMMARY AND CONCLUSIONS
High resolution can be best obtained by using capil-
lary columns (1). Selectivity may be obtained by
changing the temperature of the column or using two
columns coated with stationary phases of different
polarity. The Microsensor Gas Analyzer has all of the
above features which lends itself to excellent qualita-
tive and quantitative capabilities.
The size, two capillary columns, correlation chroma-
tography and temperature control of the columns are
features which make the Microsensor Gas Analyzer
attractive for field use. The interface of the Gas
Analyzer with the Macintosh allows for qualitative
and quantitative information with a user-friendly
system.
ACKNOWLEDGMENTS
The financial support for this work was provided by
the National Oceanic and Atmospheric
Administration, U.S. Department of Commerce,
Contract No. 50-ABNC-7-00100.
REFERENCES
(1) Freeman. R.R.: High Resolution Gas
Chromatography. Hewlett-Packard
Company, 1981.
(2) Kovats, E.; Helvitica Chemica Acta., 41 (1958),
1915.
Table 1. A selection from the retention index libraries for DB-1
and DB-1701 columns.
GAS
DB-1 701 DB-1
2.5 meters 1.75 meters
0.5 n film 0.5 n film
HEXANE 600 600
DICHLOROMETHANE 604 514
trans-1,2-DICHLOROETHYLENE 607 550
ISOPROPYL ALCOHOL 612 488
1,1-DICHLOROETHANE 640 557
ACRYLONITRILE 642 495
VINYL ACETATE 648 565
CYCLOHEXANE 669 660
ETHYL ACETATE 683 601
2,2,4-TRIMETHYL PENTANE 683 692
1,1,1-TRlCHLOROETHANE 684 630
CARBON TETRACHLORIDE 685 650
TETRAHYDROFURAN 687 615
METHYL ETHYL KETONE 690 572
CHLOROFORM 695 597
HEPTANE 700 700
BENZENE 709 645
1,2-DICHLOROETHANE 727 621
TRICHLOROETHYLENE 740 683
276
-------
DB-1701 35°C
1,1-Dichloroethylene
Dichloromethane
trans-1,2-Dichloroethylene
1,1-Dichloroethane
Chloroform
1,2-Dichloroethane
Benzene
1,2-Dichloropropane
Trichloroethylene
10
20
30
seconds
40
50
60
Figure 1. A gaseous mixture analyzed on DB-5 and DB-1701
columns of a Microsensor Gas Analyzer, model 200
(Microsensor Technology Inc., Fremont California).
700 -
ffi
o
m
4)
c
V
«••
V
DC
600
500
CCI4
(+1%)
\
1,1,1-TCE
CHCB
1,1-DCE
MEK
Vinyl Acetate
600
650
700
750
Retention Indices DB-1701
Figure 2. Retention indices calculated for a set of compounds on
DB-1 and DB-1701 columns of a Microsensor Gas
Analyzer, model 500 (Microsensor Technology, Inc.,
Fremont California).
277
-------
DEVELOPMENT OF A FIELD PORTABLE CONCENTRATOFt/PURGE AND
TRAP DEVICE FOR ANALYSIS OF VOLATILE ORGANIC COMPOUNDS IN
AMBIENT AIR AND WATER SAMPLES
Robert W. Sherman, Edward S. Collard, Michael F. Solecki,
Tom H. McKinney, Linda H. Grande and Edward B. Overton
Institute for Environmental Studies
Louisiana State University
Baton Rouge, Louisiana 70803
ABSTRACT
Volatile organic compounds (VOC) are frequently
found at very low concentrations in ambient air and
water samples. To aid in the detection of these
compounds, a portable device has been developed
which concentrates VOC analytes from field samples
on solid sorbent traps. The traps are made with a
combination of Tenax GC and Spherocarb absor-
bants which allow efficient trapping of compounds
with volatilities ranging from those of vinyl chloride to
dichlorobenzene. Relatively small volumes (200-500
mL) of ambient air samples are concentrated by
factors of 100 to 300. Additionally, this concentration
device allows the purging of analytes from water
samples and trapping the analytes on the solid
absorbents. The concentrated samples may then be
analyzed using appropriate field gas
chromatographic instruments.
INTRODUCTION
The detection of volatile organic compounds in field
samples at sub-part-per-million levels is difficult
using current laboratory instrumentation and state-of-
the-art field instrumentation. It is necessary that field
samples be concentrated to a level which is above
the instrument detection limits for these types of
analyses. A portable concentrator/purge and trap
device may be used in the laboratory or in the field to
enhance the rapid detection of volatile pollutants or
hazardous compounds.
Tenax GC and Spherocarb absorbants are both
layered in an extremely small glass lined tube (11.4
cm with 0.7 mm i.d.) and used to trap volatile organic
compounds of interest. These two absorbants retain
compounds having a fairly wide range of volatilities.
Tenax GC (1,2) and other sorbents (3) have been
used to trap organic compounds in air. It has been
determined that a Tenax GC-Spherocarb combi-
nation is well suited for field analyses of air and
water (4). These combined absorbents trap
compounds with a fairly wide range of ambient
temperature vapor pressures (2-2500 mm Hg) and
allow complete purging of the trapped compounds at
elevated temperatures.
APPARATUS
A portable, metal case is used to house the
concentration device, which is primarily composed of
two solid sorbent traps (Figure 1). Each trap has its
own heating unit, gas intake and outlet, and
appropriate valves. Controllers and indicators are
used to maintain a given temperature for a trap and
provide appropriate flow rates of carrier gas through
a trap. Since two traps are present in the concen-
trator, it is possible to accelerate the analysis time for
a sample. While one trap is desorbing, the other trap
can accept the next sample.
An air sample (200-500 mL) is injected onto a trap
using an appropriate loading device such as a
syringe, the trap is heated to 230°C, and helium is
then used to flush the trap of any desorbed organic
compounds. A concentrated sample is collected (this
volume is adjustable) into another syringe and is
ready to be used in any chromatographic instrument.
For water analysis, a cap fitted with two tubes is
placed on a vial containing a 5 mL sample (Figure 2).
One tube extends to the bottom of the vial and is
used to introduce the helium purge gas. The other
tube collects the purged gases from above the
sample. The purged gases are fed directly into the
load port of a Tenax GC-Spherocarb trap. A flow
meter indicates the amount of gas which has passed
onto the trap. From this point on, the VOC in the
water sample are concentrated in the same manner
as an air sample.
A battery of compounds has been tested for use with
the concentrator. The percent recoveries for com-
pounds with volatilities ranging from those of vinyl
chloride to dichlorobenzene have been in the range
of 40-70%.
FIELD USE
An air sample may be collected in the field and
injected directly into the concentration device. A field
sample may also be collected in a bag or some other
approved container and returned to a laboratory for
analysis via the concentrator and appropriate
279
-------
analytical instruments. Water samples may be
purged and analytes trapped using the concentrator
in the tield or in a laboratory.
The concentrator has been used on several
occasions for air monitoring and water analysis.
Figure 3 shows an air sample that has been
analyzed with and without the use of the concen-
trator. It can be seen that the concentrator enhances
the analysis of any sample by raising the concentra-
tions of volatile organic compounds above the limits
of detection of a chromatographic instrument.
Figure 4 shows an analysis of the headspace of a
typical water sample, and the analysis of the same
water sample after concentration by the purge and
trap device. Again, low levels of compounds are
detected in the concentrated sample which would not
have been noted in the unconcentrated sample.
SUMMARY AND CONCLUSIONS
The use of the concentration device in conjunction
with chromatographic instrumentation enhances the
analyses of samples by allowing lower limits of
detection for volatile organic compounds in air and
water. Since the concentrator is the size of a small
tool box, it is easily transported into the field to aid in
analytical testing at various types of sites. The
concentrator is flexible since the sample volume and
concentrated volume are not fixed. The temperature
of the traps may be varied to suit the needs of a given
analysis. In addition, sample analyses may be
obtained quickly since two traps are available. This
concentration device is convenient for detecting
levels of volatile organic compounds in air and water
that have previously been impossible for a given
chromatographic instrument.
ACKNOWLEDGEMENTS
The financial support for this work was provided by
the National Oceanic and Atmospheric
Administration, U.S. Department of Commerce,
Contract No. 50-ABNC-7-00100.
REFERENCES
(1) Pellizzari, Edo, Demian, Barbu, Krost,
Kenneth, "Sampling of Organic Compounds in
the Presence of Reactive Inorganic Gases with
Tenax GC," Analytical Chemistry. Vol. 56,
1984, pp. 793-798.
(2) Crist, Howard L, Mitchell, William J., "Field
Audit Results with Organic Gas Standards on
Volatile Organic Ambient Air Samplers
Equipped with Tenax GC," Environmental
Science and Technology. Vol. 20, No. 12,
1986, pp. 1260-1262.
(3) Williams, E.J., Sievers, R.E., "Synthesis and
Characterization of a New Sorbent for Use in
the Determination of Volatile, Complex-
Forming Organic Compounds in Air,"
Analytical Chemistry. Vol. 56, 1984, pp. 2523-
2528.
(4) Robert Cox, "Sample Collection and Analytical
Techniques for Volatile Organics in Air,"
Proceedings of the 1983 Air Pollution Control
Association Conference on Measurement and
Monitoring of Non-Criteria Contaminants in
Air.
Toggle
valves
Q Bubbler
Traps
120 Volts AC
Figure 1. Schematic of concentrator/purge and trap device.
280
-------
Collection
of Analytes
Helium purge gas
D
Loading Port
5 ml_ sample
Figure 2. Glass vial for purging and trapping water samples.
Stainless steel needles and Teflon tubing are used
for the introduction and collection of gases.
10 ppb v/v in air
Before Concentration
After Concentration
10
20
seconds
1 Dlchloromethane
2 Chloroform
3 Benzene
4 1,2-Dlchloropropane
5 Bromodlchloromethane
6 Toluene
7 Tetrachloroethylene
30
40
Figure 3.
Analyses of a 10 ppb air sample with a Microsensor
Gas Analyzer, Model 200 (Microsensor Technology,
Inc., Fremont, CA), and the same sample concentrated
500 mL/1.5 ml using the concentrator device.
Volatiles in Water
60 ng/mL
Headspace
1 Dlchloromethane
2 Chloroform
3 Benzene
4 Bromodlchloromethane
Purge and Trap
^
1
4
A
20
30 40
Seconds
50
Figure 4.
Analyses of the headspace of a water sample with a
Microsensor Gas Analyzer, Model 200 (Microsensor
Technology, Inc., Fremont, CA), and the same
sample purged and trapped using the concentrator.
281
-------
DISCUSSION
JOE SOROKA: What would you estimate is the upper limit for the concen-
trations for your purge-and-trap device? You're using only a very little amount
of Tenax* (unless you break) therefore the volume is going to be very low.
ED OVERTON: The upper limit that we recommend is the detection limit of
the analytical device, without concentrations. If we use the microchip GC, it
can detect about a part per million, without concentration. You would screen the
sample very quickly and see if you got anything. If you don't have anything,
then you can run it through a concentrator to concentrate the organics.
JOE SOROKA: Is that screening by direct headspace?
ED OVERTON: That's the only way to do it. If you don't see anything by
direct headspace, then you can do a purge and trap. Of course, you do the same
thing for volatiles in air.
AVRAHAM TEITZ: When using the concentrator, what kind of detection
limits do you have for organics? For air sampling using the concentrator, what
are the detection limits?
ED OVERTON: It strictly depends on the detector that you use. If you use a
photovac detector, the sensitivity will be lower. If you use a microchip GC, its
detection limits are on the order of one part per million, and if you concentrate
that by a factor of 300, you're down in the five to ten part per billion range.
The microchip GC that we use has a little bit lower detection limit, so you're
on the order of one pan per billion.
TOM SPITTLER: Do you have a rough idea of the cost of this instrument, and
when it's likely to be on the market?
ED OVERTON: It is made and available right now from Microsensor
Technology. It costs $8,000 for the GC. You have to use some software with it.
I think they're selling a software package that requires the knowledge of a
chemist (a timed-event software package).
Notice that the baseline was not flat, and so to get correct integrations, you have
to use timed event, a tangent skim here, or a valley. We are working on a
software package that has some intelligence built into it, that would integrate
these peaks without having a chromatographer there. That is not quite available
yet.
AL PLEVA: What do you use to plug the ends of your traps?
ED OVERTON: These things are permanently mounted, so they are plugged
by the tubing that is connecting everything in there, so you always use the same
trap. It's a basic difference between the commercially available trap device and
this device.
AL PLEVA: What do you use to hold the Tenax® in?
ED OVERTON: A small piece of wire slightly bent. You load the trap, put the
piece of wire in one end, and load in Spherocarb.® with a syringe. Then the
Tenax® is loaded and wire forced in the other end.
AL PLEVA: Do you have any problem with water condensation on the wire.
ED OVERTON: We haven't seen any problems yet. I worry a little bit about
our catalytic degradation, but so far, so good.
AL PLEVA: You're using Tenax® permanently mounted inside, is that correct?
ED OVERTON: That's correct. Tenax® and Spherocarb.® Tenax®, first plug,
Spherocarb,® second plug.
AL PLEVA: Do you have any problems with memory or build-up over time,
and how do you know when you've got to get rid of it, and replace it?
ED OVERTON: You've got to use exactly the same precautions you would
use with any purge-and-trap device - that is at the high temperature, you purge
your sample off, and then keep it at that temperature for another few minutes,
drive off any contamination and then cool it back down.
What we do is typically run it up to temperature put a mL and a half through
there, and then maybe run another 20 mL through before we lower it back down.
If you use normal, good-quality analytical procedures, you won't have those
problems. If, of course, you don't, you're going to have plenty.
AL PLEVA: How many uses do you estimate that you can get on the
concentrator before you feel that you have to replace it?
ED OVERTON: We don't know yet. I would imagine several hundred cycles,
but I don't know.
MICHAEL SOLECKI: I have several hundred runs on it, and I haven't had
anything. And I always do couple of tests on it.
AL PLEVA: And what was the top temperature that you went to and when you
went to desorption?
ED OVERTON: About 230,1 think. We tried to come close to the standard
technology, so we wouldn't have to reinvent any wheels.
DREW SAUTER: You mentioned not reinventing wheels. There is a piece of
public domain software available that does background subtraction, and it's
fairly sophisticated. It's actually written for mass spectroscopy, but you might
want to look into that for background subtraction from chromatograms. I'll
give you the listing.
ED OVERTON: This is not a trivial issue. We have run through more
algorithms than I would like to count. When you try to do chromatographic
integration without human intervention, you've got to go from highly tailed
peaks to very broad peaks. We don't want to have to put, for instance, slope
sensitivities in there. So we've got to have algorithms that take in all the things
that a normal chemist does, without having a normal chemist there and that is
a nontrivial task.
282
-------
AMBIENT AIR SAMPLING WITH A PORTABLE GAS CHROMATOGRAPH
Richard E. Berk 1ey
U. S. Environmental Protection Agency
Environmental Systems Monitoring Laboratory
Research Triangle Park, NC
ABSTRACT
The photoionization detector used in
Photovac gas chromatographs is sensitive
enough to detect certain kinds of hazard-
ous compounds in milliliter samples of
ambient air without preconcentration.
Properly operated, such instruments can
obtain data unspoiled by sampling errors
and virtually free of negative inter-
ferences, so that upper limits to true
concentrations can be estimated. Photovac
portable gas chromatographs were origin-
ally designed for industrial hygiene
analysis, typically at part per million
analyte levels. At part per billion
levels and below modification of the
instrument and operating procedures was
required. Rerouting carrier flow to con-
tinuously flush the sample loop and the
flow control valves reduced carry-over
contamination. Calibrant peak recognition
criteria were changed so the correct
calibrant peak would be recognized
reliably. A cons tant- temperature oven
stabilized retention times. Responses
were enhanced by chromatographing a larger
portion of the sample. Low-1 eve 1 calib-
ration mixtures stored in SUMMA-po1ished
cannisters reduced carry-over contam-
ination by calibrant and permitted higher
gain settings. Contamination was reduced
by flowing calibrant and sample streams
through the sample inlet for an extended
per i od of t ime.
INTRODUCTION
Estimation of toxic organic vapors in am-
bient air is usually performed in a lab-
oratory using samples which.have been col-
lected, transported, and queued pending
analysis. An instrument capable of pro-
ducing immediate results could avoid many
errors which arise from these processes.
Portable chromatographs, though less sel-
ective than laboratory instruments, can be
even more sensitive and can rapidly pro-
duce data of known quality. Photovac
portable chromatographs have a highly
sensitive photoionization detector. The
benzene detection limit has been shown to
be less than one tenth part per billion
(1,2). If operated properly they can be
used for rapid screening of haloalkenes
and substituted benzenes in air at typical
ambient levels.
EXPERIMENTAL
Initial work was done with a Model 10S50
gas chromatograph which could auto-
matically collect samples drawn through
the sample inlet system by a small dia-
phragm pump. Late-eluting compounds were
backflushed from ^ short precolumn after
the intended analytes had entered the main
column. Chromatographic operating para-
meters and solenoid valve timing were con-
trolled through a keypad by an on-board
microprocessor. Commercially available
fused-silica wall coated open tubular
columns (0.53 millimeter inside diameter)
were used. They were hand-wound on a
pasteboard spool iust smal1 enough to fit
the ambient-temperature column enclosure.
Carrier gas (ultrazero air) was supplied
either from an internal tank or an exter-
nal cylinder. Power was supplied by an
internal battery which could be recharged
when line power was available. An aux-
iliary external battery could also be
connected. The width of the peak-recog-
nition window could be set from the keypad
and was a fixed number of seconds for al 1
peaks in the chromatogram.
Calibration was done by analyzing known
samples. The identity and concentration
of each compound were entered into a cal
ibration library in nonvolatile memory.
Retention times of peaks on sample chrom-
atograms were compared with those in the
library If they fell within a preset
window the sample peak was "identified" as
the corresponding compound and quantitated
using the library response factor. Per-
iodic reca1ibration was done by analyzing
a single standard compound. Library con-
centration and retention time data for the
standard were corrected, and concentra-
283
-------
tions and retention times for all other
compounds in the library were corrected
proportionally by the microprocessor.
After evaluation, the Model 10S50 chrom-
atograph was upgraded to Model 10S70 and
equipped with a constant temperature col-
umn enclosure. Columns specially encap-
sulated in epoxy resin must be used in the
constant temperature oven. Column temp-
eratures of 20, 30, 40, or 50 C can be
set. The oven is powered by the external
auxiliary battery. The width of peak-re-
cognition windows is a percentage of re-
tention time rather than a fixed number of
seconds. The 10S70 is equipped with an
internal modem so that it can be control-
led by a remote computer. It can be op-
erated locally either through the micro-
processor keypad or through a microcom-
puter attached by RS-232 cable to the com-
munications port. Software supplied by
Photovac is used to control operating par-
ameters and store data on disk.
In field sampling the instrument was oper-
ated either as a. mobile or a stationary
sampler. Mobile operation was done in an
automobile using external auxiliary bat-
tery power and the internal carrier gas
tank. The chromatograph was placed on the
floor under the dash to shield it from
direct sunlight, and a sample probe (1/8
inch stainless steel tube) was extended
from the sample inlet through a window to
about one meter above the roof of the car.
All mobile sampling was done while the car
was parked. Except during early morning
or late evening, shaded locations were
used. Stationary operation was done
indoors using line power with an external
carrier gas tank. The sample inlet was
connected to a manifold into which exter-
ior air was being drawn by a pump or fan,
or the sample probe was extended outside
the building from the pump inlet.
RESULTS AND DISCUSSION
Attempts to operate the 10S50 unattended
at high gain were generally unsuccessful.
The microprocessor identified the calib-
rant peak by elution order, and the wrong
peak was usually chosen because many peaks
were seen at high gain, the number of them
being typically quite variable. Even-
tually it was discovered that increasing
calibrant flow time from one second to ten
seconds drastically reduced the number of
extraneous peaks and made unattended op-
eration possible, though it did not
eliminate mi sea 1 ibration entirely. Us-
ually about half the data were spoiled.
For this reason most work with the 10S50
was operator-attended mobile sampling. By
contrast, the 10S70 requires the calibrant
peak to meet both area and retention time
criteria, and misea 1ibration has not oc-
cur red .
Preliminary field sampling in the vicinity
of Research Triangle Park, NC using the
10S50 showed levels of benzene, toluene,
and tetrachloroethylene near the bottom of
the part per billion range or even lower.
Results of sampling under field conditions
were often spoiled by fluctuations in am-
bient temperature which caused peaks to
miss retention time windows. This rarely
happened to ear 1y-e1uting compounds, since
al1 windows had the same width regardless
of retention time. If the window was set
wide enough for late-eluting compounds to
be seen, then it was much too wide to dis-
criminate between ear 1 y-e1uting peaks.
Often several ear 1y-e1uting peaks were
identified as the same compound because
they al1 appeared in the same window.
Frequent reca1ibration was of limited use
in correcting library retention times.
Such corrections functioned well for
changes in flow rate and fairly well for
slight changes in column temperature, but
large temperature changes caused most com-
pounds to elute outside their retention
time windows.
A field trip to Houston, TX was made in
March, 1987. The unit was equipped with a
me thy 1si1 icone column operated in pre-
co1umn backflush mode. It was expected
that fenceline monitoring in Houston would
show substantially higher levels of ben-
zene and toluene than could be found in
Research Triangle Park, but levels of
these compounds in Houston were really not
much higher. Even close to obvious sour-
ces along the Ship Channel they were not
much higher then levels reported at TAMS
sites in nearby residential a.reas. A mix-
ture of vinylidene chloride, trichloro-
ethylene, and tetrachloroethylene in nit-
rogen, each at a concentration near 100
parts per billion, was used for field re-
calibration. Compounds which would have
eluted later than tetrachloroethylene were
backflushed. Significantly, the compounds
most often seen were the very ones present
in the reca1ibration standard. Typical
data are shown in TABLE 1. It appeared
that carry-over contamination of the sam-
ple inlet by calibrant might be occurring.
Use of 100 parts per billion of calibrant
constrained the gain setting to 100 or
lower, which was not quite high enough to
see ambient levels of most compounds.
Furthermore, the microprocessor reported
levels in units of parts per million only.
Concentrations below 0.5 part per billion
appeared as "0.000 parts per million".
Such results, though not quite untruthful,
are not quite useful.
In September, 1987 a field demonstration
study was conducted in Richmond and Hope-
weI 1, Virginia. Personnel and equipment
from EPA Region 3, Virginia Air Pollution
Control Board Region 5, and EPA/EMSL, Re-
search Triangle Park participated. During
this study the 10S50 was operated as a mo-
284
-------
bile sampler from an automobile. It was
equipped with a methylphenylsilicone
column with 1.5 micrometer phase thickness
which was operated without backflush. The
flow system was modified so that carrier
gas passed through the sample loop contin-
uously except during sampling, and the en-
tire volume of the sample loop was chrom-
atographed. Some deterioration of peak
symmetry and resolution resulted from this
arrangement, but it was offset by enhanced
sensitivity, enhanced baseline stability,
and reduced carryover contamination by
calibrant. In order to permit operation
at higher gain and minimize calibrant con-
tamination, a field recalibration standard
of 10 parts per billion tetrach1oroethy1
ene was prepared by diluting a 100 part
per billion commercial standard in a lec-
ture bottle with ultrazero air. This mix-
ture proved to be reasonably stable for
several days. It was discarded and rep-
laced after three days. Chlorobenzene,
ethy1 benzene, m-xylene, and o-xylene were
added to the calibration library list.
Standard concentrations were entered into
the microprocessor at one thousand times
their true values. This caused the micro-
processor to report concentrations in the
part per billion range as parts per mil-
lion and eliminated the annoying problem
of concentrations below 0.5 part per bil-
lion being reported as zero.
The 10S50 was operated at four sites in
Richmond and nineteen sites in and around
Hopewell. Some data early in the study
were apparently spoiled by contamination
of the inlet system from the interior of
the car The data in TABLE 2 show that
responses to most compounds were highest
in the first run at a site and trailed off
to a minimum value during the next several
runs. Increasing the sample pumping time
from ten seconds to forty-five seconds el
iminated this problem, as shown by the
data in TABLE 3 which were taken near a
large chemical plant in late evening. Ap-
parently a large release of benzene and
toluene occurred during the sampling per-
iod. Compounds eluting after tetrachloro-
ethylene were not seen because ambient
temperature was falling during the sam-
pling period, and these compounds eluted
outside their windows. Retention time
stability was a serious problem throughout
this study. Most sampling had to be done
in early morning or late evening because
daytime temperatures were much higher than
the temperature at which the library had
been created. Since optimum sampling con-
ditions were transitory, many peaks were
not recognized. It was clear that column
temperature stability would have to be
achieved before the instrument could
attain full potential.
A study of air sampling methods was con-
ducted during October. 1987 in Staten
Island, New York. Personnel from EPA Reg-
ion 2, the New York Department of Environ-
mental Conservation, the State of New Jer-
sey, and several local contractors partic-
ipated. During this study the 10S50 was
operated as a stationary sampler. It was
located inside a school building. Outdoor
air was drawn from a manifold. The site
was about 3 kilometers downwind of the New
York City Dump during the entire period of
the study. It was also about 7 to 20
kilometers downwind of numerous chemical
plants and oil refineries in New Jersey.
Use of 3 ten second calibrant flow min-
imized extraneous p'eaks and permitted one
all-night sampling period during which no
data were spoiled by misea 1 ibration.
These data are shown in TABLE 4 .
Compounds eluting after chlorobenzene were
not seen because of the retention time
stability problem. The temperature in the
room varied between about 70 80 F.
About half the data were lost because of
miscalibration during subsequent sampling.
After the Staten Island trip the 10S50 was
upgraded to Model 10S70 and equipped with
a constant temperature column oven. It
was used in field sampling in Richmond and
Hopewell, VA during June of 1988. The
recalibration standard was a 14.6 part per
billion dilution of chlorobenzene in a
SUMMA-po1ished cannister. Most data were
obtained while using the instrument as =>
stationary unattended monitor in a Vir-
ginia APCB station on Shirley Plantation.
This site was 3 to 4 kilometers north of
several chemical plants in Hopewell and
about 1 kilometer east of two or three
chemical plants at Bermuda Hundred. The
wind was generally from the southwest to
the west. Typical data are shown in
TABLE 5. No misca 1 ibrations occurred.
Many more compounds per run were seen with
t. fie constant temperature column oven in
use, especially late-eluting ones, but
early eluting compounds such as benzene
and trichloroethylene were seen rarely or
n u t at all. Certainly benzene was ac-
••ally present. Use of an earlier-eluting
calibrant or a peak recognition window
larger than + or 2 % of retention time
may solve this problem. Retention time
stability was much improved. The 10S70
reported data in parts per million or
parts per billion, as appropriate.
Standard deviations of all retention time
data obtained with the 10S50 during Sep-
tember, 1987 are compared in TABLE 6 with
standard deviations of all retention times
obtained with the 10S70 during June, 1988.
Using the 10S70, column temperature fluc-
tuated but carrier flow rates were not ad-
justed. Using the 10S70 column temper-
ature was constant at 40 C, but flow rates
were frequently adjusted in attempting to
shorten the initial stabilization period.
Nevertheless, standard deviations of re-
tention times were nearly an order of
magnitude lower.
285
-------
CONCLUSIONS
The Photovac portable chromatograph, if
properly used, can rapidly produce valid
estimates of upper concentration limits of
several classes of organic air pollutants.
The optional constant-temperature column
oven is practically a necessity for field
use under most conditions. Further work
should be done to determine the ambient
temperature limits for operation of this
unit and to fully assess the reliability
of battery-powered operation with the con-
stant temperature column oven under field
conditions. Different types of columns
should be tried so that other classes of
compounds, especially polar compounds, can
be ana 1yzed.
REFERENCES
1 . R. E. Berk 1ey,
132858.
EPA/600/4-86/041, PB87-
2. A. I. Clark, A. E. Mclntyre, J. N.
Lester, R. Perry. Intern. J.
Env iron. Ana 1. Chem. , 17. 315
(1984).
DISCLA1MER
The information contained in this article
does not necessarily reflect Environmental
Protection Agency policy.
Key words: Portable chromatographs,
photoionization detectors, air analysis.
TABLE 1. SAMPLING AIR IN DEER PARK, TEXAS WITH PHOTOVAC 10S50
3-26-87 J. P. Bonnette Junior High School, 5010 West Pasadena Boulevard.
Partly cloudy. Wind northeast light. 20 C. The Texas Air
Control Board operated a TAMS site at this location.
Time
1500
1513
1524
1537
1549
1602
1, 1-Dichloro-
Ethy1ene
Parts per billion
*
2. 0
*
<0. 5
ND
ND
by
Benzene
vo1ume
H
<0. 5
*
<0. 5
<0. 5
<0. 5
Tr i ch1 or o-
Ethy1ene
8
5
4
Calibration run.
286
-------
TABLE 2. SAMPLING AMBIENT AIR WITH PHOTOVAC 10S50 IN HOPEWELL, VA
9-15-87 403 Ramsey Avenue. In shade in front of house.
Wind light and variable. 30 C.
Time
939
959
1011
1031
1050
1110
Tri-
Ch 1 or o-
Benzene Ethene
Parts per b i 1 1 i on
6.
3.
2 .
1 .
64
46
07
09
1 . 38
0. 26
0. 21
ND
To 1 uene
by vo 1 ume
7.
6.
0
1 .
58
20
41
39
Tetra-
Chloro- Chloro-
Ethene Benzene
1
1
1
1
#
. 58
K
. 82
. 1 1
. 07
ND
0. 35
0. 32
0. 36
Ethy 1 -
Benzene
1 .
0.
0.
0.
23
31
15
13
m,p-
X y 1 e n e
2.
1 .
0.
0.
. 71
63
64
37
o -
X y 1 e n e
ND
ND
ND
ND
* This compound was calibrant in a calibration run.
ND Not detected.
TABLE 3. SAMPLING AMBIENT AIR WITH PHOTOVAC 10S50 IN RICHMOND, VA
9-24-87 South end of Commerce Street. About 500 meters downwind of a large
chemical plant. Wind very light W. 27-23 C.
Benzene
Time
1846
1907
1926
1933
1951
2013
2033
2049
Par t s per
0.
0.
7.
64.
72.
77
87
17
14
49
Tr i-
Ch 1 oro-
Ethy 1 ene
b i 1 1 i on
ND
ND
ND
6. 24
7. 10
To 1
by
2
3
12
47
46
uene
vo 1 ume
. 46
. 90
. 46
. 44
. 98
Tetra-
C h 1 o r o -
Ethy 1 ene
*
0. 36
0. 48
K
0. 19
*
ND
ND
Ch 1 oro-
Benzene
0. 26
0. 14
0. 12
ND
ND
Ethyl -
Benzene
ND
ND
ND
ND
ND
m , p-
Xy 1 ene
ND
ND
ND
ND
ND
o -
Xy 1 ene
ND
ND
ND
ND
ND
ND Not detected.
* This compound was calibrant in a calibration run.
287
-------
TABLE 4. SAMPLING AMBIENT AIR IN STATEN ISLAND, NY WITH PHOTOVAC 10S50
10-21-87 Room 457. Susan Wagner High School, 50 Br
Sampler was connected to a manifold which
Benzene
Time
0004
0019
0034
0049
0104
0119
0134
0149
0204
0219
0234
0249
0304
0319
0334
0349
0404
0419
0434
0449
0504
0519
0534
0549
0604
0619
0634
0649
0704
0719
0734
0749
0804
0819
0834
0849
0904
0919
0934
0949
1004
Parts per
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
1.
1.
1.
2.
1.
1.
1.
1.
2.
2.
1.
3.
2.
3.
3.
2.
3.
72
72
70
70
68
75
72
80
74
78
91
16
47
51
02
94
93
90
85
12
11
88
22
30
07
65
95
29
Tri-
Chloro-
Ethy 1 ene
To 1 uene
ie 1 1 e Avenue,
was importing
Tetra-
Ch 1 oro-
Ethy 1 ene
New York, NY.
outdoor air.
Ch 1 oro-
Benzene
billion by volume
ND
ND
ND
ND
ND
ND
ND
ND
0. 12
ND
ND
ND
ND
0. 07
0. 14
0.09
0. 09
0. 09
ND
ND
0. 08
0.03
ND
0.04
ND
ND
0. 10
0. 13
14.6
14.2
16.2
14.5
15.8
15. 3
15. 8
15. 5
18. 4
20.0
23.7
22. 8
28. 5
30. 4
31. 1
30. 1
26. 9
22. 4
22. 5
21. 1
20. 9
19.8
20.8
18.2
19. 2
12.6
21. 1
21.3
0. 07
ND
*
0.05
ND
X
0. 19
0. 17
*
0. 30
0. 34
X
0. 68
0.21
x
ND
ND
X
0. 24
0. 30
X
0. 39
0.07
X
0. 03
ND
X
ND
ND
X
ND
ND
X
0. 34
0.32
X
0. 39
ND
X
0.27
0. 29
ND
ND
0.04
0. 07
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND Not detected.
x This compound was calibrant in a calibration run.
288
-------
TABLE 5. SAMPLING AT SHIRLEY PLANTATION WITH PHOTOVAC 10S70
6-22-88 Sampler was located inside Virginia APCB Station 75-B. Outdoor air
was imported through the sample probe. Wind was 0-2 mph, SW to W.
Station interior temperature was 78 75 F. Outside temperature was
about 96 80 F. Plants at Bermuda Hundred were about 1.5 kilometer
west-southwest. Plants at Hopewell were 3 4 kilometer south.
Time
Tri-
Ch 1 oro-
Benzene Ethylene To 1
Parts per billion by
Tetra-
Chl oro-
uene Ethylene
vo 1 ume
Chi oro-
Benzene
m, p-
Xy 1 ene
o-
Xylene Styrene
1752
1812
1832
1852
1912
1932
1952
2012
2032
2052
2112
2132
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.
0.
0.
0.
0.
0.
0.
07
10
18
13
25
28
26
ND
ND
0.
0.
0.
0.
0.
0.
0.
0.
0.
14
16
23
15
21
27
24
22
16
*
1.08
0.79
0.68
*
0. 90
0. 70
0.67
X
0. 75
0. 77
0.62
ND
ND
ND
ND
ND
ND
ND
ND
ND
1.
23.
8.
2.
1.
24.
22.
ND
ND
74
1
18
94
78
0
5
1.04
1. 19
0. 89
ND
ND
ND
0. 81
ND
ND
ND Not detected.
* Calibrant in a calibration run.
TABLE 6. COMPARISON OF RETENTION TIME STABILITIES
USING AMBIENT-TEMPERATURE AND CONSTANT-TEMPERATURE COLUMN OVENS
To 1uene
Tetra-
Ch1oro-
Ethy1ene
Ch1 oro-
Benzene
o-
Xylene
Ambient Temperature Column using 10S50 (September, 1987)
Number of Samples 181 172
Maximum Retention Time 193.9
Minimum Retention Time (seconds) 108.5
Average Retention Time 139.1
Standard Deviation 16.4
172
278. 9
153. 3
201. 4
26. 1
112
382.2
211.8
280. 7
34.4
106
595.2
325.3
424.9
52.9
Controlled Temperature Column in 10S70 (June, 1988)
Number of Samples 81 47
Maximum Retention Time 229.2 257.0
Minimum Retention Time (seconds) 204.0 240.2
Average Retention Time 211.4 245.8
Standard Deviation 3.9 4.0
85
459. 2
423. 4
437. 8
5. 2
54
609.5
580. 3
599.8
5.8
289
-------
DISCUSSION
THOMAS SPITTLER: Looking at your data, I think they're extremely
consistent with any kind of air pollution data that are reliable and have been
published. We see a little benzene, toluene, occasionally a few chlorinated
solvents in the air, and not much else. Within the last seven or eight years, since
we've had the Photovacs® and done a lot of air studies, we still see almost
exactly the same pattern.
It leads me to believe that we're not going to get a great deal of further
information by doing air toxic studies widely and over a long period of time.
I think it's pointing us toward a lot more intensive study around specific point
sources. A lot of data showed that as you were seeing some anomalies, you
probably were looking at wind shifts blowing over specific sources, or specific
plants.
That's the value of having this kind of instrumentation. When you're out in the
field, when you see something, you can immediately move to it, you can track
it and you can identify the sources. When you don't see some specific source,
there is this very low background of a few parts per billion, something up in the
range of five to twenty of benzene, toluene, and maybe a part per billion or less
of chlorinated solvents.
RICHARD BERKLEY: I like being able to see the background when I'm out
in the field. It reassures me a little bit that things haven't gone haywire. If I'm
using an instrument that doesn't see anything below some fixed level above the
background, then I have no way to be sure that the instrument is still working
unless something suddenly shows up.
THOMAS SPITTLER: I found it particularly interesting that you had the
canisters that you could take back and run through the desorption, the mass
spectrometer, the chryofocusing and, and still see nothing significantly differ-
ent.
RICHARD BERKLEY: I would like to do a lot more of that kind of work. I
think more of it is called for to give us a greater volume of evidence. I would
like to see more work done with different kinds of methods. If we had two or
three methods agreeing with each other, it would be a tremendous burden of
proof to anybody that questioned the results.
THOMAS SPITTLER: We were doing work at some of the very notorious
waste sites up in New England, about five or six years ago. We came back
occasionally from the study having found higher levels upwind than we found
downwind of the site. When we had the time to go back and investigate, we'd
find that the trichloroethylene we were seeing higher in the upwind was simply
a matter of being closer to a dry cleaner in the neighborhood, some distance
away from the site. The sites themselves showed very, very low levels of
ambient toxics.
RICHARD BERKLEY: Yes, that is something that you can always run into
in a crowded area - the possibility of some other site upwind.
Another thing I' ve encountered several times has been upwind diffusion, where
the wind velocity was not uniform with altitude, and there would be still enough
along the ground for diffusion. We would detect an extremely high level of
something that should have been blowing away.
THOMAS SPITTLER: When you showed your slides of the house, I was
curious as to whether you had taken the opportunity to take a sample inside and
run it through GC. Did you do that?
RICHARD BERKLEY: No.
THOMAS SPITTLER: I came back from a hazardous waste site investiga-
tion, six or seven years ago, and had the Photovac with me overnight. I brought
it into my house and took some samples in the bathroom, living room, and
where we have our wood stove. I saw nothing within ten percent (10%) of the
levels inside my house at the hazardous waste site we had investigated the entire
day. We've been trying to find all these chemicals in the ambient air, where
people spend very little time, and avoiding concentrating on interior environ-
ments, which are sometimes highly toxic, or at least have higher levels of
chemicals in them.
We have diverted a lot of effort and attention into the ambient environment,
failing to realize that we 're living most of our lives in much more contaminated
inside environments. That's beginning to come out now with some emphasis
on indoor air monitoring. The portable gas chromatographs give us a tremen-
dous handle on that, because they can go anyplace.
290
-------
A PORTABLE SYSTEM UNDER DEVELOPMENT FOR THE DETECTION OF
HAZARDOUS MATERIALS IN WATER
John C. Schmidt, Philip G. Koga, and Garland C. Misener
Environmental Technologies Group, Incorporated
Baltimore, Maryland
ABSTRACT
The development of an automated biosensor
for use in the Monitor for Drinking Water
Supplies (MOWS) is described. The
biosensor is based on the inhibition of
acetylcholinesterase covalently immobil-
ized to a "clever" membrane and is
projected to be sensitive to selected
nerve agents and organophosphorus
pesticides at the low ppb to low ppm
levels. Projections for high reliability
and low consummable cost are primarily due
to the use of an improved fluidic scheme
and "clever" membranes.
The technologies underlying two potential
second generation chemical sensor modules
for the MOWS are also discussed.
INTRODUCTION
A single instrument does not currently
exist which is capable of rapidly
detecting a wide variety of pathogens,
biochemical toxins, and toxic chemicals in
potable water pipes and water reservoirs.
This paper will describe one part of the
development of such a monitor for all
three types of substances in drinking
water samples.
The Monitor For Drinking Water Supplies
(MOWS) is a modular instrument configured
to sample drinking water from pressurized
pipes, reservoirs and other water samples
in the field. The system design is
modular in order to facilitate the
incorporation of new sensor technologies
as they mature. The current design
includes five modules: a chassis module, a
sampling module and three sensor modules-
one each for chemical, toxin, and
biological threats. The modules are
enclosed within a chassis and the entire
system is controlled by a microprocessor.
This paper will describe the configuration
and preliminary performance data of the
first chemical sensor module, which is
currently under development at Environmen-
tal Technologies Group (ETG), Inc. Other
sensor technologies which may be incorpor-
ated into future chemical sensors will
also be described. The first module
detects very low levels of cholinesterase-
inhibiting pesticides and nerve agents,
and is based on the inhibition of acetyl-
cholinesterase covalently immobilized to
"clever" membranes. Since it is based on
the effect of the analyte on the
physiological target in the human body,
this sensor is subsequently referred to as
the biosensor. The use of "clever"
membranes and enhanced fluidics are
expected to result in producibility and
selectivity improvements relative to
previous enzyme-based systems. The other
chemical sensor modules being considered
for development are based on the electro-
chemical and ion mobility air monitoring
systems developed at ETG, Inc.
DESCRIPTION OF THE BIOSENSOR CHEMICAL
MODULE
The biosensor chemical module is the first
sensor module being developed for the
MOWS. It is designed to detect extremely
low levels of nerve agents, organophos-
phorus pesticides, and carbamates. The
biosensor module is based on the inhibi-
tion of the enzyme acetylcholinesterase,
the physiological target of the nerve
agents and these pesticides in the body
[1,2]. As shown in Figure 1, the sensor
consists of two syringe pumps, a sample
loop, a membrane containing immobilized
acetylcholinesterase, several Teflon
valves, and a flow-through pH sensor. All
of the fluid-wetted parts of the valves
and lines are Teflon or Delrin in order to
minimize adsorption of the analytes in the
sensor.
The operation of the module can be divided
into four phases and is summarized in
Figures 1-2. In the first phase, the
initialization phase, the buffered
acetylcholine substrate solution is passed
directly through the pH sensor and the pH
of the solution is measured. This value
291
-------
is referred to as baseline 1. While
baseline 1 is being determined, buffer is
also passing through the membrane
containing the immobilized enzyme (enzyme
membrane) to establish equilibrium.
During the second phase, the uninhibited
activity of the enzyme in the membrane
disk is measured. The buffered substrate
solution is routed through the enzyme
membrane and then to the pH sensor. Since
the enzyme hydrolyzes acetylcholine to
choline and acetic acid, the activity of
the enzyme membrane can be calculated from
the pH change of the substrate solution
before and after exposure to the the
enzyme membrane. The initial, uninhib-
ited, enzyme activity is actually the pH
value measured during the second phase
minus the baseline 1 value measured in the
fi rst phase .
Two tasks are accomplished during the
third phase. First, the baseline is
updated and, second, the enzyme membrane
is exposed to the sample. The buffered
substrate solution is once again passed
directly through the pH sensor to update
the baseline. This value is referred to
as baseline 2. The measurement of the
baseline during each odd cycle allows one
to compensate for both long-term drift of
the pH electrode and slow nonenzymatic
hydrolysis of the substrate. Simultan-
eously, the sample injection valve is
rotated such that the contents of the
sample loop are routed through the enzyme
membrane. Since the analytes of interest
inhibit the enzyme, their presence in the
sample can be determined in the next phase
when the final "exposed" or inhibited
enzyme activity is measured.
The inhibited enzyme activity is measured
during the fourth phase. As in phase 2,
this is accomplished by routing the
buffered substrate solution through the
enzyme membrane and then through the pH
sensor. The inhibited enzyme activity is
the value measured in phase 4 minus the
baseline 2 value. The percent inhibition
is calculated as follows:
Percent inhibition
100 (1 - AI/AU> where
I'
u
inhibited enzyme activity
uninhibited enzyme activity
Since the percent inhibition of the enzyme
is a function of the analyte concentra-
tion, a semiquantitative estimate of the
analyte concentration can be determined by
the algorithm in the instrument chassis.
In the absence of significant inhibition
of the enzyme, phases 3-4 are repeated
every six minutes. In the event that the
enzyme activity drops below a preset level
(either due to an exposure to analyte or a
slow natural activity loss over several
days), a new enzyme membrane is auto-
matically moved into place. The process
begins with phase one every time a new
enzyme disk is initially used.
Two aspects of the four-phase cycle are
worth noting. First, the selectivity of
the module is enhanced by the fluid path
during the third phase of the cycle.
Most previous sensors of this type routed
a mixture of the sample and substrate
solution through the enzyme and pH (or
voltammetric) module in a single step.
Therefore, they were vulnerable to false
alarms due to acid-base or redox
interferents in the sample. Since the
MOWS system described in this paper uses a
two step process in which the pH sensor is
never exposed to the sample, the
selectivity of the MOWS sensor is enhanced
considerably. Second, the sample
turnaround or cycle time of the system is
approximately 15 minutes. The cycle time
is the sum of the sample acquisition and
pretreatment time (in the sampling
module), the sensor response time, and the
algorithm response times.
Description of the Enzyme Membrane
The projected high producibility and low
production cost of the biosensor module
are primarily the result of the use of
"clever" membranes to immobilize the
acetylcholinesterase. The term "clever"
membrane refers to a relatively new type
of microporous membrane in which the
surface contains a large number of
aldehyde groups [3]. The aldehyde groups
on the surface of the membranes form
covalent bonds (Schiff bases) with amine
groups on proteins. One very attractive
feature of "clever" membranes is that many
proteins can be immobilized by simply
submerging the membranes in an aqueous
solution of the protein. This eliminates
the need for several relatively complex
and time-consuming steps normally required
during alternate covalent immobilization
schemes. Significant improvements in
reliability and decrease in consumable
production cost are projected.
The biosensor module is loaded with
several acetylcholinesterase membrane
disks which are stored dry [4] and
incorporated into a carousel. The device
is configured such that a fresh enzyme
disk can be rotated into place immediately
after any assay in which the previous disk
is depleted. The membrane material is
microporous polyvinylalcohol approximately
5 mils thick [5]. Each enzyme disk is
approximately 2 cm in diameter and
contains approximately 3 units of
acetylcholinesterase.
292
-------
Description of the Substrate
Acetylcholine chloride is used as the
substrate during the assay of the
uninhibited and inhibited activity of the
enzyme disk. The assay solution consists
of 50 millimolar acetylcholine chloride in
3 millimolar phosphate buffered saline
(PBS).
preliminary Performance of the Biosensor
Chemical Module
The preliminary sensitivity data for the
chemical biosensor module is summarized in
Table I. Figure 3 depicts the results of
sampling deionized water containing 7.0
ppb VX (0-ethyl S-(2-diisopropylamino-
ethyl) methylthiophosphonate), an
organophosphorus nerve agent. The
acetylcholinesterase activity was
inhibited approximately 28%. Since the
percent inhibition of the enzyme is a
function of the analyte concentration, a
microprocessor-based algorithm can be
trained to indicate a semiquantitative
(high, medium, low) concentration of the
analyte.
Table I. Preliminary Sensitivity Data For
Agent VX
Percent Enzyme Instrument Approximate VX
Inhibition Display Concentration
10-35%low 5-10 ppb
35-75% medium 11-99 ppb
75-100% high 100+ ppb
The sensitivity of the biosensor chemical
module will also be tested to the other
nerve agents and pesticides listed in
Table II. In general, the sensitivities
are expected to be a function of the
toxicity of the compound; the more toxic
the material, the lower the detection
limit of the sensor. The minimum
detectable level of most of the
cholinesterase-inhibiting analytes will
probably range from low ppb to low ppm
levels.
Table II. Pesticides and Other Nerve
Agents To Be Tested on the MOWS
Analyte
Agent GD
Malathion
Parathion
Paraoxon
Neostigmine
Physostigmine
The selectivity tests planned for the
chemical sensor are listed in Table III.
The percent inhibition of the enzyme due
to the interferent dissolved in deionized
water will be measured to quantitate any
false positive responses of the sensor.
False negative responses will be measured
by comparing the enzyme inhibition of 7
ppb VX to that of a solution of 7 ppb VX
and the interferent.
Data collected to date indicates the
sensor is capable of one week of contin-
uous unattended operation. Typical enzyme
disks maintain over half of their activity
after one week of continuous operation.
The substrate solution is also stable for
one week under normal conditions. Both
the reconstituted substrate solution and
enzyme disk carousel must be replaced on a
weekly interval during continuous oper-
ation. Both are stored dry, and have a
shelf life of approximately 2 years at
25°C.
DEVELOPMENT OF FUTURE CHEMICAL SENSOR
MODULES
The development of additional chemical
sensor modules is also under consideration
at ETC, Inc. Future chemical sensor
modules will likely be based on modifica-
tions to current ETC air contaminant
sensors. While these sensors were not
selected for the initial chemical sensor
because they are not ideal for the
detection of the nerve agents of primary
interest to the MOWS, they are capable of
detecting the diverse list of chemical
substances summarized in Table IV.
The sensor systems based on ion mobility
spectroscopy are described in other papers
presented during this symposium [6,7].
The electrochemical sensors produced at
ETC, Inc are amperometric sensors; an
exploded view of a typical amperometric
sensor can be found in Figure 4. A
typical amperometric sensor consists of
three electrodes in contact with an
electrolyte-saturated insulator. One of
the electrodes, the sensing electrode, is
separated from the sample by a microporous
Teflon membrane. The pores of the
membrane are small enough to prevent the
liquid electrolyte from escaping, but
large enough to permit ordinary diffusion
of the analyte from the environment to the
electrolyte. A permselective membrane on
the ambient side of the microporous
membrane provides an additional degree of
specificity. The sensing electrode is
biased at a potential sufficient to
oxidize or reduce the analyte of interest.
Electroactive compounds undergo redox
reactions directly on the sensing
electrode surface, resulting in a current
which is proportional to the analyte
concentration. Nonelectroactive
compounds, such as the organophosphorous
esters, are converted to electroactive
compounds by a reagent in the electrolyte.
Modulation of the sensing electrode
potential by several differential or
pulsed techniques has resulted in signif-
icant improvements in the sensitivity and
293
-------
specificity of this type of sensor in the
past few years. Typical modulated sensors
are capable of detecting as low as parts
per billion (ppb) levels of many organic
and inorganic air pollutants.
Amperometric gas sensors are now in
widespread use in ambient air monitoring
due to several inherent advantages of the
technology. First, the sensors are
inexpensive. All of the components shown
in Figure 4 can be injection molded from
inexpensive materials, stamped from inex-
pensive sheet stock, or otherwise
processed using very small amounts of
precious metals. Second, they are simple
and reliable. There are no moving parts
to fail, and the sensor output is usually
a linear function of concentration.
Third, they are portable. A self-
contained system containing a sensor,
associated electronics, battery, and
readout device can easily fit into a shirt
pocket.
The same advantages make amperometric
sensors attractive candidates for several
water monitoring applications. The cost,
reliability, and size advantages of
electrochemical air sensors are also
important for water contaminant sensor
applications. The configuration shown in
Figure 4 has two additional advantages for
water monitoring applications. First, the
microporous membrane adjacent to the
sensing electrode provides an air gap
which separates the water sample from the
electrolyte. This permits analytes with
relatively high vapor pressures to enter
the electrolyte and be detected. However,
large molecular weight materials such as
proteins which typically foul the
electrodes of conventional electrochemical
cells are prevented from entering the
amperometric sensor. Second, a
complicated sampling subsystem is not
necessary to separate the analyte from the
water sample as is the case with
ionization and many other sensor systems,
since water is a major component of the
electrolyte of the electrochemical sensor.
In fact, the aqueous sample simply flows
past the microporous membrane in the
simplest scheme under consideration.
Amperometric systems sensitive to cyanide,
organic arsenicals, and organic sulfides
have received the most attention to date.
Decisions regarding the development of
specific electrochemical sensor modules
are expected to be made after the initial
prototype MOWS system is tested in
September 1989.
CONCLUSION
The preliminary performance testing of the
MOWS chemical sensor module suggests that
"clever" membranes can be used to improve
the producibility and reduce the
production cost of analytical systems
based on immobilized enzymes. The minimum
detectable level of the MOWS chemical
module was approximately 5 ppb for the
most toxic oganophosphorus cholinesterase
inhibitor. A modification in the standard
flow scheme improved the selectivity of
the MDWS relative to most previous sensors
based on the cholinesterase inhibition
scheme.
When interfaced with the sampling module
and chassis module in the MDWS system, the
chemical sensor module will provide an
automatic instrument capable of continu-
ously monitoring low levels of cholines-
terase inhibitors for one week periods
without service. Several of the many
potential applications for the MDWS
include drinking water analysis, food
analysis, and screening of Superfund
sites.
Extensive expertise developed at ETC and
Allied-Signal Corporate Laboratories in
the development of trace level air
monitors is currently being applied to
develop MDWS sensor modules for
biochemical toxins and pathogens. When
the other sensor modules are completed,
the MDWS will be the first automatic
system capable of monitoring toxic
chemicals, biochemical toxins, and
pathogens simultaneously.
REFERENCES
1. Whittaker, Mary, "Cholinesterase,"
Monographs In Human Genetics, Vol. 11,
Karger, New York, New York, 1986, pp.
1-29.
2. Gray, Peter and Dawson, Raymond,
"Kinetic Constants for the Inhibition of
Eel and Rabbit Brain Acetylcholinesterase
by Some Organophosphates and Carbamates of
Military Significance," Toxic. Appl.
Pharmacol., Vol. 91, 1987, pp. 140-144.
3. Clever membranes are commercially
available from three vendors at the time
this article was written: MEMREE membranes
from Micro Membranes, Inc. in Newark, NJ;
Immobilon membranes from Millipore
Corporation in Bedford, MA; and Ultrabind
membranes from Gelman Sciences Inc. in Ann
Arbor, MI.
4. Goodson, Louis and Goodman, Alan,
"Stabilization of Cholinesterase, Detector
Kit Using Stabilized Cholinesterase, and
Methods of Making and Using the Same,"
U.S. Patent 4,324,858, 1982.
5. AM500 MEMREE membranes from Micro
Membranes, Inc (see ref 3 above) were used
in all experiments described in this
paper.
294
-------
6. Reategui, Julio et.al. "Ion Mobility
Spectrometry for the Identification and
Detection of Hazardous Chemicals," This
Symposium, Other Advanced Field Techniques
Session, 1988.
7. Reategui, Julio and Carrico, John, "A
Portable Ion Mobility Spectrometer for
Field Detection, Identification, and
Monitoring of Toxic Chemicals," This
Symposium, Poster Session Presentation,
1988.
SYRINGE PUMP •!
BUFFERED
SUBSTRATE
SYRINGE PUMP *2
BUFFER ONLY
OPERATIONAL SUMMARY
PHASE
1
2
3
4
FLUID PATH SAMPLE
VI V2 V3 V4 V5 LOOP
31 31 23 32 12 OUT
32 32 13 31 32 OUT
31 31 23 32 12 IN
32 32 13 31 32 OUT
/[
I
(V
SAMPLE
LOOP
SAMPLE
WASTE
WASTE
Figure 1. Schematic of the Biosensor
295
-------
AVja INHIBITE
ALARM WHEN
{• \ *
1 I'
/ *
s ^
r~~ j
/
\ x i
pH SENSOR
BASELINE
I
MEASURE
UNINHIBITED
ENZYME
ACTIVITY
TIME •
EXPOSE
ENZYME TO
SAMPLE
MEASURE
INHIBITED
ENZYME
ACTIVITY
Figure 2. Idealized Response of the Biosensor
z
u
o
Q.
UJ
Q
O
DC
H
O
ill
120
110
100
90
80
70
60
50
40
30
20
10
0
-10
-20
-30
-40
-50
UNINHIBITED ENZYME ACTIVITY = 112-(-13) = 125 mV
INHIBITED ENZYME ACTIVITY
->-= 82-(-8) = 90 mV
/ = 28% INHIBITION
20 30
TIME (MINUTES)
40
50
60
Figure 3. Response to 7 ppb VX
296
-------
ELECTROLYTE
RESERVOIR
C/R ELECTRODES
ELECTROLYTE
WICK
PERMSELECT1VE
MEMBRANE
Figure 4. Exploded View of a Typical
Amperometric Sensor
Table III. Summary of Interferent Test Plan
Interferent
Isopropanol
Unleaded gasoline
Trichloroethylene
Benzene
Chlorine
Zinc chloride
High [Mg+J
High [Ca+1
Tween 20
Concentration
Tested
1,000 ppra
500 ppm
500 ppm
500 ppm
1 ppm
1,000 ppm
2,400 ppm
4,000 ppm
1 pet
Interferent Tested For
False Negatives False Positives
tests planned for late 1988
Table IV. List of Potential Chemical Sensors Under Consideration
Sensor
Technology
Ion Mobility
Spectroscopy
Electrochem-
istry
Current EAS
Product Name
Advanced Chemical
Agent Detector/
Alarm (ACADA)
Chemical Agent
Monitor (CAM)
Bendix IMS
Individual Chem-
ical Agent
Detector (ICAD)
Mini Alert
Analytes Detectable in Air
Analyte Detection Limit
Organophosphorus mid ppt
Esters
Organic sulfides low ppm
and arsenicals
Selected explos- low ppb
ives
Many other organ- ppt-ppm
ic compounds
Same as ACADA
Same as ACADA
Organophosphorus mid ppb
esters
Organic sulfides low ppm
and arsenicals
Hydrazines mid ppb
Many other elec- ppm-
troactive gases percent
and vapors
297
-------
DISCUSSION
ED HEITHMAR: Why did you choose the syringe pumps for your FIA
system, instead of, say, a PD pump or something like that?
TAD BACON: That's a good question. It's still in the experimental stage, and
what I've described here is just the first stage of development for this
instrument. There is going to be a lot of miniaturization and expanding on this
technique to use the most appropriate methods.
ED HEITHMAR: There is no pressure sensitivity for your detector?
TAD BACON: No.
ED HEITHMAR: I assume that the enzyme regenerates itself after a period
of time, right?
TAD BACON: Actually, as long as it's not inhibited, and it's bound to the
membrane, then it will retain its activity for a number of days. In this particular
configuration, it will retain satisfactory activity for about three days.
Not shown in the schematic is a network of microprocessors and controllers,
which look at the information derived from the sensor. One of the things that
they look at is the response of the sensor at the baseline level, where the enzyme
is not inhibited, but when it's exposed to the acetylcholinide.
If that activity level falls below a certain point, a carousel will rotate a fresh disk
into place, with original activity. In this way, you can run the system for about
seven days. There would have to be about three or four disks.
ED HEITHMAR: Once it's been exposed to an agent, it's no longer usable.
TAD BACON: That's the other case. If inhibition occurs, the system would
signal an alert to shut down the water supply, or whatever system is being
monitored, and a new disk would be rotated in place.
298
-------
DESIGN AND PERFORMANCE OF A MOBILE MASS SPECTROMETER
DEVELOPED FOR ENVIRONMENTAL FIELD INVESTIGATIONS
Thomas M. Trainor, Ph. D. and Frank H. Laukien, Ph. D.
Bruker Instruments, Inc.
Manning Park
Billerica, Massachusetts 01821
ABSTRACT
Design goals set for a field mass spectrometer
must address several conflicting demands, includ-
ing instrument size, weight, power requirements,
and capabilities. Based on a quadrupole analyzer
and an electron-impact ionization source, the
Bruker MEM is a complete GC-MS designed specifi-
cally for field operation and deployment in a
4-wheel drive vehicle. Power requirements are
limited to 600 W, which is normally supplied by
rechargeable 24 V batteries. Mass spectrometric
capabilities include both full scan and selected
ion monitoring (SIM) data acquisition modes. A
variety of sampling accessories have been produced
for the MEM, including a unique direct air/surface
sampler, enabling direct analysis in under 15
seconds of water, soil, or air samples. In addi-
tion, a capillary GC equipped with a thermal
desorption oven is available, permitting analysis
of more complex mixtures. Applications to be
discussed will focus on actual performance of the
MEM in carrying out field screening of regulated
volatile and semi-volatile organics, and will
serve to critically evaluate the instrument's
analytical capabilities.
INTRODUCTION
The enormous demands currently placed on
organizations carrying out hazardous waste site
assessments and remedial action activities have
led many participants to consider direct field
monitoring to augment or replace their dependence
on traditional sampling/off-site analysis
protocols. This concept of real-time, on-site
analysis for environmental pollutants is becoming
an increasingly attractive alternative to the data
turnaround, sample degradation, and field mobi-
lization cost problems often encountered in rely-
ing completely on off-site laboratory analysis.
Standard organics analytical methods developed
over the last decade by the US EPA for
regulatory purposes under the major environmental
programs (CERCLA, SARA, RCRA, NPDES, etc.) all
embrace gas chromatography-mass spectrometry
(GC-MS) as the preferred approach, due primarily
to its demonstrated sensitivity and specificity
(1,2). However, the bulk of the field analytical
work conducted currently by the US EPA and private
contractors is based on non-specific GC detectors
(PID, BCD, FID, TCD, etc.) (3,4). Some attempts
have been made in terms of field deployment of
GC-MS instruments, originally designed for
laboratory use, on-site in climate-controlled,
stationary trailers which are often referred to as
"mobile laboratories" (5). For a variety of
reasons, this latter approach of introducing
full-laboratory capabilities in the harsh field
environment is often fraught with problems and is,
with today's excellent network of overnight
delivery services, of arguable benefit compared to
off-site analysis at full service laboratories.
The development of a truly mobile mass
spectrometer, that is, an instrument capable of
reliable operation under a wide range of tempera-
tures without auxiliary cooling, compressed gases,
or ac power, now permits realistic field GC-MS
activities. The Bruker Mobile Environmental
Monitor (MEM) was developed from scratch with the
following four major design goals in mind:
1. Resistance to extreme environmental
operational conditions, e. g. extremes in
terms of temperature, humidity, and
mechanical/electrical shock.
2. Simple, easy instrument operation and sample
preparation by individuals with a minimum of
GC-MS training and experience.
3. Produce unambiguous qualitative and
quantitative analytical results.
4. Provide constant monitoring of instrument
performance with appropriate error messages
reported automatically.
These design goals had to be met while simul-
taneously addressing several conflicting demands,
including instrument size, weight, power require-
ments, and analytical capabilities. The final
commercial civilian mobile mass spectrometer, the
MEM, has evolved from a related product, the MM-1,
developed earlier for the military market. The
MM-1 is now a key chemical sensor of the West
German Army for its land based reconnaissance
vehicles and is currently undergoing extensive
evaluation by the US military. The MM-1 is
deployed to detect, identify, and quantify all
chemical warfare agents in air or on surfaces, in
real-time under battlefield conditions.
The purpose of this presentation is to evaluate in
detail the unique hardware and software features
incorporated into the MEM that permit routine
trace level GC-MS analysis in the field. Results
in terms of instrument sensitivity, linearity,
accuracy, reproducibility, and stability will be
presented in the context of performing field
monitoring methods.
DESCRIPTION OF THE BRUKER MEM
General
A diagram displaying the major components of the
MEM is depicted in Figure 1. The detection unit,
or actual mass spectrometer, and the associated
electronics modules are housed together in a unit
299
-------
weighing 300 Ib. with dimensions of 30 x 38 x 32
in. (1 x w x h) . A separate control console,
connected to the MEM with a cable of varying
length, has dimensions of 21 x 17 x 13 in. The
complete system has been installed in a variety of
vehicles, with the data for this presentation
generated by a unit mounted in a Chevy Blazer
4-wheel drive truck.
The ion source is a conventional electron-impact
(El) source. A pair of filaments is included so
that operation can continue uninterrupted for
extended periods. The analyzer consists of a
specially designed one-piece glass quadrupole with
a hyperbolic metal coated surface. A mass range of
1-400 amu is provided, with a mass resolution of
one amu throughout the mass range (10% valley
definition). This single piece design has proven
to provide exceptional rigidity and stability so
as to permit the operation of the mass
spectrometer under rather severe mechanical shock
(6 g) conditions. Ion detection is carried out by
a conventional 17-stage off-axis Cu-Be discrete
dynode electron multiplier, and is coupled to a
self-scaling integrating amplifier. This ion
detection system has been shown to be capable of
providing a dynamic range of over 10°, with an
accuracy of 15% for measurements conducted over
five orders of magnitude.
By utilizing El ionization and a quadrupole
analyzer, identical to what is now used by
laboratories carrying out priority pollutant
analyses, the resulting mass spectra are com-
parable to those found in standard EI-MS library
compilations. The upper mass limit of 400 amu is a
consequence of power conservation requirements,
and not analyzer performance. In practice, for
environmental investigations this mass range is
quite adequate, since for instance none of the 146
organic compounds (including internal standards
and recovery surrogate compounds) comprising the
Hazardous Substance List (HSL) now required for
analysis under the US EPA Superfund Contract
Laboratory Analysis Program (CLP) (6) contain a
characteristic ion (primary or secondary) outside
the MEM mass range.
The high-vacuum region of the instrument is
continuously maintained at a vacuum of 10~6 torr
by means of a 80 I/sec ion getter pump. A pump of
this design, in contrast to the more common
turbomolecular or oil diffusion pumps, has no
moving parts and requires no external cooling.
These features, coupled with its maintenance free
operation, make it ideal for mobile deployment
(7). The mass spectrometer is designed to operate
without the need for an auxiliary roughing
mechanical pump, common to most laboratory MS
systems. The MEM maintains high vacuum by
automatically closing the hydraulic main valve
whenever power is removed or a major fault is
detected. At the completion of manufacturing ?nd
testing, the MEM is initially rough-pumped at the
factory with a standard mechanical pump, prior to
engaging the ion getter high vacuum pump. Once
this is accomplished, the high vacuum state should
remain intact barring a need to access the
detector region. In practice, it is not unusual
for a unit to be under daily operation in excess
of a year between rough pumps. Test units left
idle (power removed) have held vacuum for over 6
months.
A separate housing contains the complete set of
mil-spec circuit boards responsible for control-
ling the MEM. Operation of the MEM is conducted
through a separate control console containing a
keyboard, printer, video monitor, and push-button
controls. Two microprocessors provide digital
control of all mass spectrometric functions and
temperatures. A continuously operational supervi-
sion program (SELF-MONITOR) can detect more than
80 instrumental errors or malfunctions. All the
errors are counted and logged, and serious errors
are immediately displayed to the operator. Also, a
check program for all the essential operating
functions can be manually initiated. Moreover,
built in diagnostic programs allow for the
measurement of 30 analog values (temperatures,
voltages, currents) and over 100 digital switching
states allowing for detailed trouble-shooting in
the field.
Inlet System
The gas inlet system, Figure 2, controls the
direction and nature of the gas flow to the inlet
membrane interface. The inlet membrane, housed in
the heated (150 C) main valve, serves to isolate
the high vacuum region of the mass spectrometer
from the sample gas stream. Organics pass selec-
tively through this membrane into the ion source,
preferentially excluding the lighter components of
air. Normally whenever the system is on, the
carrier gas pump is on, allowing for sample flow
through the sampler line (or optional GC unit) and
past the inlet membrane surface. Alternatively, if
the sample line gets contaminated the entire gas
inlet manifold may be backflushed by turning off
the carrier gas pump and switching on the back-
flush pump. Provision has also been made in this
inlet system for the automatic delivery of a mass
calibrant compound for tuning purposes. The
operator can easily change the inlet membrane,
air/surface sampler, and capillary GC unit within
a few minutes in the field, without breaking high
vacuum.
Sampling Systems
The direct air/surface sampler consists of a
flexible, heated tube 10 feet long with a diameter
of 2 inches. The sampler head is comprised of a
silicon membrane mounted on a nickel screen.
Directly behind the membrane lies the actual
sample transfer line, a 3.5m fused-silica SE-54
bonded capillary GC column of 0.32 mm internal
diameter. The sampler is mated to the gas inlet
manifold by means of a threaded quick release
coupling, with the end of the GC column located
close to the inlet membrane. Both the head and
transfer line can be separately temperature
programmed from ambient to 260 C, and ambient to
240 C, respectively. At typical operating tempera-
tures the air flow through the sampler line is
about 2.0 ml/min.
This air/surface sampler has been found to allow
for the direct sampling of ambient air, water, and
soil for a wide range of organic compounds. For
instance, sampling can be conducted by pressing
the head directly against soil, held over a
monitoring well or sample containing jar, or
alternatively, pressed against a clean surface
(glass, aluminum foil, Teflon, etc.) to which an
aliquot of a solvent extract of a sample has been
deposited. It is this flexibility of alternate
sample introduction avenues that has permitted the
MEM to solve quickly many problems related to
field based sample preparation requirements.
The capillary GC oven accessory is a recent
addition to the MEM and was developed to provide
extended analytical capabilities as compared to
the rapid air/surface sampler. The GC oven ^is
mated to the mass spectrometer using the coupling
shared by the air/surface sampler, with exchange
of these units normally carried out in a few
minutes. The GC unit, Figure 3, has been optimized
in terms of size, weight, and power requirements,
and is controlled and powered by the MEM
electronics. Among the unique features of this GC
is an injection port designed for automatic
thermal desorption of sorbent filled sampling
tubes. Glass tubes typically filled with Tenax or
Tenax/charcoal, are utilized to trap volatile
organics directly from ambient air or, alter-
natively, from water or soil samples through a
purge and trap step. By placing these sample tubes
300
-------
in the heated injection port, a direct transfer of
the desorbed analytes can be made to the head of
the capillary column. The subsequent GC-MS run may
then be carried out under an isothermal or tem-
perature ramp program, followed by a cooling phase
in anticipation of the next analysis. During the
actual GC-MS acquisition, the sorbent tube is
automatically cleaned via a backflush of cleaned
air. Additional options available for this GC
unit include a direct (on-column) injection port
and an auxiliary nitrogen carrier gas delivery
system. Surprisingly, very good results have been
obtained with just clean air as GC carrier gas
including hundreds of injections onto a DB-624
bonded column over several months of use.
MEM Data System
A key feature of this field instrument is the
unique approach taken for automatic target
compound monitoring, which permits continuous
sampling with a minimum of operator intervention.
A typical screen display during one of the main
acquisition programs , AIR MONITOR, is shown in
Figure 4. In AIR MONITOR, up to 22 compounds may
be simultaneously monitored, through selected ion
monitoring (SIM) acquisition scans of 1 to 4 ions
per compound. When the instrument observes the
presence of the target ions in the correct ion
abundance ratios, the screen will automatically
display the name of the compound observed. Along
the x-axis of the screen reproduced in Figure 4
are the ion groups for the various target com-
pounds selected for analysis (compounds A - J) .
Ion current is plotted along the y-axis, on a
logarithmic basis, necessary for the enormous
dynamic range encompassed in potential environmen-
tal analyses. At any point in time, a complete
mass spectrum can be obtained over a prescribed
mass range by pressing a key labeled SPECTRA on
the control unit. A more detailed discussion of
the AIR MONITORING identification and quantitation
software can be found in a companion paper (8).
External PC Compatible Data System
Additional data acquisition, storage, and
manipulation capabilities have recently been added
to the MEM through interfacing a PC compatible
microcomputer. As portability/ruggedness was also
a criteria for this hardware addition, the Compaq
Portable III system has been selected, equipped
with 1 Mbyte RAM and a 40 Mbyte hard drive. The
external data system is particularly useful for
acquiring full-scan capillary GC-MS runs, as a
data storage problem for this type of application
exists with the limited memory capabilities of the
internal MEM data system. The development of
automated quantitation software, a mainstay of
todays environmental GC-MS laboratory, is under
development for implementation on the Compaq/MEM
system.
APPLICATIONS TO HAZARDOUS WASTE SITE
INVESTIGATIONS
Direct Air/Surface Sampler
The determination of the nature and extent of
contamination by petroleum products and wastes
represents a significant per-centage of the total
environmental pollution problems facing the
industry. We have found the air/surface sampler to
be a quick, effective means of screening large
numbers of soil and water samples for relative
hydrocarbon levels, with no sample preparation
required. For instance, tracking the impact of a
leaking underground gasoline tank was possible by
programming the MEM to continuously monitor for
the target compounds benzene, toluene, total
xylenes, and in addition, total aliphatic
hydrocarbons. Since a combination of shallow soil
gas wells, soil borings, and groundwater monitor-
ing wells were featured in the sampling program,
results from a large number of measurements had to
be provided in real-time. Figure 4 shows a typical
screen display, updated every 5 seconds, for
gasoline components from a water headspace sample.
Others have reported an additional advantage of
on-site analysis of volatiles with this sampler,
namely that the field results are invariably
higher in concentration to that reported by the
off-site analytical laboratories, demonstrating
rapid sample degradation to be occurring (9).
This sampler has also been found to be quite
effective in the detection and identification of
commercial polychlorinated biphenyl mixtures
(Aroclors) in soil. For these semi-volatile
components, the SURFACE-MONITOR MEM program was
employed, which permits a discrete "injection" to
be made by placing the heated sampler head (260
C) directly into the suspect soil for 30 sec,
while keeping the SE-54 capillary transfer line at
a relatively low temperature (150 C). After this
initial concentration phase, the GC column is
quickly ramped to 220 C and data acquisition is
carried out for a period of 200 sec. Under these
conditions the PCB isomers are introduced into the
MEM and monitored by ions characteristic of each
individual PCB chlorination level. Figure 5 is an
example of the MEM results from analysis of a
soil sample certified to contain 35 ppm of the
mixture Aroclor 1242 (ERA Inc., Arvada, CO). The
relative concentrations of the different PCB
chlorination levels is a characteristic of each
commercial Aroclor mixture (10) and can be an
effective means of identifying the actual Aroclor
product observed. Since no time-consuming and
expensive solvent extraction steps are required to
obtain these results in the field, this appears to
be an ideal approach to screening large numbers of
soil samples on a timely and cost-effective basis
where the data is needed most with the sampling
crew. Other classes of semi-volatile compounds, in
particular several common pesticides, polyaromatic
hydrocarbons (PAH), and polychlorinated diben-
zodioxins (11) have also proven to be amenable to
this approach.
Applications dictating the need for the GC oven
accessory are those instances where the complexity
of the matrix necessitates a more effective GC
separation (30 m versus 3-. 5 m column, selection of
unique bonded phase to affect separation, etc.)
and/or the attainment of lower detection levels
through preconcentration steps. For example, we
have found the MEM reliable for the routine
screening of water for common volatile organics at
the 50 to 100 ppb level when using the direct
air/surface sampler. In a recent application,
quantitation approaching the sub ppb level was
requested for the aromatics benzene, toluene,
ethyl benzene, and xylenes (BTEX) in surface and
groundwater. This was accomplished with field
purge and trap concentration on a 100 ml water
sample, using a commercially available purge
vessel (Alltech Assoc., Deerfield, IL) , Tenax
sorbent tubes, and ambient air as purge gas.
Figure 6 shows the total ion current chromatogram
resulting from a sample run with the target
compounds spiked at 5.0 ppb and separated on a 30
m DB-624 column (JSW Scientific). For this
program, rigorous instrument calibration was
carried out using an internal standard spiked at
50 ppb into all blanks, calibration, QC, and
actual field samples. A typical calibration curve
obtained by this method is presented in Figure 7.
Using a similar thermal desorption GC-MS approach,
volatile organics trapped from soil and ambient
air (Tedlar bags) have also been monitored in the
field by the MEM. On occasion it has also proven
useful to analyze sample solvent extracts of
semi-volatile mixtures on the GC, by injection via
an empty glass tube in the thermal-desorption
injection port, or alternatively, using the direct
on-column injection port.
301
-------
CONCLUSION
An instrument capable of providing unambiguous
GC-MS data in the field, without burdening field
sampling crews with providing complex support and
sample preparation facilities, has been shown to
be an effective tool for the characterization of
hazardous waste sites.
ACKNOWLEDGMENTS
The authors would like to thank Dr. Jochen
Franzen and Dr. Alex Loudon, of Bruker-Franzen
Analytik GmbH, for their assistance.
REFERENCES
1. Budde, W. L.; Eichelberger, J. W. ; "Organics
in the Environment", Analytical Chemistry.
vol. 51, no. 6, May 1979, 567A-574A.
2. Keith, L. H.; Telliard, W. A.; "Priority
Pollutants - A Perspective View", Environmen-
tal Science and Technology. vol. 13, no. 4,
April 1979, 416-423.
3. Fisk, J. F.; "Candidate Field Methods for
Organics Analysis for Superfund", presented
at the 1988 Pittsburgh Conference & Exposi-
tion on Analytical Chemistry and Applied
Spectroscopy; New Orleans, February 22-26,
1988.
4. Chapman, G. H.; Fredericks, S.; "The US EPA
Field Analytical Screening Project";
presented at the 5th National Conference
on Hazardous Wastes and Hazardous Materials,
Las Vegas, Nevada, April 19-21, 1988.
5. Chapman, G. H.; Clay, P.; Bradley, C.K. ;
Fredericks, S.; "Field Methods and Mobile
Laboratory Scenarios for Screening
and Analysis at Hazardous Waste Sites";
presented at the Superfund '87 National
Conference, Washington, DC, November
16-18, 1987.
6. US EPA Contract Laboratory Program,
"Statement of Work for Organics Analysis",
October, 1986.
7. Giorgi, T. A.; Ferrario, B.; Storey, B.;
"An Updated Review of Getters and
Gettering",; J. Vac. Sci. Technol., 1985,
vol. 3, 417-423.
8. Laukien, F. H.; Trainor, T. M.; "Unambiguous
Identification and Rapid Quantitation in
Field Air Monitoring Using a Fully-Mobile
Mass Spectrometer", presented at the First
International Symposium on Field Screening
Methods For Hazardous Waste Site
Investigations, Las Vegas, NV, October
11-13, 1988.
9. Dickinson, R. K. ; Hadka, M. C.; "Site
Assessment/Remediation Using a Mobile Mass
Spectrometer", AEG Newsletter, April,
1987, 20.
10. Slivon, L. E-; Shumacher, P. M.; Alford-
Stevens, A.; "Determination of Aroclor
Composition from GC/MS Level Of Chlorination
Results", presented at the US EPA Symposium
on Waste Testing and Quality Assurance,
Washington, DC, July 11-15, 1988.
11. Matz, G.; Odernheimer, B.; "Fast, Selective
Detection of TCDD Using the Mobile Mass
Spectrometer MM 1", Chemosphere.
Great Britain, 15, 2031-2034, 1986.
302
-------
co
o
co
FIGURE 1, MAJOR COMPONENTS OF THE BRUKER MOBILE ENVIRONMENTAL
MONITOR.
-------
MEM Gas Manifold
Air
Sampler 1
Membrane
,
Compounds'
Vapor
Sampler Coupling
u~r~L JH
r*
n
Calibration Gas Valve
VI / \V2
Carrier Gas Line
Flow
Meter
Carrier Gas
Pump
Exit
Filter
Calibration
Gas
Reservoir
FIGURE 2. MEM GAS INLET SYSTEM.
-------
MEM Gas Chromatograph
Oven Lid GC-Capillary Column Thermal Desorption
To The Mass Spectrometer
,0ven x Sample Tube
Heater
FIGURE 3. MEM CAPILLARY GC OVEN ACCESSORY,
Charcoal Filter
Backf lush and
Cleaned Air Pump
305
-------
AIR MONITOR
T VOC SCREEN
00:44
8-
7-
6-
5-
4-
3-
2-
H XYLENES C 5.7
C TOLUENE C 5.5
J HYDROCARBON C 5.1 100
i i i
ABCDEFGHIJKL
FIGURE 4, EXAMPLE OF THE MEM AIR MONITOR REAL-TIME
VIDEO DISPLAY.
306
-------
SURFACE MONITOR
N PCB/DIOXINS/CL
01=22
8-
7-
6-
5
4-
3-
2-
I HYDROCARBONS H4.1
B CL2-BIPHENYL H 3.9
C CL3-BIPHENYL H 4.0 165
D CL4-BIPHENYL H 3.7
E CL5-BIPHENYL H 2.6
1 'i i i i i i i i i i i i '
ABCDE FGH IJ KL
6
FIGURE 5. DETECTION OF PCB (AROCLOR 1242) IN SOIL BY
THE MEM SURFACE MONITOR PROGRAM,
307
-------
FIELD PURGE AND TRAP
5.0 ppb level
15455-j
13900-
12356-
10811 -
9267-
7722-
6178-
4633-
3089-
1544-
0
I
f
j
._. ^J\. lj
f. Benzene
2. 1,4-Difluorobenzene (Int.Std.,
3. Toluene
4. Ethyl Benzene
5. m, p -Xy/enes
6. o-Xy/ene
I
•
3
i
|
1. !
4
1
5
6
1
0.0 2.0 4.0 6.0 8.0 10.0 12.0
TIME, minutes
FIGURE 6. TOTAL ION CURRENT PROFILE FROM THE MEM ANALYSIS OF
WATER SPIKED AT THE 5.0 PPB LEVEL OF BENZENE,
TOLUENE, ETHYL BENZENE, AND XYLENES.
308
-------
7.00
0.00
TOLUENE CALIBRATION CURVE-WATER P/T
(IS = p-DIFLUOROBENZENE)
0.00 0.50 1.00 1.50 2.00 2.50
CONC. TOLUENE /CONC. IS
3.00
3.50
FIGURE 7. CALIBRATION CURVE FOR TOLUENE ANALYZED BY PURGE
AND TRAP/THERMAL DESORPTION GC-MS ON THE MEM.
309
-------
DISCUSSION
JONATHAN NYQUIST: Did you say you're using charcoal as a prefilter in
the purge and trap for the air samples?
THOMAS TRAINOR: That's one filter you can use.
JONATHAN NYQUIST: Doesn't that take out some of the organics right
there on the charcoal, before it ever gets to your instrument?
THOMAS TRAINOR: With the field purge and trap, you do not want to
introduce volatiles from the ambient air into the water or soil sample, contained
in the purging vessel. In the lab, you would use nitrogen or helium as a purge
gas. We have eliminated that and used air, but you can't assume the air that's
surrounding your waste site is clean. So you're prefiltering it.
If you're talking about the flexible membrane liner situation, that system will
identify the potential problems with your data. Where it differs from the
experts, it will provide an explanation. Say the liner is not resistant, and provide
the reasons if it generates an answer. Then it's up to the reviewer to bring these
problems up with an expert.
JONATHAN NYQUIST: You mentioned in your criteria of expert systems,
about meeting a consensus among the experts. At hazardous waste sites, we
rarely have that, particularly because site-specific issues often take control. Do
you think we'll be able to use them at different hazardous waste sites?
DANIEL GREATHOUSE: I think you'll be able to use them for specific
issues at hazardous waste sites. I don't think you'll ever be able to use one that
will just take the place of a Remedial Project Manager at a hazardous waste site
or a contractor at a hazardous waste site.
For example, we have one expert systems program at hazardous waste sites to
help screen technologies for cleaning up sites. It's simply a screen. All it does
is select from 35 technologies the ones that are most feasible to use for site clean
up, perhaps ten or six. It then focuses the attention of the reviewer on those, and
it can get more information about any of those appropriate.
I don't think they will ever get to the point of replacing an expert. There are too
many unique situations that you just cannot address with an expert system.
From our experience, the primary application of expert systems are for the more
routine problems. These are 80 out of 100 problems that do not have all these
unique characteristics.
The unique problems, which require special considerations extrapolating from
other fields, or whatever, are ones that will never be addressed by expert
systems.
310
-------
EXPERT SYSTEMS TO ASSIST IN EVALUATION OF MEASUREMENT DATA
Daniel G. Greathouse
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
26 W. Martin Luther King Drive
Cincinnati, Ohio 45268
INTRODUCTION
The Agency expends a significant proportion of its
budget to measure levels of contaminants or effects
of contamination in the environment. In addition
it mandates that other organizations, primarily
private companies, expend considerable resources
to provide measurement data to the Agency. This
requires that decisions be made concerning appro-
priate sampling methods and analytical techniques
and results in huge volumes of measurement data
that must be evaluated and interpreted by Agency
personnel. Concerns such as extent of contamina-
tion, potential health risks due to the contamina-
tion, and likelihood of adverse health consequences
by introduction of new chemicals into the market
place are just a few of the decision areas based
on these data. Clearly the potential health
consequences and economic implications of these
evaluations and interpretations are very signifi-
cant. Hence it is very important that the best
expertise be brought to bear on these decisions.
The purpose of this paper is to review the devel-
opment of expert systems by U.S. EPA to assist in
evaluation of measurement data.
The expertise required to adequately perform the
necessary evaluations and interpretations of mea-
surement data depends in part on such things as
the media that are measured, the matrix of com-
pounds in the media, and the nature of the mea-
surements. Typically, satisfactory evaluation
and interpretation will require knowledge concern-
ing the analytical methods, sampling procedures,
underlying chemistry and/or biology, effects of
the matrix of compounds in the samples, etc.
This knowledge will be gained partly from formal
education, but much will be based on extensive
relevant experience.
Scarcity of qualified personnel to design measure-
ment studies and to evaluate and interpret the
resulting data is typical in the Agency. This
problem is accentuated since these decisions have
largely been delegated to the EPA Regional Offices
and states. Expert systems are tools that have the
potential to make available the expertise required
to make these decisions. These systems will make
it possible to multiply the talent of experts.
An expert system is an automated process which
incorporates the judgement, experience, rules of
thumb, and intuition used by a human specialist
to emulate that specialist's problem solving
ability. Generally, the knowledge of specialists
is stored in a computer in the form of facts and
decision rules although many, more complex repre-
sentations are available. Expert systems have
several characteristics which differentiate them
from conventional software. Whereas conventional
computer programs rely on numeric algorithms as
the basis for their operation, expert systems are
characterized by an emphasis on symbolic process-
ing, logical inferencing, and pattern matching.
Most expert systems contain facilities which
allow the user to obtain an explanation concern-
ing why a question was asked and/or to obtain
clarification for a particular question if the
user requires more information. In addition, the
user is often supplied with the reasoning employed
in reaching a conclusion. Expert systems are
interactive; they do not run and calculate so
much as they aid thought and offer advice.
A recently increased interest in the development
of expert systems has brought to market software
tools known as expert system shells. Basically,
a shell simplifies the development of expert
systems by providing a programming environment
with a built in library of functions for common
tasks such as rule generation, definition of
objects, and interfacing with the user. This
means that the builder of the system can concen-
trate on the acquisition of knowledge rather than
on low level programming issues. The arrival of
shells in the software marketplace is in part
responsible for the success of many expert system
projects.
There are a myriad of shells available to today's
developers. Each shell has a different set of
features and capabilities. For instance, some
shells allow only text to be used for user
prompts and output while other shells are more
versatile. Some provide easy access to external
data bases. Others provide the capability to
attach the computer to certain measurement devices
with readings displayed and control functions
available graphically on the computer monitor.
311
-------
Of course, the cost of a shell is usually indic-
ative of the breadth of features although competi-
tion in this market is increasing. For some
projects a low cost/capability shell will suffice
while in other environments advanced features are
well worth their price. In general, shell selec-
tion is an important consideration which must be
made with respect to the problem domain and the
desired system functionality.
REVIEW OF EXISTING SYSTEMS
Some of the systems being developed to provide
assistance in the measurement area will be summa-
rized. For more information concerning the cited
expert systems see the list of contacts provided
in the appendi x.
Flexible Membrane Liner System (FLEX)
Flexible membrane liners are plastic sheet
materials that are used as moisture/liquid barriers
in a landfill. They are placed under wastes to
prevent leachate from leaking into the groundwater
and on top of the wastes to keep water from enter-
ing the wastes. Typical plastic formulations for
these liners are polyvinyl chloride (PVC), high
density polyethylene (HOPE), and chlorosulphonated
polyethylene (CSPE). Before one of these materials
can be used as a liner for a hazardous waste land
disposal site, the Agency requires that it be
tested for chemical resistance to the anticipated
leachate from the landfill. A series of physical
property measurements must be performed on the
liner material after exposure to the anticipated
leachate for prescribed periods of time. The
test protocol is known as the Method 9090 test.
An expert system, known as FLEX, has been developed
to assist in evaluating and interpreting these
physical property measurements. The rules are
based on the input of 6 flexible membrane liner
specialists. In essence the system determines if
the data is adequate to fit a quadratic model and
if so uses the model to determine if the predicted
trend (change in the physical property measure-
ments) over time exceeds the limits prescribed by
the experts. Standard statistical methods are
used to examine the data and to fit the quadratic
model. The system is written in Arity Prolog for
application on an IBM PC/AT class micro computer.
Currently the system is being field tested.
Geophysics Expert System
This system is being designed to aid Superfund
managers in selection of appropriate geophysical
measurement methods. Different measurement sce-
narios are recommended based on the user identified
purpose of the measurements. For example the sys-
tem proposes a measurement scheme for situations
when the goal is to assess extent of contamination
as part of a site assessment. In addition to
identifying the purpose of the measurements the
user is also requested to supply general informa-
tion concerning the site. The first version of
the system is currently being reviewed and develop-
ment of the second is scheduled for completion by
early 1989. A follow-up will be initiated when
funds become available. Each successive version
is being designed to address a wider array of
decision scenarios. The system is being developed
in Basic for application on an IBM compatible
micro computer.
Quality Assurance/Quality Control
Systems are being developed to improve the quality
of measurement data produced by laboratories and
to document this improvement. A system is being
developed by EPA Headquarters to aid in implemen-
tation of the EPA developed data quality objective
procedures/standards when designing a laboratory
measurement program. More specifically it will
provide three levels of assistance, namely (1)
define the decision, (2) specify the information
and confidence levels needed to make the decision,
and (3) design a sampling plan to gather the
information.
Radian Corporation has developed two prototype
systems to provide sampling assistance. The
smaller one aids in selection of the appropriate
analytical method that will yield the desired
level of accuracy for the compounds to be mea-
sured. If no method will provide the desired
level of accuracy the user is supplied with a
warning and information concerning the most accu-
rate method available. Each method description
includes procedures, specifications of number of
QC samples (blanks and spikes, replicates and
duplicates) to be taken, and a reference to the
EPA document from which the text was drawn.
The larger, more complex, system is designed to
aid in defining quality assurance and quality
control sampling requirements. The system ques-
tions the user to determine the kind of errors to
be controlled, for example; sampling or labora-
tory errors, biases or random errors, within or
between day variations, and so on. Based on
this, the system recommends types of QC samples
to be sent to the laboratory, i.e.,. spikes,
duplicates, replicates. The user picks recom-
mended kinds of QC s, one at a time, and
looks at them quantitatively. Based on how much
variation is acceptable, how many samples are in
a set, how rapidly (number of days of sampling)
drift is to be recognized, and the desired cer-
tainty of recognizing drift if it occurs within a
specified time frame, the program will recommend
the number of QC samples to be taken per day.
Alternatively, given the other parameters, it
will derive the level of certainty from the
number of samples.
This list of expert systems should not be con-
sidered as an inclusive list of all systems that
have been designed to assist those responsible
for decisions concerning measurement processes.
Instead they illustrate the types of systems that
can be and are being developed in the area.
DISCUSSION
Measurement of contaminants in the environment
requires knowledge concerning a wide range of
issues. Some of this information is contained in
textbooks, scientific literature, and government
312
-------
reports. Some of the general rules or principles
are only known by those who have extensive experi-
ence in the field. Expert systems provide an
excellent means of compiling and synthesizing
this knowledge and expertise in a form that can
be accessed by persons with less training and
experience. It should be emphasized however.
that these systems can not replace the expert:
they can. however, extend the expert s knowledge.
They can not provide advice concerning issues
that can not currently be resolved by an expert
or knowledgeable person in the field. For example
a system can not be developed that will give
advice concerning a new instrumentation method
that has not yet been conceived or tested. The
primary application is in the more routine situa-
tions which experts can resolve in a few hours to
a few days.
There are a number of measurement and instrumenta-
tion issues that could be used as the focus for
expert systems development. If the goal of devel-
oping an expert system is to aid other decision
makers it is important that the effort be suc-
cessful. Success is determined by three related
criteria. First the system must accurately
represent/ emulate the decision processes used by
knowledgeable persons in the field. Second, the
system must also produce sufficient increases in
decision efficiency and quality to warrant the
costs for development, maintenance, and user
training and support. Finally, the systems must
be accepted and employed by the targeted user
community. The following questions should be
asked before initiation of an expert system devel-
opment effort to assess the potential for success:
1. Do experts/knowledgeable persons exist who are
available and willing to serve as resources in
formulating the decision rules?
2. Is there consensus among the experts and/or
user community concerning the decision rules
appropriate for the application area addressed
by the system?
3. Does the system address a definable need of the
targeted user community and will knowledgeable
persons from the user community participate in
each stage of the development process?
4. Can the knowledge/expertise in the subject area
be adequately represented with available soft-
ware and hardware and will the targeted user
community have the necessary input information
for the system to provide meaningful advice?
5. Can a definite plan be developed and resources
allocated for regular maintenance of the knowl-
edge base and software and for providing user
support and training to the target user
community?
6. Are decisions in the selected area made often
enough to justify undertaking software develop-
ment and maintenance?
If these questions are answered in the affirmative
then the potential developer must explore the scope
of the problem to be addressed. If the problem is
too simple an expert system will not represent a
high enough payoff to justify the time and money
spent developing the system. If the problem is
too complex the development task will quickly
become overwhelming. A rough estimation of prob-
lem scope can be measured by the time it takes an
expert/knowledgeable person to reach a solution.
In general an expert should be able to arrive at
a decision in between one to eight hours. Addi-
tionally, the decision should be based upon
implicit or explicit rules and criteria rather
than strictly upon common sense. Consistent
responses among decision makers and over time are
also desirable features.
CONCLUSION
Expert systems offer the potential of synthesizing
the knowledge and thought processes of the instru-
mentation experts in a form that can be used by
less knowledgeable persons. In essence these
systems offer an excellent mechanism for technol-
ogy transfer. Some expert tools have been devel-
oped to aid in measurement related decisions.
There appears to be a definite need for systems
to aid specifically in instrumentation decisions
and in the array of measurement issues that must
be considered by the Agency and the regulated
community. The apparent need for these systems,
as perceived by the experts and developers, should
not, however, be used as the basis for initiating
development of an expert system in the area. The
importance of input from the targeted user commu-
nity and a thorough evaluation of system costs
and benefits prior to system design and code
development can not be over emphasized. The
targeted user community must be actively involved
during each stage of system development starting
with the conceptualization stage and proceeding
through the final stages of testing and mainte-
nance of the production system. Poorly designed
and maintained systems and/or unused systems are
not only expensive, but tend to result in a loss
in faith for a technology that offers significant
opportunities to improve the efficiency of deci-
sion makers and enhance the quality of their
decisions.
313
-------
APPENDIX
CONTACTS FOR FURTHER INFORMATION CONCERNING THE
CITED EXPERT SYSTEMS
FLEX (Flexible Membrane Liner System)
Daniel G. Greathouse
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
26 W Martin Luther King Drive
Cincinnati. OH 45268
Geophysics Expert System
Aldo T. Mazzella
Environmental Monitoring and Support Laboratory
U.S. Environmental Protection Agency
P.O. Box 15027
Las Vegas, NV 89114
Data Quality Objectives Advisor
Dean Neptune
Quality Assurance and Monitoring Support
U.S. Environmental Protection Agency
401 M Street, S.W.
Washington, DC 20460
Quality Sampling and Analysis
Lawrence Keith
Radian Corporation
P.O. Box 9948
Austin, TX 78766
DISCUSSION
W.F. ARENDALE: We've been talking about many of the problems and costs
of transferring from the research laboratory to the field. I am convinced that
expert systems have a place. But the big problem is that systems won't transfer
data. There must be something done to transfer data from one to the other.
As you monitor expert systems, do you see them being programmed directly
in LISP or PROLOG or a higher level language. Is there a collective effort to
bring this technology to the field?
DANIEL GREATHOUSE: The field is evolving very rapidly. The systems
are being developed, and some of them are in the languages, some of them are
using shells. That causes a real problem with trying to get compatibility among
them.
Here's one example of trying to do something about it, in the QA/QC area. We
did a needs assessment among our regional offices and found that there was a
real need in that area for advice on a number of different issues. There are about
eight systems being developed, and as you would expect, they are not
compatible. We are trying to pull that information together and at least provide
access to the different systems, or we're trying to work out some way to unify
the information in that area.
One of the problems is that expert systems are being developed by a number of
independent groups that aren't related. How do you get compatibility among
the products that come out? I'm not sure.
That's one of the reasons that there was an initiative approved by EPAfor expert
system development to respond to regional office needs. The objective of the
initiative is to pull together systems in various areas, and make them compat-
ible and more usable for the user community.
JONATHAN NYQUIST: I think we have all seen some of the medical expert
systems that do the diagnosis and help the doctor consider possibilities he
hadn't looked at, or might not look at otherwise. I don't think we would like to
see that expert system take the place of the doctor, and I guess that is one of the
things that makes me a little worried. When there is a lack of trained personnel,
aren't we giving them a loaded gun with these expert systems?
DANIEL GREATHOUSE: I don't think so. First, you have to realize that
those decisions are going to be made by the person without training anyway.
The idea is to try to provide available information to them, as easily as possible,
and in a way that they will not use the information in place of the decision
maker, but as a screen.
314
-------
A POSITIONING AND DATA LOGGING SYSTEM
FOR SURFACE GEOPHYSICAL SURVEYS
Jonathan E. Nyquist
Health and Safety Research Division
and
Michael S. Blair
Instrument and Controls Division
Oak Ridge National Laboratory
Oak Ridge, Tennessee 37831
ABSTRACT
The Ultrasonic Ranging and Data System (USRADS) developed
at ORNL is being adapted to work with two commercially
available geophysical instruments: a magnetometer and an EM31
terrain conductivity meter. Geophysical surveys have proven an
important preliminary step in investigating hazardous waste sites.
Magnetometers and terrain conductivity meters are used to locate
buried drums, trenches, conductive contaminant plumes and map
regional changes in geology. About half the field time of a typical
geophysical investigation is spent surveying the position of the grid
points at which the measurements will be made. Additional time
is lost and errors may be made recording instrument values in field
notebooks and transcribing the data to a computer.
Developed for gamma radiation surveys, the USRAD system keeps
track of the surveyor's position automatically by triangulating on
an ultrasonic transmitter carried in a backpack. The backpack also
contains a radio transmitter that sends the instrument's reading
coincident with the ultrasonic pulse. The surveyor's position and
the instrument's reading are recorded by a portable computer
which can plot the data to check the survey's progress. Electronic
files are stored in a form compatible with AutoCAD to speed
report writing.
INTRODUCTION
A number of surface geophysical methods developed for oil and
mining industry applications have been adapted for the
investigation of hazardous waste sites, for example: magnetic field
intensity, resistivity, time domain electromagnetics, and frequency
domain electromagnetics (1,2). While formerly the geophysical
prospector used a magnetometer to search for iron ore, now the
same instrument may be used to locate buried drums of toxic
waste (3,4). But the goal of geophysical surveying remains the
same: to learn as much as practical about the subsurface before
going to the expense of drilling. Often fewer wells are required to
characterize the subsurface as the geophysical data can be used to
choose the best well sites and to interpolate between wells. In
some cases, such as searching for buried drums, a geophysical
survey can save thousands of dollars in hit-or-miss excavation
efforts.
While geophysical surveys are relatively fast and inexpensive, time
and money are lost while surveyors set up the measurement grid.
Additionally, although most modern geophysical instruments have
a built-in data logger and can dump the survey data to a portable
computer at the end of a survey, there is no way to view the
collected data while the survey is in progress. Often the grid lines
turn out to have been too widely spaced, or the grid did not cover
a large enough area; as a result, the surveyors must be brought
back to the field to refine or extend the grid.
The need to expand or refine a measurement grid and to analyze
the data while the survey is in progress are common to all field
surveys. The same problems arose, when as part of the
Department of Energy's Uranium Mill Tailings Remedial Action
Project (UMTRAP), ORNL was requested to survey, in three
years, 8,000 properties where presence of uranium mill tailings
were suspected. To save time and money ORNL developed a
technology to automate much of the survey process and provide
tabular and graphical data display in the field or in the office for
report generation. This technological development is called the
Ultra Sonic Ranging and Data System (USRADS) (5).
Our recent work has focused on interfacing two geophysical
instruments, a proton procession magnetometer (Omni IV) and an
electromagnetic terrain conductivity meter (Geonics EM31) with
the USRADS. This work is still in progress. In this paper we will
describe the application of the two geophysical instruments and the
USRADS separately, and the work which being done to combine
them.
CASE STUDY: EVAPORATION PIT AT DYESS AIR FORCE
BASE
An evaporation pit at Dyess Air Force Base in Abilene, Texas
received liquid waste from the late 1950's to the late 1970's, when
it was backfilled and abandoned. The pit's approximate areal
dimensions are 45 x 30 m (150 x 100 ft) from base records. After
the pit was closed, base personnel continued to use a buried
38,000 liter (10,000 gal) railroad tanker car to the east of the
evaporation pit as a liquid waste repository until 1982. The tanker
has since been removed.
Although it was believed that only liquid waste was disposed of at
the site, a preliminary inspection found crushed drum fragments,
railroad spikes and other metal debris on the surface. We decided
to use a magnetometer to check for buried metal as well. A
rectangular grid consisting of 194 points on a 15 m (50 ft) spacing
took two days to survey. The region in the southwest corner was
not surveyed because small trees interfered with sighting the
transit. A team of two completed the actual magnetometer survey
in a single morning. The results show strong dipole anomalies
(paired lows and highs, Figure 1). For each pair, the buried
source, probably one or more 208 liter (55 gal) steel drums, is
located about halfway between the adjacent positive and negative
peaks. Notice that the peaks are always located on a grid point.
This is because the anomalies are smaller than the grid spacing.
The actual buried object producing the anomaly may lie as much
315
-------
-650.00 -550.00 -45000 -350.00 -25000 -15000 -5000
MAG.
NORTH
DYESS AFB HARDF11L II AND EVAPORATION PIT
MAGNETIC GRADIENT DATA. Cl = -10 CAMMAS/W
Figure 1.
The terrain conductivity data (Figure 2) adds to the picture formed
from the magnetic gradient contours. This survey took two people
an afternoon. While the magnetometer responds only to the total
mass of ferrous metal, the terrain conductivity meter responds to
anything conductive which depends on the conductivity and surface
area. It is less affected by the small buried metal, but shows a
strong anomaly centered on the former location of the buried
railroad tanker car and a plume appears to have formed to the
west. Sixteen grid points were surveyed around the tanker car,
extending the grid to the east and verifying that the tanker car was
the source.
If time had permitted, more measurements would have been made
to the east of the tanker car to completely map the plume. We
later learned that the our survey had caught only the very edge of
the evaporation pit (northwest corner of Figures 1 and 2); most of
the survey grid was over a hardfill area where metal scrap had
been dumped. The grid-should have been expanded to the west
-BSO.CO -SSO.OO -*KJ.OO -Mooo -130.00 -ISO.OO -5O.OO So 00
DYESS AFD EVAPORATION POND
EH31 QUADRATURE DATA
as well. Such surprises in the field are common and illustrate that
surveying the grid point locations and reducing the geophysical
data handicap an otherwise fast, effective and inexpensive
reconnaissance. A way is needed to analyze the data in the field
and adjust the survey lines, without mobilizing the field crew a
second time.
ULTRASONIC RANGING AND DATA SYSTEM (USRADS)
System Operation and Setup
Real-time analysis of the data is a major advantage of the
USRADS. As the surveyor walks the property, an ultrasonic crystal
in the surveyor's backpack is pulsed each second and the data from
the survey instrument are transmitted to the computer by radio.
Each second, the computer reads the time-of-flight data from
stationary receivers placed in the survey area, triangulates the
surveyor's location, plots the surveyor's location on the computer
screen, and stores all raw data. By plotting the surveyor's location
each second, the computer operator can view the surveyor's
coverage of the property at any time during the survey. In addition
to plotting the surveyor location, the computer highlights any data
point that exceeds a threshold entered by the operator, so that any
areas of concern are identified on the display, to ensure that
sufficient data have been obtained to characterize that area.
The system setup takes only about fifteen minutes. Stationary
receivers are placed so that the surveyor is in view of at least three
of them from any location on the property (Figure 3). Only the
first few receivers need to be located by a surveyor; once the
stationary receivers have been placed on the property, they are
used to calculate the speed of sound and the locations of the
remaining stationary receivers are computed automatically.
LOCATING THE USRADS SURVEYOR BY TRIANGULATION
—• - '""" ""/-^
Figure 3.
System Hardware
Figure 2.
The USRAD system consists of one surveyor's backpack, fifteen
stationary receivers, a master receiver, custom computer interface
and counter timer module, Compaq Portable II personal computer,
and a small trailer to transport this equipment. The backpack
contains the interface circuitry to receive the signal from the field
instrument (originally a portable gamma detector), an ultrasonic
transmitter and radio frequency equipment to establish a bi-
directional communication link with a computer mounted in the
trailer. The ultrasonic transmitter is a lead-zirconate-titanate
crystal that is in the form of a circular cylinder with a hollow core.
316
-------
crystal that is in the form of a circular cylinder with a hollow core.
The crystal dimensions are 5.6 cm (2.2 in) in diameter and 3.670
cm (1.445 in) in height. This crystalline material and its
dimension result in a natural resonating frequency of 19.5 kHz.
The crystal is pulsed for 10 msec each second as the data from the
portable survey instrument are transmitted to the computer via the
radio telemetry link. If the computer detects any problems, either
with the data or in determining the surveyor's location, a message
is transmitted to the surveyor and displayed on the handheld
terminal to alert the surveyor of the malfunction. The backpack
can be operated for a normal eight-hour day from a rechargeable
gel-cell.
The stationary receivers contain an ultrasonic receiver and a radio
transmitter. The dimensions of the metal box that houses the
ultrasonic receiver card, transmitter card, and rechargeable gel-cell
battery pack are 10 x 10 x 15 cm (3.9 x 3.9 x 5.9 in). Each
stationary receiver has a unique radio frequency so that the master
receiver can identify which stationary receivers heard an ultrasonic
signal. The master receiver therefore contains 15 radio receivers,
one for each stationary receiver, and a receiver and transmitter for
communication with the backpack. Both the master receiver and
the computer are powered by a gasoline-operated generator also
carried in the trailer.
System Software
A digitized schematic drawing of the property is stored in the
computer prior to the survey using AutoCAD, a commercial
computer-assisted drawing software package already widely used at
ORNL. The survey data are added to this information. The
property schematic is displayed on the computer's monitor. As the
surveyor traverses the property, his past and present position are
displayed to denote the completeness of coverage by the surveyor.
During the survey, the software checks incoming information and
alerts the surveyor (via the backpack terminal) if errors are
detected either in the survey data or position data. To ensure data
integrity, all data are stored on the hard disk every 30 seconds.
The surveyor can view the data in a number of different graphical
formats as well as obtain summary reports. The graphical formats
supported by the USRADS are Replay, Block Statistics, Contour,
and 3-D plots of the radiation data. The Replay program will
generate the same display that the surveyor viewed when the
survey of the property was completed. The data are replayed in
the same order as they were collected. The Block Statistics
routine enables the operator to select a grid block size and have
the data analyzed for each block. If the mean of the data for a
particular grid block is greater than the operator-entered threshold,
then that block is highlighted on the CRT, and the statistical
information for that grid block are stored in the summary report.
Raw data are converted to appropriate units and displayed or
printed out in tabular or graphical format. By indicating preset
thresholds, areas of contamination can be identified and statistics
can be calculated (area size, number of measurements,
measurement range, average and standard deviation). Graphical
representations are made in two and three-dimensional display.
The contour routine generates a summary report and outlines the
areas that exceeded the user input threshold. The 3-D plot
generates two different views of the data and provides a means by
which the surveyor can view the entire data obtained during the
survey. Information displayed in the field is output directly into
a report-ready format.
COMBINING USRADS AND GEOPHYSICAL INSTRUMENTS
Check for Interference
The first step was to insure that the USRAD system and the
geophysical instruments would not interfere with each other. The
USRADS backpack contains very little ferrous metal and we found
no changes in magnetometer readings taken with and without the
backpack. We also found no interference with terrain conductivity
(EM31) readings.
Interfacing the Instruments
No change is required in the geophysical instruments or USRADS
hardware. Our current efforts are devoted to rewriting the
USRAD backpack software to digitize and transmit the terrain
conductivity meter's readings. The EM31 is an analogue
instrument which can operate in one of two modes: inphase or
quadrature. The inphase mode has a non-linear response to
conductivity changes but is especially sensitive to sudden changes
in near-surface conductivity, as might be produced by a buried
drum (6). The quadrature response is directly proportional to the
conductivity of an equivalent uniform half-space. This is the mode
generally used for geologic mapping and mapping conductive
contaminant plumes (7). We are modifying the USRADS software
to record both modes simultaneously, along the range switch
setting, and to display either data set to the computer operator.
We are taking this opportunity to replace the backpack's ROM
chips with new chips coded in the C programming language instead
of assembler. This will make the backpack software compatible
with software written for the COMPAQ portable computer and
make the interfacing of USRADS with other survey instruments
easier by making reprogramming simpler. Work will begin on
interfacing a magnetometer with USRADS at the same time as the
terrain conductivity/USRAD system is being field tested.
Field Testing
We expect to begin field testing in August, 1988. There are
several locations on the Oak Ridge reservation where terrain
conductivity and magnetometer surveys have been conducted on a
grid surveyed by transit. We plan to re-survey one or more of
these areas and compare the time required to complete the
geophysical surveys and the quality of the results with those of the
earlier surveys.
SUMMARY AND CONCLUSIONS
Geophysical surveys are frequently a part of initial investigations
at hazardous waste sites. Searching for buried drums with a
magnetometer is a classic example. The slowest step in a
geophysical survey consists of locating the measurement points, and
often this step has to be repeated when the geophysical data
suggest that the survey grid needs to be expanded or refined. By
combining the USRAD system with geophysical instruments this
step can be virtually eliminated as the surveyor's position and the
instrument's reading are automatically recorded every second.
Only a few ultrasonic receivers have to be located in advance. In
addition, the computer operator can notify the surveyor when the
data displayed on the portable computer suggests that an anomaly
has been found, and additional measurements are needed. Thus
the USRAD system will add automatic position location, real-time
data processing, and automatic data transcription to geophysical
surveys.
317
-------
ACKNOWLEDGEMENTS
This work was performed by Oak Ridge National Laboratory
operated by Martin Marietta Energy Systems for the U.S.
Department of Energy under contract DE-AC05-84OR21400.
REFERENCES
(1) Glaccum, R. A., R. C. Benson and M. R. Noel, "Improving
Accuracy and Cost-Effectiveness of Hazardous Waste Site
Investigations," Groundwater Monitoring Review. Summer,
1982, pp. 36-40.
(2) Dobecki, T. L. and P. R. Romig, "Geotechnical and Ground
water Geophysics," Geophysics. Vol. 50, No. 12, 1985,
pp. 2621-2636.
(3) Tyagi, S. and A E. Lord, Jr., "Use of a Proton Procession
Magnetometer to Detect Buried Drums in Sandy Soil," Journal
of Hazardous Materials. Vol. 8, 1983, pp. 11-23.
(4) Breiner, S., Applications Manual for Portable Magneto-
meters. Geometries, Sunnyvale, CA, 1973, 58 pp.
(5) Berven, B. A., M. S. Blair and C. A. Little, "Automation of
the Radiological Survey Process: USRADS - Ultrasonic Rang-
ing and Data System," 1987 International Decommissioning
Symposium. CONF-871018, Vol. 1, ed. G.A. Tarcza, Westing-
house Hanford, Richland WA, pp. V-129-V-134.
(6) McNeill, J. D., "Use of EM31 Inphase Information," Tech-
nical Note TN-11. Geonics Ltd., Mississauga, Ontario,
Canada, 1983, 3 pp.
(7) McNeill, J. D., "Rapid, Accurate Mapping of Soil Salinity
Using Electromagnetic Ground Conductivity Meters," Tech-
nical Note TN-18. Geonics Ltd., Mississauga, Ontario,
Canada, 1986, 28 pp.
DISCUSSION
WAYNE SAUNDERS: What are the costs to develop this system? I don't
know what the frequency is of the transponders, to find yourself at a certain
location. Did that have any interference with the EM31, either the quadrature
or the in phase?
JONATHAN NYQUIST: We didn't find any interference with the radio
frequency in the EM31.1 think it will cost something on the order of $ 10,000.00
to $15,000.00.
WAYNE SAUNDERS: What does the unit weigh, the back pack?
JONATHAN NYQUIST: It weighs about 18 pounds.
HARRY McCARTY: Given that the back pack weighs only 18 pounds, did
you ever consider using Loran-C or something like that for positioning,
especially if you're going to a multi-acre site, and you don't want to have to set
up transponders? Did anything get done on that at all?
JONATHAN NYQUIST: We have thought about using some kind of little
position syntax to locate the first couple of points, on your grid on the big map,
before the small areas are surveyed. Ithinkyoucoulddothat. We haven't added
it in, because the whole thing starts getting costly.
ROY JONES: Is this system affected by industrial power distribution systems,
transmission lines, things like that? Do you have anomaly problems?
JONATHAN NYQUIST: I don't know if it's been tested right next to a
transmission line, but the system was originally developed to work in back
yards in surveying properties. It's not really affected by being near telephone
lines, power lines, houses, buildings, that kind of thing. That's one of the
reasons they went to ultrasonics for the locating, rather than a radio locating
system.
ALDO MAZZELLA: It looks like your transponder locates yourself in terms
of X,Y coordinates horizontally. Many geophysical techniques also need
topographic directions. How do you incorporate that into your system?
JONATHAN NYQUIST: The way they have been thinking about trying to do
that is by putting a second ultrasonic transmitter above the first and using the
phase difference from each of those pulses to their stationary receivers to pick
up the altitude. That is a major redesign. It is not a simple thing to add in, and
it is going to depend on somebody willing to fund that.
ALDO MAZZELLA: On the magnetometer, did you correct for the cleanness
of your system?
JONATHAN NYQUIST: Yes, we did look to see if there was any interference
from the backpack itself. It's almost all aluminum components, so there really
wasn't anything the magnetometer saw. When we measure with and without the
system being operational, we didn't see any difference.
318
-------
PROTOTYPE VOLATILE ORGANIC COMPOUND
(VOC) MONITOR
Joseph D. Wander
Environmental Sciences Branch
Environics Division, Headquarters
Air Force Engineering and Services
Center, Tyndall AFB FL
Barbara L. Lentz
Larry Michalec
Victoria Taylor
S-CUBED Corporation
La Jolla CA
ABSTRACT
The Air Force is sponsoring development
of a prototype "smart" instrument
package, centered around purge and trap
concentration and gas chromatography, to
monitor concentrations of VOCs in water.
This system will provide timely analyt-
ical support for efforts to characterize
and remediate VOC-contaminated sites on
Air Force bases.
INTRODUCTION
During the past 40 years, Air Force (AF)
aircraft operations and engine mainten-
ance have involved enormous quantities
of organic liquids, the constituents of
which are subject to regulation as
volatile organic compounds (VOCs). In
1981, the AF initiated the Installation
Restoration Program (1) (IRP), which is
a systematic effort to identify and
repair sites where soil and water con-
tamination have resulted from mission-
related activities.
IRP action at a contaminated site follows
a four-step sequence:
(1) identify the problem,
(2) characterize its magnitude and its
distribution,
(3) select or develop proper remedial
technology, if needed, and
(4) issue contracts for the remediation
effort.
Water analyses from monitoring wells are
used during the site characterization
process and in evaluating the progress
of the remediation. If pump and treat
methods are selected for remediation,
effluent water must also be certified
for discharge.
The VOC monitor was conceived to support
efforts under IRP by providing rapid,
reliable, onsite analyses of VOCs in
water samples.
INITIAL CONSIDERATIONS
The development of the prototype VOC
monitor was separated into two phases.
The objective of phase 1 was to design
and field test a system able to perform
automated analyses of a single VOC. The
objective of phase 2, which was to follow
an evaluation of the initial prototype,
is to design, build and field test an
enhanced prototype, able to perform con-
current analyses of 10 VOCs in a water
sample.
Four requirements were imposed upon
designs for all phases during development
of the VOC monitor:
(1) speed of analysis — offsite sources
for water analyses require an inconveni-
ently long turnaround time of several
days to several weeks;
(2) simplicity of operation — site per-
sonnel are presumed to have little or no
chemical training;
(3) internal control of reliability —
site personnel are presumed to have
limited capability to judge quality of
data; and
(4) ruggedness — resources for onsite
repair and maintenance are limited.
PHASE 1 PROTOTYPE
Proven, off-the-shelf purge and trap and
gas chromatography units were selected
as the analytical hardware for the phase
1 prototype; a 10-position autosampler
was included to allow unattended opera-
tion. Factors in the selection of the
specific hardware components included the
fact that all three units are serviced by
the same manufacturer's customer support
organization, and a reasonable expecta-
tion that the components would remain
available for several years.
Chromatography was performed on a 6-ft x
1/8 in stainless steel column packed
with 1 percent SP-1000™ on 60/80 mesh
Carbopack B™, in close correspondence
319
-------
with EPA Method 601; however, a flame
ionization detector (FID) was selected
for sturdiness and simplicity. Owing to
the possibility that halogen-induced
suppression of ion formation would render
the FID too insensitive for this appli-
cation, trichloroethylene (TCE), which
was the object of an ongoing IRP effort
at an AF location in Michigan, was chosen
as the analyte for the phase 1 prototype.
This selection provided both a worst-case
test for the concept and a practical
situation in which performance of the
monitor could be evaluated.
An external jack from the FID output was
connected to an analogue-to-digital
converter (A/D) and thence to a personal
computer (PC). Instrument operation,
data acquisition, data processing, re-
porting of the results of the analyses,
quality assurance, and assessment of
deviations from control or failure to
detect internal standard signals are
managed by resident software, which is
activated at the start of the analysis
by a single, menu-directed keystroke.
SAMPLE PREPARATION
The analytical procedure is detailed in
an operator's manual (2) that accompanies
the phase 1 monitor. Water samples
(collected according to EPA-specified
procedures) are delivered carefully into
an inverted 25-mL syringe. After re-
placement of the plunger, the syringe is
gently inverted. Air and excess water
are expelled until 5.0 mL of water
remains in the syringe.
The tip of a 50-jjL syringe is inserted
through the tip of the large syringe into
the water sample, to deliver a 10.0-pL
aliquot of a solution containing 5.0
ng/pL each of.three internal standards:
bromochloromethane (IS1), 2-bromo-l-
chloropropane (IS2) and 1,4-dichloro-
butane (IS3). The sample is delivered
at once into the purge vessel, where it
is secure until analyzed.
PHASE 1 SOFTWARE (3)
Output from the FID is sampled at 1 Hz;
the signal from the A/D is 12 bits, which
allows a maximum integer value of 4095.
Data from the A/D are stored in an
integer array in RAM during the chromato-
graphic run. At the conclusion of data
acquisition, the array is smoothed with
a five-point least-squares procedure (4)
prior to application of a peak-detection
routine based on a three-point, forward-
looking window (5).
The peak detection routine uses a "four-
point rising tangent" criterion to locate
the start and end of each peak, plus a
first-derivative test to assure detection
of sharp peaks. A confirmatory test (5)
based on the second derivative is applied
to eliminate false identifications.
Peaks so located are stored in an
"uncorrected" data array as sets of four
parameters: peak start point (S^), peak
end point (E^), peak maximum point
(Mi) and uncorrected peak area (Ui).
A baseline correction (5) routine is then
applied to the smoothed integer array,
which calculates a "chromatographic
baseline" by fitting a series of line
segments between maximally separated
values of S^ and E^ chosen such that
no point S^ or E^ falls beneath any
baseline segment. For each of the peaks
detected, the area (C^) beneath the
chromatographic baseline between S^
and Ei is determined and entered into
a "correction" matrix. The final
"corrected" data array describing the
true peak areas ([A^]) is generated by
subtraction of [C^] from [Uj].
ASSIGNMENT OF INTERNAL STANDARD PEAKS
The software searches for peaks in a
window from 0.5 x t2/std (a stored
value for the retention 'time of IS2) to
1.5 x t2/std- For each of the peaks
encountere'd in this window at actual
retention time t2,p' proportional
retention times (t'l,p an<^ ^-3,?'
respectively) are calculated for IS1 and
IS3. These three values are inserted
into the following function from
multivariate statistics (6):
ti,std
cos 0t= 3 3 1/2
fl t?,std.Z t?,P>
The corresponding areas (Aj^p) of the
three peaks under scrutiny are inserted
into a similar calculation of cos 0a,
for which standard areas (Ai/std) are
stored at the same time as t
Finally, a six-dimensional parameter is
calculated for this set of peaks:
cos 0 = cos
x cos 0a
The next set of peaks is examined in the
same way, and the set giving the larger
value of cos 0 is stored for the next
comparison until all such sets have been
examined. If all three IS peaks are
located and cos 0 >. 0.7, the retention
time of the analyte (TCE) is calculated
by direct proportion, and the TCE peak
is identified as the peak nearest the
calculated position (within ± 3 percent
of the value calculated). Finally, the
peak area is converted into a concentra-
tion, which is reported by the printer.
320
-------
CALIBRATION
Every 10 days, or whenever the QA manager
finds the system to be out of control,
the operator is required by the software
to perform a calibration. A water sample
and three prepared standards, at TCE con-
centrations of 1.5, 4.0, and 10.0 ng/mL,
are analyzed. The software performs a
series of three linear least squares
fits of the three concentrations of TCE
to the ratio of the corrected areas of
the analyte peaks to the corrected areas
of the peaks corresponding respectively
to IS1, IS2 and IS3. If r_ >. 0.9 for
each of the three fits, the unit accepts
the calibration and analytical operation
may be resumed; if not, the unit will
perform no analyses until the fault is
corrected and a satisfactory calibration
is performed.
CALCULATION OF CONCENTRATIONS
The value Aj from the corrected data
matrix corresponding to the retention
time Mi assigned to the analyte is
divided in turn by each of the areas of
the internal standard peaks, and the
quotients are substituted into the
respective linear expressions defined
during the calibration procedure.
The result is three values of the con-
centration of the analyte in the same
units in which the calibration was per-
formed. The software then computes the
mean (x) and standard deviation (jj.) for
the three values. If any of the values
differs from x by more than 2s., it is
declared an outlier and rejected. The
concentration is reported as the mean of
all values not rejected.
THE REPORT
Data reported for all peaks detected are
the retention time and Aj. If no error
flags were set, the concentration of the
calibrated analyte is printed; otherwise,
the message(s) defined by the flags are
printed. This format is intended to let
the printed output serve as a complete
documentation of the analytical result
and the control status of the analytical
method, which will satisfy requirements
of environmental regulatory agencies.
The results, date, time and operator are
stored in compressed form in an archival
file for verification.
QUALITY ASSURANCE
Two QA procedures were incorporated in
the phase 1 prototype. The Procedural
Guide (2) specifies that a standard (now
4.0 ng/mL) be analyzed daily at the con-
clusion of a series of samples to verify
that the calibration has not failed. If
the value reported differs by more than
2s. from the mean of the control chart
for the daily determinations, the oper-
ator is instructed to conclude that the
system is out of control and to perform
a recalibration.
The software also calls for spike and
duplicate spike analyses of every tenth
sample. A 5-fiL aliquot of a methanolic
solution containing 10 ng/|jL TCE is added
to two duplicate samples, which are then
analyzed "normally." The results of the
two spiked samples are compared with the
result of the unspiked analysis, and with
each other. If the mean recovery of TCE
from the two spiked samples is less than
75 percent or more than 125 percent, a
flag is set to print a report that ac-
curacy criteria are not satisfied and to
require recalibration. If the difference
between results for the pair of spiked
samples exceeds 25 percent of the amount
of the spike, a flag is set to report
that reproducibility criteria were not
met and to require recalibration.
PHASE 1 FIELD EVALUATION
In September 1986, The phase 1 prototype
was installed in the field, at the the
sewage plant serving Wurtsmith AFB. Some
minor renovation was required during
installation to provide bench space and
isolated electrical service. The only
standard laboratory facilities available
were a fume hood and a balance.
In this environment, the system has per-
formed satisfactorily throughout the
evaluation. The first data set gener-
ated for a split-sample evaluation was
submitted to, and accepted by, EPA
Region V as support for an application
for local alternate test procedure (ATP)
status for the system and method. Ap-
proval as a local ATP was given in
February 1988.
OBSERVATIONS
Three spontaneous mechanical failures
occurred during the test period, which
ended in November 1987. All three
(failed heater connections [two times]
and dirty contacts on a printed circuit
board in the gas chromatograph [Varian
3400] plus misalignment of a six-port
valve inside the Tekmar purge and trap
unit) required intervention by project
personnel. A subsequent incident, also
involving valving in the purge and trap
unit, was managed successfully by site
personnel, with help from Varian1s
customer support organization (provided
under a continuing annual service con-
tract, which has been in effect since
the installation).
Effective orientation and education of
site personnel was the most-difficult
321
-------
task in the phase 1 effort, and experi-
ence gained during this part of phase 1
has directed a number of changes in the
way in which the phase 2 software will
interact with the operator. Documents
accompanying the phase 2 monitor will be
expanded to include suggestions about
inventory requirements and AF sources
for supplies, and to recommend security
measures to protect the monitor from
incidental damage.
A presumably site-specific problem (under
investigation) that emerged was a chronic
(estimated to be of the order of 50 parts
per trillion) background interference
from a VOC not positively identified but
isographic with benzene. The phase 1
unit successfully discriminated against
the interferent when TCE was present at
levels above its threshold for detection
(about 0.5 parts per billion), but it
routinely identified the interferent as
TCE when the actual TCE peak was below
the detection threshold. (Because the
other VOCs in the list below have either
a higher action level or more sensitivity
at the FID, this is not expected to be a
problem to the phase 2 prototype.)
PHASE 2
Current models of the same components
were procured for assembly into the
phase 2 prototype, which was to be an
improved version of phase 1, able to
perform concurrent analyses (in indi-
vidual water samples) of as many as 10
VOCs from the following list:
benzene
chloroform
dichlorobenzenes
1,1-dichloroethane
1,1-dichloroethylene
£is.-l, 2-dichloroethylene
±_rans.-l, 2-dichloroethylene
ethylbenzene
methylene chloride
1,1,2,2-tetrachloroethane
tetrachloroethylene
toluene
1,1,1-trichloroethane
trichloroethylene
xylenes
(DCBs)
(1,1-DCA)
(1,1-DCE)
}(1,2-DCE)
(1,1,2,2-TCA)
(PCE)
(1,1,1-TCA)
(TCE)
A different PC (compatible with the
phase 1 software) was selected when the
original unit went off the market. A
chip was added to support floating-
point calculations. After assembly of
the hardware for the second prototype,
it was established that the combination
of chromatographic resolution and data
sampling rate was inadequate to provide
reliable separation of and discrimination
between 1,1,2,2-TCA, PCE and IS3.
Rather than abandon the commercially
available and accepted internal stan-
dards, we increased the data sampling
rate (7) to 9.11 Hz. This improved the
reliability of peak discrimination by
the software, but it also established
that the chromatographic separation was
insufficient. Accordingly, we replaced
the packed column with a 30-m x 0.53-mm
fused silica capillary column coated
with DB-624™.
The increased data-sampling rate also
created new problems in dealing with the
baseline, requiring reoptimization of
baseline-correction and peak-detection
algorithms, which were written for a
1-Hz data rate.
The promulgation of 40 CFR 141 created a
requirement that vinyl chloride be mea-
sured whenever two-carbon halogenated
species are detected in waters to be used
for drinking. Accordingly, vinyl
chloride was added to the list above to
preserve the option of discharging a
treated stream into a potable water
supply.
Methanol, the accepted solvent for VOC
standards, is incompletely separated on
the DB-624™ column and obscures vinyl
chloride at the FID. As EPA discourages
the use of aqueous calibration standards,
we invoked a hardware method to remove
methanol.
Nafion™ membranes are widely used in
devices to remove water from gas streams
(8-11). Loss of polar organics has been
documented in gas streams so treated
(9), and methanol is reported (10) to
cause reversible swelling of these
membranes. Baker (11) reported >90%
removal of methanol from a compressed
gas stream passed through a long-path
Nafion™ dryer.
Based on these separate reports that
methanol diffuses through Nafion™ and
that VOCs are transmitted efficiently
through such driers, we opened the
stainless steel transfer line between
the sparger and the trap and used
Swagelok™ connectors to insert a
12-in section of 1/16-in Nafion™
tubing. The tubing is temporarily
enclosed in a 1-in dia x 2-in plastic
tube that is partially filled with
calcium sulfate and purged with dry
nitrogen to sweep out solvents as they
diffuse through the membrane. Prelim-
inary evaluation of this separator
showed efficient removal of methanol.
The loss in simplicity of design and in
ruggedness of the column is compensated
by several gains:
(1) the baseline rise during temperature
programming of the packed columns ceases
to be a problem;
322
-------
(2) the VOCs are completely separated;
(3) the interference with detection of
vinyl chloride is eliminated;
(4) the chromatographic method is in
correspondence with more-current EPA
methods (524 and 624).
Other changes from phase 1 will include
decreasing to three days the maximum
period between recalibrations for ana-
lytes, replacing the Tekmar ALS-10 with
a true autosampling device that will
accept 40-mL collection bottles and
introduce internal standards during
injection, interactive assistance to the
operator in performing the analysis and
in troubleshooting, and an interactive
orientation and training program to
facilitate acceptance of the monitor by
new operator-trainees.
STATUS OF PHASE 2 PROGRAM
After the bugs have been worked out of
the software and specific procedures
have been defined for calibration and
analysis, the phase 2 prototype will be
installed at a base in California for
onsite evaluation. The target date for
the installation is January 1989, and a
six-month evaluation period is planned.
REFERENCES
1. Defense Environmental Quality Program
Policy Memorandum 81-5 (11 Dec 81).
2. Taylor, V. and Lentz, B., "Procedural
Guide for Automated Purge and Trap
Analysis of Trichloroethylene in
Water," in Taylor, V. and Wander, J.,
"Prototype Technology for Monitoring
Volatile Organics, Volume I," ESL-TR-
88-01, 1988, (DTIC AD-A195120) pp. 54
— 99.
3. Lentz, B., "Source Code," in Taylor,
V. and Wander, J., "Prototype Tech-
nology for Moniotring Volatile Organ-
ics, Volume II," ESL-TR-88-01 Vol II;
(DTIC AD-A195101).
4. Savitsky, A. and Golay, M. J. E.,
"Smoothing and Differentiation of
Data by Simplified Least Squares Pro-
cedures," Anal. Chem. Vol. 36, No. 8,
1964, pp. 1627 — 1639.
5. Woerlee, E.F.G. and Mol, J.C., "A
Real-Time "as Chromatographic Data
System for Laboratory Applications,"
J. Chromatoaraphic Sci. Vol. 18,
1980, pp. 258—266.
6. Picker, J.E. and Colby, B.N., "Aro-
clor GC Pattern Recognition," Pitts-
burgh Conference on Analytical Chem-
istry and Applied Spectroscopy, 1983;
Davis, J.C., "Statistics and Data An-
alysis in Geology," John Wiley &
Sons, New York, 1973, pp. 525—531.
7. Reese, C.E., "Chromatographic Data
Acquisition and Processing. Part 2.
Data Manipulation," J_. Chromato-
qraphic Sci. Vol. 18, 1980, pp. 249—
257.
8. Foulger, B.E. and Simmonds, P.G.,
"Drier for Field Use in the Determi-
nation of Trace Atmospheric Gases,"
Anal. Chem. Vol. 51, No. 7, 1979, pp.
1089—1090; Pleil, J.D., Oliver, K.D.
and McClenny, W.A., J. Air Pollution
Control Assoc. Vol. 37, No. 3, 1987,
pp. 244—248.
9. Burns, W.F., Tingey, D.T., Evans, R.
C. and Bates, E.H., "Problems with a
NafionTM Membrane Dryer for Drying
Chromatographic Samples," J_. Chrornat-
oqr. Vol. 269, 1983, pp. 1—9; Noij,
T., van Es, A., Cramers, C., Rijks,
J. and Dooper, R., J. High Resolution
Chromatogr. Chromatoar. Commun. Vo1.
10, No. 2, 1987, pp. 60—66.
10. Permapure Products, Inc., Bulletin
106, Farmingdale, NJ.
11. Baker, B. B. Jr., "Measuring Trace
Impurities in Air by Infrared Spec-
troscopy at 20 Meters Path and 10
Atmospheres Pressure," Am. Industrial
Hygiene J_. Vol. 35, 1974, pp. 735 —
740.
DISCUSSION
JOSEPH ROESLER: On volatile organics, the official EPA method for
sampling is a manual method. That's the only way you're allowed to sample
volatile organics. And it's written up in their procedures. Therefore, if you use
an instrument to sample volatile organics, you must have an alternate test
procedure to use the other technique.
JACOB GIBS: There are a number of commercially available chromatogra-
phic software packages that do approximately 80% of what you' ve talked about
here. What would have been the difference to you of one of asking those
vendors to modify the program for the other 20% that you needed to make it
"idiot proof," as opposed to doing everything on a clean sheet of paper?
JOSEPH WANDER: The packages didn't exist when we started (this goes
back to 1983). There is some lag time. I had considered going to a soil gas
device, using this same system, and I have been going through exactly the same
question. I'm not sure we can do it economically for what they can.
323
-------
ENVIRONMENTAL FIELD SAMPLING EXPERT SYSTEM
DEVELOPMENT OF A SOIL SAMPLING ADVISOR
R. A. Olivero, R. E. Cameron, K. J. Cabbie
M. T. Homsher, M. A. Stapanian
Lockheed Engineering & Sciences Company
1050 E. Flamingo Rd., Las Vegas, NV 89119
K. W. Brown
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
P.O. Box 93478, Las Vegas, NV 89193-3478
ABSTRACT
The generation of data of known and acceptable quality for environmental decision-making
requires the use of sound technology at every step of the measurement process. The field
sampling step can introduce an uncontrolled amount of uncertainty in the data. This
variability might be enough to compromise the attainment of the overall data quality objectives
of the project in spite of efforts to optimize instrument performance and analytical method
reliability. Computerized tutorials and decision support aids, such as expert systems, can
assist in the planning of environmental field sampling projects by providing to planners a
sound body of knowledge about sampling. This paper describes the development of a pilot expert
system that assists the sampling manager in evaluating alternative procedures in view of
project data quality objectives and hazardous site characteristics. The increased access to a
more scientific and systematic methodology for planning a sampling project will result in
significant benefits: reduced variability due to sampling errors and deficiencies, reduced
potential for sample contamination, increased representativeness, more efficient use of the
usually limited sampling resources, and an appropriate match of allowed error and required data
quality to provide the information needed for decision making. The present stage of the system
development addresses soil sampling for inorganic target analytes, which are included in the
Contract Laboratory Program, as a test subset of the complexities of the sampling problem.
Key words: Environmental sampling, soil sampling, expert systems, data quality objectives.
INTRODUCTION
The generation of analytical data of appropriate quality for monitoring, cleanup, and exposure
assessment projects requires a combination of capable instrumentation, sound methodology, and
representative sampling. Appropriate data quality is determined at the outset of a project
when data quality objectives (DQOs) are defined. The subsequent stages of sampling, chemical
analysis, data review, and data interpretation need to be planned in accordance with the
identified data quality needs. The U.S. Environmental Protection Agency (EPA) has issued
guidelines and procedures for generation of data used in environmental decision making [1, 2].
Sampling and analysis can be major sources of uncertainty in the data if appropriate quality
assurance and quality control (QA/QC) measures are not implemented. An effective data quality
review subsequent to data collection and analyses can unveil deficiencies in data quality but
cannot prevent the deficiencies. There is no substitute for conscientious planning of
hazardous site monitoring activities to ensure a rational use of resources and the attainment
of the stated DQOs.
325
-------
Traditionally, development and improvement of instrumentation for chemical analyses and
analysis protocols have received more attention than sampling procedures and their related
QA/QC aspects. Considerable improvement has been made to laboratory and field instrumentation
and methodology in the areas of precision, bias, and sensitivity. However, sampling
variability is still a difficult-to-control source of uncertainty in the measurement process.
The use of deficient sampling procedures has a severe, adverse impact on the usefulness of the
data. The improved analytical performances available today should be considered in the proper
perspective as only one component of the overall performance of the measurement process.
Sampling plans need to be specific as to the objectives of the study and the hazardous site
characteristics. This need makes the standardization of sampling procedures and protocols a
difficult task, and it may result in the use of inappropriate or oversimplified sampling
practices. In such a case the generation and use of environmental data may be based on
sampling procedures of unknown quality. In particular, personnel involved in the development
of sampling plans for environmental field monitoring may face difficulties resulting from the
lack of sufficient standardized procedures, rapidly changing technology, scarcity of experts,
lack of knowledge, insufficient training, and inadequate information.
The development of a sound sampling plan involves a large number of decision points and
requires expertise and experience derived from several disciplines such as environmental
engineering, chemistry, geology, soil science, statistics, management, and field operations. A
sampling expert must have the ability to combine all the aspects of the disciplines involved,
as well as field experience and related formal training. Written manuals do not always
adequately provide the nature and amount of background and expertise required for the
appropriate planning of a sampling project.
To compensate for deficiencies in personnel background and experience and the inadequacies of
written manuals, computerized decision support aids and tutorials are an alternative for
providing assistance in the planning of environmental field sampling projects [3]. This type
of system can make available to the planning team a sound body of knowledge compiled from
leading experts in the subject, and other recognized sources, and provide for structured
presentation in a format that delineates a natural decision pathway.
EXPERT SYSTEM FOR ENVIRONMENTAL SAMPLING
A computerized system to assist personnel in planning a sampling project, to evaluate the
relevant factors for the specific hazardous site assessment, to identify data quality
objectives, and to select the appropriate features for the sampling plan, is currently under
development. The system makes specific recommendations regarding appropriate sampling
procedures within the limitations of its current knowledge domain. These recommendations are
based on information elicited from the user through consultations with the computer. The goal
is to assist the user in devising a sampling plan that will attain the project DQOs with
appropriate technical, statistical, economic, procedural, and safety considerations.
This computer program falls within the category of expert systems [4] . Expert systems can be
defined, in general, as computer programs that incorporate the knowledge of, and simulate the
decision-making processes of, human experts in order to achieve a high-level of performance for
a particular task. Expert systems contain a body of knowledge in the form of facts and
decision rules (a knowledge base) and a mechanism for attaining logical conclusions from that
knowledge (an inference engine) The system is able to query the user for relevant
information, and recommend appropriate solutions, based on the stored knowledge base.
Preliminary studies revealed that the enormous scope of the overall sampling problem for a
Superfund study hampers the probability of success of any attempt to develop a comprehensive
initial expert system. There is a broad range of tasks (e.g., DQO determination, methodology
selection, and safety considerations), analytes (e.g., volatiles, semivolatiles, pesticides,
326
-------
inorganics, and radioactive materials), and matrices (e.g., air, water, soil, and biota)
involved. All of these variables must be considered to define and undertake a sampling plan
[5].
This paper describes a pilot expert system that is being developed to provide "intelligent"
assistance in the various aspects of soil sampling, with emphasis on determining the inorganic
target analytes included in the EPA Contract Laboratory Program (CLP). Soil was chosen because
environmental assessments involve soil sampling on a frequent basis and, compared to aqueous or
aerial systems, soil is a relatively static medium. The need to deal in this first expert
system with the complications associated with more active transport phenomena has thereby been
avoided.
The expert system prototype addresses the issues involved in sampling for inorganic target
analytes, an area for which procedures are commonly accepted. A real need for guidance,
expertise availability, and common acceptance of the proposed recommendations are key elements
for the feasibility of developing a successful expert system.
CONCEPTUALIZING THE SAMPLING DESIGN
Sampling issues, as well as analysis issues, must be addressed early during the process of
determining the DQOs. This is to ensure that the selected sampling scheme is in accordance
with DQO requirements. The main characteristics defining a given sampling process are
precision, accuracy, and representativeness and their contribution to the total error
acceptable in the decision. The details of the sampling procedure should be designed to
achieve the levels of these parameters defined by the project DQOs [6] .
Both user factors and site factors determine the design of the sampling plan. Figure 1 depicts
the interdependencies among the several factors involved. Analytes of interest, intended use
of the data, and resources available are summarized in the statement of the DQOs, which in turn
determine the analytical method requirements. The source, type, and extent of contamination,
as well as field conditions, complement the input for preparation of a sampling plan. Based on
information about these aspects, the expert system provides advice on a number of categories
that form part of the sampling plan. These categories are: requirements for attainment of
statistical confidence and representativeness (e.g., sample size, number, and location);
quality assurance and quality control (QA/QC) measures; sampling procedures, techniques, and
equipment; sample handling and shipping; documentary procedures; archiving and retrieval;
resource requirements; and safety measures [7].
SCOPE OF THE SOIL SAMPLING EXPERT SYSTEM PROTOTYPE FOR METAL CONTAMINANTS
The domain of applicability of the expert system prototype was constrained in order to allow a
clear identification of sampling situations that it is intended to address, and those
situations that fall outside of its capabilities.
The contaminants addressed are the EPA CLP inorganic target analytes. Cyanide is included in
this list together with the metal contaminants. The type of contamination source considered is
surface-level or fallout (e.g., from stack emissions). CLP inorganic target analytes in soil
have variable migration and solubility rates, although the majority of the analytes may have a
slow migration rate unless a transport medium is active. For the purpose of this prototype,
sampling is performed close to the surface for surface-level contamination sources. Also for
prototype and sampling purposes the site is assumed to be a flat, rectangular area, with
uncomplicated site and medium characteristics. Several zones with different contamination
levels or distributions can be managed.
The driving factor for the expert system is the statement of the data quality objectives for
the project. The EPA-defined DQO process requires definition of the decision to be made and
327
-------
CO
to
00
INTENDED
USE OF
DATA
RESOURCES
AVAILABILITY
ANALYTES
OF
INTEREST
DATA
QUALITY
OBJECTIVES
ANALYTICAL
METHODS
SAFETY
MEASURES
DOCUMENTARY
PROCEDURES
BUDGET
REQUIREMENTS
STATISTICAL
DESIGN IK
HANDLING
QA/QC
PROCEDURES
SAMPLING
TECHNIQUES
SITE AND MEDIUM I
[CHARACTERISTICS!
Figure 1. SAMPLING DESIGN INTERDEPENDENCES
-------
the data needed to support the decision; these identifications are made in Stages I and II of
the process. The expert system assumes that these stages have been completed and then proceeds
to gather summary information about the quantitative aspects of the decision. Some aspects are
a description of the the data to be collected, the domain (spatial and temporal) for the
decision, the statistics to be used to describe the data, and the statement of desired
performance In terms of acceptable probabilities associated with false positive and false
negative situations. This information is entered into the expert system and used by subsequent
modules to help complete Stage III of the DQO process, namely the design of the sampling aspect
of the data collection program.
The system makes suggestions concerning the sample analysis technique that is appropriate to
meet the data quality objectives. Applicable analysis techniques for EPA CLP inorganic target
analytes [8] are inductively coupled plasma-optical emission spectroscopy (ICP-OES) and atomic
absorption spectroscopy (with a cold vapor technique for mercury and a colorimetric technique
for cyanide) [9]. X-ray fluorescence spectroscopy provides field screening of metal
contaminants [10]. The expert system makes recommendations based on the level of data quality
required as determined by the DQO module.
The expert system guides the user through a site and medium (soil) characterization process.
Information is obtained regarding the physical and ecological nature and configuration of the
area and site; animal and vegetation characteristics; surface and groundwater factors; weather
and climate data; and identification of manmade structures, activities, and perturbations. The
prototype requests information about soil physical and chemical properties which help
characterize the "soil contamination system," such as grain size (texture), structure,
cohesiveness, moisture contents, color, odor, and various determinations obtained with field
monitoring equipment.
The statistical design module finds the minimum number of samples that should be collected for
a given set of DQOs. The procedures and underlying assumptions for determining the number of
samples and sampling design are outlined in Rogers et al. [11]. The present system uses
classical statistical methods. Future designs will also include geostatistical methods to
address the experimental objectives. Preliminary data are required for the calculations. The
number of samples recommended for statistical analysis depends on the experimental objectives,
the accuracy to be achieved, the nature of the pollution source, the expected distribution of
the pollutant, and logistical and budgetary constraints [12, 13]. For the prototype system,
four objectives have been identified: (1) finding the average pollutant concentration at the
site, (2) finding the spatial variability of the pollutants at the site (e.g., identifying "hot
spots"), (3) comparing the pollutant concentrations of different subareas of the site, and (4)
finding the temporal variability of the pollutants at the site. The accuracy to be achieved is
related to two types of errors that can occur in decision making, namely false negative and
false positive determinations. The nature of the contamination source (point or nonpoint) and
distribution of the contaminant (homogenecity and continuity) are also taken into account for
the design. The system accommodates the condition where concentrations may vary in subareas,
or strata, at the site due to differences in soil type, topography, or other physical
characteristics. Samples may be taken at points in a rectangular grid or at random coordinates
at the site. The system produces a listing of the local coordinates where the samples are to
be taken.
The system provides recommendations for a rigid QA/QC program to statistically evaluate the
quality of the data at each step of the analytical process and to make sure that the quality of
the data is adequate for the purposes for which the data will be used [14, 15]. The objectives
of the QA/QC procedures are to measure and obtain (1) accuracy, (2) comparability, (3)
precision, and (4) representativeness of the data. These characteristics are measured and
monitored by collecting and analyzing the appropriate number of blanks, standards, and
duplicates along with the actual environmental samples.
Recommendations are provided for the appropriate sampling tools and equipment and method of
sampling. Depending on the depth of the contaminants and soil characteristics, various types
329
-------
of equipment may be recommended, such as a trowel, scoop, auger, drive tube, and split spoon
[5,16,17,18]. The system also makes suggestions on the equipment material and decontamination
procedures.
Depending on the contaminants under consideration and the analytical technique to be employed,
the expert system provides advice on the appropriate sample handling techniques [19] (i.e.,
sample manipulation, container type, preservation, shipping, and holding times).
The user is also queried on expected and potential personnel hazards on the site. The expert
system then classifies the situation accordingly, and provides details on the required level of
personal protection as well as general safety measures and equipment needed [20, 21].
The system can supply information on standard documentation and chain-of-custody procedures.
These include sample identification, logbooks and records for samples, instrument calibration,
QA/QC, and chain-of-custody records [18]. The rationale behind each documentary procedure is
explained to the user, and lists are given with the required items of information for each
form.
The primary consideration in budgetary requirements is to meet the user's needs according to
the DQOs. Secondary considerations include, but are not limited to, safety procedures, site
conditions, climate and weather conditions, and sampling techniques. In response to input
about cost elements for sampling (e.g., travel, salary, rentals, etc.), the number of samples
(including QA/QC samples) provided by other modules, and sampling throughput estimates, the
expert system provides estimates for required dollars, personnel, and time for sampling.
SYSTEM DESIGN AND DEVELOPMENT
Most currently available computer programs that can be used in conjunction with sampling fall
into the category of data analysis aids. Many are used after the data have been gathered and
thus flaws are discovered after the fact and are costly to correct. Planning tools that offer
simulation based on statistical techniques handle numerical data only. Blind automation of
simulation, or analysis processes, introduces the risk of lending creditability to conclusions
that have been reached by applying the wrong procedures.
The expert system described in this paper handles both hard data (i.e., facts and performance
numbers) and soft data (e.g., judgments, risks, projections, intuition, and human factors).
Mathematical capabilities to handle statistical aspects of the problem are combined with
symbolic reasoning to make rule-based decisions on what sampling procedures to use. Table I is
an example of simple inference rules.
Expert system techniques are used to infer the appropriate procedures to be used in each
sampling situation. The expert system contains knowledge in the form of rules that will allow
it to search for appropriate recommendations based on facts about the sampling problem elicited
from the user (Figure 2). The knowledge entered into the expert system knowledge base has been
drawn from EPA-published manuals and other Federal agency (USDOE and USDA) and reports, peer
reviewed literature, and from interviews with environmental scientists experienced in soil
sampling. This knowledge, based on experience in both field and laboratory, is mostly in the
form of "rules of thumb" for decision making in the planning of sampling projects, of numerical
formulas, and of procedures for statistical calculations.
The IBM-PC-compatible microcomputer was used as both development and delivery environments
because it provides availability and portability advantages. The computer program is written
within an expert system development shell. The selected software package is KnowledgePro
(Knowledge Garden, Inc., Nassau, New York), a knowledge-processing software package that
combines expert system and hypertext capabilities [22]. Hypertext [23] is a method of
presenting computerized knowledge. Preselected words and phrases on the computer screen can
be highlighted and additional explanations requested about them. Figure 3 gives an example of
330
-------
DOMAIN
EXPERT
KNOWLEDGE
ACQUISITION
KNOWLEDGE
ENGINEER
I
KNOWLEDGE
REPRESENTATION
KNOWLEDGE
BASE
USER
INTERFACE
INFERENCE
ENGINE
CONSULTATION
USER
Figure 2. EXPERT SYSTEM DYNAMICS
331
-------
IF
AND
AND
AND
AND
AND
THEN
IF
AND
AND
SAMPLING DEPTH IS MORE THAN 0.6 METERS
SAMPLING DEPTH IS LESS THAN OR EQUAL TO 4.9 METERS
VERTICAL PROFILE NEEDED IS POSITIVE
VERTICAL PROFILE TYPE IS CONTINUOUS
IS NOT COARSE
IS MEDIUM OR FIRM
SOIL TEXTURE
SOIL COMPACTNESS
SAMPLING TOOL TYPE
IS VEIHMEYER SAMPLER
POLLUTION SOURCE DISTRIBUTION
CONTAMINANT DISTRIBUTION
HETEROGENICITY TYPE
THEN SAMPLING DESIGN TYPE
IS NONPOINT
IS HETEROGENEOUS
IS DISCRETE
IS STRATIFIED SAMPLING
Table I. EXAMPLE RULES
What ia the texture of the soil in the site?
COARSE
FINE
Soil
refers to the relative
proportion of the various size groups
of individual soil grains in a soil
mass. It is classified as
medium, and fine in this system.
soil is that containing particles of over
0.50mm in size according to the U.S. Soil
Conservation Service.
Figure 3. HYPERTEXT SCREENS EXAMPLE
332
-------
the use of hypertext technique. This technique allows the user to control the amount and depth
of the information that will be presented during the interactive consultation with the
computer. This feature is very desirable in a system that is intended for users with a widely
varying background or experience in terms of knowledge about planning and conducting sampling
projects. This software tool has the capability of combining decision rules for inference with
user input to control the system, all within a user-friendly environment. Turbo Pascal
routines were interfaced with KnowledgePro to provide a. more efficient handling of calculation-
intensive tasks (e.g., statistical calculations).
The knowledge was organized in a frame-based structure (see Table II) to allow the application
of an object-attribute-value scheme for knowledge representation and reasoning throughout the
system. The program structure was modularized to overcome memory and speed limitations.
CURRENT STATUS AND FUTURE DEVELOPMENT
A demonstration prototype for the sampling expert system (soil metals) has been completed and a
full prototype is under development. All the modules described in this paper have been
implemented and work continues on developing them further. Phased testing and validation are
beginning. Testing will include in-house and third party evaluations. The work described is
being done under contract to the EPA Environmental Monitoring Systems Laboratory, Exposure
Assessment Research Division, Las Vegas, Nevada.
SUMMARY AND CONCLUSIONS
The availability of a decision support and training aid, with appropriate coverage of the
issues involved in a sampling plan, including statistics, QA/QC, sampling techniques and
procedures, and safety considerations, will provide wide access to more reliable and replicable
sampling plan development. This tool should result in the generation of data of acceptable
sampling quality, with quantifiable variability possible and bias due to sampling errors and
deficiencies; a tracking system for sample contamination; increased representativeness; a more
efficient use of the usually limited sampling resources; and a greater reliability in
interpretation of results. It enables the researcher to focus on the primary objectives of the
study.
The pilot expert system ensures valid data analysis and a quantification of the quality of data
at each step, thus conserving expensive sampling and analytical resources while shortening the
time for the acquisition of data of known quality to support environmental decisions. The
pilot expert system described herein will also serve as a limited training aid for individuals
participating in soil surveys of contaminated sites. This type of system allows the user to
try "what-if" situations with information regarding site conditions and the desired confidence
in the results. The use of the system to plan sampling activities and to validate proposed
sampling plans will enable the users to detect potential errors and data deficiencies before
they occur, thereby saving time and resources and providing greater reliability for the survey
results.
The preliminary experience obtained from the development of this prototype supports the
feasibility of a larger and more detailed system. The implementation approach has been tested
and found appropriate. Changes can be readily made, and subsequently the expert system could
be continuously updated to reflect modifications and advances in environmental sampling.
ACKNOWLEDGMENTS
The authors would like to acknowledge David W. Bottrell (EPA Environmental Monitoring Systems
Laboratory, Quality Assurance Research Branch, Las Vegas) for his support of the preliminary
studies for the system, Kelly R. York, Lockheed Engineering & Sciences Company (LESC) for
333
-------
OBJECT/ATTRIBUTE:
SOIL/TEXTURE
VALUE:
PROPERTIES:
REFERENCE:
COARSE
MEDIUM
FINE
INPUT FACT
SOIL DATA COLLECTION IN SUPPORT OF
EMERGENCY RESPONSE ACTIVITIES
(PAGES 5-3 5-8)
VALUE:
DEFINITION:
CONDITION:
VALUE:
DEFINITION
CONDITION:
VALUE:
DEFINITION:
CONDITIONS:
COARSE
COARSE SOIL IS CONTAINING PARTICULES OF
OVER 0.50 mm IN SIZE ACCORDING TO THE US
SOIL CONSERVATION SERVICE
A COARSE TEXTURE SOIL ACTS AS A CONDUIT
FOR WATER AND ANY DISSOLVED CHEMICALS
FOUND IN THE WATER. LIQUID CHEMICALS
WITH A LOW VISCOSITY WILL OFTEN FLOW
THROUGH THE COARSE TEXTURED SOILS
MEDIUM
MEDIUM SOIL IS SOIL CONTAINING PARTICULES
BETWEEN 0.25 - 0.50 mm IN SIZE ACCORDING
TO THE US SOIL CONSERVATION SERVICE
A MEDIUM TEXTURED SOIL WILL INHIBIT
CHEMICAL MIGRATION OF LIQUID CHEMICALS
WITH LOW VISCOSITY
RNE
FINE SOIL IS SOIL CONTAINING PARTICULES
LESS THAN 0.25 mm IN SIZE ACCORDING TO
THE US SOIL CONSERVATION SERVICE
A FINE SOIL WILL INHIBIT MIGRATION OF MANY
LIQUID CHEMICALS
Table II. EXAMPLE FRAME
334
-------
programming the expert system, and Luis A. Herrera (LESC) for his help during the initial
stages of knowledge acquisition.
REFERENCES
(1) "Development of Data Quality Objectives," USEPA, Quality Assurance Management Staff,
Washington, D.C., 1986.
(2) "Data Quality Objectives for Remedial Response Activities. Development Process,"
Document Number EPA/540/G-87/003, USEPA, Office of Emergency and Remedial Response,
Washington, D.C., March 1987.
(3) Keith, L. H. ; Johnston, M. T.; Lewis, D. L.; "Defining Quality Assurance and Quality
Control Sampling Requirements: Expert Systems as Aids," Principles of Environmental
Sampling, L. H. Keith, Ed., American Chemical Society, Washington, D.C., 1988, pp. 85-100.
(4) Waterman, D. A.; "A Guide to Expert Systems," Addison-Wesley, Reading, Massachusetts,
1986.
(5) "Sampling for Hazardous Materials," USEPA, Office of Emergency and Remedial Response,
Washington, D.C., 1988.
(6) "Data Quality Objectives for Remedial Response Activities. Example Scenario: RI/FS
Activities at a Site with Contaminated Soils and Ground Water," Document Number
EPA/540/G-87/004, USEPA, Office of Emergency and Remedial Response, Washington, D.C.,
March 1987
(7) Mason, B. J.; "Soils Data Collection in Support of Emergency Response Activities," USEPA,
EMSL-Las Vegas, Nevada, March 1983.
(8) USEPA Contract Laboratory Program, Statement of Work-Inorganic Analysis, Multi-Media
Multi-Concentration, Washington, D.C., July 1987.
(9) Aleckson, K.A.; Fowler, J.W.; Lee, Y.J.; "Inorganic Analytical Methods Performance and
Quality Control Considerations," Quality Control in Remedial Site Investigation:
Hazardous and Industrial Solid Waste Testing, Fifth Volume, ASTM STP 925, C.L. Perket,
Ed., American Society for Testing Materials, Philadelphia, 1986, pp. 112-123.
(10) Raab, G. A.; Cardenas, D.; Simon, S. J.; Eccles, L. A.; "Evaluation of a Prototype Field-
Portable X-Ray Fluorescence System for Hazardous Waste Screening", Proceedings of the
USEPA Third Annual Symposium on Solid Waste Testing and Quality Assurance, Vol. II,
Washington, D.C., July 1987.
(11) Rogers, J., et. al., "Statistical Methods for Evaluating the Attainment of Superfund
Cleanup Standards. Volume 1: Solids and Solids Media," USEPA, Statistical Policy
Branch, Washington, D.C., February 1988 (DRAFT).
(12) Flatman, G. T.; "Design of Soil Sampling Programs: Statistical Considerations," Quality
Control in Remedial Site Investigation: Hazardous and Industrial Solid Waste Testing,
Fifth Volume, ASTM STP 925, C. L. Perket, Ed., American Society for Testing Materials,
Philadelphia, 1986, pp. 43-56.
(13) Garner, F. C.; Stapanian, M. A.; Williams, L. R. ; "Composite Sampling for Environmental
Monitoring," Principles of Environmental Sampling, L. H. Keith, Ed., American Chemical
Society, Washington, D.C., 1988, pp. 363-374.
335
-------
(14) Earth, D. S.; Mason, B. J.; "Soil Sampling Quality Assurance User's Guide," Document
Number EPA 600/4-84-043, USEPA, EMSL-Las Vegas, Nevada, May 1984.
(15) Bruner, R. J., Ill; "A review of Quality Control Considerations in Soil Sampling," Quality
Control in Remedial Site Investigation: Hazardous and Industrial Solid Waste Testing,
Fifth Volume, ASTM STP 925, C. L. Perket, Ed., American Society for Testing and Materials,
Philadelphia, 1986, pp. 35-42.
(16) Mason, B. J.; "Preparation of Soil Sampling Protocol. Techniques and Strategies,"
Document Number EPA/600/4-83/020, USEPA, EMSL-Las Vegas, Nevada, August 1983.
(17) "Characterization of Hazardous Waste Sites--A Methods Manual. Volume II. Available
Sampling Methods," Document Number EPA-600/4-83-040, USEPA, EMSL-Las Vegas, Nevada,
September 1983.
(18) "The Environmental Survey Manual," Document number DOE/EH-0053, USDOE, Office of the
Assistant Secretary-Environment, Safety, and Health and Office of Environmental Audit,
August 1987.
(19) Cameron, R. E.; "Soil Homogenization," USEPA, EMSL-Las Vegas, Nevada, August 1986.
(20) "Hazardous Materials Incident Response Operations," USEPA Office of Emergency and Remedial
Response Hazardous Response Support Division, 1983.
(21) "OSHA 40-Hour 29 CFR 1910.120 Personnel Protection and Safety Course Manual," Hazco,
Inc., Dayton, Ohio, 1988.
(22) KnowledgePro User Manual, Version 1.0, Knowledge Garden, Inc., Nassau, New York, 1988.
(23) Barrett, E., Ed.; "Text, Context, and Hypertext," MIT Press, Boston, Massachusetts, 1988.
APPENDIX CASE EXAMPLE
This appendix contains a narrative description of a fictitious hazardous site and presents an
example of a soil sampling problem for demonstration purposes. This example gives the reader
an idea of the type and detail of both input and output information handled by the expert
system prototype described in this paper. The recommendations produced by running a
consultation session with the Expert System for Soil Sampling Prototype follow the description.
SITE AND PROBLEM NARRATIVE DESCRIPTION
An abandoned storage yard for a battery reprocessing plant in Gray City, Nevada, has been
characterized in a preliminary study. Analytical level II quality data on the concentration of
lead at the site was gathered on-site with the use of a portable X-ray fluorescence (XRF)
instrument. Analyses were performed for other suspected metals, but none were found at any
significant concentration.
Stages I and II of the DQO process have been completed. The decision to be made by the state
authority is whether or not the site is in compliance with a percentile-based regulation. It
has been proposed to determine if a pre-established percent of collected and analyzed samples
exceed certain criterion for lead concentration. For this purpose it is necessary to gather
data on the concentration of this contaminant in the soil. The required data type is the
concentration of lead in the soil in parts per million (ppm). State authorities have
established the action level for lead at 1,000 ppm and the percentile testing threshold at 75
percent (it is to be determined if more than 25 percent of the samples have lead concentrations
336
-------
above 1000 ppm or not) . The true proportion of samples exceeding the action limit for which
the site can be reliably declared as clean is 15 percent. A false negative rate of 10 percent
and a larger false positive rate of 20 percent have been established for the determination as
tolerance limits.
Contamination is known to have begun at least two years ago. Knowledge about the vertical
migration of lead at this site is not available, but the migration rate in similar situations
have been determined to be almost negligible. Thus the contaminant should be found no deeper
than 6 inches, as confirmed during preliminarily portable XRF measurements.
The site has an area of 5,000 square meters. It is located in an arid area with clear skies
and negligible to light winds almost year around. The soil is low in moisture content and
texture is medium (particulates between 0.25 mm and 0.50 mm dia.). Soil compactness is medium.
The relief at the site is flat and there are no discernible depressions. Portions of the site
are covered with grass. Fifty-five-gallon drums marked "SULFURIC ACID" are visible at the
southwest corner of the site. Localized high levels of sulfuric acid contamination are
suspected as indicated by the distressed affects on the vegetation around the drums and acidic
soil pH measurements.
Two areas (strata) of known contamination were identified from the preliminary study by XRF.
One is 15 by 26 meters, and the second is 20 by 30 meters in size. From
preliminary data, the suspected proportion of contaminated samples greater than 1000 ppm is 90
percent for both areas. Sampling will be confined to these areas.
It is considered that one sampling team of three members is available for this sampling phase.
It is estimated that It is feasible to collect and analyze approximately 25 samples per day.
The team has determined that a sample loss rate (unanalyzable and uncollectable samples) of 5 %
can be expected. The cost components associated with the sampling team operation are as
follows:
Travel costs $ 1500
Supplies costs $ 500
Service costs $ 2000
Shipping costs $ 1500
Average salary per day $ 125
Per diem rate $ 80
EXPERT SYSTEM PROTOTYPE RECOMMENDATIONS
The level of analytical quality should be Level III (Non CLP) or Level IV (CLP-quality). An
appropriate analytical technique for analysis of lead at this level is graphite-furnace atomic
absorption. Historic performance data for the method for lead analysis are:
Precision (% RSD): 9.2%
Accuracy (% BIAS): -2.2%
Concentration Range: 11.5 714 ug/Kg
Grab sampling Is recommended. A stratified sampling design is appropriate, with 17 samples for
stratum one, and 26 samples taken for stratum two, for a total of 43 samples. This number
accounts for the expected sample loss rate due to unanalyzability and uncollectability.
The following QA/QC samples are recommended:
Field blanks 2
Sample bank blanks 1
Reagent blanks 3
337
-------
Laboratory controls 3
Calibration checks 3
Spiked samples 3
Duplicate samples 3
Total recoverable 2
It is recommended that the site be cleared of vegetation and surface litter before sampling.
The appropriate sampling tool is a scoop and/or trowel. The tools should be wrapped in
industrial-strength aluminum foil when not in use and cleaned by the USEPA standard
decontamination protocol.
On-site sample sieving is not recommended. Sample homogenization is needed. Samples should be
stored in 250 mL jars pre-cleaned by the USEPA jar-cleaning protocol C. The recommended
maximum sample holding time is six months.
Level C personnel safety protection is recommended. The decontamination location should be
outside of the sampling zone.
The estimated total cost for sampling is $ 11,640. The estimated cost per sample is $ 155.
Sampling time is estimated to be three days.
NOTICE
Although research described in this article has been funded wholly or in part by the United
States Environmental Protection Agency under contract number 68-03-3249 to Lockheed Engineering
& Sciences Company, it has not been subjected to Agency review and therefore does not
necessarily reflect the views of the Agency and no official endorsement should be inferred.
Mention of trade names or commercial products does not constitute Agency endorsement of the
product.
338
-------
DISCUSSION
TOM PRITCHETT: What you' re doing is extremely critical. We have people
at EPA in the field crying for this information and some type of guidance. We
literally need it yesterday in the field. We can't afford to wait a year or two years
to get it through approval processes. What you're doing is outstanding work.
1 just hope we can see it in three months.
ROMAN OLIVERO: We are an EPA contractor, and how to get this through
the system is up to EMSL-Las Vegas. That's why we have a phased approach
to testing. You might get it in two months, after in-house testing, or in six
months after review. The product quality of what you get will depend on the
testing level.
DON FLORY: Do you have built into this decision-making system typical
precisions for different matrices for the different analytical methods?
ROMAN OLIVERO: We are addressing metals and cyanide in a soil matrix,
and we have published typical precision and accuracy for the metals.
There are no data on sampling precision and accuracy. The other half of the coin
in still remiss, and EMSL is working toward that. But just knowing how good
your analytical method is in the lab may not be enough, your sampling error
could outweigh that. So that doesn't mean much, really.
DON FLORY: Could I change the data base, then, and put some new numbers
in?
ROMAN OLIVERO: Since we have a lack of data in many areas, we're
thinking of providing a way for people to change, update those numbers, or
provide those numbers we don't have, as they become available. It's not in the
system right now, but we have pretty much decided we will do that.
DON FLORY: Let me suggest strongly that you do that. For one example, if
we have a demonstration phase for a project, and we take a lot of samples and
get QA/QC data, and then we want to put those into the remediation plan, we
can have real data-real precision and accuracy for those kinds of samples in
those matrices. We should be able to put those in, so we can make that next step.
The other question is can we put the cost analyses in and compare data quality
objectives, then, to total cost? You're calculating all of the numbered samples
and everything we need to run, and we would need to be able to do that, also.
Thai would be very helpful.
ROMAN OLIVERO: Yes, two of the models that we're developing last are
DQO and budget, the most important ones. The techniques were there in the
documents. We plan to provide a way for users to, in an interactive way, change
their DQO's and the money available, so they can get mem to match.
DON FLORY: Did you say how big a computer, how big a memory, what kind
of computer this will run on?
ROMAN OLIVERO: It was developed on Knowledge Pro, which is an IBM-
PC based development tool for expert system. It runs on an IBM-PC or any
compatible. We have it in a self-powered portable here, which can be taken into
the field. You need 640K to run it.
Running off a hard drive is a lot faster, even though we run it off here, and it runs
on color or monochrome. You need a printer to get recommendations. They are
shown on the screen, but of course you want to print it, and you can use an
optional mouse pointing device.
DON FLORY: You have been running it off the floppies?
ROMAN OLIVERO: We have it on hard drive but it can be run off floppy
disks. There are no royalties that we need to pay to either Bolan or to
Knowledge Garden for Knowledge Pro. You get a one-time system, and you
don't need anything else in your computer, any other software, or anything to
run it. It's self contained.
DELYLE EASTWOOD: We are having a meeting to discuss an ASTM
guidance document on field operations this evening. It's a very preliminary
document. I'm sure there are a lot of mistakes in it, but at least it's a start. EPA
is also working on guidance documents. Obviously there is a need for both
expert systems and guidance documents, and I hope that there is some way to
update the data in these expert systems. I did see one or two things on your slide
that I would disagree with.
Some of the points are debatable, but I think there's a need for both guidance
documents and expert systems, and 1 hope people will come to the discussion
of this ASTM field and laboratory operations document, because we need all
the input we can get.
ROMAN OLIVERO: Yes, we need documents and expert systems. We need
the input of the people. We obtained all the documents we could, and we have
some good experts, but we're just a little tiny piece of knowledge available out
there. If you can contribute with either document references, even unpublished
information, or your own expertise somehow, please contact us. We would like
to work with you.
JONATHAN NYQUIST: Once you've gotten your recommendation, can the
user get a dump of the rules that we used to reach that decision?
ROMAN OLIVERO: We're working on implementing a feature. That's
called a Y feature. It's very hard on this tool, but definitely we do need that.
339
-------
INTRODUCTION TO THE SESSION ON
OTHER ADVANCED FIELD
TECHNIQUES
RONALD MITCHUM, CHAIRPERSON
The other techniques in this session are a very diverse group of topics, going
all the way from purge and trap, GC/MS, to short-term bioassessments, or
biomonitoring at Superfund sites, all the way down to EPA field screening
projects.
I would talk to you just a minute or two about the philosophy that I have
regarding field screening techniques. That might set the stage for these other
categories, the other techniques that maybe next year will be even bigger.
I had some definite ideas about field screening and what it should be. One of
the problems that we have is it may be fast and cheap, and not give an answer.
In field screening techniques, we need to look at the purpose for having them.
We did this so we can obtain data in the field, at a somewhat faster rate than we
could by sending it out to some laboratories that service us.We did it so we
might even have some of the same quality assurance that we have in some of
these other laboratories, so the data would have some meaning.
But what is the question that we're trying to ask in the field? In many cases, the
question is not a quantitative question. The question is a yes-no question. In
many cases if 1 could say that there is not one of the NPL compounds at this site,
we could save ourselves a lot of trouble.
But unfortunately, many of our screening techniques are very specific tech-
niques, rather than the general technique that would answer that broad question.
What's more important? To us in the EPA, zeros are very important, because
most of the numbers we get are zeros. Our ability to measure zeros is very
important. Our ability to measure gangs of zeros is even more important.
An example of that is the Love Canal habitability study. We took 1300 dioxin
samples around houses and in ditches. We analyzed all 1300, and we had one
positive hit.
That tells you something about the way that we look at samples coming out of
the field. If we had a very good way to obtain zeros, then we would have had
a very good way to do 1300 samples very effectively, and a very lousy way to
have done one.
In many cases, that's very important, and in many cases, this is the question you
might consider in the future. How do we get good zeros? What are some of the
methods for doing that?
Another philosophical issue about field screening techniques is what kind of
technologies do we put out there? We see some very good, innovative
approaches to doing field measurements - that's fieldable technology, versus
package or single-person technologies.
In many cases, we're dealing with answering the same questions in a little
different way. What's the most cost effective way for us to do that? It's more
cost effective for me to take a GC in the field in a truck than it is for me to fund
the development of a GC I can carry in my hand. You've got to consider those
kinds of things when you look at fieldable technologies.
Give me a method that will measure me a zero, and I'll make you a millionaire.
Give me a method that will measure one compound, and I'll guarantee you to
fail, because we've never asked one question at a time. We always ask more
than one question in chemistry, so you've got to always consider that, also, in
your development efforts. How many questions do we ask?
In many cases, you've also got to take a look at how and what was the history
behind the development of technologies and chemistry, and that's what we're
talking about.
In many cases, chemistry was developed for a different reason than for field
applications. So you take a look at the route of the technique, and say, now how
do I address that and make it go to the field? Don't take it as such and say, I've
got to take this $150,000 mass spectrometer to the field some way. It wasn't
designed to do that, and it's also not answering the question you want it to
answer.
341
-------
EVALUATION OF A FIELD-BASED, MOBILE, GAS CHROMATOGRAPH-MASS
SPECTROMETER FOR THE IDENTIFICATION AND QUANTIFICATION OF
VOLATILE ORGANIC COMPOUNDS ON EPA'S HAZARDOUS SUBSTANCE LIST
Albert Robbat, Jr. and George Xyrafas
Tufts University, Chemistry Department
Trace Analytical Measurement Laboratory
Medford, Massachusetts 02155
ABSTRACT
A gas chromatograph-mass spectrometer (GC-MS)
has been evaluated for the analysis of volatile organic
compounds (VOCs) on EPA's Hazardous Substance List.
The gas chromatographic separation of thirty-five VOCs
on a DB624 capillary column using ambient air as the
carrier gas was excellent. A wide linear dynamic range
for the mass spectrometer was obtained for each
compound with minimal detectable quantities (lOng) at
the level required for hazardous waste site investigation.
The GC-MS was transported in a Chevrolet Blazer for
field investigations of suspected chlorinated VOC
contamination in tap water and in groundwater. Field
and laboratory measurements are intercompared.
Key words: field gas chromatograph-mass spectrometer,
groundwater contamination, on-site field
investigations.
INTRODUCTION
In November, 1986, the citizens of Massachusetts
voted by referendum overwhelmingly to amend the
state's Superfund law and require the state's Department
of Environmental Quality Engineering (DEQE) to list by
January 1987, 400 hazardous waste sites that will be
investigated; at least 600 additional sites by January
1988, and a minimum of 1000 new hazardous waste sites
each year thereafter. The law stipulates well-defined
timetables for classification, remediation and disposition
of listed sites. For example, "priority sites" must be
fully evaluated and a permanent remedy must be
completed within four years (technology permitting), or
be temporarily secured, and/or capped. "Nonpriority
sites" must be fully evaluated and have an action plan
instituted within seven years of site listing. To meet
this most difficult and costly timetable, it is becoming
increasingly important to develop new, on-site,
analytical instrumentation which can identify and
quantify environmentally important compounds. Field
technology that can provide unambiguous identification
and quantification of chemical contaminants should
provide site managers with analytical data necessary to
make immediate decisions. Moreover, on-site decision-
making should result in improved deployment of field
personnel and equipment and should result in increased
cost savings. This is due to the fact that commercial and
state laboratories currently experience a two-to-four
month lag time between field sample collection and
production of laboratory results.
A new gas chromatograph-mass spectrometer (GC-
MS), designed specifically for field investigations, has
been evaluated for volatile organic compounds (VOC)
found on the U.S. EPA's Hazardous Substance List
(HSL). In this report, the high resolution, capillary
DB624, separation of the HSL VOCs are presented. In
addition, the linear dynamic range, the minimum
detectable quantities, and the selectivity of the field
GC-MS toward VOCs present in aqueous solutions have
been investigated. Finally, field GC-MS, laboratory
GC-MS, and laboratory GC-ECD results are
intercompared for the determination of trichloroethene
and 1,1,1-trichloroethane found in home drinking water.
EXPERIMENTAL SECTION
The Bruker Instruments GC-MS used in this
investigation has been described in detail in another
paper presented at this symposium (Trainor and
Laukien). The GC was equipped with an automated
thermal desorption oven.
In this study a 30m x 0.32mm i.d. fused silica
capillary column coated with 1.8pm film of DB624 was
used. The optimum gas chromatographic operating
condition for on-site measurements during summer
conditions was established for the separation of all HSL
VOCs simultaneously: viz.; 28 °C isothermal for seven
minutes followed by linear temperature programming at
7 "C/min to 120 °C. The final temperature was held
constant for 7 min. The carrier gas was ambient air
flowing at 1 ml/min. The mass range scanned was
between 45 amu and 260 amu with a 2 sec scan time.
The area under each VOCs primary m/z ion was used in
the quantitation calculations as specified by EPA
procedures. Data acquisition was delayed for 3.7 min
from the start of compound desorption into the GC.
The linear dynamic range was established by using
known concentrations of standard solutions from
Supelco. Six standard solutions were made (in methanol)
containing compound concentrations of:
standards 1) 1000 ng//il; 2) 4000 ng/pl; 3) 5000 ng/pl;
343
-------
4) 10,000 ng//il; internal standard 5) 2000 ng/fil full
scan MS; and 6) 4000 ng/jd SIM. Standard solutions
were injected into a tube containing 150 mg of Tenax.
The tube was placed into the GC desorption oven
maintained at 220 °C for 45 sec. The actual GC
experiment began immediately after the 45 sec sample
desorption period. Standard solutions were serial diluted
and appropriate quantities injected into a new tube. The
procedure was continued until GC-MS signals resulting
from these compounds were no longer observable.
Figure 1 illustrates the sample preparation technique
employed in the drinking water experiments for VOC
analysis. The water solution was purged with ambient
air using a Gillian (model 513) air sampling pump with
the VOCs trapped in a tube. Eighty ml drinking water
samples were used (sparged for 5 min at 0.5 1/min) with
the VOCs collected on 150 mg of 50/50 Tenax/charcoal.
Figure 1.
Purge and Trap Sanpllng Veasel.
Home drinking water investigations were performed
in cooperation with DEQE and the town's water
department. Personnel from DEQE collected the tap
water according to standard practice. The drinking
water samples were analyzed by us (field GC-MS), town
(laboratory GC-ECD), and DEQE's Lawrence
Experiment Station (laboratory GC-MS). DEQE
performed VOC analyses of the water samples as
prescribed by EPA method 524. Headspace (volume 1
cc) GC-ECD experiments were performed by town
personnel on an instrument manufactured by Analytical
Instrument Development Inc. linked to an integrator. A
6 ft x 1/8 in stainless steel column containing 1% SP1000
was maintained at 185 °C. The carrier gas consisted of
5% methane in argon at a flow rate of 40 ml/min.
Sample preparation consisted of agitating a 30 ml aliquot
of the tap water sample for 1 minute before analysis.
Since the GC-MS and sample sparging technique
utilized air from the site as the carrier gas, background
signals at the site were evaluated by performing GC-MS
experiments on Tenax/charcoal collected samples from
80 ml of purified water (blank) and standard solutions
containing known concentrations of chlorinated VOCs in
80 ml of purified water at the beginning, middle, and
end of the days' experiments. The GC-MS operating
conditions and sample preparation procedures have been
described above with the exception of the GC column
temperature employed. The GC oven was maintained at
30 °C for 11.7 min, linearly ramped 10 °C/min to a final
temperature of 124 °C. The final temperature was held
constant for five minutes. The data acquisition delay
time was 2.8 min.
RESULTS AND DISCUSSION
Well-defined GC-MS operating conditions (see
experimental section) were established keeping in mind
the need for developing on-site analytical tools capable
of identifying the wide molecular diversity of the VOCs
on EPA's hazardous substance list. Figure 2 illustrates
the high resolution gas chromatographic separation of 31
of the 35 VOCs. The peak numbers correspond to
compound numbers identified in Table 1. The
concentration of each compound injected was 200 ng.
The excellent chromatographic separation was obtained
on a 30m x 0.32mm i.d. fused silica capillary column
coated with a l.S^m film of DB624. Although some of
the compounds coeluted, all of the VOCs can be
differentiated based on their mass spectrum.
Figure 2.
Field OC-MS of a mettanol solution containing
a standard mixture of volatile 1ISL
compoundson a DB624 capillary colum.
Peak numbers correspond to compounds
listed In dynamic range table.
The GC-MS full scan linear dynamic range toward
the VOCs was evaluated within wide concentration
ranges by preparing standard solutions containing
different initial concentrations as specified in Table 1.
The selected ion monitoring (SIM) linear dynamic range
for some chlorinated VOCs revealed that the dynamic
range could be extended by a factor of about ten.
Generally, the full scan MS detection limits are within
the useful analytical range for hazardous waste site
assessment. The minimum detectable quantity (at
S/N=3) for identification purposes was between lOng
and 40ng of compound injected for full scan MS and
344
-------
Table 1. HSL VOC Identity and Linear Dynamic Range for Field GC-MS.
# Compound
full scan MS(ng)
SIM (ng)
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
1.
2.
3.
Chloromethane
Bromomethane
Vinylchloride
Chloroe thane
Methylene chloride
Acetone
Carbon Disulfide
1 , 1 -Dichloroethene
1 , 1 -Dichloroethane
1,2 Dichloroethene
Chloroform
1 ,2-Dichloroe thane
2-Butanone
1,1,1 -Trichloroethane
Carbon tetrachloride
Vinyl acetate
Bromodichloromethane
1 , 1 ,2,2-Tetrachloroethane
1 ,2-dichloropropane
trans 1,3-Dichloropropene
Trichloroethene
Dibromochloromethane
1 , 1 ,2-Trichloroethane
Benzene
cis- 1 ,3-Dichloropropene
2-Chloroethyl Vinyl Ether
Bromoform
2-Hexanone
4-Methyl-2-pentanone
Tetrachloroethene
Toluene
Chlorobenzene
Ethylbenzene
Styrene
Total Xylenes
internal Standards
Bromochloromethane
1 ,4-difluorobenzene
Cl-benzene-d5
initial concentration (ng//jl) of
al) 1000
b2) 4000
C3) 5000
d4) 10000
e5) 2000 (internal standard)
f6) 4000
100-10,000d
100-10,000d
100-10,000d
100-10,000d
40-400a
40-4,000b
40-4,000b
40-400a
40-1,0003
100-l,000b
40-1000a
40-1000b
40-4,000b
100-1000b
100-10003
did not quantify
40-5000C
100-5000C
40-1000a
40-5000C
40-10003
40-lOOOa
40-1000a
40-1000b
40-5000C
did not quantify
400-5000°
40-4000b
40-4,000b
40-1000a
40-5000C
40-4003
40-5000C
did not quanitify
did not quantify
20-20006
20-20006
20-20006
standard solutions before
0.999
0.994
0.999
0.996
0.998
0.993
0.975
0.997
0.996
0.999
0.996
0.999
0.991
0.999
0.997
0.999
0.998
0.990
0.998
0.983
0.995
0.997
0.999
0.997
0.999
0.976
0.990
0.983
0.998
0.998
0.996
0.998
0.998
0.991
dilution
10-1000f
4-1000f
10-1000f
4-1000f
10-1000f
4-1000f
10-1000f
0.998
0.999
0.998
0.995
0.999
0.996
0.998
4-1000f
0.996
345
-------
about twenty-five times more sensitive by SIM. The MS
displayed a linear response within the concentration
ranges studied. It should be recognized that the upper
limits and minimum detectable quantities are strongly
dependent upon the GC operating conditions and
primary ion MS signal strength. For some compounds,
the upper limit was restricted due to poor
chromatography. Because HSL compounds represents a
target list of environmentally important organic
compounds having widely different molecular structures,
such diversity of compounds will have varying
selectivities on a given stationary phase under differing
experimental conditions. The dynamic range data was
based on a single GC-MS operating condition for all of
the VOCs studied (versus a number of operating
conditions optimized for various compounds, e.g.,
chlorinated hydrocarbons or gasoline constituents,
benzene, toluene, and xylenes) and was specifically
optimized for on-site, field investigations which might
be performed in varying climatic conditions.
The GC-MS response factor, RF, and percent
standard deviation for twelve chlorinated VOCs
problematic in contaminated waters were determined.
The average value reported in Table 2 was calculated
from experiments performed at five different
concentrations between the range of 20 and 200/j.g/l.
Each set of experiments was repeated three times.
Response factors were calculated using 50 ppb of
1,4-difluorobenzene as the internal standard. The data
reveal that the GC-MS responded within the error
tolerances accepted for commercial laboratory analysis.
Figure 3 illustrates a typical GC-MS total ion current
chromatogram of a standard chlorinated VOC mixture
containing 100 jig/1 of compound sparged from purified
water.
Table 2.
Average Response Factor and Percent
Standard Deviation for Some Chlorinated VOCs.
#
L
i
3.
4.
5.
6.
7.
8.
9.
1C.
11
12.
Compound
vinyl chloride
14-dkhloroetbene
methylene chloride
trans-l^dichloroethene
14-dichloroethane
ds-l^-dichloroethene
chloroform
144-trfchloroethane
carbon tetrachloride
1^-dJchloroethane
trichloroethene
tetrachloroethene
RF
0.042
0302
0452
0392
0337
0.418
0271
O109
0062
0432
O429
O722
%SD
1&5
93
15.7
12.7
2LO
12.6
144
1Z1
2Z6
163
155
17.0
M-difluorobenzene internal standard, IS
Ttafl chloride
M-dlchloroclhim
mrthf lent eWorld*
L
L
S.
4.
S. IJ-dlchloroeihuie
«. di-U-dlchloratUMM
7. chloro/orm
(. l.l,14ricJUorocUiiiM
». ariMnumchloHdt
19, U-dJchlororUunt
IL (richloTMIbcnc
U. kbwhlorotlhaw
JL
Figure 3.
GC-MS response toward a standard
mixture of twelve volatile chlorinated
HSL compounds.
The GC, MS, and column (over several months)
exhibited remarkable stability and consistently excellent
separation and quantitation. These characteristics were
obtained despite the use of semipurified ambient
(charcoal filtered vs. cylinder) air as the carrier gas and
the drastically changing climatic conditions, e.g., rain,
high humidity and temperature.
On-site evaluation of the mobile GC-MS was
performed in cooperation with DEQE personnel,
homeowners having chlorinated VOC contamination in
their drinking water, and the town's chemist. Drinking
water is supplied by each homeowner's well. Based on
data from an earlier study, DEQE field personnel
collected samples (July. 1988) from about a dozen homes
anticipated to have chlorinated VOC concentrations
above the maximum allowable levels for trichloroethene
(TCE) and 1,1,1-trichloroethane (TCEA). Split samples
were analyzed by DEQE's laboratory (GC-MS), the
town's laboratory (GC-ECD), and by the authors. The
results are presented in Table 3. The trichloroethene
intercomparative measurements are in excellent
agreement. The comparative measurement for samples in
which we detected TCEA was quite good however, we
did not detect TCEA in every sample that the laboratory
techniques identified this compound in. The
nondetection of TCEA in some samples was quite
surprising and unexpected since concentration levels
were approximately the same as those samples found in
the other samples. Moreover, the TCEA dynamic range
encompassed the concentration levels of TCEA in the
samples. Further studies with the field GC-MS at the
laboratory involving TCEA revealed no apparent reason
why we should not have detected this compound in
homeowner samples. Figure 4 represents a typical GC-
MS total ion current chromatogram of the tap water
obtained from one of the homes investigated (DEQE
sample #18). Inspection of the figure illustrated the
346
-------
Table 3.
Comparison of Field and Laboratory
Measurements for the Analysis of Trichloroethene
and 1,1,1-Trichloroethane (PFB).
TRICHLOROETHENE
field measurements suggests that this instrument can be
used to perform site characterization studies of VOCs in
aqueous solutions and in air. Work is in progress to
assess the Bruker GC-MS capabilities for HSL VOCs
present in soil and sludge. Studies related to
semivolatile organic compounds in various matrices have
been initiated.
DEQE*
15
ISA
18
20
21
22
24
36
37
38
39
53
Field
GC-MS
27
ND
-48
6
31
21
5
2
12
25
10
ND
Lab
GC-MS
27
ND
45
4
28
14
8
2
12
27
10
ND
Lab
GC-ECD
30
<1
51
5
33
17
10
2
13
30
10
<1
Ttojl chloride
1,1-dlchloroHbene
mnhjttriK eWorld*
tr»nj U-dlcWorortbrat
L
J.
carbon MncMoridl
I*. 1,2-dlchlonKlluuit
IL trlehlonxlhnw
11. ktradikrTKlhait
dj-ljH)lthkproetbra«
chloroform
I A JL
1,1,1-TRICHLOROETHANE
DEQE*
15
18A
18
20
21
22
24
36
37
38
39
53
Field
GC-MS
8
ND
40
ND
12
9
ND
ND
ND
ND
ND
ND
Lab
GC-MS
9
ND
49
<1
8
4
5
<1
5
8
4
ND
Lab
GC-ECD
8
<1
40
1
10
5
3
<1
18
8
4
<1
presence of additional chlorinated hydrocarbons:
1,1-dichloroethene (2 ppb), 1,1-dichloroethane (14 ppb),
cis-l,2-dichloroethene (8 ppb), and tetrachloroethene (4
Ppb).
The field GC-MS measurements allowed decision-
makers to make same day recommendations (results from
the laboratory were obtained September, 1988). It is
evident from the data presented (see also reference 2)
that the field GC-MS instrument, under the
experimental conditions specified, is well-suited for
identifying and quantifying HSL VOCs on-site. The
overall remarkable agreement between laboratory and
Figure 4.
GC-MS of VOCs found In tap water,
DEQE smple #18.
ACKNOWLEDGEMENT
The authors thank Bruker Instruments for use of the
GC-MS instrument and Chevrolet Blazer as well as the
many technical discussions related to this work. The
authors also thank Lynn Chappel, DEQE, for
coordinating the site investigation, and Alan Khafkart,
Littleton Water Department, for access to their data.
The authors greatly appreciate the help and interest of
many other DEQE personnel for allowing us to
participate in this and other environmentally important
site investigations. Without their help it would have
been nearly impossible to gain access to appropriate
sites.
REFERENCES
1. Trainor, T.M. and Laukien, F.D. "Design and
Performance of a Mobile Mass Spectrometer for
Environmental Field Investigations", Field Screening
Methods for Hazardous Waste Site Investigations, First
International Symposium, October 11-13, 1988, Las
Vegas, Nevada.
2. Robbat, A., Jr. and Xyrafas, G. "On-Site Soil
Gas Analysis Of Gasoline Components Using a Field Gas
Chromatograph-Mass Spectrometer", Field Screening
Methods for Hazardous Mobile Waste Site Investigations,
First International Symposium, October 11-13, 1988, Las
Vegas, Nevada.
347
-------
DISCUSSION
JOE SOROKA: How did you define your minimum detection limits?
AL ROBBAT: The minimum detection limit was simply determined by doing
the dilution series and getting to the point where we could no longer observe
the signal at a signal-to-noise ratio of three to one. What that means is that those
points did not necessarily fall on the dynamic range. We use it for two different
purposes. One is for screening and secondly to determine presence or absence.
JOE SOROKA: So the detection limits you have listed were three times the
signal to noise, whereas the linear dynamic range is significantly higher?
JOE ROBBAT: I wouldn't say significantly higher. I would say maybe one
more signal to noise higher.
RONALD MITCHUM: Regarding the HNU you described, was that a GC?
Did you measure surface emissions with this Bruker GC/MS, and the little
probe, and try to match that with what you saw in the wells in that one study?
What was the source of the chlorinated hydrocarbons in the lab study that you
did?
AL ROBBAT: To the first two questions, no and no. The source was a
semiconductor company that occupied that particular site. It's purported that
they were dumping solvents into the stream. That company has since gone out
of business.
348
-------
ION MOBILITY SPECTROMETRY FOR IDENTIFICATION
AND DETECTION OF HAZARDOUS CHEMICALS
Julio Reategui
Project Engineer
Tad Bacon Glenn Spangler
Senior Engineer Principal Engineer
Environmental Technologies Group, Inc.
1400 Taylor Avenue
Baltimore, Maryland 21284-9840
Joseph Roehl
Marketing Mgr.
ABSTRACT
This paper describes an Ion Mobility Spectrometry
(IMS) System which has been designed to detect
organic vapors in ambient air as might be required
to survey and characterize hazardous waste sites.
The system allows real-time identification of
chemical vapors and determination of their concen-
tration by generating and interpreting spectral
data. The array of applications suitable for this
detection system will be discussed.
INTRODUCTION
Growing public concern and regulatory issues re-
lating to air, ground and water quality have
created a need to accurately and rapidly assess
exposure to hazardous chemicals. Similarly, the
process industry continually seeks ways to control
purity of materials using real-time trace chemi-
cal detection techniques to adjust parameters.
New developments in biotechnology, integrated
circuit manufacturing, communications, and
materials manufacturing have resulted in trace
chemical detection requirements that did not exist
five years ago. These trends are expected to con-
tinue through the next decade.
Presently, a number of analytical technologies
such as gas chromatography, mass spectrometry,
and various ionization and photo absorption
techniques are used to detect, identify and
quantify these chemicals. Each of these tech-
nologies, however, has limitations, and no single
instrument is suitable for all monitoring and de-
tection applications. Furthermore, the choice of
instrument or technique depends upon the applica-
tion, which sometimes involves conflicting re-
quirements. This increasingly complex problem
presents a formidable challenge to the chemical
detection industry. Together with these needs,
the customer desires small, cost-effective and
easy to operate instrumentation.
ION MOBILITY SPECTROMETRY HOW IT WORKS
IMS was developed in the late 60's by the
Franklin GNO Corporation as a Laboratory instru-
ment for analysis of trace concentrations of
organic compounds. Early on, the potential of
this technique was recognized as a method of
identification for extremely low levels of chemi-
cals. Advances in technology since that time have
resulted in miniaturization of the IMS detector
cell, and the required electronics/data processing
systems. These improvements, along with the in-
troduction of a membrane inlet system and a recir-
culating air purification system, have made IMS
practical for ambient air monitoring. The result-
ing instrument is small, rugged and requires
little routine maintenance.
The principle of IMS is illustrated in Figures 1
to 4. Ambient air is drawn into the instrument
and past a semi-permeable membrane on the outside
of the cell by use of a sampling pump. The mem-
brane allows materials of interest to pass into
the detection cell, while attenuating many pos-
sible interferents. Purified dry air from a self-
contained scrubbing system sweeps the membrane on
the inside of the cell and delivers the sample to
the reaction region. There the sample, consisting
of one or more components, is ionized by reactions
with a weak plasma of positive and negative ions,
formed by ionization of the purified air by a
radioactive source. The ionized sample molecules
and reactant ions drift through the cell under the
influence of an applied electric field. A shutter
grid allows periodic introduction of the ions into
a drift tube where they separate based on charge,
mass, and shape. Smaller ions move faster than
larger ions through the drift tube and arrive
first at the detector. The ability of an ion to
move through another gas is called ''mobility".
Because different ions have different mobilities,
the ions arrive at the collector with different
drift times. The current created at the detector
is amplified, measured as a function of time, and
a spectrum is generated. The identity of the
molecules can then be determined using pattern
recognition algorithms using a computer or micro-
processor to analyze and compare features of the
IMS signature with information stored in memory.
The electric field is periodically reversed so
that ions of both polarities can be studied. A
general purposes IMS based detection system
(GPIMS) is illustrated in Figures 5 to 9.
Specificity is a function of the success with
which the detection algorithm recognizes specific
compounds by their unique mobility spectrums.
Additional specificity might also be obtained for
349
-------
certain compounds by membrane selection or using
dopant vapors in the carrier gas, changing the
ionization chemistry.
Figure 10 illustrates the ability to detect a
specific compound in the presence of high concen-
trations of possible interferences. In this ex-
ample, TDI is given as the target compound,
detected in the negative mode. Even saturated
headspace vapors of chlorobenzene, 2-propanol, and
ammonium hydroxide do not interfere with the de-
tection of TDI. The response of these materials
occurs in the positive mode (not shown) and could
be identified. Phosgene is detected in the nega-
tive mode, and so also appears along with the TDI,
and is identified. In field screening applica-
tions, the instrument can be operated as a general
class detector or as a specific detector. In the
general class mode, the unit distinguishes, for
instance, halogenated compounds which produce
negative ions, and compounds such as ketones,
esters, alcohols, ethers, etc. which produce posi-
tive ions. In the specific mode, the instrument
can identify the specific compounds on the basis
of their mobilities. Semi-quantitative results
are available for both modes of operation. Under
controlled conditions, the instrument also may be
calibrated to produce accurate quantitative re-
sults .
Since IMS responds to a variety of chemical vapors,
an IMS detection system can be reprogrammed to de-
tect new chemicals without any hardware modifica-
tion and reject new interferents as they arise.
This could provide a great savings for those who
use a detection system that needs to respond to
changing requirements or threats.
As a result of the ion mobility separation pro-
cess, IMS is more specific than other types of
ionization detectors, and less prone to false
alarms. IMS combines the simplicity and sensi-
tivity of ionization detectors with the additional
degree of specificity gained from interpretation
of IMS' spectral data through detection algorithms.
It, therefore, bridges the gap between the non-
specific screening devices and elaborate analyti-
cal instruments.
It is a versatile and a sensitive real-time trace
vapor detector. Because of its small size and
minimal system requirements, IMS is ideal for use
as a cost-effective fixed or portable chemical
detector. Detectors based on IMS provide real-
time analysis (usually within 5 seconds); they are
small enough to be hand-held; inherently more
rugged than a GC or MS; very sensitive (in the
sub-ppb range); and can be trained to detect,
identify and estimate concentrations of many
chemical compounds. These relationships are illus-
trated in Table 1.
IMS systems are unique among chemical detectors in
that their detection capabilities can be enhanced
for specific applications through simple operating
parameter modifications. Overall, IMS systems
offer technical and logistical advantages that
make them attractive as chemical vapor detectors
for a wide variety of application in commercial,
industrial and military markets.
APPLICATIONS OF ION MOBILITY SPECTROMETRY
The potential commercial and industrial applica-
tions for IMS technology include the following:
0 Field detection and identification of prede-
termined target chemicals for chemical process-
ing industries.
° Pre-screening of samples prior to laboratory
analysis to determine relative levels of con-
tamination. This would speed up sample prepara-
tion times, and reduce down time due to over-
saturated analytical instruments.
0 Screening detector for quick examination of
hazardous waste dumps, for specific chemicals or
groups of chemicals.
0 Environmental monitoring systems, such as net-
works of detectors positioned around chemical
factories and chemical agent demilitarization
sites.
° Trace contaminant detection in controlled chemi-
cal processes (used in clean rooms and chemical
production facilities).
° Portable, hand-held chemical leak and spill de-
tectors suitable for use in emergency situations.
° Tandem detectors, such as GC/IMS and IMS/MS, to
be used as analytical tools to support research
performed in government laboratories, universi-
ties and institutions.
Chemicals that can be detected by IMS systems are
listed in Figures 11 and 12.
SUMMARY
In summary, IMS is an especially attractive tech-
nology for instruments used in field screening
applications. It may be used as a class detector
or to identify specific chemicals. The instrument
provides sensitivity in the PPB to PPM range in
real-time (<5 sec.).
ACKNOWLEDGEMENTS
Special thanks are extended to Dr. Joe Epstein;
Dr. David Lubman of the University of Michigan,
Ann Arbor; and Dr. Len Luskus from Brooks AFB,
Texas, who reviewed the text and provided valuable
comments during the preparation of this paper.
The authors also wish to express their apprecia-
tion to Ms. Rita Rosenberger for her skillful typing
of the manuscript and Messrs. Bob Allen and Barry
Rodgers for their contributions with the illustra-
tions of this paper.
350
-------
REFERENCES
(1) Spangler, G. E, Carrico, J. P. and Kim, S. H.,
"Membrane Inlet Studies with Ion Mobility
Spectrometry", Paper 402, 9th Annual Meeting
of the Federation of Analytical Chemistry and
Spectroscopy Societies, Philadelphia, PA,
September 1982.
(2) Lubman, D. M. and Kronick, M. N. , "Plasma
Chromatography with Laser Produced Ions",
Anal. Chem. 54(9), 1546 (1982).
(3) Hill, Jr., H. H. and Bairn, M. A., "Ambient
Pressure lonization Detectors for Gas Chroma-
tography. Part II: Radioactive Source
lonization Detectors", Trends in Anal. Chem.
1(10), 232 (1982).
(4) Spangler, G. E., Vora, K. N. and Carrico,
J. P., "Miniature Ion Mobility Spectrometer
Cell", J. Phys. E. (Scientific Instruments)
li, 191 (1986).
(5) Carrico, J. P. Drake, A. W., Campbell, D. N. ,
Roehl, J. E., Sima, G. R., Spangler, G. E.,
Vora, K. N. and White, R. J., "Chemical
Detection and Alarm for Hazardous Chemicals
using Ion Mobility Instrumentation", Amer.
Lab. ^8(2), 152 (1986).
(6) Eiceman, G. A., "Atmospheric Sensing of
Hazardous Organic-Compounds Using Ion
Mobility Spectrometry", Abstr. ACS 192, 3
(September 1986).
(7) Lawrence, A. H., Nanji, A. A. and Michael,
N. Z., "Use of Skin Surface Sampling and Ion
Mobility Spectrometry as a Preliminary
Screening Method for Drug Detection in an
Emergency Room", Journal of Toxicology-
Clinical Toxicology Z5(6), 501 (1987).
(8) Schellenbaum, R. L., "Air Flow Studies for
Personnel Explosives Screening Portals",
Proceedings of the Carnahan Conference on
Security Technology (Report SAND-87-0822C;
Conf-870743-1), Atlanta, GA, July 1987.
NTIS reference: DE87007772/XAB.
(9) Karasek, F. W., "Plasma Chromatography of
the Polychlorinated Biphenyls", Anal. Chem.
43(14), 1982 (1971).
SAMPLE
IN
ELECTRIC FIELD
'Ni REACTION SHUTTER/' DRIFT APERTURE COLLECTOR
IONIZER REGION GRID / REGION GRID
COLLECTOR
CURRENT
I
10
D 30 4O
MILLISECONDS
50
60
Figure ^. Theory of Operation of IMS Cell
351
-------
POSITIVE IONS
NEGATIVE IONS
Proton Transfer
RH+ + P -» R + PH+
Nucleophilic Attachment
R+ + P -* RP+
Hydride or Hydroxide Abstraction
R+ + PH -* RH + P+
Oxidation
R+ + P -» R + P+
Complex Rearrangement
Charge Transfer
R- + P -» R + P"
Dissociative Capture
R' + AP -* R t A' +
Proton Abstraction
R" + HP -» RH + P-
Electrophilic Attachment
R- + P •* RP-
Figure 2. Ion/Molecule Reactions
16 N m
f 27r
L—
1/2
WHERE
e IONIC CHARGE
m IONIC MASS
N = MOLECULAR NUMBER DENSITY
M = MOLECULAR MASS
k = BOLTZMANN CONSTANT
T = TEMPERATURE
rm = POSITION OF MINIMUM POTENTIAL FOR INTERACTION
n(1,1)* FIRST ORDER COLLISION INTEGRAL
A = CORRECTION TERM FOR HIGHER APPROXIMATIONS
Figure 3. Mason-Schamp Theory For Mobility
DRIFT VELOCITY (Vd)
Vd = KE
MOBILITY (K)
REDUCED MOBILITY (KQ)
K - K--
K° ' K 760 T
Figure 4. IMS Equations
352
-------
AIR
SAMPLE
IMS MODULE
DETECTOR
HOST COMPUTER
MODULE
DATA ACQUISITION,
ANALYSIS. STORAGE.
RETRIEVAL AND
DISPLAY
SOFTWARE
-— ' " —-
USER(S)
Figure 5. General Purpose Ion Mobility Spectrometry System (GP-IMS)
for Real-Time Field and Laboratory Evaluation of Chemical
Vapors in PPT to PPM Range
MICROPROCESSOR
MODULE
SENSOR INTERFACE MODULE
INTERFACE
AND
SIGNAL
BUFFER
CELL MODULEl
PRE/POST
AMPLIFIER
• SYSTEM CONTROL
• SIGNAL PROCESSING
• DATA OUTPUT (SERIAL)
PNEUMATICS
MODULE
Figure 6. Block Diagram of the GP-IMS
353
-------
I INLET:
(SAMPLE AIR"
——1—"
1
ACQUIRE SAMPLE |
SAMPLE AIR
MEMBRANE:
PRE-SCREEN SAMPLE ]
SAMPLE AIR
Mi":
IONIZE SAMPLE]
IONIZED SAMPLE
DRIFT REGION:
ELECTRIC FIELDS:
SEPARATE IONS
ION FLOW
I ELECTROMETER:
DETECT IONS |
r, ANALOG SIGNAL
uP BOARD:
DATA (SIGNAL) t
PROCESS SIGNAL |
t
<
1 DIGITAL CONTROL
' i
uP BOARD: NOTIFY REMOTE| INTERFACE/CONTROL
P°^r- ^OACLKLS
DIGITAL DATA
(SERIAL)
TO
IMS CELL-4-
, , DIGITAL (SERIAL)
ANALOG SIGNAL
( SIGNAL )
SIGNAL J
Figure 7. Ion Mobility Spectrometry System Theory of Operation
Figure 8. Modular Commercial Detector That Can Be Set
to Recognize a Wide Variety of Chemicals
354
-------
CABLE AND
STORAGE AREA
(BEHIND LD)
COMPUTER
OUTPUT
OSCILLOSCOPE
OUTPUT
FLOW
ACCESS/
BYPASS
SYSTEM
STATUS
DISPLAY
CONTROL
SWITCHES
AND
INDICATING
LIGHTS
FLOW
CONTROL
110V AC
POWER INPUT
AND SWITCH
SAMPLE
INLET
SAMPLE
EXHAUST
Figure 9. Ion Mobility Spectrometer
355
-------
3 ppb toluene diisocyanate
(TDI) in pure air
3 ppb TDI + saturated headspace
A|R vapors of chlorobenzene
PEAK
TDI
PEAK
3 ppb TDI + saturated headspace
vapors of 2-propanol
AIR
PEAK
3 ppb TDI + saturated headspace
vapors of ammonium hydroxide
TDI
PEAK
3 ppb TDI + 1 ppm phosgene
Figure 10. Detection of TDI in the Presence of High
Concentration of Possible Interferences
356
-------
CHEMICALS THAT CAN BE DETECTED BY IMS
• TOLUENE DIAMINE (TDA) •
• DINITROTOLUENE (DNT) •
• TRINITROTOLUENE (TNT) •
• TOLUENE DIISOCYANATE (TDD
• METHYLENE BIS PHENYL ISOCYANATE
• TOLUIDINES
• METHYLENE DIANILINE (MDA) •
• VINYL ACETATE •
• TETRAHYDROFURAN (THF) •
• FORMALDEHYDE •
• ACRYLONITRILE •
• ACETALDEHYDE •
• CYCLOHEXANONE •
• ACETONE •
• ALCOHOLS •
• KETONES •
• HALOGENATED COMPOUNDS •
• NITRO-COMPOUNDS, EXPLOSIVES •
• AMINES •
• ESTERS •
ORGANOPHOSPHORUS COMPOUNDS
DRUGS
PESTICIDES
(MOD
PHENOLS
ETHYL ETHER
METHYL ETHYL KETONE
PYRIDINE
PIPERIDINE
HYDROGEN CYANIDE (HCN)
PHOSGENE
HYDROCHLORIC ACID (HCI)
BENZYL CHLORIDE
HYDROIODIC ACID (HI)
PHOSPHOROUS TRICHLORIDE (PCI3)
HYDROGEN BROMIDE (HBr)
MANY OTHER NIOSH/OSHA REGULATED CHEMICALS
Figure 11. Chemicals That Can Be Detected By IMS
PROTON AFFINITY SCALE
CLASSES OF COMPOUNDS WHICH IMS CAN DETECT
POSITIVE I OH NODE:
Those which give a strong response:
Pyridines
Unsaturated Anilines
Aliphatic Amines
Phosphines
a,«t-Disubstitute
-------
TABLE 1. COMPARISON OF IMS WITH OTHER PORTABLE AMBIENT AIR MONITORS
INSTRUMENT
Photo-
Datecto r
(PID)
loni zation
Detecto r
Ga B
Mass
IMS
Detect 8 most
o rganics .
Does not
identify .
Nonspecific, None PPB-PPM
Detects *1 1
o rganics .
ident i f y .
Specific Mixtures.
matrix .
Specific
can be con- tion of a GC.
trolled by
ants, polarity.
Specific
RESPONSE
pounds" . Only indicates
if 'something" ia
p re Beat .
organics'. Only indi-
there. Ho re complicated
than PIO.
tify compounds. Re-
sensitivity .
second. ,...,.. S50-200K V.ty good ....itivity.
quires high vacuum pump a.
Long set-up time.
us e package ,
DISCUSSION
HAL STUBER: Is there a molecular weight range for which the field portable
instruments or techniques are applicable?
JULIO REATEGUI: Yes, the theoretical molecular weight range is between
15 and 500. However, the instrument works better if the compound is between,
say 50 and 200.
TOM PRITCHETT: The coronal discharge atmospheric pressure chemical
ionization is probably the best source for ionizing polar compounds - the
compounds we have the most trouble analyzing by any other way in air
matrices. Toluene isocyanate, for example, is something we make evacuation
decisions at detection limits that are three times higher than theTLVs. There are
some other polar compounds that technology, although in its infant stage, has
the capabilities for, that we just can't touch with an HNU or any other type of
real-time monitor, including the GCs, because you'll never get them past the
injection ports.
JULIO REATEGUI: I agree that the coronal source is good for this applica-
tion. In the example I showed, I used an ionization source that is not as good.
It was a nickel-63 source. I'm sure when we try with a coronal, we can probably
do better than that.
CHARLES MANN: You mentioned that using absolute peak heights wasn't
effective for quantitative analysis and the improvement you got in looking at
ratios. Are you intending to pursue this line? You mentioned having an
extensive software development. It looks as though that might be a fruitful
avenue. Do you intend to go that route?
JULIO REATEGUI: Yes. There are some things you can't patent here, so we
decided to maintain some proprietary information. But my guess is that a few
months down the road, the details of this technology will become more and
more available.
358
-------
UTILIZATION OF SHORT-TERM BIOASSESSMENTS AND
BIOMONITORING AT SUPERFUND SITES
David W. Charters
U.S. Environmental Protection Agency
Environmental Response Team
Edison, New Jersey 08837
The Environmental Protection Agency's Superfund program is requiring
thai more extensive environmental assessments be performed at both Removal
and Remedial sites in accordance with the provisions of SARA. Bioassessment
and biomonitoring are presently being utilized as field screening methods at
several sites to meet these conditions. The tests and studies discussed have been
chosen based on their ability to be cost effective, rapid, and not delay the cleanup
procedures. Field screening studies include collection of small mammals, fish,
benthic invertebrates, and plants with attention paid to alterations in community
structure, population dynamics, bioaccumulation of toxicants and histopathol-
ogy as well as other parameters which are site specific. Laboratory tests include
aqueous acute and chronic, and both elutriate and solid phase tests. Terrestrial
tests include contact tests and tests for phytotoxicity. Combinations of these
tests are then correlated with chemical and physical parameters collected in the
field to give a more comprehensive environmental assessment of the site and its
impact on the surrounding area.
Several sites will be discussed to illustrate how the bioassessments and
biomonitoring are directly applied to the cleanup of sites at both the RI/FS stage
and at the Removal Phase.
DISCUSSION
JAMES DELEVAN: Would you care to comment on introduced species for
the purpose of a bioassay. compared to collection of indigenous species?
DAVID CHARTERS: We run both types of studies. In the case here, the lab
killed all the controls, however the actual samples worked well.
Too frequently, for toxicity testing, people are very interested in running
indigenous species - but it doesn't give you the needed information when
running standards.
If you are dealing with an introduced species, as in taking water out and running
species in that, it's a very good idea.
I'm opposed to doing exotics in silu. For example, if fat head minnows ex-ergot
loose in Region X there would be a lot of problems. You have to be very careful
if you are introducing any exotics.
JONATHAN NYQUIST: Sometimes it takes a while for the chemicals that are
released to get anywhere, as through the ground, before they hit the biosphere,
so you may not see that as a stress right away. How long does the contamination
have lo be there before you will expect to see bioeffects?
DAVID CHARTERS: You never do any of these studies in a vacuum. In this
case, you've got a site that's been there since 1949. In other cases, you've got
a site that's been there since Tuesday. You have to look at the exposure routes.
If the stream is impacted, it's important to know that. If it's not impacted, that's
information, also.
If you ask the right questions, and you have the right data quality1 objectives, you
will answer the question regarding the time involved. Whether contamination
reaches the stream in two years or five years is a hydrogeologic question.
You can't do chemistry in a vacuum. Y'ou have to do biology in the same
framework. If contamination is there, it's important to assess the environmental
impact. If it is not there, that is just as important information.
JONATHAN NYQUIST: Suppose the effect is being caused by something
other than the contaminants you're looking at? If acid rain, or something else
is causing the problem, how do you separate that from the dump site?
DAVID CHARTERS: You separate it by using appropriate reference areas.
Frequently in these cases, we do have that problem, and some people are using
the same watershed as a reference.
If you're more than a kilometer from the site upstream, you're too far, assuming
there are no impacts there. You want to be as close as possible lo that site, and
then everything is related to that reference area. There are no pristine areas left.
basically.
If there is a hazardous waste site there, you've got problems. If you use the
appropriate upstream controls, or general reference areas lor something like
small mammal collection, you can answer those questions, and it's extremely
important to make sure you tie these things to that site. It doesn't do any good
if you can't tie it to the site.
JONATHAN NYQUIST: If, as you suggest, you've had a site that may have
been impacted since 1949 conceivably, what is the risk of using generic clumps.
like piercers, grazers, and so on? These could be in a natural ratio, but the
species changes would be different over long periods of time, because of
acclimation and taking advantage of a situation?
DAVID CHARTERS: The difference is, in this case, we're talking hundreds
of yards of stream. Nothing more. Invertebrates tend not to go upstream, they
go downstream. They are the bottom of the food chain. If there are fish there.
the fish are going to eat them. It's a constant turn over in population, and the
population tends to work downstream.
In these particular cases, you want to take similar habitats. You want to take
them close together, and you want to be very sure of where you're getting these
things.
If I just took functional feeding group analysis, it does not stand alone. None
of this stands alone. You need to do other things. If I give a chemist functional
feeding group analysis, he won't know what to do with it. So you run multiple
biological tests. You will be able to obtain what you wanted in the first place:
an overwhelming preponderance of data, indicating there is an impact.
KEN HANKS: Did you look at any effects of bioconcencration through the
food chain, say from the invertebrates to the fish, through the aquatic birds?
We do go up. Taking the analysis from fish to birds is a big jump, because you
can't nail the birds down to that site, and if you can't do that, don't sample the
bird.
359
-------
HIGH-PERFORMANCE LIQUID CHROMATOGRAPH AS A VIABLE FIELD SCREENING METHOD
FOR HAZARDOUS WASTE SITE INVESTIGATIONS
Vanavan Ekambaram and James B. Burch
Senior Project and Senior Staff Scientist
Woodward-Clyde Consultants
4582 South Street Parkway, Suite 1000
Denver, Colorado 80237
Abstract
A field-operable high-performance liquid chromato-
graph (HPLC) was employed for the chemical analysis
of water and soil at a former wood treating site.
Since a gradient pump was not available, an
isocratic pump was employed using a two-step
elution procedure in the analysis of polynuclear
aromatic hydrocarbons (PAHs), which are the primary
contaminants at this site. A 50:50 mixture of
acetonitrile and water was used for the low
molecular weight PAHs, and a 85:15 mixture was used
for the higher molecular weight PAHs. This two-
step HPLC analysis provided an efficient column
characteristics, similar to the commonly used
gradient chromatographic systems. Good resolution
(R > 1.0) was obtained with this system for all
compounds except the benz(a)anthracene chrysene
pair. The capacity factor (k') for the column used
was 2 to 16 with approximately 2500 theoretical
plates. Method detection limits of 0.03 to 3.0
parts per billion was achieved for PAHs in water.
The results were compared with data obtained from
commercial laboratories on split samples and the
comparison shows that the field HPLC provided
comparable data that can be used in site
characterization. Field-operable HPLC, if operated
by a trained chemist, can be a viable low-cost
detection and monitoring tool and can provide rapid
turnaround that facilitates evaluation of
remediation designs.
Introduction
Creosote, a high-temperature distillate derived
from coal tar, is the most extensively used
industrial wood preservative in the United States
(Runker, 1). It is a complex mixture of liquid and
solid aromatic hydrocarbons: 85 percent (by weight)
poly nuclear aromatic hydrocarbons (PAHs),
12 percent phenolic compounds, and 3 percent
heterocyclic N, S and 0 compounds. The abandoned
and former wood treatment facilities are often
sites of widespread soil and ground water contami-
nation of the PAHs.
Certain PAH compounds are known or suspected car-
cinogens. The chronic toxicity of PAHs is related
to the number of rings they contain; the two and
three-ring PAHs are relatively non-toxic while the
four and higher ring PAHs are extremely toxic.
Because of their toxicity, they play an important
role in the risk evaluations of ground water and
soil contamination. Therefore, it is essential to
obtain accurate and precise chemical data on the
levels of PAHs in the environment.
PAHs are a class of compounds that are particularly
suitable for analysis via HPLC. A HPLC was
installed in the field laboratory at a NPL site,
and was developed for the analysis of ground water
at the parts per billion level. Procedures were
developed so that quantitative results could be
obtained on sixteen PAH compounds. The
chromatographic procedure used to measure the PAH
levels in ground water, and the performance
criteria for the HPLC are described in this
paper. The advantages of this using HPLC in the
field is that it cuts down on the turnaround time
and cost that are normally associated with
laboratory analysis and facilitates acquisition of
chemical results on several samples.
Sample Extraction
Extraction procedures for water samples can be per-
formed by using a Sep-Pak liquid-solid extraction
technique or liquid-liquid extraction. One draw-
back with using Sep-Pak is that it tests only a
limited amount of water. Thus, samples with high
levels of contamination may saturate the system
resulting in poor recovery of the desired com-
pounds. Therefore, separatory funnel liquid-liquid
extraction was adopted following EPA guidelines for
the analysis of PAH compounds (40 CFR, EPA Method
610, 2). This method takes advantage of the
hydrophobic properties of PAH compounds
(octanol/water partition coefficients of 2xl03 or
greater) which insures a high percentage of par-
titioning of PAHs into the organic phase. This
method also allows for larger volumes of sample to
be processed (up to 2000 times in test analyses)
effectively decreasing the method detection
limit. Up to 1 liter of water samples were
extracted with 100 ml of methylene chloride and the
extract concentrated down to 0.5 ml.
HPLC Procedures
Since the column packing material is non-polar in
nature, a relatively polar solvent was needed to
effectively portion the sample compounds. A mix-
361
-------
ture of water and either acetonitrile or methanol
are generally used in reverse phase chromato-
graphy. Acetonitrile is less viscous and more com-
patible with PAH compounds, and therefore, a mix-
ture of acetonitrile and water was used as the
mobile phase for both water and soil analyses.
Since PAH compounds have a wide range of retention
properties in chromatographic columns, a gradient
pump that can change the proportions of acetom-
trile and water in the mobile phase is often used
in PAH analyses (40 CFR, Method 610, 2). However,
since a gradient pump was not available at this
site, a two-step elution procedure was adopted. A
50:50 mixture of acetonitrile and water was used
for the low molecular weight PAH compounds:
napthalene, acenaphthalene, acenaphthene, fluorene,
phenanthrene, fluoranthene, anthracene, and
pyrene. An 85:15 acetonitrile/ water mobile phase
was used for the high molecular weight PAH com-
pounds: (a)anthracene, chrysene, benzo(b)fluor-
anthene, benzo(k)fluoranthene, benzo(a)pyrene,
dibenzo(a.h) anthracene, benzo(g,h,i)perylene, and
indeno(l,2,3-cd)pyrene. PAH standards were run to
determine the retention times (time from injection
to elution), using the two mobile phases.
When the 50:50 acetonitrile/water mobile phase is
used, the high molecular weight PAHs elute after
30 minutes, so that analysis times are extremely
high and the peaks are wide. Therefore, it was
necessary to change the mobile phase to 85:15
acetonitrile/water mobile phase which facilitates
the faster elution of high molecular weight PAH
compounds yielding shorter retention times. When
the solvent system is changed, it is run through
the column for at least 30 minutes before injecting
a sample such that high molecular weight PAH's that
may be insoluble at the head of the column are
dissolved and flushed out of the column by the wash
procedures. The column clean up was accomplished
by several injections of 100% methanol. Several
sample extracts were run with 50:50 mobile phase
before the mobile phase was changed and the samples
were then rerun using the 85:15 mobile phase. When
85:15 acetonitrile/water mobile phase is used,
lighter PAH compounds elute in less than four
minutes, but they are not resolved well enough and
therefore not used for quantification. An
ultraviolet absorption and a fluorescence detectors
were used for measuring the signals.
HPLC Performance Results
An efficient chromatographic separation is achieved
by striking a balance between resolution, speed and
capacity. The analyst who wishes to optimize HPLC
partitioning can alter one of these parameters only
at the expense of the other two. There are several
theoretical parameters, based on data collected by
the analyst, which are used to evaluate the
efficiency of the analytical conditions. In this
discussion, the resolution (R), number of theo-
retical plates (N), height equivalent to theoreti-
cal plates (HETP), capacity factor (k1) and method
detection limits (MDL) are used to evaluate current
conditions.
Resolution: Resolution is the degree to which two
compounds are separated by the column. The degree
of resolution between two peaks depends on the
width of each peak and the distance between them.
The resolution between adjacent PAH compounds
analyzed by the on-site HPLC are presented in
Table I. A resolution value of 0.5 or greater is
generally needed to distinguish two closely
resolved peaks of equal response whereas a value of
1.0 or greater should be obtained for quantifiable
analyses. Table I indicates that excellent resolu-
tion values were obtained for all compounds except
between benzo(a)anthracene and chrysene. The
R value for these two compounds was 0.74, slightly
below the cutoff for optimal resolution.
Capacity Factor: The capacity factor (k1) is a
measure of the column's ability to retain com-
pounds. Small k1 values indicate that sample com-
ponents are poorly retained, with short retention
times. Conversely, large k1 values indicate that
peaks will have long retention times and wide,
poorly defined peaks. Studies indicate that column
efficiency is optimized at k1 values between 2 and
6 (Majors, 1984). But values between 1 and 15 are
more practically acceptable (Johnson and
Stevenson, 1978).
Capacity factors for PAH compounds in this analysis
are listed in Table II. The k1 values for 85:15
compounds are all within the 1-15 working range as
are all 50:50 compounds except for pyrene, which
has a k' value of 16.86. Although the retention
time for pyrene is high (23.7 minutes), its reso-
lution is also high (1.79).
Theoretical Plates: The concept of theoretical
plates (N) is used as a measure of column effi-
ciency. Column efficiency can be increased by
increasing column length, which increases the
probability that solutes will interact with the
column matrix. The value of N is given by:
t 2
N = 5.5 N— 0
W,
(1)
where: tv
retention time from t=0 to the peak
maximum for a given peak (minutes)
the peak width at half the peak height
(minutes).
The equation that incorporates Wj,, is considered
more accurate since it eliminates of peak
tailing. Table II provides the capacity factors
and the number of theoretical plates.
Another measure, the height equivalent to theoreti-
cal plates (HETP), allows comparison of the
efficiencies of columns of different lengths.
The smaller the HETP value, the better the column
efficiency. HPLC columns generally have HETP
values of 0.01 1.0 nun. HETP values of 0.055 and
0.052 were calculated for the 85:15 and 50:50
analyses, respectively, and are in good agreement
with the above range indicating good column effi-
ciency.
Method Detection Limits: The method detection
limit (MDL) is defined as the minimum concentration
of a substance that can be measured and reported
with 99% confidence that the value is above zero
362
-------
(Glasser et al., 3). (The instrumental detection
limit, as opposed to method detection limit, is the
minimum concentration which can be detected
directly by the instrument. It does not take into
account sample preparation and extraction proce-
dures.) The spiked samples were run through the
extraction and analytical procedures in exactly the
same manner as other samples and the mean and
standard deviation are determined for each com-
pound. The MDL is calculated by the formula:
MDL = t
where:
x s.
(2)
(df = n-1, 1-a = .99)
t = The students t statistic for a one-
tailed test
df Degrees of freedom
n Number of observations
a the probability of a Type I error
s = Standard deviation
In these calculations df = 6 and t = 3.143 (Pearson
and Hartley, 1970). The MDL values for on-site
HPLC analysis of PAH compounds using ultraviolet
and fluorescence detectors are presented in
Table III. The range of MDL's for PAH compounds is
0.16 to 3.38 ug/1. The results for the lighter
PAHs were calculated using ultraviolet and for the
heavier PAHs using the fluorescence detectors.
These results are comparable to reported MDL's for
PAH's, from EPA Method 610, which range from 0.01
to 4.0 vg/1 (Cole, et al., 4). Higher MDL's are
noted for 50:50 compounds (napthalene through
pyrene). In practice, PAH compounds were at times
reported below the method detection limits down to
a concentration of approximately 0.1 ppb. The
sample-specific detection limits for a given sample
varied depending upon how the sample was treated
(i.e., sample concentration or dilution).
Extraction Efficiency: Percent recovery ranges
were established by analyzing twenty spiked samples
and calculating the standard deviation(s) of the
percent recoveries. One liter of reagent grade
water was spiked with two milliliters of a known
mixture of standard concentrations for each PAH
compound. These samples were then extracted, con-
centrated, and analyzed according to HPLC proce-
dures for water samples cited above. The mean per-
cent recovery and standard deviation for each com-
pound was then calculated. The percent recovery
range was set at two standard deviations above and
below the mean recovery (X + 2s). Results of these
calculations suggest that good recoveries of PAHs
were obtained.
laboratories. The calculated relative deviations
were, in general, 0.5%, which could be due to
differences in analytical conditions, extraction
efficiencies, etc. The method blanks and method
spikes analyzed via HPLC also provide results which
are encouraging.
The method detection limits range from 0.03 yg/1
for benzo(k)fluoranthene to 3.0 ug/1 for
fluorene. These values are, in general, comparable
to the recommended detection limits for the EPA
Method 610. Even though, one can quantify concen-
trations below these limits, the values would not
be at a high confidence level.
The reproducibility by HPLC also appears to be well
controlled. Several replicate analyses were per-
formed on the spike samples (Table IV) and the mean
(X) and standard deviations (s) indicate that the
average coefficient of variation (s/X) is, in
general, <30%.
In summary, HPLC offers several advantages in the
hazardous waste site investigations and in the
remedial designs development. Hundreds of samples
can be processed at low cost and rapid turnaround
t i me.
References
(1) von Rumker, Rosmarie, Lawless, E.W., and
Meiners, A.F., "Production, Distribution, Use
and Environmental Impact Potential of Selected
Pesticides" U.S. Environmental Protection
Agency, EPA 540/1-74-001, 1976, 439 p.
(2) Federal Register, "Environmental Protection
Agency Regulations on Test Procedures for
Analysis of Pollutants" 40 CFR 136, 1986,
pp. 131:4288-4296.
(3) Glasser, J.A., Forest, D.C., McKee, G.D.,
Quave, S.A. and Budde, W.L., "Trace Analysis
for Waste Waters," Environ. Sci. Tech., 1981,
15
(4) Cole, T., Riggin, R., and Glasser, J.A.
"Evaluation of Method Detection Limits and
Analytical Curve for EPA Method 610,
Polynuclear Aromatics" Proc. of the Fifth
International Symp. for Polynuclear Aromatic
Hydrocarbons, 1980, Battelle Columbia
Laboratory, Columbus, OH.
The overall performance of the field HPLC proce-
dures yielded useful results which are encouraging
for future application in the evaluation of
remedial actions.
Quantitative data were obtained for all PAH com-
pounds in the ug/1 range. For the off-site ground
water samples, where the concentrations of PAH's
are, in general, in the ug/1 range, the HPLC
provided quantitative information that are in
reasonable agreement with the data from commercial
363
-------
TABLE I
RESOLUTION OF PAH COMPOUNDS BY ON-SITE HPLC
TABLE III
HPLC METHOD DETECTION LIMITS
Adjacent Compounds
Naphthalene /Acenaphthalene
Acenaphthalene/Acenaphthene
Acenaphthene/Fluorene
Fluorene/Phenanthrene
Phenanthrene/Anthracene
Anthracene/Fluoranthene
Fluoranthene/Pyrene
Benzo(a)anthracene/Chrysene
Chrysene/Benzo(b)f luoranthene
Benzo(b)f luoranthene/Benzo(k)f luoranthene
Benzo(k)fluoranthene/Benzo(a)pyrene
Benzo(a)Pyrene/D1benzo(a,h) anthracene
D1oenzo(a,h)anthracene/Benzo(g,h,1)perylene
Benzo(g,h,1)pery1ene/Indeno(l,2,3-cd)pyrene
TABLE II
HPLC COLUMN EVALUATION
Compound Capacity Factor
(k1)
50:50 Mobile Phase
Naphthalene 3.15
Acenaphthalene 4.17
Acenaphthene 5.99
Fluorene 6.63
Phenanthrene 8.64
Anthracene 10.66
Fluoranthene 14.61
Pyrene 16.86
85:15 Mobile Phase
Benzo{a)anthracene 2.48
Chrysene 2.74
Benzo(b)f luoranthene 4.24
Benzo(k)fluoranthene 5.06
Benzo{a)pyrene 5.94
D1benzo(a,h)anthracene 8.15
Benzo(g,h,1)perylene 9.41
Indeno(l,2,3-cd)pyrene 10.64
Resolution
Factor (R)1
2.86
4.58
1.38
2.98
2.50
3.80
1.77
0.74
3.61
1.68
1.53
3.88
2.23
1.69
Number of Theoretical
Plates (N)
2381.4
3136.0
4096.0
2365.0
2818.6
2571.9
2820.8
2837.5
1901.1
1656.1
2168.9
1956.8
2571.1
4213.9
3832.2
3620.0
(
1
Compound
Spike
Concen-
tration
(ug/L)
Mean'
Concentration
Recovered
(ng/L)
Standard
Deviation
Method2
Detection
Limit
(ug/L)
HPLC EPA 610J
Naphthalene
Acenaphthalene (n=5)
Acenaphthene
Fluorene (n=5)
Phenanthrene
Anthracene
Fluoranthene
Pyrene
Benzo (a) anthracene
Chrysene
Benzo (b)f luoranthene
Benzo (k)f luoranthene
Benzo(a)pyrene
D i benzo (a , h ) ant hracene
Benzo (g,h,1)perylene
I ndeno ( 1 , 2 , 3-cd ) pyrene
4
2
4
3
2
4
4
4
4
2
0.1
0.05
0.1
0.4
0.4
0.4
2.46
1.39
2.92
2.96
1.71
3.91
3.60
3.26
3.49
1.92
0.087
0.044
0.086
0.341
0.353
0.346
0.3254
0.571
0.8121
0.898
0.2451
0.5621
0.4601
0.2334
0.1706
0.1313
0.0151
0.0095
0.0143
0.0521
0.0587
0.0558
1.0
1.9
2.6
3.0
0.79
1.8
1.4
0.73
0.54
0.41
0.05
0.03
0.05
0.16
0.19
0.18
1.8
2.3
1.8
0.21
0.64
0.66
0.21
0.27
0.013
0.15
0.018
0.017
0.023
0.030
0.076
0.043
1 Number of samples analyzed = 7, except where noted otherwise.
2 40 CFR, EPA Method 620, 1986, NIOSH Method 5506, 1985.
364
-------
DISCUSSION
GIGI BEAR: Is there any reason you are using HPLC, instead of SFC (super
critical fluid chromatography) for the PAHs. In the last year, it's been proven
as very efficient, very fast, and you don't have to use solvents.
VAN EKAMBARAM: These conditions were primarily set up because the
HPLC was already on site. They were set up, to some degree, to do the
extractions with the solvents.
RONALD MITCHUM: Did you find any chlorinated PAH's? Did you find
any metabolites due to microbial degradation in the soil - like hydroxylated
metabolites or methylated PAHs? Could you describe what kind of HPLC you
used? And what was the detector? And could you make a comment about using
afluorescence detector here instead of, maybe, a UV detector, if that's what you
used?
VAN EKAMBARAM: We did not see any chlorinated PAHs. There are
several things that elute. I think those are the volatile organic compounds, and
pentachlorophenol, which elutes before that. Since pentachlorophenol was one
of the compounds used on site, that did not surprise us.
In fact, we also tried to change the chromatography to get quantitative data on
the pentachlorophenol by changing the solvent strength a little bit, as well as
some of the other compounds. B ut time and budget ran out, so we did not pursue
that.
We also tried to institute a bioremediation program on site to degrade the PAHs,
and we have been fairly successful in that. We have been injecting hydrogen
peroxide, and we see fairly good degradation of the PAHs. We were able to
create nontoxic zone, and two of them look like they were cleaned up. They
initially showed, before the peroxide injection, about 500 ppb total PAHs. Over
a period of six months we did not see any degradation. But after six months.
there was a breakthrough of oxygen, and correspondingly, the PAH concentra-
tions decreased quite a bit.
As part of this study, we wanted to use HPLC to look at the transformations we
are going to make. Maybe you are cleaning up naphthalene and other PAHs that
are in the ground water. But what are we producing? Are there organics which
could be present in the transformation products because of these bacteria. I
reviewed all the chromatograms, and they look very clean, even though there
are some peaks that I could not identify because I didn't have a standard. I think
by two or three months after we think it has been cleaned up, it's fairly clean.
So there may be have been some transformation products, but they're all
cleaned.
As to the kind of HPLC used, don't recall the exact number. We had a UV lamp
on the UV detector. The fluorescence detector was an add-on item. We looked
at fluorescence and UV detectors signals together. And as one would expect, the
higher molecular rate PAHs had a very high cross section for fluorescence. We
looked at both the UV signals and the fluorescence signal, and the correlation
was pretty good. Typically, what we did on those 300 samples was to run the
sample at 50, using UV-1, because for the lighter PAHs fluorescence is not that
good. For the heavier PAHs, like a 85-15 mixture, we would use both the UV
and fluorescence detectors. Finally, the data that we published were of the best
data that we got between the different injections.
RONALD MITCHUM: The creosote area is one that is overlooked quite a bit.
Still, the treatment of wood products with creosote goes on. There are some
changes being made in the wood treating industry, but we still have a lot of
treatment in that area. Not only sites are contaminated, but also current sites that
are active. There's a process, I think, that needs some looking at.
365
-------
SPECIFIC DETECTION OF ANY GAS CHROMATOGRAPHABLE
ELEMENT IN SEDIMENT EXTRACTS
Michael Szelewski and Michael Wilson
Hewlett-Packard Company
Avondale, Pennsylvania 19311
Abstract
A novel microwave-induced helium plasma detector has
been developed for gas chromatography. Atomic
emission spectroscopy (AES) is used as the detection
technique. The system can detect the component
elements of any compound amenable to analysis by gas
chromatography. This includes elements which have
been impossible or difficult to monitor with other GC
detectors.
The technique has been applied to the analysis of
sediment extracts. Compound and class identification is
improved compared to less specific or more tedious
analysis schemes. Research has shown that elemental
response may be independent of analyte molecular
structure. As a screening technique, GC/AES has the
ability to show the presence or absence or any gas-
chromatographable element, such as chlorine, bromine,
oxygen, nitrogen, phosphorus or sulfur.
Introduction
There is a wide variety of gas chromatograph detectors
currently used for the analysis of hazardous wastes. One
of these, the flame-ionization detector, is chosen for its
universal capabilities in screening samples. Other more
selective detectors, such as the electrolytic conductivity
and flame-photometric detectors, are chosen for their
compound specificity. This specificity is used in
differentiating analytes from one another and from
matrix interferences. Additionally, the sensitivity of
various detectors must be considered when dealing with
environmental samples.
Most recently there has been a shift in GC detector usage
toward hyphenated techniques. These techniques use
mass spectrometry or Fourier transform infrared
spectrometry coupled to GC resulting in GC/MS and
GC/FTIR respectively. GC/MS is the tool most
commonly used for unequivocal identification of
hazardous compounds. GC/FTIR has also been used for
structural information elucidation and compound
identification of hazardous wastes v1/. Combining
information from both GC/MS and GC/FTIR has been
done to improve the number of compounds that coulcLbe
positively identified In a hazardous waste site sample P'.
Presented here is a newer hyphenated technique which
combines gas chromatography with atomic emission
spectroscopy (GC/AES). Two recent review articles
provide a thorough discussion of microwave helium
induced plasma (MPD) utilized in this technnique (•**).
The detector in this work is of experimental design ,as
shown in Figure 1 and has been previously described ^
'). The GC/AES has the ability to detect any gas
chromatographable element, with a high degree of
selectivity and sensitivity shown for those elements that
are commonly found in hazardous waste compounds.
Work up to this point shows that GC/AES response for
any element may be independent of the compound
structure that contains the element. This compound
independent response can yield element ratios, based on
internal standards, that supplement structural
information gained by GC/MS and GC/FTIR. These
element ratios can be used to reduce time in spectral
interpretation of unknown compounds found during
hazardous waste site screening.
The combined GC/AES/Data system has similar size
and weight to current field laboratory instrumentation
(GC/MSD, GC/FTIR). Utility requirements (power,
gases) and mobile laboratory room conditions
(temperature, ventilation) are also similar.
Experimental
Gas Chromatography
HP 5890A Gas Chromatograph
Split/splitless injection port at 275°C
Split ratio = 25:1
HP 7673A Automatic Liquid Sampler
Injection volume = 1 ul
5% Phenyl methyl silicone FSCC
Initial Temp = 40°C
Initial Time = 0.5 min
Program rate = 10°C/min
Final Temp = 280°C
Final Time = 20 min
Atomic Emisssion Spectrometer
Expermental equipment utilizing a microwave
induced helium, masma and movable diode
array detector*--5'"-'.
Mass Spectroscopy
HP 5970B Mass Selective Detector
367
-------
Fourier Transform Infrared
HP 5965A Infrared Detector
Sample
Sediment was gathered from a river bank
within an abandoned chemical dumpsite in
New York State. For a complete description
of sample preparation see page 1820 of
reference #1.
Results
The element specific chromatograms are shown in Figure
2a and 2b for the sediment that was analyzed. Various
combinations of up to four elements can be detected
simultaneously. Examples of element combinations and
the corresponding emission wavelengths are listed in
Table 1.
Table 1. Element Wavelength Groups
Element
Groups
Emission
Wavelenght (nm)
Carbon
Hydrogen
Chlorine
Bromine
Carbon
Nitrogen
Sulfur
Fluorine
Oxygen
247.9 (2nd order)
486.1
479.5
478.6
193.1
174.2
180.7
685.6
777.2
Having element specific chromatograms gives excellent
information about unknown samples. Figure 3 shows the
content of nine elements in the compounds in the dump
site soil extract that have been separated by gas
chromatography and detected by atomic emission
spectroscopy.
For screening purposes the presence of gas
chromatographable compounds that contain specific
elements can readily be determined. There exists the
possibility for interference in element detection from
another element present in large quantities, such as
carbon.
Element selectivities and sensitivities are shown in Table
2. Also included are sensitivities for other common GC
detectors. The AES detector compares favorably with
these GC detectors while combining the specificity of
more than one of these.
During remediation of a hazardous waste site, where
contaminant levels should be decreasing, the GC-AES
can be used to detect low levels of these contaminants.
Additionally qualitative information on the compounds
of interest and any new ones that may appear can readily
be obtained from the element specific chromatograms.
A second area in which the GC/AES can contribute in
the screening process is that of site characterization. A
commonly accepted technique for compound
identification is GC/MS. For target compound analysis
automated library search and quantitation work well.
For non-target compound analysis, however, GC/MS can
present problems. Detected analytes may not be present
in on-line search libraries or the quality of the library
match may be too low to positively identify the
compound. Other hyphenated techniques such as
GC/FTIR can provide additional structuaJ information
for unknowns in environmental samples ( '. There still
exists the possibility for disagreement between GC/MS
and GC/FTIR library searches for the aforementioned
reasons.
Table 2. Typical GC/AES Results with
Comparison to Other GC Detectors
GC Detectors AES Detector
Select MDL Atom Select MDL
FID
TCD
NPD 35K
70K
1000
0.5
0.1
BCD 1-100K 0.02
FPD 10K 2
0.5
ELCD 100K 1
C
H
N
P
Cl
Br
S
P
Cl
0.5
1
25K 15
50K 1
25K 15
10K 15
10K
50K
25K 15
Select. = relative to C
MDL = pg/sec of element
Manual interpretation of the spectra can be time
consuming. The GC/AES can speed up the
interpretation process by providing the analyst with:
1) information on the specific elements that are
present or absent, and
2) approximate ratios of the elements based on
an internal standards.
Figure 4a illustrates chromatograms from all three
spectral detection techniques. The carbon and chlorine
specific shromatograms from atomic emission of Figure
4b are used in the following discussion in comparing
information from MS, FTIR and AE.
368
-------
As an example, Peak #1 was identified by GC/MS and
GC/FTIR to be o-chlorotoluene. Known internal
standards could have been added to the extract prior to
GC/AES analyses of course, but this peak had already
been identified. Peak #2 was not detected by GC/FTIR.
The GC/AES data gave a C:C1 ratio of 1:0.29 which
supports the second best match in the GC/MS library
search. Oxygen was not present which eliminated the
possibility of the best library match by GC/MS.
In previous work^ \ peak #3 had been tentatively
identified as an alkane. GC/AES showed the presence
of chlorine and the lack of other heteroatoms such as
O,F,N, S or Br. The C:C1 ratio of 1:0.1 supports the best
GC/MS library match found when the sample was used
with the HP 5970B in the current work.
Conclusions
The GC/AES has the following advantages and benefits
over current GC detectors for hazardous waste site
characterization and monitoring.
1) Element specific detection
* potential time savings in sample
preparation
2) Multiple element detection
* reduces space needed at on-site
monitoring laboratories by limiting
number of GCs needed with multiple
detectors
3) Compound independency for response factors
* saves time by reducing the number of
standards that have to be analyzed
4) Provides element ratios
* reduces time for interpetation of GC/MS
and GC/FTIR data for non-target
compounds
References
(1) Gurka, D.F. and Betowski, L.D. Analytical Chemistry.
1982, 54, p 1820.
(2) Gurka, D.F., Hiatt, M.H. and Titus, R.L.," Nontarget
Compound Analysis of Hazardous Waste and
Environmental Extracts by Combined
FSCC/GC/FT-IR and FSCC/GC/MS", Hazardous
and Industrial Waste Testing: Fourth Symposium,
ASTM STP 886, J.K. Petros Jr., W.J. Lacy and R.A
Conway, Eds., American Society for Testing and
Materials, Philadelphia, 1986, p 129-161.
(3) Uden, P.C., "Element-Selective Chromatography
Detection by Atomic Emission Spectroscopy .
Chromatography Forum. Nov-Dec 1986, p 17-26.
(4) Ebdon, L., Hill, S. and Ward, W., "Directly Coupled
Gas Chromatography-Atomic Emission
Spectroscopy: A Review". Analyst. Ill, October
1986, p 1113-1138.
(5) Quimby, B.D. and Sullivan, J.J., "An Improved
Microwave Cavity, Discharge Tube, and Gas Flow
System for GC-AED", R.M. Barnes, Ed., Proceedings
1988 Winter Conference on Plasma
Spectrochemistry, San Diego, January 1988, p 225.
(6) Sullivan, J.J. and Quimby, B.D., "Characterization of
a GC-AES Using a Novel Photodectctor,
Spectrometer and Computer System". R.M. Barnes,
Ed., Proceedings 1988 Winter Conference on Plasma
Spectrochemistry, San Diego, January 1988, p 225.
(7) Wylie, P.L., Quimby, B.D. and Sullivan J.J.
"Application of an Experimental Microwave-Induced
Helium Plasma Detector for Gas Chromatography",
R.M. Barnes, Ed., Proceedings 1988 Winter
Conference on Plasma Spectrochemistry, San Diego,
January 1988, p 181.
369
-------
Auto-
matic
Sampler
Figure 1. Diagram of Experimental GC/AES System
370
-------
2a
Chemical Dump Site Soil Extract
Carbon, Nitrogen, Sulfur Screen
I 1_
L
X(nm)
JJ
Carbon 193.0
i. .ilj.^ 1.
Nitrogen 174.2
10
2b
i L
1 i i
T1 me
Sulfur 180.7
— 1.1 . . , j
P* " pT 'i i" i-n i i'> I.KI n.i. «. >iin«i v' ' i ' •
15 20 25
( mi n . )
Carbon, Chlorine Screen
I,
i.
.
L.
I
LJ
u.
x(nm)
,
11
ll.
Carbon 495.7
.l.L ,1 hV .i .
Chlorine 479.5
i 1 \
.Ul_iLJL.i A..I...J.,. ....
10 15
T1 me (mln.
20
25
Figure 2. Examples of Screening Chromatograms
371
-------
Chemical Dump Site Soil Extract
ILJ
1 L 1 1 . . . 1. .. .
ll. . 1 L.I... . ..
1
L.n . .. .1 li.i. l.
Carbon
Hydrogen
Chlorine
495.7
Bromine
L 1 Fluorine
i . Oxyqen
i Sulfur
478.6
685.6
777.2
180.7
Phosphorus 178.1
Nitrogen 174.2
10 15
T 1 me ( m i n .
20
25
Figure 3. Nine Element-Specific Chromatograms from GC
Analysis of Chemical Dump Site Soil Extracts
372
-------
Chemical Dump Site Soil Extract
a 1
1 L
Compound Identification
1
, L
3
I
Jl.LL_JL
AE
LJL___
Carbon
495.7 nm
4b
_JUUV_A
10 15
T1me (m 1 n . )
\A * A
IR
TRC
MS
TIC
25
Carbon 495.7 nm
31
Chlorine 479.5 nm
-i—• • i—i—r-
I—I—
16
10 11
12 13
T 1 me C m i n
14 15
Figure 4. Chromatograms from GC/AE, GC/MS and GC/IR
373
-------
DISCUSSION
JOHN EVANS: I was curious about the 40 to 1 split. It's fairly high. Why you
did not discuss the hydrogen results and your ratios?
MICHAEL SZELEWSKI: The split was run primarily because of the high
concentration of the extract, and not wanting to dilute it down too many times
and subject it to potential contamination with other things. The extract is so
highly concentrated that without running it split on the atomic emission
detector, we would be outside the linear range on the carbon channel. The only
reason I didn't show the hydrogen channel is that it looks very similar to the
carbon channel.
JOHN EVANS: You must have a ratio of some sort. You were just showing
partial formula. And that's all. Why didn't you ratio the hydrogen?
MICHAEL SZELEWSKI: No particular reason. The hydrogen ratios on
these particular compounds so work out fairly well.
ED HEITHMAR: I notice that for things like bromine, you only have a linear
dynamic range of about 1,000. Is that due to self absorption or to energy-
limiting factors in the plasma?
MICHAEL SZELEWSKI; I don'treally know. I don't think it's due to energy
levels in the plasma. I think there are interactions inside the plasma within the
walls of the tube in the cavity. That has been the biggest limiting factor up to
this point. Getting the proper combination of "secret sauces" and reagent and
make-up gases has been the single biggest thing that has improved dynamic
range.
ED HEITHMAR: So with high concentrations, you don't see a great deal of
plasma poisoning for coeluting compounds - hydrocarbons on top of some-
thing else.
MICHAEL SZELEWSKI: No.
UNIDENTIFIED PARTICIPANT: One of the problems that I've always
noticed is that your ability to run heteroatoms is great. Our ability to get
heteroatoms to the GC is not so great. Unfortunately, a lot of sulfur-containing
compounds and phosphorus-containing compounds chromatograph very well.
Can you comment about the future of HPLC interfaced to this thing?
MICHAEL SZELEWSKI: At this point, it's not even available in the GC
version. As far as HPLC on the priority list of things to look at, it is there. It is
the next obvious choice, but there is no direct research going on right now. I
would say that it's probably number one on the list of things to do next, as far
as this specific detector is concerned.
374
-------
THE U.S. EPA FIELD ANALYTICAL SCREENING PROJECT (FASP)
G. Hunt Chapman
Ecology and Environment, Inc.
Arlington, Virginia
Scott Fredericks
Hazardous Site Evaluation Division
U.S. Environmental Protection Agency
Washington, D.C.
ABSTRACT
In response to the concerns raised by Congress
and the public, the U.S. Environmental Protection
Agency (U.S. EPA) is evaluating means of
improving charaterization of hazardous waste
sites through enhancing on-site analytical
capabilities for the investigation of hazardous
waste sites. The Field Analytical Screening
Project (FASP) has been established in an effort
to address these concerns. FASP will be used to
generate quick turnaround screening data where
the rigorously qualified CLP data is not
required. It is designed to provide screening
data for specific compounds known to be on the
site from previous CLP data. FASP will be used
mainly during Listing Site Inspections (LSIs) to
help determine the extent of contamination and in
choosing samples to document a contaminant
release for the Hazard Ranking System (HRS).
A Base Support Facility (BSF) is located in each
region to serve as a staging area for all field
activities. A support vehicle is equipped to
perform screening for target compounds at the
site. Instrumentation will be stored at the base
support facility and transferred to the support
vehicle as needed for specific sites.
INTRODUCTION
Changes in the pre-remedial process since the
enactment of SARA have increased the need for
on-site screening. The development and
implementation of a national field screening
program depends on a framework that provides a
consistent approach to produce data of a known
quality. However, this framework must be
flexible enough to meet specific regional needs.
The FASP approach described in this paper
provides a structure that meets this dual
requirement.
THE FASP PROGRAM
The FASP Approach and Structure
U.S. EPA goals have been used to develop a FASP
concept paper which is used to define the
approach, structure, and role, as well as to
assist in the implementation of FASP. In the
concept paper, FASP Data Quality Objectives
(DQOs) define the levels of data quality required
for specific tasks performed during a site
inspection. These tasks are discussed later in
this paper.
A FASP working group has been established to
develop and expedite FASP. It consists of
knowledgeable and experienced U.S. EPA and
contractor personnel from each region. These
include the National FASP project manager and the
Zone I and II Screening Managers (ZSMs) to
coordinate and implement the national management
plan, and Regional Screening Coordinators (RSCs)
from each region to implement regional programs.
The RSCs have conducted regional feasibility
studies to determine detailed plans and assist
final decisions regarding the level and type of
field screening to be undertaken in each region.
This process began by asking regional data users
what quality of data they needed for specific
pre-remedial tasks. Defining this level of data
quality facilitated the determination of the
appropriate levels of instrumentation, and the
regional approach and application of field
screening.
A national FASP framework is necessary to provide
the consistency and coordination needed to
support U.S. EPA pre-remedial goals. This
structure is a very important aspect of the FASP
concept. It is managed through centralized
organization and coordination by the Zone
Screening Managers (ZSM) and the FASP working
group. All aspects of FASP are scrutinized and
discussed by working-group members to insure that
the over-all project goals are met and that they
are standardized for all regions. The nationally
coordinated FASP program includes the following:
Analytical methods consistency of the
methodology is extremely important. The use of
standardized methods insure a consistent
capability and a predictable level of data as
defined by the DQOs. A FASP methods manual
containing approved analytical methods is
distributed to each region.
QA/QC protocols - A FASP QA/QC manual is provided
to each region detailing a comprehensive QA/QC
protocol. It includes proper calibration
375
-------
techniques, use of matrix spikes and duplicates,
personnel training, documentation procedures,
maintenance schedules, use of analytical-grade
reagents, etc.
Safety procedures A FASP safety manual is
issued to all regions. This includes safety
protocols to be followed during all FASP
activity.
Personnel training A training program is
established in conjunction with the appropriate
product vendors as well as with experienced FIT
chemists. A FASP training center is established
where FIT chemists can receive additional
hands-on training if needed.
Analytical instrumentation - national
coordination of instrumentation purchases ensures
that FASP goals are met. This does not mean that
all regions must use the same instrumentation.
Data Quality Objectives (DQOs) FASP DQOs are
used to define the data quality required for
specific on-site tasks. They are to ensure that
specific, consistent standards are maintained for
on-site screening during various pre-remedial
activities in all U.S. EPA regions.
All of these aspects of the centrally managed
FASP project are essential in providing
consistent, reliable data that can be used with
confidence by the U.S. EPA in all regions.
In general, the FASP emphasis is the screening of
target compounds known to be onsite from previous
sample analysis. This screening data is used to
complement CLP data as required by data users.
This approach allows for customized screening of
only the compounds of interest, which shortens
the time required for screening. Customized
screening methods are developed for many organic
compounds on U.S. EPA's Target Compound List
(TCL). Methods and QA/QC protocols are defined
for all FASP screening to provide a defined data
quality that meets established DQOs. FASP is
presently implemented in six "pilot-study"
regions. If the pilot studies prove to be
beneficial, FASP will be implemental in the
remaining U.S. EPA regions.
The Role of FASP
FASP is intended to support site investigation
activities during a Screening Site Inspection
(SSI) or a Listing Site Inspection (LSI) where
target compounds are identified from prior
analysis. Samples are screened specifically for
these target compounds and the methods are
optimized for them. FASP is not well suited to
screen for a broad range of unknown contaminants.
Uncharacterized sites and unknown contaminants
require the usual CLP analysis which includes
mass spectral confirmation for organic compounds
and Inductively Coupled Argon Plasma (ICAP) for
metals. FASP is best utilized on large sites
that require screening a small number of
contaminants in many samples. As mentioned
above, the screening methods are customized for
these target compounds, which allows for rapid
screening at a lower cost. In some cases, the
use of specific instrumentation to optimize
screening methods for target compounds allows for
detection limits below Routine Analytical
Services (RAS) CLP results. A portion of the
FASP samples can be split and sent to the CLP for
confirmation.
FASP can provide several important benefits to
the inspection team on the site. One of the most
important of these is on-site feedback. In the
past, Field Investigation Teams (FITs) had to
rely on visual inspection or intuition to choose
sampling points during inspections. This has
often resulted in insufficient site
characterization or the analysis of inappropriate
samples and return trips to the site for
resampling. On-site feedback of FASP screening
results can aid the FIT in site characterization.
This includes finding "hot spots" on site and
determining the extent of contamination. This is
especially useful at large sites where an equal
number of CLP samples would otherwise be cost
prohibitive. Once the FASP data is reported, the
FIT team can reliably determine the appropriate
samples to be sent to the CLP for confirmation.
These CLP samples are then used for litigation
purposes or other uses requiring more rigorous
protocols.
On-site feedback is also used to direct on-going
work, redirect sampling efforts and modify work
plans. This includes the installation of
monitoring wells where screening data can
determine the depth of well screen placement.
Another application is in removal operations
where screening data can determine if enough
contaminated material has been removed while the
removal equipment is still on site. When FASP
results indicate that enough material has been
removed, a sample sent to the CLP can be used for
confirmation.
Another benefit of on-site screening is the
optimization of air monitoring programs. Air
monitoring for volatile organic compounds is
performed with composite samples using absorbant
tubes such as Tenax, or with grab samples using
Tedlar bags or some other suitable container.
When using Tenax, the time between sampling and
analysis must be minimized. This is because
Tenax is such a strong adsorbant, it can become
contaminated while in storage. Other factors
such as artifact formation and break-through
often complicate and/or invalidate the results.
In addition, grab samples taken on site have very
short holding times and must be analyzed on site.
Therefore, it is best to perform air monitoring
on site where holding times are minimized and
sampling problems can be corrected while the
samplers are still there.
Soil-gas sampling for organic contaminants is a
technique that can best be used in conjunction
with field screening. It is a technique used
primarily to delineate the horizontal profile of
volatile organic groundwater contaminant plumes.
It improves the placement of monitoring wells
used to document groundwater contamination. The
soil-gas sample can either be a composite sample
376
-------
(Tenax or other adsorbant) or a grab sample.
These samples should also be analyzed on the site
for the same reasons discussed for air samples.
FASP Implementation
FASP has been implemented through a phased
approach, i.e. building blocks. Each phase
represents a specific TCL fraction or group of
compounds. The phased approach is used to allow
regions to select analytical capabilities based
on their particular requirements. Each phase's
implementation includes instrumentation,
screening methods, training, and QA/QC protocols.
This comprehensive approach ensures that all FASP
results meet the required level of data quality
established in the DQOs. Each phase establishes
specific capabilities and builds on the previous
phase. The phased approach addresses the need to
implement a basic field screening capability
while ensuring that steps are taken to administer
the program properly, thereby meeting the goals
of proper training, documentation, and
instrumentation.
Phase 1 involves screening for volatile organics
in water, soil, and grab air samples. The water
and soil samples are screened using a portable
Photovac gas chromatograph (GC) using the
head-space technique. The air samples are
screened directly from the Tedlar bag using the
Photovac. The Photovac GC was chosen for Phase 1
because it is portable and can obtain very low
detection limits for many TCL compounds. The
portability is needed so that FIT chemists can
easily go to the field with minimum support. A
field kit was fabricated by the FIT to support
the Photovac in the field. It contains all
necessary support equipment needed to screen
samples on site. This approach for Phase 1
allows for the quickest and easiest
implementation of a field screening capability
while the FASP support facilities are being
procured and/or constructed.
The Base Support Facility (BSF) is also part of
Phase I. This is used as a staging area for all
field screening activity. It is used for
calibration and maintenance of instrumentation,
storage of supplies, method development, and to
screen samples from smaller sites (with fewer
samples) when mobilization of the support vehicle
is cost prohibitive.
Phase 2 involves screening for PCBs in water and
soil. This phase involves the use of compact
transportable GCs and equipment required to
perform basic sample preparation and extraction.
PCB analysis requires more sophisticated
equipment and sample preparation facilities such
as utilities, shelter, and adequate working
space. This requirement is the justification for
the procurement of the support vehicle which is
part of Phase 2. This support vehicle is also
required for all field screening performed in
subsequent phases.
The support vehicle is equipped to handle
simplified sample preparation and/or extraction
as well as the space needed for several small
GCs. Because of the limited space in the support
vehicle and the customized screening at each
site, the support vehicle will be equipped with
only the instruments and equipment needed for the
target compounds at that site. The Base Support
Facility is used to keep all instruments and
other equipment in "field-ready" condition when
needed, When the Support Vehicle is not in the
field, it will be parked next to the Base Support
Facility to increase the total lab working space
and allow for rapid outfitting for the field.
The remaining phases are:
o Phase 3 - Screening for chlorinated
pesticides in water and soil.
o Phase A - Screening for selected
base/neutral organics in water
and soil.
o Phase 5 Screening for selected
acid-fraction organics in water
and soil.
o Phase 6 Screening for selected metals
in water and soil.
o Phase 7 - Special screening, i.e.
composite air sampling, GC/MSD,
herbicides, Purge and Trap,
etc.
Phases 2-5 all use similar types of
transportable, compact GCs equipped with
different columns and specialized detectors.
Phase 6 can use one of two methods for metals
analysis. First, Atomic Absorption
Spectrophotometry (AAS) is the method more
similar to CLP methods. However, there are
several disadvantages for field screening. The
samples must be acid digested to get acceptable
results for most samples. This can be both
dangerous and time consuming in the confined
space of the Support Vehicle. Also, only one
metal can be screened at a time. The biggest
advantage of AAS is that the method detection
limits are comparable to CLP. The second choice
is X-Ray Fluorescence (XRF) Spectrophotometry.
This method does not require acid digestions and
most TCL metals can be screened simultaneously.
Its disadvantages are that the detection limits
are very high and that matrix corrections are
often required for an accurate result. This can
be a problem when screening differing and unknown
environmental matrices.
CONCLUSION
In conclusion, it should be noted that the
capabilities of sample screening and laboratory
analysis in general are governed by instrumental
sophistication and capabilities. As FASP
continues and instrumentation develops further,
the quality of screening data will also improve.
The importance and benefits of on-site feedback
and matching of the screening methods to site
DQOs cannot be overstated. Proper management and
implementation of FASP can greatly benefit the
entire pre-remedial program in terms of
efficiency, cost savings and time savings.
377
-------
CONCLUDING REMARKS
LLEWELLYN WILIAMS
These past three days you have been exposed to a fairly broad
range of technologies; perhaps some of you were surprised with
the diversity. I think you may also have been surprised that some
of these technologies are as far along as they are, that they are out
there, and data are being produced with them.
Some of you were also probably surprised that some of the
technologies we have heard about for years are not quite where
we thought they were, and that some of them are going to require
new developments and breakthroughs in materials and our
approaches, before they can become true monitoring, or measure-
ment tools that we can use in the field.
What was important was to bring this kind of group together in
one place. We have a balance of academics, we have the people
from the private sector, from commercial operations. We certainly
have a good representation from the Federal sector and States, as
well.
It's important that we know where this technology is. Don't be
disappointed that it isn't where we hoped it would be. Let us
know where it is, so that we can do our planning accordingly. This
is very important for all of us, certainly for us as regulators.
We saw some opportunities to perhaps save money in the way
we do our operations in the future. Cost savings are great, but not
if they are at the expense of the environment or human health and
safety. We need good quality assurance to be associated with our
field methods. We need good data quality objectives. We need to
know what kind of data we must have to make critical decisions.
Also, we may want to think about taking those cost savings
and turning them back into better monitoring and better measure-
ment. Rather than think only in terms of saving money, let's get
more confidence in the decisions that we make, upon which our
environmental and human health and safety depend.
We have seen opportunities to perhaps avoid some of the high
cost of environmental zeros, using lower-cost technologies that
enable us to key on those samples that are more appropriately
taken to a laboratory for high tech analysis.
We know of programs in the past that could have saved
hundreds of thousands of dollars if they had a screening method
available to them — where the high cost of conventional method-
ology ran us into the ground.
Some of these operational hurdles will be overcome. But it
takes time to see these technologies develop, and to see the
changes. And since some breakthroughs are necessary in some of
the technologies, it will take time.
We are in a position in many cases to be able to help. Can we
do it all, do we have the kind of bucks to support everybody's
activity? No. We need to be looking for those kinds of technolo-
gies that are going to help us do our job in the near term, but with
an eye toward those special high-potential technologies that are
going to help us all in the future.
379
-------
QUALITY ASSURANCE PLAN USED AT THE LOVE CANAL
EMERGENCY DECLARATION AREA INDOOR AIR
ANALYSES BY THE TAGA 6000E MASS SPECTROMETER/
MASS SPECTROMETER
Thomas H. Pritchett
U. S. EPA Environmental Response Team
Edison, New Jersey
David B. Mickunas & Nicholas Kurlick
International Technology Inc. (REAC)
Edison, New Jersey
ABSTRACT
A quality assurance/quality control (QA/QC) plan
was developed for a Trace Atmospheric Gas Analyzer
(TAGA) used in monitoring indicator chemicals in
ambient air during the Love Canal Emergency
Declaration Area habitability study. The QA/QC
requirements were instituted to assure the data
would be technically sound and legally defensible.
Eight data quality objectives were defined for the
study (Table I). Six of these objectives related
directly to the instrument's performance, which
set criteria for accuracy, precision, detection
limits, sensitivity decay, and sampling
efficiency. The remaining objectives were
concerned with the comprehensiveness of
documentation and sampling.
INTRODUCTION
In the Autumn of 1986, the US EPA Region II
requested the assistance US EPA Emergency Response
Team (ERT) by making the TAGA 6000E Mass
Spectrometer/Mass Spectrometer (MS/MS) available
during the indoor air analyses for the Love Canal
Emergency Declaration Area (EDA) Habitability
Study. As one of three environmental studies
cited in the "Love Canal Emergency Declaration
Area Proposed Habitability Criteria , the
purpose of this study was to report the results of
the air assessment for indicator chemicals. These
results, together with the other environmental
studies, are intended to provide data to assist
the New York Commissioner of Health in determining
the habitability of the structures within the
Emergency Declaration Area.
The air assessment study was designed to monitor
residential structures in the EDA for chemicals
whose presence in indoor air would strongly
suggest that they originate from the Love Canal.
Two compounds and their structural isomers,
referred to as the Love Canal Indicator Chemicals
(LCICs), were chlorobenzene and chlorotoluene.
The EDA sampling strategy included seasonal
variations, diurnal variations, rainfall and
groundwater levels, occupied versus unoccupied
status, and sampling randomness. Furthermore, the
sampling strategy asserted that the presence of
any LCIC in the indoor air of a residence is
adequate indication that the habitability of the
residence requires further evaluation. The latter
was the recommendation of the Technical Review
Committee (TRC), which is composed of
representatives from state and federal agencies
involved with the Love Canal. This requirement
was a result of the "Pilot Study for the Love
Canal EDA Habitability Study"2 conducted in 1986
where the number of detections was too limited for
statistical analysis and, therefore, would require
an excessive number of controls to be sampled.
Consequently, the TRC made the decision to sample
each physically-sound, residential structure in
the EDA, assuming permission could be obtained
from the people living in these structures, in
lieu of using a control population.
PROCEDURE
Appendix A of the "Love Canal Full-Scale Air
Sampling Study Quality Assurance Project Plan"3
contained a detailed discussion of the standard
operating and reporting procedures used by the
TAGA group during the study. Those operational
procedures directly related to the TAGA quality
assurance plan are:
1) Initial Tuning At the start of each analysis
period, the instrument was readied.
2) Instrument Calibration At the start of each
analysis period, before each residential
sampling, and prior to final QA/QC analyses,
the TAGA was calibrated by using the LCICs.
3) Measurement of Sampling Line's Transport
Efficiencies At the start and end of each
analysis period, but after the instrument was
tuned and calibrated, the transport
efficiencies of the sampling line for the
LCICs were determined.
4) QA/QC Sample Analyses At the start and end
of each analysis period and during external
audits, QA/QC sample were analyzed to measure
the accuracy, precision, or detection limit
and quantitation limit verification.
5) Sample Air Flow (SAF) Transducer and Mass Flow
Controller (MFC) Calibration At the start
and after the end of each sampling phase,
381
-------
every week during a phase, and after repairs,
the SAP transducer and the MFC were calibrated
using an NBS traceable standard.
RESULTS AND DISCUSSION
Overall project management coordination was the
responsibility of CH2M Hill, Inc., the Love Canal
remedial contractor, under the direction of the US
EPA Region II. CH2M Hill responsibilities
included designing the overall sampling program,
generating the overall Quality Assurance Project
Plan (QAPP), and coordinating of sampling process
and its review. The ERT assumed responsibility
for developing the TAGA analytical methodologies
and for developing the quality assurance plan for
the TAGA-generated data. Northrop Services, the
operations contractor for the Quality Assurance
Branch of the EPA's Environmental Measurements
Support Laboratory at Research Triangle Park, NC
(EMSL/RTP), assumed responsibility for providing
external field QA audits and TAGA performance
evaluation (PE) samples during all phases of the
study. The TRC had final approval authority over
all phases of the sampling program, including the
TAGA quality assurance plan.
The design and data quality objectives (DQOs) of
this project received intense examination. The
QAPP, which incorporates the TAGA Standard
Operating and Reporting Plan (SORP) and the
sampling program, was not approved until it was
thoroughly reviewed by all agencies represented on
the TRC and by the Quality Assurance Branch of
EMSL/RTP. The DQOs were instituted to assure data
acquired would be technically sound and legally
defensible. These goals were achieved by:
1) ensuring that the required sensitivity was
met;
2) ensuring that the required precision and
accuracy were met;
3) ensuring that the response factor drift was
within specifications;
4) ensuring that the ancillary equipment was
functioning properly;
5) ensuring that the data was reviewed in
accordance with the QAPP; and
6) ensuring that the sampling was complete.
Associated with the DQOs were one of the three,
following levels of action. These levels were:
1) R Requires sampling or resampling of a
residence;
2) A Requires correction of problem before
sampling can continue; and
3) G Requires additional PE analyses, but
sampling can continue.
The "R" criteria were applied to the sampling and
documentation completeness. The "A" criteria were
applied to instrument and ancillary equipment
performances involved with the analyses (with the
exception of the accuracy data quality
objective). The "G" criteria were applied to the
accuracy DQO (due to the experimental nature of
certification of PE Summa canisters).
The accuracy DQO was modified for the initial
segments of the study. Due to problems with
certification, no data was available for Phases 1
and 2 for the 6-liter Summas or for Phase 1 for
the 16-liter Summas. However, a Scott gas
cylinder, other than the one used for calibration,
was analyzed as an internal check. The measured
error exceeded 25% only once in the program and on
that day, three separate accuracy checks were
performed with two within tolerance. The
verification of detection and quantitation limits
by spiking the sample 1 ppb and 2 ppb,
respectively, above their limits never failed.
Furthermore, the Northrop Services provided for
each phase external audit Summa canisters to
assure the required accuracy condition was met.
The analyses of these canisters by the TAGA always
had the mandatory accuracy.
The precision data quality objective required
that, within a phase, standard deviation from the
mean be less than 25%. In all cases this
objective was met.
The results of the detection limit data quality
objective are summarized in Table II. This
specification demands that the detection limit be
less than 4 ppb for both LCICs. All of the
reported LCIC detection limits met this criteria
when rounded to the closest whole number. Only
one detection limit, 4.2 ppb, for one of the LCICs
was greater than 4.0 ppb. Even though this
detection limit satisfied the DQO when rounded
down, the house involved was resampled during that
phase as a within-phase replicate house analysis.
Additionally, it was chosen as one of the houses
that was analyzed during each of the remaining
phases as an overlay analysis The four additional
sets of detection limit data met the detection
limit specification.
The data quality objective requiring that average
response factor decay percent between consecutive
calibrations for both LCICs be less than or equal
to 15% was exceeded several times for one or both
LCICs during the study. In all cases the
appropriate corrective action was taken and the
sensitivity stabilized. Additionally, in all
cases the response factors acquired after the
excessive decay were sufficient to give the TAGA
the required sensitivity to continue operations.
The calibration system is diagrammed in Figure 1.
The accuracy of the TAGA calibrations, as
determined by the accuracy of the flow rate
measuring devices, was a data quality objective.
Two different flow measurements were used in
calculating the LCIC calibration concentrations.
The mass flow controller (MFC) measured the flow
rate of the calibration gas and the sample air
flow (SAF) measured the flow rate of the ambient
air dilution stream. The specification for the
accuracy for all of the SAF and MFC calibration
points mandated that the error be less than or
equal to 10%. Additionally, during each phase the
Northrop Services conducted an audit of both the
SAF and MFC measuring devices using an NBS
traceable laminar element to check their
accuracy. All SAF and MFC calibration points
audited for all phases of the study had an error
less than 10%.
382
-------
The data quality objective requiring that the
total loss of LCICs in the sampling line be less
than or equal to 15% translated into a transport
efficiency percentage of greater than 85%. The
equipment for determining transport efficiency is
diagrammed in Figure 2. All percent transport
efficiencies for all phases of the study were
greater than 85% for both LCICs. During the first
phase a heated transfer line was used to maintain
the line temperature greater than or equal to
70°F. This was done in an effort to prevent
losses of LCICs by condensation. An experiment
conducted between the first and second phases
demonstrated that an unheated Teflon line could
transport the LCICs at temperatures as low as
32°F, An unheated transfer line was used for
the duration of the study. Additionally, the
ambient air temperature was monitored for each
transport efficiency test. The lowest temperature
observed was 31°F. The results of the transport
efficiencies for Phases 2-4 are plotted vs.
temperature in Figure 3.
Sampling 100 percent of the EDA structures for
which permission to gain entry was granted and for
which the structures was determined safe to enter
was a data quality objective. CH2M Hill was
responsible to assure this DQO was met. During
the course of this study, 562 EDA residential
structures were sampled.
The DQO requiring that the documentation be
complete and consistent for sampled residences was
met. Acceptable data packages were generated for
all residences sampled by the TAGA. The
consistency and completeness of the data packages
were checked by two different groups, internal and
external review teams, under the direction of the
ERT QA/QC coordinator. All of the data packages
generated during this study passed these two
reviews and all significant inconsistencies or
omissions were corrected.
The internal data review group was responsible for
ensuring that the documentation package was
complete, all data entries were consistent for the
documentation package, and all data entry errors
were corrected. The external data review group
was responsible examining the documentation
package for the completeness and consistency and
documenting on a comment log sheet to the ERT
QA/QC coordinator any anomalies. The ERT QA/QC
coordinator was responsible for reviewing the
comments by the external data review group and
categorizing the discrepancy as:
1) the anomaly does not affect the data quality
no action required;
2) the anomaly does affect the data quality
action by the data review group is required;
3) the anomaly does affect the data quality
action by the ERT QA/QC coordinator is
required; or
4) the anomaly does affect the data quality the
house must be resampled.
CONCLUSIONS
Technically sound and legally defensible data can
be acquired with the TAGA technology if the proper
QA/QC protocol is invoked. The QA/QC protocol had
redundancies to ensure anomalies in the data did
not corrupt the quality.
REFERENCES
1. "Love Canal Emergency Declaration Area;
Proposed Habitability Criteria," New York
State Department of Health, Albany, NY, 1986.
2. "Pilot Study for the Love Canal EDA
Habitability Study," CH2M Hill Southeast,
Inc., Reston, VA 1986.
3. "Love Canal Full-Scale Air Sampling Study
Quality Assurance Project Plan," CH2M Hill
Southeast, Inc., Reston, VA 1987.
TABLE I. TAGA Data Quality Objectives for the
Love Canal EDA Air Habitability Study.
Objective
1. Overall TAGA Accuracy as
Determined by Performance
Evaluation Analyses (% error
of reported concentration)
2. Overall TAGA Precision within
a Phase by Periodic Analyses
of Same Cylinder (% error of
reported concentration)
Criteria
Criteria Key
<25% G
<25%
3. Detection Limits for LCICs <4
(ppb)
4. Allowable Decay Between <15%
Consecutive
5. Accuracy of TAGA Calibration
(% difference, reported vs.
actual )
Sample Air Flow <10%
Mass Flow Controller <10%
6. Total Loss of LCIC in <15%
Sampling Line
(% loss of ion signal )
7. Residential Structures Sampled 100
(% of EDA structures for which
permission to gain entry was
granted and for which the
structure was determined safe
to enter)
8. Documentation Complete and 100
Consistent for Sampled
Residences (% complete)
1) R Requires sampling or resampling of a
residence;
2) A Requires correction of problem before
sampling can continue; and
3) G Requires additional PE analyses, but
sampl ing can continue.
A
A
383
-------
TABLE II. Summary of Detection Limits Used For
House Analyses During The Love
Canal EDA Full Air Study
Chlorobenzene
(ppb)
Phase 1
Average Detection Limits 1.0
Standard Deviation 0.5
Minimum Detection Limit 0.5
Maximum Detection Limit 3.7
Phase 2
Average Detection Limits 0.7
Standard Deviation 0.2
Minimum Detection Limit 0.4
Maximum Detection Limit 1.6
Chlorotoluene
(ppb)
N = 129
N = 133
1.2
0.5
0.6
4.2
0.8
0.3
0.4
2.3
Chlorobenzene
(ppb)
Chlorotoluene
(ppb)
Phase 3
N = 148
Average Detection Limits 0.9
Standard Deviation 0.2
Minimum Detection Limit 0.5
Maximum Detection Limit 1.8
Phase 4
Average Detection Limits 0.7
Standard Deviation 0.2
Minimum Detection Limit 0.4
Maximum Detection Limit 1.5
N 155
1.2
0.4
0.6
2.7
0.9
0.3
0.4
2.1
Summary
Chlorobenzene Chlorotoluene
(ppb) (ppb)
Average Detection Limits 0.8
Standard Deviation 0.3
Minimum Detection Limit 0.4
Maximum Detection Limit 1.5
N 565
1.0
0.4
0.4
2.1
T/F TEFLON TU8INC
(< f LENGTH)
AIR SAyPUNO PUI*>
FIGURE 1. Equipment Set-up for Performing TAGA Calibrations
Using Standard Gas Cylinders.
384
-------
CYUNDER OF
STANDARD GAS
UKTURE
AIR SAUPUNG PUMP
FIGURE 2. Equipment Set—up for Checking the Transport
Efficiency of the TAGA Sampling Lines.
Unhtated Sampling OHM. PhoM* 2-4
112-
108-
92 -
90 -
A
X
*
1
1
X
A
V *
*
•
X '
X
fi
A A
*
X
«
1
Ax{
X
A
X (
A
A
*
X •
A
A
X
A '
X •
X
XA :
A
A
A
A *
X
B w '
X
A
X
*
30
40
60
70
Twnptnrturt In *F
K Chlorob«nz«nc A ChlorotoJu*ne
Figure 3. % Transport Eft*, vs. Temp.
385
-------
THE KWIK-SKRENE ANALYTICAL TESTING SYSTEM
Description of a tool for remediation of PCB spills.
G.R.Woollerton,S.Valin,J.P.Gibeault
Syprotec Inc.,Pointe Claire,Quebec.
INSTRUCTION
Soils and oils contaminated with PCB
(polychlorinated biphenyl) threaten our
environmental and economic well being.
Public and private sectors of most North
American communities strive to control
PCB disposal and minimize the risks they
pose to public health and property.
Remedial reclamation of property is an
important link in the chain of control.
These efforts require tools. One of
these is the KWIK-SKRENE ANALYTICAL
TESTING SYSTEM, utilized to detect PCB
contaminated soils and direct "clean up"
efforts. KWIK-SKRENE gives the user an
opportunity to economise on time and
money. This paper presents the
procedures, contents, and utility of the
system.
The KWIK-SKRENE directs technicians to
locate PCB contaminated soils and reduces
the time and expense for laboratory
analysis. It is not uncommon that
conventional laboratory analyses and
reports require more than one week to
complete. Results from the KWIK-SKRENE
may require only one hour, and as many as
40 test results can be obtained in an
eight hour period.
The KWIK-SKRENE is a qualitative test
that evaluates a soil extract for 10 ppm
of Aroclor 1260 (a PCB) and provides
information about the severity of the
contamination. The test fits into
essential analytical procedures of
remedial action teams and has
applications for oil, incinerator ash,
and solid surface contaminations.
THE PROCEDURES AND CONTENTS OF THE KWIK-
SKRENE .
The procedures for evaluating soils with
the KWIK-SKRENE separate into two parts.
After collecting samples, the technician
must 1) extract the PCB and oil from the
samples, then 2) analyze the extracts for
PCB.
1)
SOIL EXTRACTION
Sampling procedures will vary depending
upon the protocols aadopted by different
contractors, but it is assumed that the
soil sample is in the form of a cylinder
weighing 200 to 500 grams taken from the
ground with a conventional core sampler.
It is usually wet and contains stones and
other debris. To properly test the soil,
it must be dry and in a powdered state.
This is achieved by cutting the sample
down the long axis and, using one side,
the sample is broken up, and stones,
twigs, insects, and other debris removed.
The sample is then put into a dish,
dispersed and oven dried. A microwave
oven dries a sample in two 2 minute
cycles at 500 watts. Samples dried in
convection ovens will require two hours
at 105°C. The dried soil is then crushed
with a mortar and pestle. 20 grams of
soil is then mixed with 20 grams of
extraction fluid and shaken on a
mechanical shaker for 30 minutes. The
extraction fluid is passed through a
filter column and filter disc to remove
chemical interferences aand particles.
The filtrate that passes through is a
clear oil ready for analysis.
387
-------
2)
EXTRACT ANALYSIS
Analysis of the extract is performed with
the standard KWIK-SKRENE colorimetric
test for PCB in dielectric oils. 500 ul
of Halogen reagent are added to 3 ml of
extract, following by 1.5 ml of indicator
buffer. Simmultaneously, the technician
tests a "control" (supplied with the
KWIK-SKRENE) that simulates a 10 ppm
concentration. The control test is
observed for a color change from yellow
to mauve at which time the sample extract
is observed. If the sample test remains
yellow the extract contains less than the
equivalent of 10 ppm Aroclor 1260. If it
is brown, mauve, purple, blue or clear,
it contains more than 10 ppm Aroclor
1260. The ultimate color of the sample
test and the speed of color change
depends upon PCB concentration.
CONTENTS OF A SOIL SYSTEM
The KWIK-SKRENE ATS for soils is supplied
as a package containing all of the
necessary hardware to extract and analyze
soil samples including a mechanical
shaker that is used for both test tubes
or extraction bottles. 10 soil
extractions and analyses can be performed
simulataneously.
Other items included in the package are:
REUSABLE
pipettors,
safety goggles,
test tube racks,
timer,
mortar and pestle,
vacuum manifold and
weighing scale
case.
DISPOSABLE
halogen regent
Indicator buffer,
test tubes,
control oil,
pipet tips,
pump gloves,
plastic bags,
CHEMISTRY OF THE KWIK-SKRENE
The physical chemistry:
PCB and oil are extracted from the soil
using an aliphatic fluid(proprietary) .
PCB is very soluble in this fluid; more
than 95% of the PCB is recovered from a
fresh spill in 5 minutes. It does not
dissolve inorganic salts or water than
may interfere with the indicator
chemistry of the test, but this insoluble
matter may be carried in the fluid as a
fine suspension or colloid. The fluid
will dissolve organic acids and bases
that may interfere with the indicator
chemistry. To elimate most of these
interferences the fluid is passed through
an absorption column(proprietary) and a
0.45um pore size filter disc. The
absorption column collects soluble
interferences and the filter disc blocks
suspended matter.
Wet Chemistry:
The KWIK-SKRENE indicator chemistry has
two basic steps, 1.) The halogen reagent
combines with PCB releasing chlorine and
sodium atoms. 2.) The aqueous indicator
buffer extracts the sodium chloride; then
an oxidizing powder converts the chloride
anions to chlorine which reacts with the
yellow indicator turning it blue. The
rate at which the color changes depends
upon the concentration of PCB. For
example, the 10 ppm control test changes
color after five minutes whereas the
change occurs in less time with more
concentrated samples. It is possible for
the experienced technician to qualify the
concentration of a sample according to
this rate of color change.
388
-------
A NEW METHOD FOR THE DETECTION AND MEASUREMENT
OF AROMATIC COMPOUNDS IN WATER
JOHN D. HANBY
HANBY ANALYTICAL LABORATORIES, INC.
HOUSTON, TEXAS
ABSTRACT
A Field Test Kit procedure for the rapid analysis
of petroleum aromatic hydrocarbons over a wide
range of concentrations in water and soil has pro-
ven of extreme utility in accurate assessments at
spill sites, hazardous waste areas and underground
storage tank removal locations. The kit was used
to evaluate diesel oil concentrations at the
Ashland Oil spill on the Monongahela and in con-
junction with EPA studies of the oil concentrations
in the Ohio river at Wheeling, West Virginia. A
study of gasoline-in-soil at concentrations ranging
from 5 ing/Kg to 10,000 mg/Kg was conducted. The
colormetric results from the procedure in both
water and soil are determined by comparison to a
color chart.
UV/VIS spectrophotometric studies were performed
to determine the relationship of concentration to
color intensity (reflectance).
INTRODUCTION
In the introduction to his encyclopedic treatise
Friedel-Crafts Alkylation Chemistry - A century of
Discovery, Royston M. Roberts makes the statement,
"probably no other reaction has been of more prac-
tical value." Professor Roberts goes on to say,
"Major processes for the production of high octane
gasoline, synthetic rubber, plastics and synthetic
detergents are applications of Friedel-Crafts chem-
istry" (1984). It is fitting that, over a century
after this monumental discovery, a technique for
the analysis of environmental contamination caused
by products of this reaction has been developed
which employs the same chemistry.
The analysis of organic ccmpunds in aqueous
solution has long been recognized as problematical
for many reasons. Primary among them of course is
the fact of the limited solubility of such non-
polar compounds in such an extremely polar solvent.
In a recent laboratory study undertaken for the
American Petroleum Institute on the solubilities of
petroleum hydrocarbons in groundwater, it was
pointed out that it was not possible to obtain
linear response when trying to directly inject
water standards of various aromatic hydrocarbons
into a gas chromatograph (American Petroleum
Institute 1985). This irreproducibility in
analysis of water samples has been a source of con-
sternation to proponents of gas chromatography for
a long time. A fairly comprehensive review of the
sort of problems associated with the gas chromato-
graphy of water samples is presented by Grob in
Chapter 5 of Identification and Analysis of Organic
Pollutants in water in their argument for the use
<">f capillary versus packed column GC. A statement
from that reference is particularly appropriate,
"Environmental chemistry includes probably the most
extreme branch of analytical chemistry environ-
mental samples should be analyzed with means and
methods to provide maximum separation efficiency
and resolutions" (Keith 1981).
Certainly chromatography of all types has proven
to be a technique of "maximum separation efficiency
and resolution". With the advent of capillary
columns of thousands of theoretical plates of sep-
aration efficiency the ability to resolve picogram
quantities of substances is available. However,
the problem of obtaining representative samples
and their subsequent quantitative as well as qual-
itative analysis remains as perhaps the dominant
problem in environmental assessment. Among the
criteria involved in sampler design discussed by
Johnson, et al, in a recent article in "Groundwater"
are those which would "prevent changes in the
analyte concentration due to: (1) sorption of de-
gradation in the well; (2) changes in temperature
or pressure; (3) cross-contamination between mon-
itoring wells due to the sampling equipment"(1987).
Each of these criteria might also be applied not
only to the collection of samples but to their
analyses as well. In the subsequent laboratory
analysis of a sample which may have been very well
collected, preserved, and transported to the lab-
oratory each of the above factors plays an analog-
ous role: (1) sorption or degradation in the samp-
ling container and analytical transference device,
e.g., syringe, pipet, or beaker; (2) changes in
temperature or pressure (particularly applicable to
the extreme pressure/temperature changes occuring
in the syringe and then the GC itself); (3) cross-
contamination of syringes, purge and trap devices,
sample lines, injectors, columns, and detectors.
The problem of sorption of organics in sampling
devices and in the passage of samples through
analytical tubing, was addressed in an article by
389
-------
Barcelona, et al, in Analytical Chemistry. In
that discussion, the sorption of various organic
liquids in different organic materials is well
documented (1985). This problem is seemingly one
of a particularly Sisyphean nature, i.e., the con-
tainment of a substance within a like substance is
akin to rolling a stone up a hill only to have it
fall immediately down the other side. Certainly
the problems encountered by the industry in
attempting to contain petroleum products in unlined
fiberglass tanks attests to this dilemma. The dev-
elopment of permeation tube calibration systems is
based on the phenomenon (O'Keefe 1966).
The present method addresses all of the problems
mentioned in that, to put it most succinctly, it
combines immediacy and simplicity of analysis.
That is, it is easily transportable to the field
which eliminates problems of sample transfer and
storage, and it provides an immediate analysis of
a large volume sample which speaks generally to the
problem of representativeness.
THE EXTRACTION^ODLORIMETRIC TECHNIQUE
The Hanby Field Test Msthod for aromatics in
water described here comes in the form of a kit
complete with necessary reagents and apparatus to
perform immediate analyses at the groundwater well
site. It is contained in a rugged plastic case
with enough reagents to perform thirty field anal-
yses. Within the case are contained: a 500 ml
separatory funnel, a tripod ring stand, a 10 ml
graduated cylinder, 2 reagent (liquid) bottles,
one dessicant jar with 30 reagent (powder) vials,
a color chart depicting test results for eleven
typical aromatics, plastic safety glasses and 12
pairs of gloves. Upon arrival at the site the kit
is opened and the tripod ring stand is assembled.
A 500 ml water sample is introduced into the sep-
aratory funnel which is placed in the ring stand.
Next, 5 ml of the extraction reagent is poured
into the separatory funnel using the 10 ml gradu-
ated cylinder. The sample is vigorously extracted
for two minutes with occasional release of the
slight pressure build-up which occurs. The funnel
is placed back in the ring stand and the extraction
phase is allowed to separate to the bottom for five
minutes. After phase separation is complete the
lower extraction layer is drained into a test tube,
allowing a small amount of the extraction solvent
to remain in the separatory funnel. Then one of
the reagent vials is opened and the contents immed-
iately poured into the test tube. The tube is
shaken for two minutes allowing the catalyst to be
dispersed well throughout the extraction reagent
so that color development, which is concentrated
in the powder, will be uniform. Hue and intensity
of the color of the catalyst which has settled in
the tube is now compared to the standard aromatics
pictured in the color chart.
The wide range of intense colors produced in
Friedel-Crafts reactions has been observed since
the discovery of this reaction. A brief descrip-
tion of the chemistry of the reaction, as well as
the color involved, is given by Shriner, et al, in
their widely used book (1980). In this novel adap-
tation of Friedel-Crafts aUkylation chemistry, one
of the reactants, the alkyl halide is used as the
extractant. The alkyl halide extractant plus the
aromatic compound present in the water sample are
caused to form electrophilic aromatic substitution
products by the Lewis acid catalyst which is added
in great enough amount to also act as the necessary
dehydrant to allow the Friedel-Crafts reaction to
proceed. These products are generally very large
molecules, i.e., phenyl groups clustered around the
alkyl moiety, which have a high degree of electron
delocalization. These two factors are the principle
reasons for the extreme sensitivity of this proced-
ure. That is, large molecules are produced which
are very intensely colored.
In the field conditions where this procedure is
by and large carried out the reaction is exposed to
sunlight. This means that there will be a window
in which to observe the color that is produced.
This is because of the general instability of the
reaction products to photochemical oxidation.
Strong sunlight will cause most of the colors pro-
duced to fade to various shades of brown within
just a minute or two; therefore, it is advisable to
perform the test in a shaded area.
APPLICATIONS OF THE METHOD
Obviously, this method will have a wide variety
of applications in field investigations. In fact,
utilization of the Hanby Field Test by an environ-
mental testing company has been going on since
August, 1987. Site investigations of hazardous
waste-containing landfills and underground stor-
age tank leaks have been conducted in several
states thus far and use of the kit has greatly fac-
ilitated sampling site locations. The first field
use of the kit was in the establishment of ground-
water monitoring well locations at an organic chem-
ical processing unit. An article describing this
first field use of the method is in preparation.
Recent regulations for the monitoring of under-
ground storage tanks require that soil/groundwater
investigations be carried out regularly to insure
that no leakage has occured. It can be seen that
the use of this technique shich is easily learned
and can be performed at an extremely low cost will
provide an immediate and defenitive answer to these
requirements.
OHIO RIVER STUDY
In the evening of January 2nd, 1988, the collapse
of a tank containing approximately 3.5 million
gallons of diesel fuel precipitated one of the
worst inland oil spills in the country's history.
Approximately one million gallons of the oil washed
in a huge wave over the containing dikes around
the tanks at the Ashland Oil Plant at West Eliza-
beth, Pennsylvania and into the Monongahela River.
Monday morning, two days after the spill, I con-
tacted Mike Burns of the Western Pennsylvania
Water Company in regard to using the Hanby Field
Test Kit at the company's water treatment facility
on the Monongahela south of Pittsburgh. Mike asked
me to bring one of the kits to the plant. The next
day I flew into Pittsburgh and was met by Mike at
the West Perm Water Works Treatment facility where
I demonstrated the use of the kit for the personnel
390
-------
at the plant. Then Mike suggested I call John
Potter, the chief chemist at the Water Treatment
Plant in Wheeling, West Virginia which was the next
major facility taking water from the Ohio river.
John said the kit sounded like it would fill a real
need for a rapid analysis of the river water at the
plant's intakes. The next day, Wednesday, I was
demonstrating use of the kit to the personnel at
Wheeling. It was immediately put into use on an
around-the-clock basis when they realized that in
just a few minutes they could get visual indication
down to 100 ppb of the diesel aromatic components.
The next morning I met with the West Virginia
Department of Natural Resources personnel who were
in Wheeling to monitor the oil spill. That after-
noon I was invited by the office of the EPA in
Wheeling to join EPA chemist Bob Donaghy, West
Virginia Department of Natural Resources Inspectors
Sam Ferris and Brad Swiger, and the Ohio River
Valley Water Sanitation Commission Coordinator of
Field Operations Jerry Schulte on the river tug-
boat Debbie Sue to make a run up the Ohio River
from Wheeling to try to locate the front of the
spill.
The investigators net at the Debbie Sue at noon
where it was tied up at the docks on the south
side of Wheeling. A light snowfall had begun and
the temperature was around 10 °F as the boat pull-
ed out into the Ohio, headed up stream.
Due to the fluctuating voltage from the tug's
generator the fluorometer readings exhibited a
fairly wide swing during the ensuing measurements.
As for the measurements performed with the Field
Test Kit, I was primarily concerned with the sen-
sitivity of the test in relationship to the near-
zero temperature of the water. Reference to the
API study of solubilities of petroleum hydrocarbon
components in groundwater had indicated rather
large decreases in partitioning of these components
into water at lower temperatures.
There was no time, however, to spend worrying
about these matters of close quantitation. The
boat was soon into the channel and Jerry Schulte
was bringing aboard the first bailer of water. On
the first sample taken, just minutes after leaving
the dock, an obviously detectable coloration was
seen in the catalyst material of the Field Test.
Reference to the color chart indicated presence of
aromatic constituents at something less than 0.5
ppm diesel. Having no reference colors or data at
these temperatures I arbitrarily chose this inten-
sity to represent 0.1 ppm. The fluorometer was
bouncing between zero and four on its iroveable rHal
indicator. (It was an old Turner model arbitrarily
nunbered from 0 to 100.)
We continued approximately eighteen miles up the
Ohio taking samples from the surface and the bottom
on the West Virginia side, mid-channel, and the
Ohio side. As table 1 indicates the results from
the EPA f luorometer and the Hanby Field Test Kit
tracked each other fairly consistently at each
point.
TABLE 1
Ohio River Sampling For Diesel Oil
January 7, 1988
Ohio River
Mile Point
89.0
85.5
85.5
85.5
85.0
85.0
84.5
84.5
84.5
82.0
81.0
80.0
79.0
77.0
76.0
75.0
74.0
70.0
SOIL STUDIES
Fluorometer
Reading
4
8
11
6
10-15
10
20
25
30
33
57
43
35
48
29
Hanby Field
Test (ppm)
0.10
0.15
0.15
0.20
0.20
0.20
0.20
0.50
0.20
.00
.00
.50
.50
10.00
8.00
7.00
5.00
3.00
Site investigations at leaking underground
storage tanks, hazardous waste sites, surface
spills, etc. normally involve at least prelimary
evaluations of the soil prior to the installation
of monitoring wells. The use of the Hanby Field
Test Kit for soils has found wide application in
this regard which has proven to be far superior to
methods such as soil gas techniques involving the
use of OVA's or field portable chromatographs to
sniff the headspace of samples.
The technique developed at Hanby Analytical
Laboratories for use of the kit is as follows:
1. Place 200 grams well-crumbled soil in a
quart jar.
2. Add 500 mil of distilled water and shake
well for 15-20 minutes so that water/soil
is very well mixed.
3. Pour mixture into a 1 liter Imhoff cone.
Sprinkle 2 level tablespoons powdered
ferrous sulfate into soil water mixture.
Allow mixture to clarify for thirty minutes. *
4. Decant clarified portion into the Field Test
Kit separatory funnel and perform the FTK
extraction/colorimetric test.
* Place a piece of Parafilm or aluminum foil
over each Imhoff cone to keep volatilization
losses minimized.
A study of two different soil types was undertaken
using super unleaded gasoline at various concentra-
tions. The studies illustrated in the slides are:
391
-------
1) a typical Texas "Gumbo" topsoil containing: 0
(Blank), 10, 50, and 200 mg gasoline mg/Kg 2) a
deeper sandy soil containing: 0, 5, 25, 100, 500,
2500, and 10,000 mg/Kg.
Soil samples were first sifted and homogenized to
remove twigs, leaves, rocks, etc. 200 gram portions
of each soil type (4 for the top soil and 7 for the
sandy soil) were weighed into quart mason jars.
Appropriate amounts of gasoline were injected di-
rectly onto the soil using microliter syringes with
the solvent flush technique. The soil/water mix-
ture was shaken in the capped jars over a ten
minute period. 500 ml of de-ionized water was
added to the jars which were then shaken period-
ically over a thirty minute period. The soil/
water mixture was poured into the IJmhof f cones and
5 ml. finely ground ferrous sulfate heptahydrate
was shaken into each. The ferrous sulfate was very
effective in settling the suspended matter and
clarification of approximately 300 ml of the super-
natant water was achieved in thirty minutes. 250
ml. of the supernatant was employed in the subse-
quent Hahby Field Test Method.
UV/VIS SPECTROPHOTOMETER STUDIES
Determinations of principle wavelengths and
reflectance data were made in correspondence with
aromatic compounds depicted on the color chart.
These investigations were conducted by preparing
a range of concentrations of selected aromatic com-
pounds, performing the Hanby extraction/colorimetric
procedure and then immediately measuring the reflec-
tance of the catalyst.
METHOD
Ten parts-per-million (vol/vol) solutions of
benzene, toluene, o-xylene, special unleaded gas-
oline, naphthalene and diesel were prepared by in-
jecting 20 microliter amounts of each compound
into 2.0 liters of de-ionized water at 20°-21°C
and stirring for one hour. Dilutions from the
stock solutions were prepared to 0.01, 0.02, 1.0
and 5.0 ppm. The extraction/colorimetric procedure
employed with the Field Test Kit was modified to
fit the requirements of the UV/VIS reflectance
apparatus. Four ml of the extraction reagent were
used to extract the water samples for two minutes.
The extraction solvent was then drained into a
cuvette. Two grams of the catalyst material was
added to the cuvette which was covered with its
teflon cap, and the mixture was shaken vigorously
for three minutes. The cuvette was placed in the
spectrophotometer and scanned over a range of 350
nm to 600 nm.
INSTRUMENTAL PARAMETERS
For this study a Varian DMS 300 UV/VIS spectro-
photoneter was utilized. Instrument settings were:
Slit width 2 nm, tungsten source, scan rate 50 nm/
min. All of the scans were corrected to 100% trans-
mittance baseline using a blank sample which was
scanned in reference to a barium sulphate reflec-
tance disk. The sample compartment was fitted with
a diffuse reflectance accessory which was modified
by blocking out the top portion of the light path
so that only the catalyst in the bottom half of the
cuvette would be scanned.
DISCUSSION
Figure 1 shows the spectrograms for each of the
six substances scanned. The concentration for
each of the plots is as follows (ppm by volume):
A=0.1, B=0.2, C=1.0, D=5.0, E=10.0. These concen-
trations exhibit well defined differences in the
traces of their reflectance curves for each of the
substances.
Figure 2 shows the reflectance trace for diff-
erence concentrations of benzene. These concen-
trations are (ppm by volume): A=0.01, B=0.05, C=
0.25, D=1.0. These runs were scanned on the spec-
trophotometer from 250 nm to 700 nm at different
instrument settings: slit width = 1.0 nm, scan
rate = 20 nm/min., smoothing constant = 5 (sec).
The different instrument parameters plus the fact
that a special cuvette was constructed to cover a
larger area of the reflectance opening contributed
to the greatly enhanced differences in the traces
at these, even lower, concentrations.
CONCLUSIONS
The development of a field method for the analysis
of organic contaminants at sub-part-per-million
levels in water has been proven to be a valuable
tool in the establishment and the sampling of
groundwater monitoring wells. The accuracy of the
method has been proven to far exceed that of direct
injection gas chromatography. A rapid soil-wash
method has also been developed employing the Hanby
Field Test Kit technique which has proven to be
effective on top and deep soils over a range of 5
mg/Kg to 10,000 mg/Kg gasoline in soil. Develop-
ment of instrumental spectrophotometric techniques
will allow even greater sensitivity and qualitative
analysis of aromatic contaminats in soil and ground-
water.
A variation of the procedure involving the extrac-
tion of a sample with an aromatic solvent and then
addition of the Lewis acid catalyst allows for the
determination of the presence of alkyl halides,
e.g. trichloroethylene. In this version of the test
a reflectance adaptor for the spectrophotometer is
not necessary since the color is not concentrated
in the catalyst but is developed in the extractant
solvent.
392
-------
REFERENCES
1. Roberts, R.M., Khalaf, A.A., 1984. Friedel-
Crafts Alkylation Chemistry; A Century of
Discovery. Marcel Dekker, Inc., New York,
pp 790.
2. TRC Environmental Consultants, Inc., 1985.
Laboratory Study on Solubilities of Petroleum
Hydrocarbons in Groundwater, American Petro-
leun Institute, Washington, DC.
3. Keith, L.H., 1981. Identification and Analysis
of Organic Pollutants in Water, Ann Arbor
Science Publishers, Inc., Ann Arbor.
4. Johnson, R.L., Pankow, J.F., Cherry, J.A., 1987.
Design of a Ground-Water Sanpler for Collecting
Volatile Qrganics and Dissolved Gases in Small-
Dianeter Wells. Ground Water, V. 25, pp 448-
454.
5. Barcelona, J.J., 1985. Sample Tubing Effects
on Ground-Water Samples. Analytical Chemistry,
V. 57, pp 460-465.
6. O'Keefe, A.E., 1966. Primary Standards for
Trace Gas Analysis. Analytical Chemistry, V.
30, pp 760-768.
7. Shriner, R.L., Fuson, R.C., Curtin, D.Y.,
Morrill, T.C., 1980. The Systematic Identifi-
cation of Organic Conpounds. John Wiley &
Sons, New York.
To]ucne
o-Xylenc
i'! < 51) I S!l I n (II I
Super Unleaded Gasoline
Diesel
Naphthalene
Figure 1
393
-------
Figure 2
394
-------
DEVELOPMENT OF A TEMPERATURE PROGRAMMED
MICROCHIP, HIGH RESOLUTION GAS CHROMATOGRAPH/MASS
SPECTROMETER FOR VOLATILE ORGANIC COMPOUND ANALYSIS
E. B. Overton, E. S. Collard, H. P. Dharmasena, P Klinkhachorn and C. F. Steele
Institute for Environmental Studies
Louisiana State University
Baton Rouge, Louisiana 70803
ABSTRACT
The interfacing of a temperature programmed
microchip gas chromatograph with a mass
spectrometer has developed to produce a field
deployable analytical system capable of rapid,
laboratory type analyses of ambient air samples for
volatile organic compounds. The addition of a
microprocessor based temperature controller for
temperature control and temperature programming of
the analytical column (utilizing a small thermoelectric
heat pump) has enhanced the performance of this
system for quantitative analysis. Programming the
solid-state microchip gas chromatograph column
oven below ambient temperature at the beginning of
the analytical run and then upwards to a higher than
ambient temperature during the analytical run yields
acceptable wide peak widths at the beginning of the
analytical run and sharper peaks during the course
of the analysis when compared to isothermal
analyses. The temperature programming and control
of the gas chromatographic liquid phase improves
the sensitivity and peak shapes of all of the
compounds during an analytical run.
INTRODUCTION
A microchip gas chromatograph produced by using
silicon micromachining technology has been
discussed in another paper (1). The development of
these small devices has produced analytical
instruments which can perform complex separations
of volatile mixtures with analysis times of two to three
minutes. On-site environmental operations, whether
hazardous waste-site clean-up or hazardous
materials spill mitigation, require extracting the most
information from a sample in the least amount of
time. The need to produce laboratory quality
analyses with short sample turn around times
prompted us to consider the interfacing of a
microchip gas chromatograph with a mass
spectrometer to produce an analytical system for the
analysis of volatile organic compounds.
Temperature programming of the analytical column,
a technique which has long been recognized as a
method of improving analytical performance, was
used as a means of reducing run time and optimizing
chromatographic peak widths. Programming of the
solid-state microchip gas chromatograph column
oven below ambient temperature at the beginning of
the analytical run then upwards to a higher than
ambient temperature during the analytical run
produces acceptable wide peak widths at the
beginning of the analytical run and sharper peaks
during the course of the of the analysis when
compared to isothermal analyses.
MATERIALS AND METHODS
A modified microchip gas chromatograph
(Michromonitor 500, Micro Sensor Technology, Inc.,
Freemont, California) was interfaced with an ion trap
mass spectrometer (Finnigan Model 700 Ion Trap
Detector, Finnigan MAT, San Jose, California) to
function as a small, rapid gas chromatograph/mass
spectrometer (GC/MS) system. The combination of
these two instruments allows for rapid mass spectral
identification and quantification using extracted ion
current profiles of volatile organics in environmental
samples.
The mass spectral operating parameters,
including the tuning procedure, were as outlined in
the operating manual for the Ion Trap Detector. The
scanning rate of the mass spectrometer was set at
the maximum rate of 4 scans per second from 45 to
246 AMU. The lower limit of 45 AMU was chosen to
eliminate the problem of scanning over components
in the air peak. The upper limit of 246 AMU was
used to allow adequate spectra acquisition for most
(10 volatile priority pollutants and similar
compounds) of the compounds which the microchip
gas chromatograph is capable of analyzing. The
microchip gas chromatograph was adjusted to
provide a flow of approximately 4 ml per minute
through the analytical column. The analytical column
(1 meter X 0.23 mm i.d. DB-1) was wound into an
aluminum column jig which was mounted onto the
chromatographic module with a specially developed
zero-dead volume fitting. The free effluent end of the
column (which normally goes to the detector of the
microchip gas chromatograph) was attached to the
open split interface of the mass spectrometer.
A Peltier' device (thermoelectric heat pump)
(2) was chosen to supply the heating and cooling
required for temperature programming of the column.
A microprocessor based (Intel 8751 single chip
microprocessor) controller was designed and built in
our laboratory to provide electronic control of the
395
-------
Peltier1 device. The temperature programmer
(Peltier1 device and controller) uses only electrical
current and does not require liquid cryrogenic
coolants to control the column temperature below
ambient temperature. The Peltier' device along wilth
its heat sink was mounted on one side of the column
jig (Figure 1). A temperature sensor (Analog Devices
AD-590) was mounted on the other side of the
column jig to measure the temperature of the column
assembly and provide information for the
temperature controller. The Peltier' device along
with the controller is capable of set point temperature
control (± 0.05°C) from -5°C to 90°C or temperature
programming (+0.6°C) of the analytical column from
2°C to 80°C. Temperature ramping speeds range
from 0.1 °C to 1 °C per second with good
reproducibility, however, 0.8°C per second is as fast
as practical for true linear programming (Figure 2).
RESULTS AND DISCUSSION
The isothermal gas chromatograph mass
spectrometer with the microchip gas chromatograph
functioned well while operating at ambient
temperature (approximately 25°C), however, the
early eluting peaks were so narrow (3-4 mass scans
across the peak) that precision of the quantitative
data was not acceptable even though the mass
spectrometer was scanning as rapidly as possible (4
scans per second from 45 to 246 AMU) over the
mass range selected for the analysis (Figure 3). It
was belived that if the chromatography was
controlled such that the early eluting compounds
would exhibit peak widths on the order of 7-8 scans
(approximately 2 seconds) the quantitative
information for these compounds would then become
more precise. Complete chromatographic
separation of each component in the mixture was not
considered as important as improved quantitative
ability since standard mass spectral data
manipulation techniques can be used to identify
components and quantitation can be obtained from
extracted ion current profiles.
Qualitative identification of the compounds run
on the system was good in that the mass
spectrometer was able to distinguish between two
compounds eluting in the same chromatographic
peak. Several combinations of temperatures and
ramping speeds as well as delay times (starting the
temperature ramp at some time after the injection is
made) were investigated. The best results were
obtained by starting the analytical column
temperature at 2°C and ramping the temperature at
0.8°C per second to 80°C starting immediately at the
time of injection. Comparing the chromatographic
runs with and without temperature clearly shows
increased peak widths for early eluting peaks (Figure
4). This increased amount of time in which the mass
spectrometer is allowed to obtain data from a given
chromatographic peak provides more accurate
integration of the peak area. More accurate
quantitative information from the analysis is possible.
CONCLUSIONS
The interfacing of the microchip gas chromatograph
and the ion trap mass spectrometer has proved to be
an excellent system for the rapid analysis of volatile
organic compounds in environmental samples. The
sensitivity of the system is approximately O.Sppm
(volume/volume in air) for most of the volatile
compounds studied. Pre-analysis concentration
techniques such as trapping on adsorbents with
thermal desorption can increase the detection limits
to 1 ppb. Temperature programming of the analytical
column of the gas chromatograph increases the
precision of the quantitative information for all
compounds. Rapid sample turn around time as well
as laborabory type analytical results are possible
with this system. Special sample treatment is not
necessary and there are no adsorbants or solvents
used with this system. Actual use of the system in an
air monitoring program has demonstrated excellent
results that allows sample turn around times of
approximately five minutes.
ACKNOWLEDGMENTS
The financial support for this work from the National
Oceanic and Atmospheric Administration, U.S.
Department of Commerce, Contract No. 50-ABNC-7-
00100, is gratefully acknowledged.
REFERENCES
(1) Angell, James B., Stephen C. Terry and Philip
W. Barth, "Silicon Micromechanical Devices,"
Scientific American. April 1983, pp 44-55.
(2) Shields, J.P., "All About Thermoelectric
Coolers," Radio Electornics. May, 1988,p 61.
396
-------
COLUMN MOUNTING
FOOT
Figure 1. Construction of Microchip GC Column Oven
10000
8000-
6000-
4000-
2000-
CD
X
0
PELTIER RAMPS
1.3
1.5
0 SECONDS 100 200
Figure 2. Temperature Ramps of Peltier' Device
300
397
-------
5
D
E
T
E
C
T
O
R
•
Ft
E
S
P
0
N
S
E
&2|
6 9
344
SCA
1
2
3
4
10 13 I
7
8
9
10
1 1
1 2
7*° II 1 3
NS
TIME
II 14
12
I I
'[
II
15
16
17
14 15
*f\
II
ffl 200 380 4(
0:26 0:51 l:i? 1:
1,1-Olchloroethylene
Methylene chloride
Trans-1 ,2-dlchloroethylene
1,1-Dlchloroethane
Chloroform
1,2-Dlchloroethane
Benzene
Carbon tetrachlorlde
1,2-Dlchloro pro pane
Trlchloroethylene
2-Chloroethyl
vinyl ether
1,1,2-Trlchloroethane
Toluene
Tetrachloroethylene
Ethyl
benzene
o-Xylene
1,1,2,2-Tetrachloroethan*
f-j—rtt&t
)0
43
1 6
1
A
1
' W
5(
2:
7
S»^
ie
38
Figure 3. 17 Component Volatile Organic Mixture
from Microchip GC/MS
m
57
Methylene chloride
1NX
83.
.,.,.... I . ,..,,.,, I ,,,.
, , , | ,,,. | ,,,,,,,,, | ,,,,
20 39
8:85 9:88
1 • ' ' i ' 1
Chloroform .
, , . , | , , , FT f . , , |
48 58
8:11 8:14
m.
TOT
1,1-Dlchloroethylene
Methylene chloride
20
30
~\ A
50
0;14
0:1°
7«
Isothermal @ 27 *C
Temperatur* Programmed
0-9CTC @0.e°C/S»cond
Figure 4. Comparison of Peak Widths
398
-------
APPLICATIONS OF THE PYRAN THERMAL
EXTRACTOR-GC/MS FOR THE RAPID
CHARACTERIZATION AND MONITORING OF
HAZARDOUS WASTE SITES
C.B. Henry and E.B. Overton
#42 Atkinson Hall
Institute for Environmental Studies
Louisiana State University
Baton Rouge, LA 70803
C. Sutton
Ruska Instruments
P.O. Box 742688, Houston, TX 77274
The Pyran Thermal Extractor-GC/MS instrument is, in
essence, a robotic device that both extracts and
analyzes semivolatile analytes in solid matrix samples.
It is capable of many analytical applications that
currently require CLP-type analytical techniques. The
instrument is designed to be rugged enough for field
deployment. In this paper, we provide data to
demonstrate the applicability of thermal extraction
techniques using the Pyran system on soils/sediments
and semi-solid sludges. The instrument is a qualitative
and quantitative tool for rapid chemical analysis of a
variety of sample types. The detection limit for the PAH
phenanthrene in soil/sediment samples is estimated at
10ppb(ng/g).
INTRODUCTION
Site investigations and cleanup activities under Super-
fund often require the rapid analysis of samples for
semivolatile hazardous substances in order to make on-
site decisions. Historically most Superfund analytical
data has been generated through the Contract Labortory
Program (CLP) whose methods involve the complex
extraction, cleanup, and analysis and are outlined in
EPA's CLP protocols under Superfund and SW-846
analytical protocols under RCRA. These methods are
time consuming and expensive. The instrument
investigated here has an analytical turn- around time for
base-neutral and acid-extractable compounds of slightly
longer than one hour, and no sample preparation other
than weighing the sample is required.
In this paper, we experimentally investigate the possible
use of thermal extraction-GC/MS techniques for the
rapid extraction and analysis of semivolatile organic
compounds in soil and sediment samples. The
instruments used were the Pyran Level I Thermal
Extractor/Flame lonization Detector (FID) and the Level II
Thermal Extractor/Mass Spectrometer. Each has a
chemically inert all quartz thermal extractor. The Level I
extractor is interfaced directly to a FID, while the Level II
is interfaced to an all quartz cold trap and a fused silica
capillary GC column which elutes into an ion trap
detector. The instruments were developed to meet the
rugged analytical needs of petroleum exploration. We
feel this makes them ideally suited for field applications
with only slight modifications1.
The Level I unit produces a thermal extraction profile,
plot of temperature vs detector response, and is a useful
tool for rapid screening of samples. The Level II
provides quantitative and qualitative information about
samples which we will demonstrate is generally
comparable to data from conventional Contract Lab
Program (CLP) type analyses.
A 25 component calibration standard was analyzed to
demonstrate the sensitivity and quantitative linearity of
the instrument; 22 of the 25 compounds are EPA Target
Compounds. In addition, samples from various
environmentally impacted sites were analyzed to
demonstrate the applicability of the instrumentation in
the rapid analysis of "real world" (as opposed to
laboratory spiked) samples. Wet and dry sediment
samples were analyzed to asses the affect of moisture
content on the thermal extraction process.
Pentachlorophenol contaminated creosote sludges from
2 sites (referred to in this paper as sludge A and sludge
B) and a sediment sample contaminated with diesel oil
were analyzed on both the Level I and Level II
instruments. Thermal extraction analyses of the
sediment sample were compared with samples
analyzed by established analytical methods.
EXPERIMENTAL METHOD
Soil and sludge samples were prepared for analysis by
weighing a small aliquot of the sample into a porous
quartz "cup" and placing it into the pyrocell loader. No
other sample preparation was required. A small amount
of the sediment sample was air dried before analysis by
placing the wet sediment in a scintillation vial which was
then placed in a plain glass desiccator filled with CaSO4
overnight. None of the sediment samples were treated
in any manner to enhance the homogeneity of the
sample. The calibration standards were in solution with
399
-------
methanol and were added to the lid of the quartz cup by
a syringe technique. The 6-point standard curve
ranged from 6.4 ng to 600 ng applied to the quartz cup.
The level II instrument was configured with a 15 meter
DB-5 capillary column (0.25 mm ID and 0.25 micron film
thickness from J&W Scientific Co.). The column was
interfaced to a Finnigan MAT Ion Trap Detector (ITD) via
an open split interface, and the ITD was operated in the
scanning mode over a mass range of 47-507 amu using
automatic gain control (AGC) software.
The Level I instrument was temperature programmed
from 30°C to 500°C at a temperature rate of 30°C/min.
Flow rates for the He carrier gas, H2, and zero air were
60 mL/min, 55 mL/min, and 550 mL/min, respectively.
Nelson 2600 chromatographic software was used for
data processing of the FID response. Quantitative
analyses was by an external standard method using
response factors estimated from phenanthrene or n-
C32-
The Level II instrument was operated by the following
technique: The pyrocell was set at an initial temperature
of 30°C then increased to 240°C at 30°C/min to
thermally extract the analytes then cooled to 30°C at
30°C/min. The capillary column was initially set to -35°C
and held for 13 min to cold-trap the analytes then
temperature programmed to 300°C at 5 C/min and held
for 10 min. The ITD was turned on 21 min into the
analysis. The trap and split were maintained at 330°C
and 300°C, respectively. The split ratio was set for a
1:40 split. The total analysis time was 84 min.
Quantitations were by an external standard method
using authentic standards when available.
When wet sediment samples were analyzed in the Level
II instrument, the pyrocell temperature program was
slightly modified to hold the pyrocell at 60°C for 5 min
with the vent open. This is required to prevent ice
formed from closing off the capillary column.
The conventional GC/MS analyses were performed by
an organic extraction method outlined by the National
Oceanic and Atmospheric Administration (NOAA) for use
in the National Status and Trends Program2. The
nomenclature "conventional" refers to solvent extraction
techniques in this paper. The extracts were analyzed on
a Hewlett-Packard 5890 GC equipped with a 30 M J&W
DB-5 column directly interfaced to a Hewlett Packard
5970B mass spectrometer. The report values were
corrected for recovery using the surrogate standard
hexamethylbenzene.
RESULTS AND DISCUSSION
Standards
The 25-component, 6-point standard curve was linear
and demonstrated good sensitivity. Table 1 lists the
compounds analyzed in the standard, the mean
retention time in scan numbers for each component, the
percent relative standard deviation (%RSD) of the
retention time for 6 analyses, and the correlation
coefficient (R) of the detector response vs amount of
analyte (response factor). The range of R was 0.82-
1.00, and the mean value was 0.96. Representative
standard curves are shown in figure 1. Figure 2 shows
the Total Ion Chromatogram (TIC) of a calibration
standard analysis from scan 400 to scan 900. The
instrument has a detection limit for phenanthrene of less
than 1 ng. If this value is extrapolated for a 100 mg
sediment sample, the detection limit for phenanthrene
would be in the 10 ppb range.
Sludge Samples
From the thermal chromatograms produced by the Level
I analyses and shown in Figure 3, it appeared that
sludge B contained higher molecular weight
contaminates than sludge A. The level II analyses of the
same sludges supported the results from the Level I
screening analyses (Figure 4). In both analyses, the
compound specific determination of the semivolatile
contaminants was rapidly assessed. Sludge A
contained a typical unresolved complex mixture with
resolved normal aliphatic hydrocarbons and a high
concentration of pentachlorophenol. Octachloro-
dibenzo-p-dioxin and the two heptachlorodibenzo-p-
dioxin isomers were also detected. Sludge B contained
a much greater concentration of polynuclear aromatic
hydrocarbons than A. These two preliminary analyses
indicate that the Level II instrument is a useful tool in the
rapid assessment of hazardous waste sludges typical of
those associated with Superfund and Nonhazardous
Oilfield Waste (NOW) sites.
Diesel contaminated sediment sample
The Level I analysis of the diesel contaminated
sediment (sediment A) is shown in figure 5 compared to
a control soil sample which contains only normal
background organic constituents, which we believe are
primarily pyrolysis products of high molecular weight
organic compounds. The concentration of hydrocarbons
in the diesel contaminated sediment was estimated at
5.8 |ig/mg by an external standard quantitation method
using a response factor calculated from n-C32- The
heavier molecular weight components, those extracted
after 13 min of analysis time and typically associated
with the natural organic content of the sediment, were
not quantitated.
Figure 6 contrasts a conventional solvent extraction
followed by GC/MS analysis to a Level II analysis of dry
and a wet diesel contaminated sediment (48% moisture
by weight). Each analyses contained an unresolved
complex mixture (chromatographic hump) with resolved
normal aliphatic hydrocarbons from n-Cis to n-C23-
The precision of the analyses using the Level II
instrument was compared to that of conventional solvent
extraction GC/MS analyses. Four duplicate Level II
analyses of the sediment were compared to two
duplicate conventional extractions and GC/MS
analyses. Each extract was analyzed twice. The mean
400
-------
recovery of the conventional analyses was 77%. The
quantitative results for selected components are given in
table 2. The mean %RSD of the thermal extraction
analyses was 20.60 compared to 19.51 for the
conventional analyses. From this limited set of
analyses, it appears that the precision of the thermal
extraction technique using the Pyran system is
comparable to that obtained through conventional
solvent extraction GC/MS techniques. A significant but
unknown portion of the deviation between sample
analyses is the natural variability of environmental
samples. The sediment was not treated in any manner
to enhance its homogenity. It was intended to be
representative of moderately contaminated
environmental samples.
The mean value of phenanthrene determined using the
Level II instrument was 3.1 ng/mg compared to 4.6
ng/mg as determined by a quadropole mass
spectrometic analysis. From repeated analyses of the
sample aliquot, we estimate that 95% of the
phenanthrene was removed in the initial thermal
extraction.
There seems to be no substantial loss of the lighter
molecular weight components in the wet sediment when
analyzed by the Level II instrument. In general, the
calculated quantities were less than the mean
determined from the dry sediment analyses by 27%.
However this may in part be due to physical obstruction
of the carrier flow when the sediment dries as a solid
plug in the quartz cup encapsulating part of the
sediment. The wet sample analysis was intended to
represent a worst case scenario, as the sample cup was
pack completely full with 180 mg wet sediment. The
percent moisture was determined by standard methods
to be 48%.
REFERENCES
(1) Overton, E.B. and S.J. Martin; "A field deployable
analytical instrument for the analysis of semivolatile
organic compounds of Superfund sites," Proceedings of
the Third Annual US EPA Symposium on Solid Waste
Testing and Quality Assurance. 1987. Vol. II, p. 8-55.
(2) MacLeod Jr., W.D., and etal. Standard Analytical
Procedures of the NOAA National Analytical Facility,
1985-1986. NOAA Technical Memorandum NMFS
F/NWC-92.
CONCLUSION
In conclusion, the Pyran Thermal Chromatograph (Level
I) and Thermal Extractor-GC/MS (Level II) is capable of
providing rapid semivolatile organic analysis of most
solid and semi-solid samples that are encountered at
Superfund sites. The robotic characteristics of the
instrument reduce the analysis time associated with
conventional solvent extraction methodologies. The
instrument has a sample turn around time of
approximately 1.5 hours with detection limits in the low
parts per billion range. Further methods development
and evaluation is required to fully asses the potential of
this technique for both laboratory and field deployable
analytical applications.
401
-------
100*
TOT-
1.
UJ
co
O
a.
co
u
ft
cc
O
o
UJ
I-
LLI
a
UJ
CO
1
1/1
UJ
CC
cc
o
UJ
a
UJ
CO
1
CO
UJ
cc
oc
o
o
UJ
phenanthrene
ng
pentachlorobenzene
ng
2,4-dichlorophenol
0 100 200 300 400 500 BOO
ng
Figure 1 Representative standard curves.
1. 2-chlorophenol
2. phenol
3. 2-nitrophenol
4. 2,4-dimethylphenol
5. 2,4-dichlorophenol
2. 6. naphthalene
7. 1,2,4-trlchlorobenzene
6.
4.
3-
-. '_
6:41
CHRO)
500 600
8:21 10:01
' " I '^T^^J' '
700 800
11:41 13:21
900
15:01
Figure 2 TIC of Pyran Level II calibration standard analysis
(scans 400-900).
402
-------
600 n
$ 500
O. 400
HI
Q-
CC
o
300 -
200 H
HI
UJ 100
Q
0 5
time
SLUDGE A
10
15
20
III
0 5
time
SLUDGE B
10
15
20
Figure 3 Level I thermal extraction profiles.
1037 SLUDGE A
TOI-
PCP
OCDD
1007,
101-
SLUDGE B
PCP/PHENANTHRENE
10:01
1200
20:01
30:01
2400
40:01
50:01
Figure 4 Pyran Level II analyses of sludge samples.
403
-------
control soil
.. ooo
e -O-H
diesel contaminated
sediment
nC-32 (0.019 mg)
Figure 5 Level I thermal extraction profiles.
Pyran Level
101.
sediment A
180 mg wet wl.
(48% moisture)
101-
sediment A
31 mg dry wt.
i i ' i
4W 6N
6:41 18:81
888 m 1289
13:21 16:41 28:81
TIC of
1 . 0E5-
B.0E4-
7.0E4-
6.0E4-
5.0E4-
i . 0E4-
3. 0E4-
c .0EI4-
10000-
0
V3 : DMUD RI . D
1 LjJ^
1
1
F
1
ill
,
HI
conventional analysis
Wjj
\
^X— 1 III.
10 15 20 25 30
Figure 6 Comparison of Pyran to conventional analysis of sed. A.
404
-------
Table 1 6-point Standard Curve
COMPOUND (AMW)
2-chlorophenol (128.6)
1,4-dlchlorobonzene (147.0)
phenol (94.1)
2-nltrophenol (139.1)
2,4-dlmethylphenol
1,2,3-trlchlorophenol (181.4)
2,4-dlchlorophenol (163.0)
naphthalene (128.2)
1,2,4-trlchlorobenzene (181.4)
4-chloro-3-methylphenol (142.6)
1,2,4,5-tetrachlorobenzene (216)
2,4,6-trlchlorophenol (197.4)
acenaphthylane (152.2)
acenaphthena (154.2)
pentachlorobenzene (250.3)
fluorene (166.2)
hexachlorobenzene (284.8)
phenanthrene (178.2)
anthracene (178.2)
carbazole (167.2)
(luorenthrene (202.3)
pyrene
benz(a)anthracene (228)
chrysene (228)
benzo(a)pyrene (252)
MEAN RT
acan *
474.0
506.8
508.8
722.6
803.8
81 1 .0
812.7
819.0
863.2
1019.0
1050.7
1092.7
1 221.0
1278.5
1338.8
1420.5
1596.8
1 668.0
1 682.0
1733.8
1999.5
2050.8
2399.5
2407.0
2754.8
%RSD
n = 6
2.6750
2.5260
2.4690
.5330
.6380
.5200
.6720
.6030
.5640
.2390
.1160
0.9984
0.7025
0.8615
0.6398
0.5749
0.4690
0.3848
0.3567
0.2948
0.21 63
0.191 1
0.1079
0.1084
0.0741
R
0.91
0.87
0.99
0.88
0.96
0.96
0.99
0.82
0.97
0.95
0.99
0.98
1.00
0.99
1.00
0.82
0.99
1 .00
0.99
0.99
0.99
1.00
0.98
0.99
0.98
Table 2
Comparison of Pyran Level II analyses of a diesel contaminated
sediment to conventional solvent extraction-GC/MS analyses
sample size nC-16 phytane nC-19 nC-20
PYRAN LV II mg dry wt. ng/mg ng/mg ng/mg ng/mg
wet sed.
(180) 94
74
53
CONV METH
35
28
dry sed.
dry sed.
dry sed.
dry sed.
MEAN
STDDEV
%RSD
1
2
3
4
4
3
2
3
2
5
0
1
1
87.
19.
22.
58
95
95
00
00
48
39
43
58
53
8 9
60.75
19.84
32.66
57
8
14
4 5
58
6 3
6 3
.25
.50
.85
47
5
12
40
4 5
5 3
5 1
.25
.91
.51
1a
1 b
2a
2b
MEAN
STDDEV
%RSD
5.90
5.90
5.89
5.89
92
98
120
120
107,50
14.64
13.62
48
36
56
56
49.00
9.45
19.29
6 1
39
64
7 1
58.75
13.82
23.52
6 1
37
60
6 1
54.75
1 1 .84
21 .63
405
-------
FIELD DEPLOYABLE INSTRUMENT FOR THE ANALYSIS
OF SEMIVOLATILE ORGANIC COMPOUNDS
E.B. Overton and C.B. Henry
#42 Atkinson Hall
Institute for Environmental Studies
Louisiana State University
Baton Rouge, LA 70803
C. Sutton
Ruska Instruments
P.O. Box 742688, Houston, TX 77274
Introduction
Chemical analyses play crucial roles in virtually
all types of environmental chemical hazard
assessments. For example, data used to determine if a
site is contaminated with toxic chemicals are obtained
from analysis of environmental samples taken from the
site. The extent of contamination is determined from
results of chemical analyses. The effectiveness of
mitigative strategies can be determined from analytical
data. Health effects may be determined from
collaboration with chemical analysis data. Finally, legal
claims of environmental contamination must be
supported by indisputable data from chemical analyses.
There are standard methods and procedures
such as those promulgated in SW846, which are
designed to provide a recipe for chemical analyses of
specific types of analytes in various sample matrices.
Implicit in these promulgated procedures are techniques
to ensure quality of the analytical results. All current
methods are designed and intended for use in a
conventional laboratory. Virtually no field screening
methods for semivolatile organic analytes have been
developed that do not require the services of a
conventional trace chemical analysis laboratory. The
types of conventional analytical procedures are outlined
in Figure 1.
There are many applications where laboratory
bound analytical procedures are not optimally suited.
For example, site evaluation requires rapid turn-around
of analytical results while personnel are deployed in the
field. Also, excavation of contaminated material can be
facilitated by rapid field screening analysis of samples.
In this paper, we describe an analytical instrument that
has potential to provide rapid field screening analyses of
solid and semi-solid samples for semivolatile organic
components.
Discussion
The Pyran Thermal Chromatograph,
manufactured by Ruska Instruments of Houston, Texas,
is an instrument that was specifically designed and
developed to meet the analytical needs of petroleum
exploration and development activities. It is a self-
contained thermal extraction system and GCMS
analyzer that is relatively compact, rugged and designed
for field applications. It is constructed of quartz to
provide the chemical inertness and stability that is
needed for reproducible thermal extraction and pyrolysis
of organic compounds from source rock samples. Since
quartz does not absorb radiant energy, the quartz
construction allows precise temperature control at both
subambient and elevated temperatures using a
computer controlled combination of cryogenic cooling
(liquid CC>2) and radiant heating. The Level I analyzer
includes a thermal extraction unit that is interfaced
directly to a flame ionization detector (figure 2). The
Level II unit includes a thermal extractor module that is
interfaced, with all quartz components, to a specially
designed capillary column gas Chromatograph that has
no moving parts (figure 3). Again, all quartz construction
of the chromatographic oven and column allows precise
and reproducible temperature control and programming
from sub-ambient to several hundred degrees
centigrade. The chromatographic effluents are detected
by an Ion Trap Mass Spectrometer and analyzed by
conventional data treatment software.
The Level I analyzer thermally extracts organic
components and detects the substitutents without any
chromatographic separation using flame ionization
detection. It is designed to permit rapid
screening of samples and has analysis times of
less than fifteen minutes. The Level II analyzer has
a thermal extraction module that is interfaced to a GCMS
analyzer. The GCMS unit has the capability to
identify and quantitate specific substances that
are thermally extracted from the sample. Total
analysis time, including extraction, is generally on the
order of one hour. The Level II analyzer is designed to
identify specific chemicals in a sample and to measure
their concentrations with analytical turn-around times
that are significantly faster than is available from
conventional solvent extraction laboratory GCMS
analysis of environmental samples. The thermal
extraction procedure includes weighing a small aliquot
of sample (10-100 mg) in an all quartz crucible. The
crucible is then placed in the pyrocell compartment of
407
-------
the Pyran analyzer. The thermal extraction unit is then
flushed with helium. If the sample is wet, it can be dried
by raising its temperature to 60 to 80°C and venting the
exhaust vapors prior to their entering the GCMS
analyzer compartment. After drying (if needed), the
sample temperature is raised rapidly to 250 to 300°C
under precise temperature control. Sample analytes
that have appreciable vapor pressures at 250 to 300°C
are swept into the initial portion of a cryogenically
cooled fused silica capillary column. The sample is then
analyzed by conventional GCMS procedures. The
drying and thermal extraction step takes between fifteen
to twenty minutes. Figure 4 is an example of the
temperature profiles of the pyrocell and GC columns
during an analysis cycle.
Because the pyrocell's temperature can be
precisely controlled at different set points, analytes with
different vapor pressures can be selectively enhanced in
the thermal extracts. This process has been called
"thermal slicing." For example, if the pyrocell is heated
to 175°C, the more volatile analytes will be driven out of
the sample matrix and analyzed leaving unextracted
those analytes that have extremely low vapor pressure
at 175°C. Alternatively, more volatile analytes can be
vented at 175°C and then the pyrocell's temperature
raised to 250°C to enhance extraction of less volatile
organic compounds. The precise control of the
extracting conditions (temperature, helium flow rate, vent
time and split ratios) provide a thermal extraction and
analysis system with great flexibility.
The Pyran thermal extraction-GCMS analyzer is
designed to analyze compounds that are in a solid or
semi-solid sample matrix. Liquid or gaseous samples
may be analyzed by passing the sample over some type
of solid adsorbent, (tenax, activated carbon, XAD resin,
etc.) and then thermally extracting the solid adsorbent
with either a Level I or II Pyran analyzer.
It must be emphasized that thermal extraction
removes all volatile components from the solid matrix
not simply the analytes of interest. Special care must be
exercised to identify and quantitate analytes of interest
in the complex mixtures of organic matter that are
routinely found in environmental the samples. Also,
different types of solid materials have varying affinities
for various types of analytes. A single thermal extraction
sequence may not be suitable for the same analytes in
different sample matrices. Examples of thermal
extraction analyses are presented else where in these
preceedings.
Conclusion
The Pyran Level I and II Analyzers have potential
for valuable applications in many environmental
analyses. Of particular importance is their application to
the field screening for specific semi-volatile organic
compounds in a variety of solid and semi-solid
environmental samples. In addition to field screening
analytical capability, these techniques have total
analysis times of 60 to 90 minutes.
Organic
Solvent
100ml
Extraction
Rotovap
Purge and Trap
Concentrated Sample
n 20-50ul
Clean-up
Column
GC-MS Control and
Data Acquisition System
GC-MS
Figure 1. Conventional Laboratory Methods
408
-------
LCCX
He
CAPILLARY
GC COLUMN
-70°to 400° C
TC/MS
Quartz Analyzer
=r= •«- Liquid CO2 Cooing
COLD-TRAP 1
-70° to 600°C
PYROCELL
0*to 600°C
SAMPLE LOADER -
Splitter Purge
SpStter Vent
•*- Heium Carrier Oas
Figure 2. Level I-FID analyzer schematic
Figure 3. Level ll-thermal extractor-MS analyzer schematic
°C
TRAP! (ISOTHERMALAT330*0)
PVROCELL
/.
\/"\
7
COLUMN
10 15 20 25 30
time (min.)
Figure 4. Typical Pyran temperature program for wet sediments.
409
-------
EVALUATION OF MICROWAVE DETECTION TECHNIQUES TO PREPARE SOLID
AND HAZARDOUS WASTE SAMPLES FOR ELEMENTAL ANALYSIS
Peter M. Grohse, David A. Binstock, and Alvia Gaskill, Jr., Research
Triangle Institute, Research Triangle Park, North Carolina 27709; Howard 1!
Kingston, National Bureau of Standards, Gaithersburg, Maryland 20899; and
Charles Sellers, Office of Solid Waste, U.S. Environmental Protection
Agency, Washington, DC 20460
ABSTRACT
The techniques that are typically used to
prepare Resource Conservation and Recovery Act
(RCRA) wastes for analysis for metals and other
elements are generally relatively time consuming,
requiring several hours to several days to com-
plete. They often involve the use of acid diges-
tions and thermal decomposition steps which may
result in analyte losses, incomplete recoveries, or
sample contamination. These limitations are well
known to the analytical community and to the end
users of these data in the U.S. Environmental Pro-
tection Agency (EPA), States, and industry. The
inefficiency of these techniques reduces laboratory
sample throughput, drives up the cost of analytical
testing, and impedes decisionmaking. Given these
concerns, the hazardous waste industry and the EPA
Office of Solid Waste Methods Section are interest-
ed in developing cost-effective sample preparation
techniques for metals and other elements in envi-
ronmental and process waste samples. Once devel-
oped, these techniques can then be written as meth-
ods for inclusion in EPA-OSW "Test Methods for
Evaluation of Solid Waste SW-846" and made avail-
able to the user community.
This paper reports on the evaluation of sever-
al microwave assisted sample preparation methods
for determining elements in solid waste. The meth-
od was evaluated for microwave-assisted digestion
of sediments, sludges, soils, and oils.
INTRODUCTION
One particularly attractive sample preparation
technique that is now receiving considerable atten-
tion is microwave-assisted sample dissolution. A
typical example of this technique involves placing
a sample in an acid solution in a closed inert
vessel equipped with a pressure relief valve. The
vessel is then subjected to microwave energy in a
modified microwave oven. The conditions of high
temperature generated in the container, coupled
with the rapid heating of the sample via direct
microwave energization of the acid molecules, can
result in significantly reduced preparation time,
from several hours in a conventional convection
oven, hot plate, or steam bath to several minutes
in the microwave oven.
Previously, work was reported on the evalua-
tion of a commercially available microwave oven
sample preparation system (Binstock et al., 1987).
The effect of sample preparation conditions, in-
cluding the acid matrix, heating time, and pres-
sure, were evaluated for toxic or hazardous ele-
ments in particulates, ashes, oils, and oil fuels.
Based on in-vessel temperature and pressure
profile studies conducted by the NBS, microwave
oven preparation conditions for oils and soils have
been determined and written as a. draft method.
These involve the use of concentrated nitric acid
as the digestion medium. The intent is not to
completely solubilize all elements in the sample.
Rather, it is to solubilize those most likely to be
made environmentally available.
NBS Standard Reference Materials, representa-
tive of oils and soils were prepared in this labor-
atory by several microwave-assisted digestion meth-
ods. Analyses were carried out by Inductively
Coupled Plasma Spectrometry and Graphite Furnace
Atomic Absorption.
EXPERIMENTAL METHODS
Microwave Oven
The MDS-81D Microwave system (GEM Corporation,
Indian Trail, NC) was used for this study. The
oven resembles a standard microwave oven, but is
equipped with additional features to facilitate
sample preparation. For example, the Teflon-coated
microwave cavity has a, variable speed corrosion
resistant exhaust system and three safety inter-
locks. A precise microwave variable power supply
is controlled by a programmable micro-processor
digital computer. Other elements of the system
include a rotating turntable, Teflon vessels with
caps and a patented pressure relief valve, a capp-
ing system, and a cooling tank. The 120 m Teflon
sample vessels and caps are designed to withstand
pressures up to 100 psi and temperatures up to
200 °C.
Inductively Coupled Plasma Emission Spectrometry
fICPES)
Analytical measurements were performed using
an Instrumentation Laboratory Plasma 200 ICP
411
-------
(Franklin, MA) or a Leeman Labs plasma Spec I ICP
(Lowell, MA) Both instruments are sequential
ICPs.
Atomic Absorption Spectrophotometry (AAS)
All AAS measurements were made using a Perkin
Elmer Zeeman 3030 with graphite furnace atomiza-
tion.
Reagents
All inorganic acids used were of "Ultrex"
quality, from J. T. Baker Chemical Co. Other chem-
icals were of analytical reagent grade quality.
Deionized (D.I.) water of 18 Mfl/cm specific resis-
tivity was used.
Standard Reference Materials
The microwave method evaluation was carried
out using the following materials:
• NBS SRM 4355—Peruvian Soil
NBS SRM 2704—Buffalo River Sediment
NBS SRM 1085—Wear Metals in Oil
NBS SRM 1634b—Trace Elements in Fuel
Oil.
In addition, to simulate a contaminated soil, a 1:1
mixture of 1634b and 2704 was prepared and ana-
lyzed.
Microwave Preparation Method
The method described below was developed for
two vessels in the microwave oven and is optimized
for temperatures and pressures that would produce
efficient chemical decomposition of the sample. A
higher power setting (574 W) would be utilized for
six vessels.
A 0.25 g sample was heated in the microwave
oven with 10 mL concentrated HNOs for ten minutes
at a power setting of 344 watts. Two sample ves-
sels at a time were placed in the microwave oven
carousel with accompanying vapor trap vessels.
Samples were diluted to 50 mL with D.I. water.
To reduce the likelihood of analyte loss due
to volatilization, a, configuration was employed
which utilizes a second vessel to trap the hot acid
vapor and any aerosol expelled when the pressure
valve opens. A PFA Teflon tube connects the diges-
tion vessel to a second vessel with a double-ported
cap. The second port on the catch vessel remains
open to the atmosphere preventing pressure buildup
in the second vessel. The acid and any sample
condensed in the second vessel are washed back into
the sample vessel at the end of the microwave pro-
cedure and made to 50 mL volume with laboratory
pure water. The contents of the sample vessel are
then analyzed. Ten replicates of each of the four
NBS SRMs, and the mixture, were digested and ana-
lyzed.
Samples were analyzed for 19 elements by ICP.
As and Se were determined by graphite furnace AA.
Initially 0.5 g of Peruvian Soil was digested;
however, this was found to cause undue pressure
buildup and venting within the sample and overflow
vessel. For subsequent runs, the sample weight was
reduced to approximately 0.25 g. A tube was also
added, connecting the overflow vessel with the
center well on the carousel, to capture potential
venting from the overflow vessel.
A study examining the overflow/capture solu-
tions for any appreciable recoveries was per-
formed. The condensate was collected and analyzed
separately for four replicates each of Peruvian
soil and wear metals in oil.
By comparison, a series of microwave diges-
tions were performed that employed all reagents
utilized in Method 3050 including nitric acid
(HNOs), hydrochloric acid (HC1), and hydrogen per-
oxide (H202)• Samples that underwent these diges-
tions included NBS 2704 (Buffalo River Sediment),
the 1:1 mixture of NBS 2704 and NBS 1634b (Trace
Elements in Fuel Oil), and NBS 1085 (Wear Metals in
Lubricating Oils).
RESULTS
Since the microwave method is a non-rigorous
acid-leach digestion (Tables 1, 2, and 3) like
3050, analytical results for the digestates of two
soil/ sediment SRMs and the mixture were compared
with those obtained using SW-846 method 3050.
In general, good agreement was obtained be-
tween the two methods. For most elements, compara-
tive values were within 25 percent. Exceptions
were Al and V in Peruvian Soil; Al, Ba, Be, and V
in Buffalo River Sediment; and As and Be in the 1:1
mixture of Buffalo River Sediment and Fuel Oil. As
predicted, using these non-rigorous digestions, the
soil values were usually below the expected levels.
Graphite furnace analyses, for As and Se, were
hindered by an apparent interference due to the
high acid concentration of the digestate (approxi-
mately 20 percent). This was the apparent cause of
a large erratic background signal.
Results for digestion of the two oil SRMs,
along with analysis of spiked samples, are shown in
Tables 4 through 7. In the case of SRM 1085,
excellent agreement was obtained with the NBS Cer-
tified Values (Table 4), with all certified elemen-
tal concentrations within 14 percent of NBS values.
SRM 1634b, Trace Elements in Fuel Oil, is a more
difficult material because of increased viscosity,
a. larger amount of aromatic compounds, and lower
elemental levels. Spike recoveries for both oil
SRMs were excellent (Tables 6 and 7) with the
exception of As and Se (see above).
An examination of the overflow/capture vessels
for any appreciable elemental recoveries indicated
that,with the exception of Zn in Peruvian soil,
(with a mean condensate concentration of 6.4 /ig/g)
no recoveries above background level were observed
for four replicates each of Peruvian Soil and Wear
Metals in Oil. A small quantity of condensate was
observed in four of the ten vessels.
For those samples undergoing the microwave
412
-------
digestion that utilized HC1 and R'fl'z in addition
to HN03, results indicated that there was no appar-
ent improvement over those results obtained with
HNOs alone. Selected results are shown in Tables
8, 9, and 10.
CONCLUSIONS
Evaluation of a draft microwave digestion
method, for determining elements in solid waste,
indicates that the HN03 microwave method should
prove a suitable alternative for SW-846 method 3050
with a substantial time/cost savings. It also
provides satisfactory results for microwave diges-
tion of oils.
Current studies are being conducted with the
goal of optimizing sample weight and microwave
power parameters. A collaborative study will then
be conducted.
REFERENCE
Binstock, D. A., P M. Grohse, P. L. Swift, A.
Gaskill, Jr., T. R. Copeland, and P. H.
Friedman, 1987, Evaluation of Microwave
Techniques to Prepare Solid and Hazardous
Waste Samples for Elemental Analysis, Solid
Waste Testing, and Quality Assurance, 3rd
Annual Symposium.
TABLE 1. ICP ANALYSIS OF NBS SRM 4355 PERUVIAN SOILa
Element
Al
As
Ba
Be
Cd
Ca
Cr
Co
Cu
Fe
Pb
Mg
Mn
Mo
Ni
Se
Ag
Sr
V
Zn
Mean + S.D. (n=10)
2.12 + 0.20 %
46.4 + 1.5b
140 + 7
0.758 + 0.039
1.91 + 1.01
10,500 + 700
13.7 + 3.0
10.6 + 0.8
64.5 + 2.0
2.40 + 0.16 %
131 + 10
7,250 + 300
531 + 20
NDd
10.1 + 2.6
NDb,e
NDf
85.3 + 4.9
65.3 + 4.3
396 + 23
3050 (n=3)
3.34 + 0.33
51.6 + 3.5b
182 + 17
0.959 + 0.059
NDC
11,000 + 900
13.0 + 1.6
10.4 + 1.2
60.3 + 3.9
2.80 + 0.24 %
149 + 9
7,480 + 640
565 + 39
NDd
10.3 + 2.0
NDb.e
NDf
112 + 9
93.0 + 8.3
356 + 30
% Difference
-36
-10
-23
-21
-4.5
+5.4
+ 1.9
+7.0
-17
-12
-3.1
-6.0
-1.9
-24
-30
+11
IAEA values
8%
90
600
2
2
2%
30
10
80
4%
100
2%
900
2
10
1
2
300
20
400
aResults in
Determined by GFAA.
CD.L. 1.0
dD.L. 2.75 /ig/g.
eD.L. 0.2 /
-------
TABLE 2. TCP ANALYSIS OF NBS 2704 BUFFALO RIVER SEDIMENT*
Element
Al
As
Ba
Be
Cd
Ca
Cr
Co
Cu
Fe
Pb
Mg
Mn
Mo
Ni
Se
Ag
Sr
V
Zn
Mean + S.D. (n=10)
1.25 + 0.08 %
11.6 + 0.5b
79.3 + 3.4
0.689 + 0.110
NDC
187 + 1.04 %
69.4 + 3.4
9.42 + 1.26
89.1 + 3.6
2.91 ± 0.11 %
153 + 19
7,990 + 240
465 + 15
NDd
37.8 + 3.2
NDb>e
NDf
30.9 + 2.8
25.1 + 1.5
392 + 19
3050 (n=3)
2.50 + 0.19
12.8 + l.lb
132 + 10
1.05 + 0.05
NDC
1.88 + 0.01 %
78.9 + 2.9
10.8 + 0.5
88.5 + 1.7
3.29 + 0.07 %
169 + 8
9,080 + 150
486 + 4
NDd
41.8 + 0.6
N])b,e
NDf
41.4 + 1.0
49.4 + 2.8
403 + 4
% Difference
-50
-9.4
-40
-34
-0.5
-9.5
-13
+0.7
-12
-9.5
-12
-4.3
-9.6
-25
-49
-2.7
aResults in /ig/g. dD.L. 2.75 /ig/g.
bDetermined by GFAA. eD.L. 0.2 /ig/g.
CD.L. 1.0 /Jg/g. fD.L. 3.0 /ig/g.
414
-------
TABLE 3. ICP ANALYSIS OF 1:1 MIXTURE: NBS 2704—BUFFALO RIVER
SEDIMENT NBS 1634b—TRACE ELEMENTS IN FUEL OILa
Element
Al
As
Ba
Be
Cd
Ca
Cr
Co
Cu
Fe
Pb
Mg
Mn
Mo
Ni
Se
Ag
Sr
V
Zn
Mean + S.D. (n=10)
6,550 + 680
5.52 + 0.32b
45.7 + 4.0
0.741 + 0.110
NDC
1.04 + 0.08 %
41.1 + 13.8
NDC
42.1 + 5.6
1.40 + 0.08 %
72.1 ± 13.4
4,000 + 310
225 + 16
NDd
26.6 + 5.6
NDb.e
NDf
ND
41.5 + 1.9
211 + 18
3050 (n=3)
8,720 + 1,760
3.84 + 0.99b
58.2 + 8.0
0.559 + 0.053
NDC
1.18 + 0.04 %
45.1 + 2.8
5.99 + 0.29
51.6 + 2.9
1.66 + 0.09 %
83.4 + 5.4
4,650 + 250
252 + 11
NDd
35.4 + 1.4
NDb'e
NDf
15.1 + 1.6
47.1 + 5.0
231 + 8
% Difference
-25
+44
-21
+32.6
-12
-8.9
-18
-16
-14
-14
-11
-25
-12
-8.6
aResults in /ig/g. dD.L. 2.75 /lg/g.
bDetermined by GFAA. eD.L. 0.2 /ig/g.
CD.L. 1.0 ^g/g. fD.L. 3.0 /ig/g.
415
-------
TABLE 4. ICP ANALYSIS OF NBS SUM 1085 WEAR METALS IN
LUBRICATING OILa
Element
Al
As
Ba
Be
Cd
Ca
Cr
Co
Cu
Fe
Pb
Mg
Mn
Mo
Ni
Se
Ag
Sr
V
Zn
Mean + S.D. (n=10)
337 + 22
b
0.928 + 0.890
NDC
2.07 + 0.92
67.3 + 91.6
310 + 10
NDd
316 + 11
320 + 11
305 + 19
300 + 15
0.837 + 0.325
265 + 9
310 + 11
b
309 + 27
NDd
NDe
8.12 + 4.89
( ) Not certified
aResults in /Jg/g.
bTo be determined by GFAA.
NBS values !
296
—
—
298
295
300
(305)
297
292
303
(291)
«0.3)
CD.L. 0.6 /Jg/g.
dD.L. 2.4 yug/g.
eD.L. 1.6 /ig/g.
% Difference
+14
+4
+7
+7
0
+ 1
-9
+2
+6
416
-------
TABLE 5. IGP ANALYSIS OF NBS SRM 16Mb TRACE
ELEMENTS IN FUEL OILa
Element
Al
As
Ba
Be
Cd
Cr
Co
Cu
Fe
Pb
Mg
Mn
Mo
Ni
Se
Ag
Sr
V
Zn
( ) Not certified
aResults in pg/g.
bDetermined by GFAA.
CD.L. 0.4 pg/g.
dD.L. 0.9 pg/g.
eD.L. 0.6 pg/g.
Mean + S.D. (n=10)
29.7 + 8.1
NDb.c
3.68 + 0.36
0.220 + 0.023
NDd
NDe
NDf
W)g
38.7 + 4.4
NDh
13.7 + 1.6
0.492 + 0.224
NDi
29.3 + 3.0
NDb>°
NDJ
NDk
57.7 + 2.2
3.28 + 3.69
fD.L.
SD.L.
hD.L.
iD.L.
JD.L.
kD.L.
NBS values
16
0.12
(1.3)
(0.7)
0.32
31.6
(2.8)
—
0.23
28
0.18
54.4
3.0
2.4 pg/g.
3.6 pg/g.
S.2 pg/g.
10.0 pg/g.
4-8 pg/g.
3.0 pg/g.
417
-------
TABLE 6. ICP ANALYSIS OF SPIKED NBS OILS
SRM 1085—WEAR METALS IN LUBRICATING OILa
Element
Al
As
Ba
Be
Cd
Ca
Cr
Co
Cu
Fe
Pb
Mg
Mn
Ni
Se
Ag
Sr
V
Zn
Expec.
2.00
0.200
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
0.200
2.00
2.00
2.00
Found
1.91
0.121b
1.90
1.93
1.85
1.74
1.84
1.93
1.86
2.00
1.85
1.83
1.97
1.80
0.120b
c
1.97
1.94
1.86
% Rec.
96
60
95
96
92
87
92
96
93
100
92
92
98
90
60
98
97
93
Unspiked
concentration
1.47
ND
0.004
ND
0.009
0.293
1.35
ND
1.38
1.40
1.33
1.31
0.004
1.35
ND
ND
ND
0.005
aResults in /Jg/mL.
bDetermined by GFAA.
cSpike unsuccessful.
418
-------
TABLE 7. ICP ANALYSIS OF SPIKED NBS OILS
SRM 1634b—FUEL OILa
Element
Al
As
Ba
Be
Cd
Ca
Cr
Co
Cu
Fe
Pb
Mg
Mn
Mo
Ni
Se
Ag
Sr
V
Zn
Expec.
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
Found
2.06
b
1.84
1.87
1.82
1.77
1.89
1.92
1.88
2.05
1.87
1.90
1.94
1.77
1.92
b
c
2.02
1.94
1.84
% Rec.
103
92
94
91
88
94
96
94
102
94
95
97
88
96
101
97
92
Unspiked
concentration
0.127
0.016
0.001
ND
0.364
0.011
ND
ND
0.166
ND
0.059
0.002
ND
0.126
ND
0.247
ND
aResults in /ig/mL.
bNot spiked.
cSpike unsuccessful.
419
-------
TABLE 8. ICP ANALYSIS OF NBS 2704 BUFFALO RIVER SEDHffiNTa
Element
Al
Ba
Be
Cd
Ca
Cr
Co
Cu
Fe
Pb
Mg
Mn
Ni
Sr
V
Zn
HN03
Mean ± S.D. (n=10)
1.25 ± 0.08 %
79.3 ± 3.4
0.689 ± 0.110
<0.9
1.87 ± 1.04 %
69.4 ± 3.4
9.42 ± 1.26
89.1 ± 3.6
2.91 i 0.11 %
153 ± 19
7,990 ± 240
465 ± 15
37.8 ± 3.2
30.9 ± 2.8
25.1 ± 1.5
392 ± 19
HN03/HC1/H202
Mean ± S.D. (n=5)
1.59 ± 0.18 %
117 ± 9
0.634 ± 0.052
2.57 ± 0.32
2.09 ± 0.10 %
103 ± 6
8.68 ± 1.00
110 ± 2
2.96 ± 0.10 %
133 ± 12
8,990 ± 360
478 ± 12
41.3 ± 3.6
41.7 ± 3.6
30.4 ± 4.2
432 ± 49
3050 (n=3)
2.50 ± 0.19 %
132 t 10
1.05 ± 0.05
<0.9
1.88 ± 0.01 %
78.9 ± 2.9
10.8 ± 0.5
88.5 ± 1.7
3.29 ± 0.07 %
169 ± 8
9,080 ± 150
486 ± 4
41.8 ± 0.6
41.4 ± 1.0
49.4 ± 2.8
403 ± 4
NBS Values
6.11 %
414
—
3.45
2.60 %
135
14.0
98.6
4.11 %
161
1.20 %
555
44.1
(130)
95
438
aResults in /Jg/g.
TABLE 9. ICP ANALYSIS OF 1:1 MIXTURE: NBS 2704—BUFFALO RIVER
SEDIMENT NBS 1634B—TRACE ELEMENTS IN FUEL OILa
Element
Al
Ba
Be
Cd
Ca
Cr
Co
Cu
Fe
Pb
Mg
Mn
Ni
Sr
V
Zn
HN03
Mean ± S.D. (n=10)
6,550 ± 680
45.7 ± 4.0
0.741 ± 0.110
<0.9
1.04 ± 0.08 %
41.1 ± 13.8
<1.0
42.1 ± 5.6
1.40 ± 0.08 %
72.1 ± 13.4
4,000 ± 310
225 ± 16
26.6 ± 5.6
<2.4
41.5 ± 1.9
211 ± 18
HN03/HC1/H202
Mean ± S.D. (n=5)
6,510 ± 2,380
55.5 ± 13.6
0.197 ± 0.069
1.29 ± 0.53
0.864 ± 0.047 %
43.6 ± 3.5
3.95 ± 0.32
44.2 ± 3.6
1.18 ± 0.07 %
56.0 ± 4.9
3,570 ± 320
184 + 9
27.4 ± 0.7
22.8 ± 5.8
32.1 ± 4.9
148 ± 8
3050 (n=3)
8,720 i 760
58.2 ± 8.0
0.559 ± 0.053
<0.9
1.18 ± 1.04 %
45.1 ± 2.8
5.99 ± 0.29
51.6 ± 2.9
1.66 ± 0.09 %
83.4 ± 5.4
4,650 ± 250
252 ± 11
35.4 ± 1.4
15.1 ± 1.6
47.1 ± 5.0
231 ± 8
aResults in /ig/g.
420
-------
TABLE 10. ICP ANALYSIS OF NBS SUM 1086 WEAR METALS IN
LUBRICATING OILSa
Element
Al
Ba
Be
Cd
Ca
Cr
Co
Cu
Fe
Pb
Mg
Mn
Mo
Ni
Ag
Sr
V
Zn
HN03
Mean ± S.D. (n=10)
337 + 22
0.928 ± 0.890
<0.6
2.07 ± 0.92
67.3 ± 91.6
310 ± 10
<2.4
316 ± 11
320 ± 11
305 ± 19
300 ± IS
0.837 t 0.325
265 ± 9
310 ± 11
309 ± 27
<2.4
<1.6
8.12 ± 4.89
HN03/HC1/H202
Mean ± S.D. (n=5)
239 ± 17
0.250 ± 0.153
<0.6
<0.9
37.8 ± 16.4
275 ± 17
<2.4
229 ± 14
250 ± 12
209 ± 18
196 ± 16
1.20 ± 0.16
168 ± 15
237 ± 9
76.2 ± 78.0
<2.4
<1.6
9.06 ± 1.53
NBS Values
296
—
—
—
298
—
295
300
(305)
297
—
292
303
(291)
—
«0.3)
( ) Not certified.
aResults in
421
-------
Rapid Screening of Organic Contaminants Using a
Mobile Mass Spectrometer in the Field
Michael C. Hadka, PhD.
Walter B. Satterthwaite, Assoc.
720 N. Five Points Road
West Chester, Pa. 19380
Randall K. Dickinson, PG.
United Engineers
30 S. Seventeenth St.
Philadelphia, Pa. 19103
ABSTRACT
The source of organic contaminants emanating from a
landfill were determined by on—site screening using a
mobile mass spectrometer (MM—1). The rapid screening
provided by the MM—1 was an effective tool in providing
immediate sample analysis for quick response and decision
making in the field. In addition, positive identification of
the contaminants was obtained from the mass spectra
provided by the instrument. The MM—1 was used to screen
soil, soil gas, and ground water for volatile organics down to
the 10 ppb level with an analysis time of ten seconds.
Volatile chlorinated organics were discovered in water from a
storm sewer located beneath a landfill which was used to
dispose of incinerator ash. Water in the storm sewer was
first screened by the MM—1 and the source was isolated.
Contaminated water was leaking into the joints of one 6"
section of the sewer. A grid pattern was laid out and the
contamination was delineated by obtaining continuous split
spoon soil samples with immediate analysis by the MM—1.
The MM—1 data was plotted on the sampling grid pattern,
thus delineating the source location.
Compounds identified by the MM—1 include
1,1,1—trichloroethane, trichloroethene and dichloroethenes.
Selected samples sent to a laboratory verified the MM—1
results.
INTRODUCTION
Rapid on—site analysis is the most economical method for
site investigations. It allows the problem to be defined in
the field instead of waiting for laboratory analysis in order to
determine the site condition. For organic analysis, on—site
analytical techniques usually involve photoionization
detectors, portable gas chromatographs or a mobile
laboratory in a trailer. This equipment is either low in
sensitivity, poor in selectivity, or expensive to set up and
operate.
As an innovative approach to rapid on—site analysis, a
mobile mass spectrometer was used during an on—site
investigation to pin point the source of contamination and
delineate the contaminant plume in soil and water. The
mobile mass spectrometer was chosen for its high
sensitivity, unambiguous identification, high dynamic range,
freedom from interferences, and ruggedness. The
instrument provided rapid identification of the
contaminant(s) and semi-quantitative results which
facilitated the field scientists' ability to make on the spot
decisions to determine the next step of the investigation.
This procedure resulted in a more thorough site
investigation while at the same time completing the job in
days instead of weeks.
INSTRUMENTATION
The instrumentation used consisted of a battery operated
mobile mass spectrometer (MM—1) manufactured by Bruker
Instruments and mounted in a four wheel drive vehicle
(Figure 1). The three main sections of the instrument are
the sample inlet system, the quadrupole mass spectrometer
and the monitor. The sample inlet system is the most
unique feature of the instrument, allowing the direct
sampling of air, water, and solids with little to no sample
preparation.
The sample inlet system consists of a 3.5 meter fused silica
column inside a flexible hose (Figure 2). One end of the
sample inlet system is connected to the mass spectrometer.
A large silicone membrane is attached to the sampling end
of the inlet system. Organic vapors in the air are
continuously drawn through the heated silicone membrane
and capillary column with a suction pump and into the mass
spectrometer.
The instrument can be run in one of two modes. In the first
mode the sample inlet system is operated isothermally, up
to 230 C. In the second mode the sample inlet is operated
in a temperature program mode allowing the temperature to
be ramped from a low temperature to a higher temperature.
The temperature program mode allows a limited separation
of the compounds by the capillary column.
Compound identification is performed by one of two
methods. The first method uses a selected ion monitoring
procedure for specified target compounds. Up to four ions
per compound and up to 22 organic compounds can be
monitored simultaneously. In this mode the instrument's
423
-------
microprocessor identifies each compound by determining if
the monitored ions are above a minimum detection level and
their relative intensities matches the library reference. The
monitor displays the intensity of the ions monitored for each
compound and what compounds were identified. The
second method of identification is performed by acquisition
of the full spectrum (up to mass 400). The instrument's
library is searched for a match and the name of the
compound is displayed if a match is found. If no match is
found the spectrum can be interpreted and the compound
identified by a mass spectroscopist/chemist.
The intensities of the ions are proportional to the
compound's concentration. The ion intensity for each
compound identified is displayed on the monitor as a log
function. Thus, every increase in one log unit is a tenfold
increase in concentration. Calibration of the instrument is
performed by using standards of known concentration. The
instrument is capable of measuring up to eight orders of
magnitude from 10 parts per billion (ppb) in air or water, up
to the percent range. The detection limit for soils is around
100 ppb. Compound identification and quantification can be
performed in five to ten seconds.
ANALYTICAL METHODOLOGY
At the beginning of each day the instrument is calibrated
and background concentrations are obtained. The mass
calibration is performed automatically by the instrument
using a fluorohydrocarbon (FC—77) standard. A
background is taken by sampling the air in a clean
environment. The instrument automatically sets the
minimum detection limit during the background acquisition.
During the operation the instrument uses argon in the air as
an internal standard to continuously monitor the
performance of the mass spectrometer. The capillary
column and silicone membrane are set to operate at 180
degrees C. After setup and preparation, the instrument is
driven to each sample location.
Sampling is done directly using the sample inlet system with
little or no sample preparation. With the instrument's
mobility and the 3.5 meter reach of the sample inlet system,
most samples can be measured in—situ by one of the
following procedures:
o For soil and solids, the heated sample probe is pressed
against the sample. Volatiles and semivolatiles are heated
and vaporized into the sample inlet system. Thus even
PCBs can be analyzed directly by the MM—1.
o For aqueous samples, a headspace technique is used. The
sample is collected in a 500 ml wide—mouth bottle allowing
a 4 cm headspace below the lid. The bottle is sealed and
shaken. The lid is removed and the sample probe
immediately placed over the mouth of the bottle.
o For soil gas samples the probe is inserted down the hole of
the soil boring.
For this work soil gas, soil and aqueous samples were
measured. The soil was analyzed on the surface and at
various depths from samples collected in 2 inch diameter
split spoons.
SITE BACKGROUND
The site investigation was performed at a municipal landfill
in eastern New York. The landfill is roughly rectangular in
shape with the long axis oriented north—south. The total
filled area encompasses approximately 26 acres. Since its
inception in the 1930's, the facility has been operated by the
local city public works authority. According to available
records, no private haulers or disposal companies have ever
used the landfill.
Originally, the landfill was used for the disposal of ash
residue from the city operated sanitary waste incinerator.
Since the termination of incineration operations in the late
1970's, the landfill has been used for the composting of
leaves and grass cuttings which continues to the present
time. The leaves and cuttings are shredded, mixed with
silty sand, and sold as topsoil substitute.
The ash material varies in depth from 9 to 18 feet and is
grayish—black to black similar in nature to silt and sandy
silt. Bulk material such as wood, carpeting, and tires were
disposed of along with the ash and during the site
investigation were found sporadically throughout the
deposit. A greenish—gray clay underlies the ash material
and is the uppermost naturally occurring deposit at the site.
The clay layer is 20 feet thick at the northwestern end of the
landfill, gradually pinching out, and is not present in the
southeastern portion of the landfill. The clay overlies the
Manhattan Formation which consists of a mica rich,
schistose gneiss underlying the entire area of the landfill.
Locally, the upper portion of the bedrock may be highly
weathered to essentially a platy, fine to medium grained
sand. The weathered portion may be up to 10 feet in
thickness.
In 1984 and 1985, several storm sewer and sanitary sewer
pipelines were constructed across the landfill (Figure 3).
The storm sewer lines drained surface water runoff from a
residential area on the west side of the landfill. The runoff
water was discharged to an unnamed stream that formed
the west border of the landfill. The pipelines were installed
at least several feet below the bottom of the ash fill. In
some areas, channels were excavated in the bedrock to lay
the pipelines where competent rock was encountered.
In August and October of 1986 and March 1987 the local
environmental agency in conjunction with the state
environmental agency sampled and analyzed several of the
storm sewer lines that traverse the landfill. Samples were
obtained from the manhole inlets along the pipelines and the
discharge outlets to the stream. Sample analyses indicated
the presence of chlorinated solvents. Table 1 is a summary
of the sample analytical results, and sample locations are
shown on Figure 3. The principal compounds identified
were 1,1,1—trichloroethane, 1,1—dichloroethane,
1,1—dichloroethene, 1,2—dichloroethenes, trichloroethene,
and vinyl chloride. The greatest concentration encountered
424
-------
was 4600 ug/l (ppb) for 1,1,1-trichloroethane.
PHASE I - SITE SCREENING
The initial phase of the investigation was to verify the
presence of contaminants at the locations sampled by the
agencies. The two pipelines that traversed the southern
portion of the landfill were sampled and analyzed with the
MM—1. Samples of the water in the pipelines were obtained
from the manhole inlets. Additionally, the pipeline outfalls
to the stream as well as sections of the stream were
sampled and analyzed. The results of these analyses are
shown on Table 2.
The analyses indicated that the contamination was
emanating from the southernmost pipeline. The sample
locations are shown on Figure 3. The contaminants were
encountered at the greatest concentration in manhole (MH)
11. The same contaminants were detected at a lesser
concentration at the outfall of the MH —11 pipeline and in
the stream, downstream of the same outfall.
PHASE II - CONTAMINANT LOCATION
Phase I of the investigation indicated that the contamination
was entering the pipeline somewhere between MH—10 and
MH—11. Assuming a point source, Phase II was to
determine the exact location of the contamination along the
pipeline.
This section of the pipeline was constructed of precast,
3—foot inside diameter, reinforced concrete, sectioned in 6
foot lengths. Each section was fitted with a rubber gasket
at the connection ends.
Based upon discussions with municipal department
employees, this section of the pipeline was located beneath
the ground water table. This was a localized condition
because of an adjacent area to the pipeline which was
formerly used as a drainage basin for surface runoff within
the landfill. Consequently, the pipeline was inspected from
the inside where three joints were found to be leaking from
the top. The analyses of these samples did not detect any
contaminants.
Further discussions with the landfill operators revealed that
some small amount of liquids of unknown origin were once
disposed of in the former drainage basin.
Three test pits were excavated in the former drainage basin
adjacent to the pipeline (see Figure 3). The pits were
excavated to a depth of approximately 8 feet. Standing
water in the pits was approximately three feet below the
ground surface. Several samples were obtained from each
pit but none of the chlorinated organics compounds were
detected by the MM-1.
At this point, the pipeline was investigated once again. A
water sample was obtained immediately down flow of each
section joint. Every sample was analyzed by the MM—1.
Starting at MH —10, twelve of the joints downstream of the
manhole were sampled and tested negative for the
chlorinated organic compounds. The pipe joints upstream
of MH —11 were sampled and tested. The third pipe joint
upstream from MH—11 was analyzed with the greatest
concentration (3000 ppb). The fourth joint upstream of the
manhole encountered chlorinated organic compounds at the
limit of detection (see Table 3). The location where the
contaminants were entering the pipeline was 18 feet
upstream from manhole MH-11.
PHASE III - SOURCE DELINEATION
As an initial attempt to locate the contaminant source, a
test pit was excavated down to the storm sewer line at the
joint located in Phase II. The excavation pit indicated that
the bedrock was removed to create a channel for the pipeline
in this section of the line. The 3—foot pipeline was located
approximately 13 feet below the ground surface and the
ground water was approximately 15 feet below the surface.
Soil and water samples were obtained from the pit and
analyzed indicating elevated concentrations of chlorinated
organics. The greatest concentrations detected were up to
500,000 ug/L.
The highest contamination was encountered in a tan colored
sand deposit that apparently was oriented nearly
perpendicular to the pipeline and approximately eight feet
below the ground surface.
This deposit was saturated and was draining into the
excavated pit. The sand provided a conduit for the
contaminated water which drained into and collected in the
channel excavated in the bedrock. The water then was
entering the pipeline at one of the joints which was
improperly sealed. The competent bedrock was relatively
impermeable which precluded the vertical migration of the
contaminated water.
A 50 foot by 50 foot borehole grid pattern (Figure 4) was
laid out in the area most likely to contain the source of the
contamination. Grid point Bl was the location where the
contamination was entering the pipeline. Soil borings were
drilled at each coordinate location. Soil samples were
obtained using continuous split—spoon samplers. Each
sample was immediately analyzed by the MM—1 to
determine the presence of volatile organic compounds.
Each boring was completed to auger refusal at the top of
competent bedrock. Soil samples could not be recovered
from a number of sampling intervals because of subsurface
conditions encountered. Wood, tires, and the localized very
loose, non—cohesive nature of the ash material prevented
recovery of samples in several of the split spoons samples.
The sequence of drilling on the grid pattern was dictated by
the results of the boring sample analysis by the MM—1 as
each boring was completed. Once the approximate location
of the source had been determined, additional boreholes
were drilled off the grid pattern to further delineate the
source of the contamination. As shown on Figure 5, the
area of the contamination is limited to that encompassed by
a diagonal from coordinate Bl to C2, extending slightly
beyond the C2—B2 gridline, and southward to the B gridline.
425
-------
Figure 6 is a fence diagram depicting the borehole locations,
corresponding depth with analytical results. As can be seen
from the figure, the contaminated area varied in depth from
between 10 to 16 feet below the ground surface and was up
to 6 feet thick. Based on the analytical results, the source
extended approximately 80 feet in a northeast direction from
the pipeline and was roughly 20 feet wide. Assuming these
dimensions, the volume of the contaminated area was 360
cubic yards of soil/fill material.
VERIFICATION OF MM-1 RESULTS
In order to verify the MM-1 results a select number of
samples were sent to a certified laboratory for volatile
organic analysis. The samples were analyzed using the EPA
Contract Laboratory Program (CLP) procedures utilizing gas
chromatography/mass spectroscopy. To minimize the loss
of volatiles from the soil the CLP procedure was modified by
collecting the sample in bottles containing a known volume
of methanol. Thus the methanol served as a preservative
and an extraction solvent for the volatiles. Methanol is the
extraction solvent used in the CLP procedure.
The results for the soil verification samples, Table 4,
confirm the identifications of dichloroethenes,
1,1,1—trichloroethane and trichloroethene by the MM—1.
The MM—1 could not distinguish the differences between
1,1—dichloroethene and 1,2—dichloroethene in the operation
mode used. The MM—1 did not detect low levels of
1,1—dichloroethane and 1,1,2,2—tetrachloroethane.
Since the ion intensity is directly proportional to the
concentration, the log ion intensity from the MM—1 can be
used to calculate semiquantitative results. The laboratory
results for the soil samples are compared to the MM—1
output in Table 4 and plotted in Figure 7. This plot was
used to estimate the concentration of the organics for each
sample. Since the greatest variable in measurement is the
sample size, the log—log plot in Figure 7 is sufficient to
estimate semiquantitative results. For soil samples when
the sample probe is pressed against the sample such factors
as packing density of the soil and the depth to which the
organics are stripped effects the results by this technique.
In the ground water matrix, shown in Table 5, the results of
verification samples show agreement with the MM-1
results. Vinyl chloride and 1,1-dichloroethane were not
detected in Sample 3 by the MM—1 due to interference
from other chlorinated organics present. This is because in
the isothermal mode, all of the compounds are detected at
once and the ions monitored for vinyl chloride and
1,1-dichloroethane are common to the other chlorinated
compounds.
COST COMPARISON
Table 6 is a cost summarization for the three phases of work
that were completed for the site investigation. The actual
costs incurred using the MM-1 are compared with the
estimated costs expected if the investigation was conducted
using standard EPA protocol (1). The actual costs for the
field work were $52,400 as compared to an estimated cost of
$248,000 using EPA procedures. The laboratory costs were
calculated based on the analysis of the number of samples
obtained during the actual investigation. Travel expenses
were considered to be equivalent even though the EPA
procedure would be of longer duration and, therefore, more
costly.
In addition to the lower costs, the other advantages of
utilizing the mobile mass spectrometer include the following:
o Instantaneous results confined the extent of the study area
to just that which was contaminated. Every sample
analyzed continually defined the limit of contamination as
the field work progressed. This subsequently reduced the
number of samples required as compared to having all the
samples laboratory analyzed.
o The MM—1 was used as an air monitor for personnel
health and safety during field operations. The air
monitoring results were used as the trigger to upgrade
personal protective clothing and equipment. A higher work
productivity was achieved using this procedure than if a
defined level of protection would have been required to be
worn at all times during the field work. The air monitoring
capabilities of the MM—1 not only identified conditions
when upgraded levels of personnel protection were required
but also provided approximate contaminant concentrations
to determine the specific level of protection.
o The most significant advantage using the MM—1 is time.
The actual project was completed four and one—half weeks
including oral presentation of the data and final written
report submittal to the agencies. Under EPA protocol, a
minimum of 4 to 6 weeks would be required after each phase
of work before the analytical data could be obtained from
the laboratory. Consequently, the project would be
estimated to be completed in 16 to 22 weeks using standard
EPA procedures.
o The necessary manpower is reduced since only a minimum
number of samples are analyzed by a laboratory. Bottle
preparation, sample handling, chain of custody, sample logs,
decontamination procedures and shipping of the samples are
drastically reduced.
These items are indirect cost factors which were not
incorporated into the calculations detailed on Table 6.
Therefore, the estimated cost of $248,000 to complete the
EPA conducted investigation is conservative.
CONCLUSION
The use of the MM—1 for the site investigation provided a
cost effective method to accurately analyze environmental
samples in the field. Positive identification of the organic
contaminants present and their concentrations was provided
to the investigating team on a sample by sample basis.
This allowed site personnel to make immediate judgments
as to where to sample next and quickly determine the source
of the contamination and delineate the contaminant plume.
With the rapid turnaround afforded by the MM-1, less time
was spent at the site for manpower and equipment while at
426
-------
the same time keeping laboratory costs minimal. Thus,
costs of the investigation were only one—fifth of a
conventional investigation. Significantly more samples may
be tested using this technique than would be with a
conventional investigation. In addition, the sampling crew
was not randomly sampling the site and sending
unnecessary samples to the laboratory, increasing costs.
As a field investigating tool the MM—1 is extremely
versatile; capable of testing air, water and soil with little or
no sample preparation. Its sensitivity is better than most
field equipment and it gives positive compound identification
that most field equipment is not capable of doing. Although
the MM-1 did not identify every compound present in the
operation mode used, the MM—1 was able to identify the
major contaminants and solve the problem in the field. If
the identification of additional compounds is required gas
chromatography separation is available with the sample
probe by using temperature programming.
The MM—1 also serves as a valuable quality assurance tool.
Since it measures most samples either in—situ or
immediately thereafter, the MM—1 can be used to verify
laboratory results which are prone to errors due to
cross—contamination, loss of volatiles during transport of
the samples, or internal laboratory contamination during
sample preparation and analysis.
References
1. U.S. Environmental Protection Agency "Remedial Action
Costing Procedures Manual" EPA/600/8-87/049, October
1987
Figure 1. MM-1 mounted in a four wheel drive
vehicle with the sampling inlet system used to
measure the headspace of a well.
427
-------
Air
Sampler {
Membrane—*?^
Compounds
Vapor
!-.. -Sampler Coupling
^N/Membrane Housing
• Mass Spectrometer
"
to Vacuum Pump
Figure 2-Diagram of Sample Probe
SANITARY SEWER
PIPELINE
STORM SEWER
PIPELINE
Scale
Q 200 feel
SANITARY SEWER
PIPELINE
STORM SEWER
PIPELINE
I I I I BOREHOLE GRID
L-L-L-J PATTERN
Scale
O 200 feet
Figure 3 - Storm Sampling and Test Pit Locations
Figure 4-Location of Grid Pattern for Drilling
428
-------
C1
e
Q3/0.3/ND
ND/ND/ND
ND/ND/ND
0.1/2/ND
Figure 5 - Contaminant Delineation with MM-I
LEGEND
TCA/DCE/TCE (concentration in ppm)
0 BOREHOLE
,—.— STORM SEWER
CONTAMINANT PLUME
Scaie(inPPM)
0 KDfeet
LEGEND
O Boring Location
NR Soil Samples Not Recovered
BR Bedrock Encountered
— Approx. Location of Storm Sewer
ND Volatile Organics Not Detected
83 Total Volatile Organic Concentration
(Expressed as PPM of Total Volatile Organics Present)
-NR
-NR
-ND
-O.I
-0.5
-ND
-ND
Figure 6 - Vertical Contaminant Delineation
429
-------
IUUU
100
|
1|
•§1'°
^l
\
O.I
//
D./y /
//
//
///
/
- /
•'/ LEGEND
• TCE
^ ***%^
234567
MM-I Result
Log Ion Intensity
f
|
8
Figure 7- Comparison of MM-I Ion Intensity
to Laboratory Data for Soils
TABLE \
AGENCY ANALYTICAL RESULTS
(ug/L)
LOCATION
A B C D
chloroform
chloroethane
vinyl chloride
1 , 1-dichloroethane
1 , 1-dichloroethene
1 , 2-dichloroethene
1 , 1 / 1-trichloroethane
trichloroethene
NF NF 1.6 3.2
NF 20 NF NF
NF 130 3.8 NF
NF 100 3.1 NF
NF 100 1.1 NF
NF 1800 28 NF
NF 4600 45 NF
NF 750 9.3 NF
E F G
NF 15 NF
NF NF NF
NF NF NF
NF NF NF
NF NF NF
5.5 NF NF
1.8 NF NF
5.2 NF NF
430
-------
TABLE 2
MM-1 RESULTS FOR MANHOLE WATER SAMPLES
(ug/L)
LOCATION
MH 8
NF
NF
NF
NF
MH 10
NF
NF
NF
NF
MH 11
NF
NF
1000
100
Outlet B
NF
NF
300
50
1,1-dichloroethane
dichloroethenes*
1,1,1-trichloroethane
trichloroethene
NF = not found
* includes 1,1-dichloroethene and cis- and trans-l,2-dichloroethene
TABLE 3
MM-1 RESULTS FOR STORM SEWER SAMPLES TAKEN AT
CONSECUTIVE JOINTS UPSTREAM OF MANHOLE 11
(ug/L)
1,1-dichloroethane
dichloroethenes*
1,1,1-trichloroethane
trichloroethene
Joint
1
NF
NF
800
100
Joint
2
NF
NF
1000
100
Joint
3
NF
NF
3000
400
Joint
4
NF
NF
5
NF
Joints
5-7
NF
NF
5
NF
Joint
8
NF
NF
NF
NF
NF = not found
* includes 1,1-dichloroethene and cis- and trans-l,2-dichloroethene
TABLE 4
COMPARISON OF LABORATORY RESULTS FOR SOIL SAMPLES IN MG/L
TO THE MM-1 RESULTS IN LOG ION INTENSITY
Sample 1
Tab
0.91
19.4
.78
.46
MM-1
NF
5.6
5.9
5.9
Sample 2
Lab
NF
30.3
646
262
MM-1
NF
6.4
7.1
6.9
Sample 3
Lab
1.1
4.8
13.1
0.6
MM-1
NF
5.1
5.1
NF
Sample 4
Lab
NF
12.0
39.0
11.0
MM-1
NF
5.5
5.7
5.2
1,1-dichloroethane
dichloroethenes*
1,1,1-trichloroethane 178
trichloroethene
NF = not found
* includes 1,1-dichloreothaene and cis- and trans-l,2-dichloroethene
431
-------
TABLE 5
COMPARISON OF LABORATORY RESULTS FOR GROUNDWATER SAMPLES IN UG/L
TO MM-1 RESULTS IN LOG ION INTENSITY
1, l-dichloroethane
dichloroethenes*
vinyl chloride
1,1,1-trichloroethane
trichloroethene
NF = not found
Sample 1
Lab MM-1
NF NF
NF NF
NF NF
NF NF
NF NF
Sample 2
Lab MM-1
NF NF
58,500 5.9
NF
6.5
NF
172,000
5.5
Sample 3
Lab MM-1
94
215
157
29
122
33,000
* includes 1,1-dichloroethene and cis- and trans-l,2-dichloroethene
NF
4.2
NF
2.9
3.9
TABLE 6
SITE INVESTIGATION
COST COMPARISON
PHASE I
SITE SCREENING
(15 SAMPLES)
PHASE II
CONTAMINATION LOCATION
(43 SAMPLES)
PHASE III -
SOURCE DELINEATION
(195 SAMPLES)
GRAND TOTAL:
ACTUAL COSTS
PEOPLE DAYS TOTALd)
2 0.5 $1,350
2 1.0 14,050
2 10.0 $47,000
$52,400
ESTIMATED EPA COSTS (*)
PEOPLE DAYS LABOR(2) LABORATORY(3) DRILLING(4) TOTAL
3 1 $1,200 $12,000 N/A $13,200
1.5 $2,400 $34,400 N/A
$36,800
15 $24,000 $156,000 $18,000 $198,000
$248,000
NOTES:
1 Includes 10 laboratory verification samples based on $800/sample.
2 Based on $50/hr/person.
3 Based on $800/sample for method No. 624-625 analysis.
4 Based on $1200/day/dri 11 rig.
* US EPA, "Remedial Action Costing Procedures Manual" EPA/600/8-87/049, October 1987.
432
-------
DETERMINATION OF CHLORDANE IN SOIL BY ENZYME IMMUNOASSAY
Rodney J. Bushway1, Wayne M. Pask2, Joan King1,
Brian Perkins' and Bruce S. Ferguson
Professor and Research Associates, Department of Food Science
„ University of Maine, Orono, ME 04469
Associate Analytical Chemist, Indiana State Chemists Office
Department of Biochemistry, Purdue University, West Lafayette, IN 47907
President, ImmunoSystems, Inc.
Biddeford, ME 04005
ABSTRACT
An enzyme Immunoassay (EIA) method has been
developed to screen soils for chlordane residues.
Soils are extracted with a 10 min son I cat I on in
methanol/water (90:10). A 100 ul aliquot Is
removed from a 1:4 dilution of the sample and
added to a polystyrene tube coated with chlordane
antibody. Next, 160 ul of chlordane-enzyme
conjugate is added to the tube. The chlordane
in the sample "competes" with the enzyme-tagged
chlordane for the antibody Immobilized on the
tube. Tubes are Incubated at room temperature
for 5 min before rinsing with distilled water to
remove unreacted sample and enzyme conjugate.
Finally, 160 ul each of substrate and chromogen
are added and the colored reaction product Is
allowed to develop for 5 min before the reaction
Is "stopped" with 2.5N sulfurlc acid. Samples
can be quantified using a hand-held battery
powered photometer which makes possible analysis
In the field. Qualitative ("yes-no") results are
even simpler to achieve. The linearity range for
chlordane Is 2.5 to 80 ng/tube. Samples greater
than 80 ng can be diluted. Agreement between
this method and gas chromatography indicate that
the Immunoassay procedure would be a good
screening test for chlordane residues in soil.
Other cyclodlenes, Including heptachlor,
heptachlor epoxlde, dleldrln, aldrln, endrln,
endrln ketone, chlordene and endosulfan, cross-
react making this screening method suitable as a
broad spectrum test. ElA's such as this, which
are fast, simple, sensitive, and Inexpensive are
highly suited for rapid, on-site screening of
environmental contaminants, particularly at a
site where specific residues are suspected.
INTRODUCTION
Chlordane (1,2,4,5,6,7,8,8-octachloro-2,3-
3a,4,7,7a-hexa hydro-4,7-methano-1H-indene) an
organochiorine insecticide belonging to the
cyclodlene group had been used for approximately
40 years against soil and animal pests. However;
In 1974 because of Its potential health effects
(1), the United States Environmental Protection
Agency (U.S. EPA) revoked all uses of chlordane
except subterranean termite control. Finally,
after studies showing that chlordane vapors were
present In living areas of treated homes (1-4),
the EPA in 1987 cancelled all uses. Even though
chlordane Is no longer applied, It still poses a
substantial environmental threat because of its
persistence (5), past widespread distribution
(1,6) and potential chronic toxlclty (1,4,6).
Thus the need to monitor soil around treated
houses and toxic waste sites still exists.
Methods presently available to measure chlordane
residues In soil are electron capture gas
chromatography (7,8) and gas chromatography/mass
spectroscopy (6). Both techniques employ very
expensive equipment that require experienced
technicians. Furthermore analysis is lengthy and
cannot be performed in the field. Recently new
technology (Immunochemlcal) has been applied to
the analysis of pesticides (9). Immunoassay
techniques have the advantages of being quicker,
less expensive and on site adaptable.
This paper describes an Immunoassay for
determining chlordane in soil which can be
modified slightly for field applications. The
method Is an excellent screening procedure and
shows good agreement with gas chromatographic
analyses. Because of the cross reactivity of the
chlordane antibody, this immunoassay test has
applications for other cyclodlene Insecticides.
MATERIALS AND METHODS
Preparation of Pesticide Standards
Chlordane and all other cyclodiene pesticides
were obtained from the EPA. Stock solutions of
all pesticides were prepared by accurately
weighing approximately 10 mg of each into 100 ml
volumetric flasks and bringing to volume with
methanol. Intermediate standard solutions were
obtained by pipetting 0.125 ml endrin, endosulfan
and endrin ketone and 0.25 ml of aldrln,
dleldrln, chlordane (mixture of alpha 5.9 mg and
gamma 4.2 mg/100 ml), heptachlor, heptachlor
epoxide and chlordene into 5 ml volumetric flasks
and using methanol to bring to volume. Actual
working standards were prepared by removing 1,
2.5, 5, 10, 20, 40, 80, and 160 ul allquots from
each of the Intermediate standard solutions and
adding the allquots to separate 1 ml volumetric
flasks. These standards were brought to volume
with 0.75 ml of 0.067M phosphate buffer pH 7.2
433
-------
containing 0.25? Tween 80 and 25? methanol.
Working standards were used to test linearity,
cross-reactivity and to make standard curves.
Preparation of Antibody
Chlordane antlserum was prepared by dertvatlzlng
chlordane at carbon atom 2 and covalently
conjugating It to bovine gamma globulin through a
modified carbodllmlde crosslInking procedure.
Final molar ratio of hapten to globulin was 30:1.
Antlserum was prepared In rabbits by multiple
sub-cutaneous Injections over several months.
Blood was collected from the rabbits on a monthly
schedule and the serum separated and stored
frozen at -10 C.
Preparation of Tubes and Test Kits
Antibodies to chlordane were coated to the walls
of polystyrene test tubes by a proprietary method
developed by ImmunoSystems. Shelf-life of the
dried and stabilized antibody-coated tubes was
greater than one year. Horse-radish peroxldase
was covalently bound to chlordane ("the Enzyme
conjugate") also by a modified carbollmlde
conjugation technique and Is stable In liquid for
over 1 year at 4 C. The substrate and chromogen
were stabilized buffer preparations of hydrogen
peroxide and tetramethyIbenzldlne (TMB),
respect!vely.
Preparation of Sol I
Soil samples for chlordane extraction were first
prepared by air drying a sample on Whatman 3MM
chromatography paper overnight at room
temperature; followed by sieving through a # 10
sieve. The mixed soil was placed In a clean pint
mason Jar for storage. Sample extraction was
performed one of two ways depending upon which
analysis was to be used.
Extraction of Chlordane from Soil for Immunoassay
For Immunoassay determination 10 g of soil was
weighed Into a 125 ml Erlenmeyer flask to which
50 ml of (90 + 10) methanol:water was added.
This mixture was sonicated for 10 mln. A 200 ul
aliquot was removed and placed In a 7 ml
scintillation vial containing 600 ul of phosphate
buffer/Tween 80 solution. Recovery was 90? or
better.
Extraction of Chlordane from Soil for GC
Chlordane extraction In soil for gas
chromatography (GC) was performed by following
the procedure of Saha, 1971 (8). This method
could also be used for the Immunoassay technique
except that a 1 ml aliquot of the acetone:hexane
extract must be evaporated to dryness under
nitrogen and replaced with 1 ml of methanol.
Analysis of Chlordane by Immunoassay
Soil analysis of chlordane by Immunoassay was
done by adding 100 ul of the sample from the 7 ml
scintillation vial to one of the antibody coated
test tubes followed by 160 ul of chlordane
"enzyme conjugate". The mixture was allowed to
Incubate 5 mln at room temperature before rinsing
(4 times) the unreacted mixture away with water.
Substrate (160 ul hydrogen peroxide) was added,
followed by 160 ul of chromogen (TMB). After 3
mln of Incubation the reaction was stopped with 1
drop of 2.5N suIfuric acid. The amount of yellow
color was measured by reading the difference In
optical density (ADD) between the control and
each sample at 450 nm with a hand-held battery-
powered differential photometer from Artel, Inc.
As many as 6 samples plus a control can be run
simultaneously without losing accuracy. A
control sample (no chlordane present) must be run
with each set of tubes since Its OD value Is used
to measure the AOD/OD values (where ADD Is the
difference In optical density of the samples or
standards from the control divided by the optical
density of the control read against water) of the
standards and samples.
Analysis of Chlordane by GC
Gas chromatographlc conditions were as follows:
gas chromatograph, Varlan 3700; column, 1/4" x
6', 2 mm I.D. Pyrex glass; liquid phase, 4? SE-
30/6? SP-2401; solid phase, 100/120 Supelcoport;
oven temperature, 185 C; Injector temperature,
270 C; detector temperature, 350; detector type,
electron capture (NI63); carrier gas, nitrogen;
flow rate, 30 ml/mln; Injection volumes, 1-3 ul;
Integrator, Spectra Physics 4100.
RESULTS AND DISCUSSION
The Immunoassay shows a linear relationship
(Figure 1) from 25 to 800 ng/ml (2.5 to 80
ng/tube) which was observed between the logarithm
of the chlordane concentration and the AOD/OD at
450 nm. For samples containing greater than 800
ng/ml a dilution must be made. A ADD of 0.88 or
greater Indicates a sample concentration of more
than 800 ng/ml which means the sample should be
diluted and the analysis repeated.
ReproduclblIIty results of the chlordane
Immunoassay can be seen In Tables I and II.
Table I shows the consistency data obtained from
analyzing standards, different days and over a
period of 6 months. Percent coefficients of
variation (? CV) range from 17.1 to 5.4, which
are excellent for a residue method, but are even
greater considering the Immunoassay Is a
screening technique. As one would expect the %
CV are lower as the optical density values
Increase. Although the standards are very
reproducible over a long time, It Is still
recommended for quantification purposes that a 3
point standard curve be performed each day at the
5, 40 and 80 ng chlordane/tube level.
Table II gives results of a reproduclblIIty study
on actual chlordane soil samples. Samples range
from 0.8 to 897 ppm which Is In the normal level
of chlordane In soil. It can go as high as 3000
ppm. Like Table I, there was one analysis
performed per day, but only for a two week
period. The ? CV are good with the exception of
434
-------
the 0.8 ppm sample which was 42.9?. However when
one considers that all samples are diluted
Initially 1:4, this makes a sample containing 0.8
ppm near the lower detection limit where higher %
CV would be expected. It should be possible to
el Imate the dilution step on samples less than 1
ppm and thus obtain more consistency. In general
the % CV In Table II are slightly higher than the
standards (Table I) but both studies Indicate
that the clordane Immunoassay Is very
reproducible from day to day.
Comparisons were made between the GC and
Immunoassay methods for chlordane determination
In soil (Table III). Seven soil samples were
analyzed by both techniques and only for one
sample (soil #6) was the agreement between values
off by a lot. However, since the Immunoassay
test would be used as a screening tool all
agreements between the seven samples are
acceptable. The only criterion that a screening
method needs to meet to be successful Is to be
able to determine the presence or nonpresence of
the test compound and be within a certain range
of magnitude. Once this Is determined then a
classical analytical method would be used to
determine the exact amount. The chlordane
Immunoassay meets that criterion.
Perfect correlation between the two chlordane
methods Is not likely to exist since chlordane Is
applied as the technical material which contains
many different cyclodlenes that have different
cross-reactlvltlves. Furthermore, It Is possible
to have compounds In some soil samples that may
Interfere nonspeclfleally with the antibody
antigen reactions causing problems.
As mentioned earller, the Immunoassay has cross-
reactivity with other cyclodlenes (Table IV).
The reactivity of the antibody seems to be
dependent upon the spatial orientation of the
hexa chlorine portion of the cyclodlenes since
kepone, ml rex, gamma chlordene and alpha
chlordene do not cross-react while the other
dlenes do. Of the cyclodlenes that cross-react
endosulfan, endrln and endrln ketone are the most
sensitive. Other noncyclodlene organochlorlne
pesticides were also tried (like DDT and Lindane)
but showed no cross-reactivity.
CONCLUSION
The enzyme Immunoassay offers an excel lent
screening method for chlordane In soil that Is
reproducible, quick, Inexpensive and field
adaptable. Furthermore, because of the cross-
reactivity of the chlordane antibody, the
Immunoassay has the potential to be a broad
spectrum cyclodlene test. Although only soil was
analyzed. It should be possible to modify this
assay so that other matrices such as food, air
and water could be analyzed for cyclodlenes.
REFERENCES
(1) Fenske, Richard A. and Sternback, Todd,
"Indoor Air Levels of Chlordane In
Residences In New Jersey", Bull. Environ.
Contam. Toxlcol., Vol. 39, 1987, pp.903-910.
(2) Livingston, J.M. and Jones, C.R., "Living
Area Contamination by Chlordane Used for
Termite Treatment", Bull. Environ. Contam.
Toxlcol., Vol. 27, 1981, pp. 406-411.
(3) Wright, C.G. and Lelby R.B., "Chlordane and
Heptachlor In the Ambient Air of Houses
Treated for Termites", Bull. Environ.
Contam. Toxlcol., VoJ. 28, 1982, pp. 617-
623.
(4) Louis, J.B. and Klsselbach, Jr., K.C.,
"Indoor Air Levels of Chlordane and
Heptachlor Following Termltlclde
Applications", Bull. Environ. Contam.
Toxlcol., Vol. 39, 1987, pp.911-918.
(5) Brooks, G.T., "Action of Chlorinated
Insecticides", Chlorinated Insecticides Vol.
II Biological and Environmental Aspects, CRC
Press, Inc., Cleveland, Ohio, 1974, pp. 67-
68.
(6) Taguchi, S. and YakushlJI, T., "Influence of
Termite Treatment In the Home on the
Chlordane Concentration in Human Milk'1,
Arch. Environ. Contam. Toxlcol., Vol. 17,
1988, pp. 65-71.
(7) Brooks, G.T., "Insecticides of the Dlene-
Organochlorlne Group", Chlorinated
Insecticides Vol. I Technology and
Application, CRC Press, Inc., Cleveland,
Ohio, 1974, pp. 155-158.
(8) Saha, J.G., "Comparison of Several Methods
for Extracting Chlordane Residues from
Soil", JAOAC, Vol. 54, 1971, pp. 170-174.
(9) Newsome, W.H., "Potential and Advantages of
Immunochemlcal Methods for Analysis of
Foods", JAOAC, Vol 69, 1986, pp. 919-923.
ACKNOWLEDGEMENTS
We thank the Maine Experiment Station for their
support. This paper Is #1315 of the Maine
Agricultural Experiment Station.
435
-------
1.0 •
0.8-
0.6-
0.4-
0.2-
0.0 •
1 0
100
ng/ml
1000
Figure 1- Typical Standard Curve Chlordane in Buffer
Table I- Reproducibility of the Chlordane Immunoassay-Standards in
Buffer
Number of Standard
Samples Analyzed
% Coefficient of
Variation
ng/ml
50 100 200 400 800
13 47 47 47 47
17.1 16.0 13.8 9.9 5.4
Table II- Reproducibility of the Chlordane Immunoassay on Actual
Soil Samples Containing Chlordane
Number of
Samples
analyzed
Amount of Chlordane in ppm
0.8 3.8 6.9 15.4 81.6 118.3 260.0 630.0 897.0
10
8
8
8
% Coeffient 42.9 17.3 21.4 18.8 16.7 12.1 11.8 16.9 11.1
of Variation
436
-------
Table III- Comparison of the Immunoassay and GC Methods
for Determination of Chlordane Residues in Actual Soil
Samples
-ppm Chlordane Found-
Sample Immunoassay GC
Soil-1
Soil-2
Soil-3
Soil-4
Soil-5
Soil-6
Soil-7
37 4
23.5
12.1
23.4
10.8
573.0
1523.0
50.0
20.9
17.0
20.9
16.6
989.7
1600.0
Table IV- Summary of Chlordane Cross-Reactivity Data
--ng/ml
Compound Lower Limit of Detection
Dieldrin 25
Aldrin 25
Heptachlor 25
Heptachlor Epoxide 25
Endrin 10
Endosulfan 10
Endrin Ketone 1 0
Chlordane 25
Chlordene 15
437
-------
DEVELOPMENT OF A PROTOCOL FOR THE ASSESSMENT OF
GAS CHROMATOGRAPHIC FIELD SCREENING METHODS
M. T. Homsher, V A. Ecker, M. H. Bartling, L. D. Woods,
and R. A. Olivero, Lockheed Engineering & Sciences Company,
Las Vegas, Nevada 89119
D. W. Bottrell and J. D. Petty.
United States Environmental Protection Agency,
Environmental Monitoring Systems Laboratory- Las Vegas, Nevada 89114
ACKNOWLEDGMENT
We gratefully acknowledge F C. Garner, M. A. Stapanian, D. Eastwood,
P Wylie, S. Levine, and J. Y. Aoyama for their help on this paper.
ABSTRACT
Advanced field monitoring methods are designed to meet the expanding need
for rapid, low-cost field measurements while maintaining data quality that
is characterized and adequate to support decision making. The costs and
time necessary to acquire environmental data reflect the total of the in-
dividual components that include sampling, preparation, analysis, quality
assurance, and documentation. The goal of field activities is to minimize
or combine functions in order to significantly decrease the time necessary
for analytical determination, reporting, and data validation. Specific
field techniques that are currently available, such as field gas chromato-
graphy, may exhibit data quality characteristics that represent an improve-
ment in information content and integrity over traditional laboratory
analytical measurements. The objective of all environmental measurement
systems, field and laboratory, is to meet the data quality objectives
necessary to support decisions. All analytical systems can be evaluated and
compared to identify specific data quality characteristics that determine the
applicability and sufficiency for a particular analytical problem based on
sensitivity, precision, and accuracy, provided that representativeness, and
completeness are equal.
The purposes of this study are to develop a standard procedure for the
evaluation of field instrumentation and to apply this procedure to six gas
chromatographic systems. In addition, the project will include the investi-
gation of field quality control method components and minimum data documenta-
tion. Three-dimensional graphic presentations of the data generated to
support this effort have been developed and are demonstrated in this paper.
These automated displays allow rapid examination, evaluation, and comparison of
data quality characteristics (precision, accuracy, sensitivity, etc.). This
procedure is consistent with and supports the USEPA's Data Quality Objective
concept that is required for field measurement projects.
BACKGROUND
The Superfund Amendments and Reauthorization Act of 1986 (SARA) is expected
to expand the number of National Priority List Sites from the initial 400 to
439
-------
1600 [1] The sampling and analytical capacities of the national analytical
laboratories cannot meet this increasing demand for services through the ap-
plication of currently available Contract Laboratory Program (CLP) resources.
In anticipation of this problem, the Analytical Operations Branch (AOB) of the
Office of Solid Waste and Emergency Response initiated activities to expand
the Program'": analytical capabilities. One aspect of these activities in-
volves the investigation of rapid screening methods and field analyses.
Procedures utilizing field portable gas chromatographic instruments are gaining
widespread acceptance. Regional and contract personnel routinely use field
equipment for rapid, on-site investigations [2] AOB has established work
groups designed to standardize and promote the use of field methods. These
groups are involved with evaluating both instrumentation and analytical pro-
cedures. The purpose of this document is to describe the activities in pro-
gress to assess the status of currently available gas chromatographic field
screening instruments and techniques appropriate for the analysis of volatile
organics in soil.
INTRODUCTION
In the CLP, soil samples have traditionally been sent from the sampling site
to remote laboratories for purge and trap Gas Chromatograph/Mass Spectrometry
analysis. In response to the increasing number of sample analyses required,
the AOB is currently integrating mobile laboratories and field screening tech-
niques into their analytical repertoire [3]
Field techniques, especially those associated with the analysis of volatile
organics, require individual sample/matrix considerations, but utilize common
analytical systems. Field analyses have the inherent advantage of providing
immediate information relevant to making a specific decision. Results gener-
ated from on-site measurements may, however, be subject to a decrease in
certainty and defensibility due to variations in Quality Control measures or
the use of procedures which have yet to be validated. For many field situa-
tions, a rapid response to a problem at a previously characterized environ-
mental site may be more appropriate than investing the increased time and
money required to obtain a high degree of certainty in an individual measure-
ment .
As a result of the Regional EPA need to address this situation, the investi-
gation of rapid, field screening methods for volatile organics was initiated.
To provide support as quickly as possible, the first phase of the project was
restricted to the evaluation of gas chromatographic (GC) systems. This report
is likewise restricted to investigative efforts for evaluating field GCs in
the analysis of volatile organics.
GUIDANCE FOR FIELD EVALUATION
The objective of this study is the preliminary evaluation of field tech-
niques and field instrumentation. A draft evaluation plan has been outlined
that is similar and complementary to parallel projects conducted at the
Environmental Monitoring Systems Laboratory- Research Triangle Park (EMSL-RTP)
and the Environmental Response Team, Edison, NJ [4], [5]. These studies will
provide the information necessary to utilize current instrumentation to gen-
erate and document data which are acceptable for specific, defined purposes.
The field investigation will identify gaps in current technologies that pre-
vent maximum quality and efficiency of field technologies for on-site field
440
-------
measurements. Manufacturers are active participants in current -evaluation
efforts. Industry competition will undoubtedly result in the incorporation
of instrumental modifications suggested in the course of the field evaluations.
RELATED ACTIVITIES
This study is part of a program which involves multiple related projects,
groups, and report requirements. The overall goal is to address many of the
problematic sampling and analytical issues associated with the measurement and
related confidence interval of volatile analytes. The related studies will be
coordinated by a Work Assignment Manager to increase the impact of the
individual projects and minimize redundant research efforts.
To date, soil sampling procedures and the effects of alternative preparation
techniques have not been extensively investigated. Environmental Monitoring
Systems Laboratory- Las Vegas is responsible for the development of a guidance
document that will address the areas of sampling techniques and preparation
options appropriate for the identification of volatile organics in soil. The
resulting document is scheduled for completion in December 1989 Study design
issues have been undertaken by the Environmental Research Center of the Uni-
versity of Nevada Las Vegas. Implementation of the study and laboratory sup-
port functions will be performed by Lockheed Engineering & Sciences Company
(LESC) These activities fit the time frame of the EMSL-LV Field GC Evalua-
tion Study. As a result, soil sampling/preparation options will be evaluated
simultaneously with the evaluations of field instrumental techniques and
systems. Both sampling and field analytical assessments will be referenced to
standard CLP analyses for volatile organics [6] . Considerations basic to this
study are the issues of sample storage and preanalytical holding times. These
issues have been addressed at Oak Ridge National Laboratory [7]. Local pro-
jects will incorporate the findings of this study to maximize long-term rele-
vance and data quality These studies will provide data for correlation that
have not been available previously and will accomplish this with a common
reference set of analyses. Each study, if done separately, would require the
same degree of reference CLP support. This approach is cost effective, as
well as technically valuable.
EXPERIMENTAL DESIGN, FOR
GAS CHROMATOGRAPHIC FIELD SCREENING METHODS
Replicate analyses of collocated soil samples are a basic concept in the plan
to evaluate four field gas chromatographs (GC). Collocated samples are col
lected so that they are equally representative of a given point in space and
time. At each of two sites, soil samples will be taken at 10 to 20 different
locations. The homogenized sample from each location will be split into a
number of subsamples sufficient to provide three aliquots of soil for each
field and laboratory-based instrument. Each of the three subsamples will
then be analyzed in triplicate, providing nine replicate analyses per in-
strument. The data provided by the replicate analyses will be used in the
assessment of the field GCs and their evaluation relative to laboratory-based
GCs and a gas chromatograph/mass spectrometer (GC/MS) To the extent poss-
ible, the same type of chromatography columns will be used in all the instru-
ments. While the ideal experimental design would use completely identical
columns, detectors, integrators and operating conditions, this is not
possible with the equipment currently available. Table I describes the
instrumentation to be utilized in this study.
441
-------
Prior to mobilization, optimum operating conditions, sensitivity, detector
stability and linear range are determined for each field GC. During this
bench testing period and on the first day of field activities at each site,
five-point calibration curves are created. The ratio of response to con-
centration (response factor or RF) will be calculated for each target analyte.
The Percent Relative Standard Deviation (% RSD) of the RFs from the three
middle standard levels must be <- 10% for each analyte. The % RSD of the RFs
for all five standards must be £. 25% for each analyte.
RF = mass of analyte/peak area
% RSD = RF Standard Deviation x. 100
mean RF
The calibration curve will be verified each working day by the measurement of
three calibration standards at high, medium, and low levels. The RF of each
analyte in the continuing calibration standards must be ^ 25 percent Relative
Percent Difference (RPD) with respect to the initial calibration.
% RPD = IRFi _ RFcl x 100
RFi
RFi = Mean RF from initial calibration
RFc — RF from continuing calibration
In order to provide equivalence, all GC analyses will be performed using the
same lots of single component USEPA Quality Assurance Materials Bank (QAMB)
Reference Standards. Prior to the initiation of bench and field testing,
these standard lots will be analyzed on GC/MS using QAMB standard mixes to
demonstrate qualitative and quantitative comparability of the two standard
types.
As a means of tracking and documenting the ongoing sensitivity, stability, and
precision and accuracy of instrumental measurements, Performance Evaluation
Material (PEM) will be prepared using uncontaminated soil native to each of
the respective sites. Figure 1 shows the PEM preparation scheme. The com-
pounds to be spiked onto the matrix are dependent upon the particular analytes
present at each site. These analytes and their approximate concentrations at
each location on the site have been determined in preliminary studies. PEM
analyses, at a frequency of one per day and sample location by GC/MS, and three
per day and sample location by each GC, will be performed concurrently with
each day's sample analyses.
The four field instruments will be evaluated in three different scenarios:
1. Bench testing in the laboratory using standards and performance
evaluation material.
2. Field testing using standards, PEM, and soil samples from a local
desert site .
3. Comparison testing using standards, PEM and soil samples from a
remote, hazardous waste site [8]. Data from this additional site
is being included since local desert conditions are not represen-
442
-------
tative of the majority of hazardous waste sites.
At each phase of the testing, replicate samples and PEM will also be analyzed
on the instruments described in Table I.
Comparison of the results of all these analyses establishes the precision,
accuracy, and relative sensitivity of the field instrumentation under the
described conditions, and the frequency of false positive and false negative
rates. Documentation of the analytical conditions will establish guidance
for the use of these instruments under field conditions [9].
Figure 2 is a description of a manual headspace [10] technique using an
ambient temperature water bath. The use of various aqueous salts to decrease
the solubility of analytes in the liquid phase, thus increasing vapor
phase concentration may be incorporated into the scheme [11] A variety of
injection techniques [12] may be investigated in an effort to increase the
sensitivity of this method while decreasing the variability normally associ
ated with manual headspace injection techniques [13]
Figure 3 is a description of an automated headspace technique using a Hewlett
Packard 19395A system coupled to a Varian 3400 gas chromatograph with a photo-
ionization and electron capture detector in series [14] This type of analysis
may represent the future quick-turnaround methods currently under development
[15]
Figure 4 is a description of the purge and trap gas chromatograph/mass spectro-
meter technique modified for the use of a DB-624 0.53mm internal diameter 30
meter column [16], [17] This reference analysis provides the ability to
detect false negatives and false positives [18] which are more probable with
the gas chromatograph- only systems, and also provides articulation for compar-
ison with laboratory-based methods.
DATA REPORTING
A series of data reporting forms has been devised to facilitate the accurate,
complete and timely reporting of the data [19] These forms will be field
tested during the second and third phases of the assessment. Eventually,
the forms may be suitable, after modification during actual use, for the
real time transmission [20] of validated analytical results [21], associated
quality assurance and quality control data from the laboratory to the field
and to appropriate sample control centers.
DATA EVALUATION/DATA QUALITY VECTOR CONCEPT
The purpose of this section of the paper is to demonstrate that quantitative
data comparability is possible and may permit rapid, cost effective evaluation
of data from several sources at high throughputs while maintaining/improving
data quality requirements and identifying data aspects requiring further in-
ves tigation.
Since each phase of the study (bench and two field locations) will feature
analysis of soil samples by three techniques using the same lot of standards,
identical performance evaluation samples, and triplicate analyses of collo-
cated samples, sufficient data will be available to effectively compare the
three techniques [22]. Review of the literature and the USEPA policies for
443
-------
collecting environmental data center around data quality objectives (DQOs),
the precision, accuracy, representativeness, completeness, and comparability
(qualitative only) characteristics of the data acquired [23]
In this discussion, the following terms will be used with the understanding
that variation in usage from organization to organization does exist. The
variability in the definition and use of these terms is a policy issue be-
yond the scope of this paper.
Accuracy The degree of agreement of a measured value with the
true or expected value of the quantity of concern.
Bias A systematic error inherent in a method or caused by
sample, some artifact or idiosyncrasy of the measurement
system or matrix. Temperature effects and extraction in-
efficiencies are examples of the first kind. Blank conta-
mination, mechanical losses, and calibration errors are
examples of the latter kinds. Bias may be both positive or
negative, and several kinds can exist concurrently so that
net bias is all that can be evaluated, except under special
condi tions.
To date no official universally accepted detection limit definition has been
available nor required by USEPA [24]
Precision The degree of mutual agreement characteristic of independent
measurements as the result of repeated application of the
process under specified conditions.
Limit of De tec tion (LOP) The smallest concentration or amount of
some component of interest that can be measured by a single
measurement with a stated level of confidence. This is
decided by continually diluting a standard solution until
no response significantly above noise is noted.
Limit of Quantitation (LOQ) The lower limit of concentration or
amount of substance that must be present before a method is
considered to provide quantitative results. By convention,
LOQ = 10 So where So is the estimate of the standard devia-
tion at the lowest level of measurement.
Limit of Linearity (LOL) The upper level of reliable measurement
for all practical purposes. Thus, the useful range of the
methodology is that range of concentrations between the LOQ
and the LOL.
Sens itivity The ability of the instrument to discriminate between
samples having differing concentrations or containing dif-
fering amounts of an analyte. This is evaluated by deter-
mining the smallest change in concentration which will give
a significantly different assay response.
Completeness A measure of the amount of valid data obtained from
a measurement system compared to the amount that was expected
to be obtained.
Representativeness The degree to which data accurately and pre-
cisely represent a characteristic of a population, parameter
444
-------
variations at a sampling point, or an environmental condi-
tion.
Comparability A measure of the confidence with which one data set
can be compared to another [25].
As illustrated with the analysis of this data, it is proposed that an estimate
of the relative sensitivity of data sets and possibly each sample be used on
an analyte-by-analyte basis to provide a measurable data quality vector. The
data quality vector will be comprised of an estimate of sensitivity based on
the ratio of sample - to -standard area for the same compound, precision expressed
as percent relative standard deviation, and percent bias. The frequency of
false negative and false positive rates will be examined as well.
Comparison of the three methods used: (1) Manual headspace GC, (2) automated
headspace GC, and (3) purge and trap GC/MS will be performed on a compound-
by-compound basis in a three-dimensional plot where percent bias is displayed
on the x-axis, the ratio of sample area to standard area is displayed on
the y-axis, and the percent relative standard deviation (%RSD) is displayed on
the z-axis. The influence on the sample- to -standard area ratio will result
in an apparent higher concentration standard use than in the absence of an
interference from the matrix. On an individual sample basis, the matrix
effect could be readily indicated in the three-dimensional plot. The use of
this plot will facilitate the rapid visual examination and evaluation of the
data against data quality objectives for precision, bias, and the estimate of
sensitivity. The location of each point (or a set of points) in this three-
dimensional space can be described by the Euclidean distance and direction to
a point (or the centroid of a set of points) from the origin or by an appro-
priate vector notation.
The precision as percent RSD and the accuracy are required attributes of data
quality. The ratio of sample - to -standard area provides a quick reference for
each sample and analytical method that is unitless but representative of the
actual performance of the individual equipment employed in the measurement
process. The data quality vector concept can be used to compare
various sampling methodologies; an objective of this study as well. The con-
cept of a. data quality vector will facilitate the evaluation of sample data
against data quality objectives which in turn can reduce the data evaluation
and data validation time requirements necessary to support environmental
decis ions.
Scaling factors may be used to represent an equal or variable contribution for
each axis to equitably evaluate the data quality vector for all samples. This
scaling is suggested since changes in the precision and accuracy with proxim-
ity to the detection limit are expected, as documented in the literature [26]
All quantitation will be performed above the quantitation limit from calibra-
tion information for each respective, collocated set of samples for each in-
strument and method. The use of data only in the region of quantitation [27],
may provide an additional feature to increase the timeliness of the data
reporting while increasing the information content of the analysis. For the
purposes of this study, values below the quantitation limit may be reported as
less than the quantitation limit and be excluded from the comparison section
of this study. For research purposes, when it is desirable to obtain informa-
tion near the detection limit, the reporting of estimated values and the area
ratio of sample area to the standard area, together with the precision, will
be helpful in increasing study information. Conversely, values above the
limit of linearity can be reported as greater than values and again be ex-
445
-------
eluded from the comparison, while being scheduled for reanalysis after dilu-
tion or resampling, as appropriate. This approach could provide timely infor-
mation for remediation purposes as well as limiting the field evaluation to
only quantitative values. The exception to this approach would be the false
negatives and false positive values. A record of false negatives and false
positives will be maintained for potential correlation on an analyte-by - ana -
lyte and method-by-method basis. The information on false negatives and false
positives will be derived from extensive use of known performance evaluation
materials and GC/MS analysis. A practical limitation to the assignment of
false positives is the greater sensitivity of GC/ECD in comparison to GC/MS.
A larger sample may be used for GC/MS analysis to overcome this limitation.
DATA QUALITY VECTOR EXAMPLES
Consider an example when data from the analysis of identical standards using
similar analytical techniques occupies the same space, to some known degree
of confidence, for precision, bias, and the ratio of sample -standard area. In
this situation, the representativeness and completeness are the same for the
two analytical techniques since a common lot of standards is in use and com-
pleteness is monitored. If these consistent conditions can be provided, then
a quantitative comparability statement can be made about the analysis of the
same standard by the two analytical techniques.
In practice, this type of approach is used daily in the analysis of environ-
mental samples with the same types of instrument and the same method to
provide consistent data of known quality. If this process of comparison
of the same standard analyses on different instruments/methods shows promise,
then a quantitative comparability statement is possible not only with stand-
ards, but also with homogeneous soil performance materials of the same matrix.
The use of the data quality vector plot in these examples based on standard
analysis also includes, for illustration purposes, strawman data quality
levels, Table 2, based on the past experience of laboratories analyzing
thousands of samples [28] These levels are plotted on Figure 5. In the
first example, Figure 5, the detection limit is displayed on the vertical
axis; the sample - to -standard ratio replaces detection limit in a later ex-
ample. The replacement provides a standard ratio unique to each analysis
and allows a uniform plot of data from sources with different definitions of
detection limit. The percent bias and relative standard deviation axes are
labeled. The three data quality levels are displayed as planes for clarity.
when in actuality each data quality level is a rectangular solid.
These three-dimensional plots are displayed by an in-house developed
computer program written in Pascal language. Interactive manipulation of the
display for rotation around any axis or projection onto a plane is possible
via keyboard commands or a mouse pointing device. These manipulations provide
ready examination of data clusters or data anomalies (outliers)
Data in Table 3 is the basis of a comparison plot, Figure 6, of capillary and
packed column volatile GC/MS analysis in a water matrix [29] Six volatile
compounds are listed in Table 3 by percent bias, percent relative standard
deviation (%RSD), and detection limit. It would have been preferable to use
soil volatile analysis data for this comparison, however, insufficient soil
standard analysis data is currently available. The multiple analyte plot of
this data in Figure 6 display the capillary column data as squares and the
packed column data as triangles. In this plot the general grouping of the
446
-------
data is observed in the vicinity of the data quality level I in Figure 6.
Closer inspection of Table 3 and Figure 6, however, reveal that in several
instances, this data does not meet the percent bias and percent RSD criteria
for data quality level I, and actually is data quality level II for these two
attributes. In addition, disparity between percent bias and percent RSD for
specific analytes in the two types of analysis becomes more apparent in
Figure 7 In this figure, a projection along the detection limit axis plots
only percent bias and percent RSD. The individual points have been labeled
sequentially according to their order in Table 3 for comparison purposes.
The capillary column data is again depicted with a square shape, while the
packed column data is depicted by a triangular shape. For this data, gen-
erally the capillary column data displays a lower % RSD since it is single
laboratory data. The most startling disparity is the difference between the
percent bias for compound number 2 (1,2 -dichloroethane) . The capillary column
value (-30%) almost exceeds the data quality level II criteria for percent
bias. A reversal of bias is also seen for compound 3 (trichloroethene) In
both instances, this observed difference is statistically significant at the
one percent (1%) significance level. Since the single laboratory data only
consists of six points, insufficient data is present for conclusion without
further verification. The three-dimensional plot does bring out the direc-
tion and magnitude of the differences in a graphic fashion for further in-
ves tigation.
In the second example, data from Table 4, the results of repetitive GC stan-
dard analysis by purge and trap and static headspace analysis at a 20 micro-
gram per liter concentration is displayed in a three-dimensional plot for a
single analyte tetrachloroethene in Figure 8. The sample - to -standard data
has been modeled from ratios reported by Wylie, et al. in this comparison, and
area ratio replaces the detection limit axis seen in the first example.
Both the percent bias and percent RSD axes are labeled in Figure 8. The purge
and trap data is depicted as a triangle and the headspace data is depicted as
a square.
The printed page shown as Figure 8 lacks the perspective and versatility of
the computer monitor; however, two distinct clusters are observed with approx-
imately equal percent bias range. The discriminating feature in the display
is the lower percent RSD for the headspace analysis
This distinction is more apparent in Figure 9, a two-dimensional projection
of Figure 8. The ability to achieve comparable area ratio and percent bias,
with superior (lower percent RSD) precision by headspace analysis was possible
by using a larger sample volume in the headspace vial, an elevated equilibrium
temperature, addition of a salt solution to the standard, and use of equal
equilibration times for each headspace vial.
In the third example, the preliminary field data in Table 5 from the analy-
sis of the same volatile soil performance evaluation material by five field
gas chromatographics and one laboratory gas chromatograph is considered.
The three-dimensional plot in Figure 10 displays percent RSD, percent bias,
and sensitivity. A two-dimensional projection of the data in Figure 11 on
the precision and bias axes, together with the Strawman data quality values,
displays the relative performance of these systems for a single analyte in
soil .
The three-dimensional plot of this data can be rotated in space by the compu-
ter program for different perspectives as well as be examined by statistical
techniques [30], principal component analysis, pattern recognition techniques
[31], and projected onto two axes for easier two-dimensional comparisons.
447
-------
With the inclusion of the overlay of the strawman data quality objectives
(Figure 10), an individual compound or element data point or a series of data
points may be examined for possible relationships simultaneously. The use of
this plot for sampling and analysis data display, together with the data qua-
lity vector concept, represents an opportunity to bring quantitative compara-
bility and powerful data analysis techniques to provide data of known quality
to the data user in a more timely and useful format.
CONCLUSIONS
Recent reported advances in the literature such as 30-second [32], and 5-
second GC analysis [33], and the use of automated and robotic systems for
volatile analysis will require advances in data reporting, validation, com-
parison and interpretation. The use of procedures such as this for rapid,
accurate data validation and interpretation may optimize the information
content of analytical results and decrease the time required to supply
validated environmental data to the data users. The procedures described
here have been applied to the evaluations of currently available field in-
strumentation, sample preparative options, and analytical techniques. This
work collectively supports the development of an integrated approach to
directly present field screening alternatives in the context of data quality
obj ectives.
NOTICE
Although research described in this article has been funded wholly by the
United States Environmental Protection Agency under contract number 68-03-
3249 to Lockheed Engineering & Sciences Company, it has not been subjected
to Agency review and therefore does not necessarily reflect the views of
the Agency, and no official endorsement should be inferred. Mention of
trade names or commercial products does not constitute Agency endorsement of
the produc t.
11-1-88
448
-------
BIBLIOGRAPHY
1. LIngle, S.; "SARA Implementation Plans," USEPA Seventh
Annual Superfund Analytical Services Conference; Las Vegas,
Nevada; July, 1987.
2. Szilagyi, A. P.; Gesalman, C. M.; and Bennet, D. A.;
"Catalog of Field Screening Methods," Proceedings of the
Third Annual USEPA Solid Was te Tes ting and Quality Assurance
Symposium: Washington, D.C.; July, 1987; pp. 8-47 to 8-53.
3. Chapman, H. and Clay, P F.; "Field Investigation Team
(FIT) Screening Methods and Mobile Laboratories
Complementary to Contract Laboratory Program (CLP)," TDD HQ-
8507-01; October 17, 1986 (Draft); pp. 7-1 to 8-3.
4. Pritchett, T. H.; "Technical Guidance for the General Scope
of Work for the Emergency Response Team (ERT) Substance
Detection Investigation (SDI) Study," Memorandum dated
March, 1987; pp. 1-4.
5. Pritchett, T. H.; "Technical Summary of the Experimental
Design for the EPA Environmental Response Team's Substance
Detection Investigation," dated April 14, 1987; pp. 1-6.
6. Homsher, M. T.; Ecker, V A.; Bartling, M. H.; Kumar, K.
S.; Wade, S. A.; and Lewis, T. E.; "Quality Assurance
Project Plan for Field Gas Chromatographs and Laboratory
Analyses Supporting Soil Sampling Documentation," June,
1988; Section 1, pp. 1-3
7 Maskarinec, M.; Johnson, L.; and Holladay S.;
"Recommendations for Holding Times of Environmental
Samples," Proceedings of the Fourth Annual USEPA Waste
Tes ting and Quality As surance Sympo s ium. Washington, D. C. ;
July, 1988; pp. H-29 H-45.
8. IBID 6; Section 7.0; pp. 3-5.
9 Homsher, M. T.; Cappo, K. A.; Malley, P. A.; Olivero, R. A.;
Woods, L. D.; Lockheed Engineering and Management Services
Co.; Bottrell, D. W.; and Petty, J D.; USEPA, Las Vegas,
Nevada; "Assessment and Evaluation of Field Screening
Techniques," presented at the Ninth Annual Rocky Mountain
ACS Meeting; Las Vegas, Nevada; March 28, 1988.
10. Spittler, T. M.; "Use of Portable Organic Vapor Detectors
for Hazardous Waste Site Investigation," National Conference
Management of Uncontrolled Hazardous Waste Sites; October
15-17, 1980; Washington, D.C.
11. Croll, B. T.; Sumner, M. E.; and Leathard, D. A.;
"Determination of Trihalomethanes in Water Using Gas Syringe
Injection of Headspace Vapours and Electron Capture Gas
Chromatography." Analyst. January, 1986; Vol. Ill; pp. 73-76
449
-------
12. IBID 3; pp. 3-10.
13. Chapman, H. G.; "Method 201 Screening for Volatile
Organics in Soil/Sediment," Field Analytical Screening
Project Standard Operating Procedures, Ecology and
Environment (Draft); June, 1987; pp. 1-11.
14. Wylie, P. L.; Comparison of Headspace with Purge and Trap
Techniques for the Analysis of Volatile Priority
Pollutants," Journal of the American Water Works. 1988,
Vol. 80, pp. 66-72 .
15. "EPA Region IV Mobile Laboratory Protocol, Screening Method
for Volatile Organic Compounds;" Revision 3; January, 1988;
USEPA Region IV; Athens, GA.
16. Markelov, M. and Seitz, B. R.; "Analyses of Water and Soils
for Trace Organic Contamination via Headspace and Purge and
Trap Techniques Using Robotics," Proceedings of the Tenth
Annual Analytical Sympos ium, USEPA Office of Water; Norfolk,
Virginia; May 13-14, 1987; pp. 249-263.
17 USEPA Contract Laboratory Program; "Statement of Work for
Organic Analysis Multimedia, Multiconcentrat ion," USEPA;
Washington, D C.; 1988 Invitation to Bid WA 87-236/237;
Exhibit D.
18. Jutliner, F ; "Quantitative Analysis of Monoterpenes and
Volatile Organic Pollution Product (VOC) in Forest Air of
the Southern Black Forest; Chemosphere : 1988; Vol. 17, No.
2; pp. 309-317
19. IBID 6; Section 8.0
20. Bogen, D. L., Robertson, G. L., Homsher, M. T., Lockheed
Engineering and Management Services Co.; Bondelid T.,
McDonald, P., Horizon Systems; and Petty, J D., USEPA, Las
Vegas, Nevada; "Real Time Quality Assurance through
Telecommunications," presented at the Ninth Annual Rocky
Mountain ACS Meeting; Las Vegas, Nevada; March 29, 1988.
21. Shumann, C. R., Olivero, R. A., Fitzgerald, K. E., and
Homsher, M. T., Lockheed Engineering and Sciences Company;
and Petty, J. D., USEPA, Las Vegas, Nevada; "Automation of
Regional Data Usability Studies," Proceedings of the Office
oj; Solid Waste Symposium. Washington, D. C.; July, 1988; pp.
E-l to E-13.
450
-------
22. IBID 9.
23. "Data Quality Objectives for Remedial Response Activities
Development Process; USEPA Office of Solid Waste and
Emergency Response; EPA/540/G- 87/003 ; March, 1987
24. Garner, F C. and Robertson, G. L.; "Evaluation of
Detection Limit Estimators," Chemometrics and Intelligent
Laboratory Systems. Vol. 3, No. 1-2; February, 1988; pp. 53-
59 .
25. Taylor, J.K.; "Quality Assurance of Chemical Measurements," Lewis
Publishers, Inc., Chelsea, Michigan; 1987.
26. Horowitz, W.; "Sampling An Eternal Problem," in the Draft
Proceedings of The Quality Assurance in Environmental
Measurements Meeting: Baltimore, Maryland; May 25-26, 1988;
pp. 1-6.
27 Keith, L. H.; Crummett, W.; Deegan, J ; Libby, R. A.; Taylor, J. K.;
and Wenter, G.; "Principles of Environmental Analysis," Analytical
Chemistry Vol. 55; 1983; pp. 712A-724A.
28. Flotard, R. D.; Homsher, M. T.; Wolff, J. S.; and Moore, J M.;
"Volatile Organic Analytical Methods Performance and Quality Control
Considerations;" Quality Control in Remedial Site Investigations:
Perket, C.L., ed., American Society for Testing and Material, Standard
Technical Publication 925, pp. 194-195.
29. Clark, R. R.; Zalikowski, J. A.; "Comparison of Capillary Column and
Packed Column Analysis for Volatile Organics;" Proceedings of the
Third Annual USEPA Solid Was te Tes ting and Quality As surance Sympo s ium:
Washington, D.C.; July, 1987; pp. 6-137 to 6-150.
30. Youden, W. J.; "Statistical Techniques for Collaborative Tests;"
Association of Official Analytical Chemist. Washington, D.C.; 1978;
pp. 36-41.
31. Wold, S. and Sjostrom; "SIMCA: A Method for Analyzing Chemical Data
in Terms of Similarity and Analogy," Chemometrics: Theory and Appli-
cation: B. R. Kowalski, Ed.; ACS; San Francisco, CA; 1976; pp. 243-
282 .
32. Overton, E. B.; Sherman, R. W.; Collard, E. S.; Kinkhacorn, P; and
Dharmasena, H. P.; "Current Instrumentation for Field Analysis of
Organics;" Proceedings of the Pittsburgh Conference: February 24, 1988;
Abstract 618.
33 Lanning, L.; Sacks, R.; Levine, S. P ; and Mouradan, R.; "Improved
Instrumentation for High Speed Gas Chromatography," Analytical
Chemistry (in press).
451
-------
PERFORMANCE EVALUATION MATERIAL
PREPARATION SCHEME
Performance Evaluation Materials (PEMs) being used in the study are site-
specific. Each matrix is native to the sampling site, and each spiking solu-
tion contains analytes determined, through CLP analyses, to be present at the
site.
MATRIX PREPARATION
o Soil Source Determination
Uncontaminated location on sampling site
Physical characteristics similar to contaminated soil
o Soil Homogenation
o Soil Aliquotting According to Analytical Method
INST/METHOD SAMPLE CONTAINER SAMPLE WT.
GC/Manual Headspace 40 mL VGA Vial 12 g
GC/Automated HS 20 mL Crimp-top Vial 10 g
GC/MS 40 mL VGA Vial 5g
MATRIX SPIKING
o Spiking Levels 25 ng/g and I25 ng/g
Analyte concentrations chosen to be above the practical quantitation limit and within
the linear range of all instrumentation.
o Spiking Compounds
Site-specific spiking solutions include various combinations of the following analytes:
Benzene Chlorobenzene
Toluene Trans-1,2-Dichloroethylene
Ethylbenzene 1,1,1 -Trichloroethane
P&M-Xylenes (co-elute) Trichloroethylene
O-Xylene Tetrachloroethylene
Chloroform 1,1,2,2-Tetrachloroethane
PEM STORAGE AND SHIPMENT
Immediately after spiking, each container is tightly capped and replaced in the original
shipping boxes. To reduce analyte loss and/or contamination during shipment, the boxes are:
o Heatsealed in Scotchpak bags
o Stored at 4 degrees C until shipment
o Shipped and stored on site in clean ice chests containing freeze gel packs
FIGURE 1
452
-------
SAMPLE PREPARATION (Triplicate Subsamples for Each Instrument)
Sample Size: 12 g 0.5 g
Sample Container: 40 ml screw-top VOA vial with PTFE septum
Surrogate/RT Standard Addition: 10 ml of carbon-free water containing a non-
target compound RT standard at a concentration equivalent
to 50 ug/Kg is injected through septum. (One of several
possible aqueous salt solutions may be substituted for
chlorine-free water.)
Sample Equilibration: Sample vial is shaken for 15 seconds, placed in a
water bath at ambient temperature for 15 minutes and
shaken for 15 seconds prior to each injection.
SAMPLE ANALYSIS
Sample Injection: 500 uL of sample headspace is withdrawn through the
septum with a gas-tight syringe and injected into the GC.
Instrument Parameters: Varian 3400, Column - J&W 30 m x .53 mm DB-624
FSCC. Carrier gas flow - He at 8 mL/min
Make-up gas flow - N2 at 40 mL/min
Injector Temperature: 130 degrees C
HNU Model PI-5202 PID
PID Temperature: 200 degrees C
PID Lamp Intensity: 45% full scale
ECD Temperature: 200 degrees C
GC Temperature Program: Initial Temp. - 60 degrees C
Initial Time - 5 min.
Program Rate -10 degrees C/min
Final Temp. 130 degrees C
Final Time -3 min.
Parameters for Field GCs will be determined during bench testing.
DATA REDUCTION
Integrators: Hewlett-Packard 3396A
Data Processing: Zenith Portable Computer Model ZFL-181 -93 formatted
with Lotus Spreadsheet.
FIGURE 2
MANUAL HEADSPACE SAMPLE ANALYTICAL SEQUENCE
453
-------
SAMPLE PREPARATION (Triplicate Subsamples for Each Instrument)
Sample Size: 10 g 0.5 g
Sample Container: 22 ml headspace vial, PTFE-faced silicone rubber septum,
aluminum crimp cap
Surrogate/RT Standard Addition: 10 mL of chlorine-free water containing an RT
standard at a concentration equivalent to 50 ug/Kg is injected
through the septum. (One of several possible aqueous salt
solutions may be substituted for chlorine-free water.)
Sample Equilibration: Samples are equilibrated for 30 minutes in a 60
degree C heated carousel prior to analysis.
AMPLE ANALYSIS
HP 19395A Headspace Sampler Settings: Sampler Bath Temp. - 60 degrees C
Sample Loop Size -1 mL
Sample Loop Temp. 100 degrees C
Carrier Gas HE at 6 psi
Auxiliary Pressure 1 psi
Headspace Injection Sequence and Set Point Display: Probe Down 01
Start Pressurization 03
Stop Pressurization 13
Start Venting 14
Stop Venting 25
Start Injecting 26
Stop Injecting 227
Probe Up 228
Instrument Parameters: As for manual headspace analysis using Varian 3400
GC and J&W DB-624 column 30 m x .53 mm
GC Temperature Program: As for manual headspace analysis
DATA REDUCTION
As for manual headspace data reduction.
FIGURE 3
AUTOMATED, HEATED HEADSPACE SAMPLE ANALYTICAL SEQUENCE
454
-------
SAMPLE PREPARATION
Sample Size: 5g 0.1 g A smaller sample weight may be used if
screening analyses indicate a high level of volatile
contamination.
Sample Container: 40 mL screw-top VOA vial with PTFE septum. The final
sample aliquot is weighed into a purge device.
Surrogate/Internal Standard Addition: 5 ml of reagent water containing
Internal Standards and Surrogate Standard as indicated
in the EPA protocol [17] is added to the purge device.
SAMPLE ANALYSIS
Purge and Trap Conditions: Trap Configuration -1 cm. Supelco 3% SP-2100
on 60/80 chromosorb, 15 cm Tenax, 8 cm 35/60 silica gel
Purge Flow Rate - 30 ml_/min.
Purge Time -11 min.
Desorb Temperature - 200 degrees C (4 min.)
Instrument Parameters: Finnigan 4021-D GC/MS
J&W DB-624 FSCC 30 m x .53 mm
Carrier Gas - He at 15 mL/min.
Makeup Gas - He at 40 mL/min.
Separator Oven Temperature - 260 degrees C
Injector Temperature - 260 degrees C
Ionizer Temperature - 250 degrees C
Scan Time: 2 sec/scan
GC Temperature Program: Initial Temperature -10 degrees C
Initial Time - 3 min.
Program Rate #1 -10 to 40 degrees at 5 degrees C/min.
Program Rate #2 - 40 to 140 degrees at 10 degrees C/
min.
Final Temperature-140 degrees C
DATA REDUCTION
Data System: Finnigan INCOS Model 2400
FIGURE 4
GC/MS SAMPLE ANALYTICAL SEQUENCE
455
-------
DET. LIM.
-------
X RSD
30
<] Packed
QCapi 1 lary
0]
15
B
E
m
-30
-10
10
X BIAS
30
FIGURE 7
TWO-DIMENSIONAL DATA QUALITY PLOT (Multiple VOA Analvtes)
P & T
Headspace
AREA
RATIO
ZRSD
-15
-10
-5
10
15 ZBIAS
FIGURE 8
COMPARISON THREE-D PLOT OF PURGE 8 TRAP AND HEADSPACE QC DATA
457
-------
AREA RATIO
« P & T
o Headspacc
XRSD
10
15
20
FIGURE 9
TWO-DIMENSIONAL DATA QUALITY PLOT P S T VS. HEADSPACE GC
SENSITIVITY
0.-
RSD
-90
-50 -25
25 50
90 XBIAS
FIGURE 10
3-DIMENSIONAL PLOT OF INSTRUMENT COMPARISON DATA
458
-------
RSD
30
115
7. BIAS
-30
-10
10
3D
FIGURE 11
TWO-DIMENSIONAL DATA QUALITY PLOT - MULTIPLE INSTRUMENT COMPARISON
459
-------
TABLE I
INSTRUMENT CONFIGURATIONS
TEMP INJECT COLUMN
MANUFACTURER MODEL NO. DETECTOR(S) PROG TYPE TYPE
HNU SYSTEMS
INC.
SHIMADZU
SCIENTIFIC
INSTRUMENTS
321
MINI-2E
FIELD GAS CHROMATOGRAPHS
ECD
PHOTOVAC 10S50
INTERNATIONAL
SCIENTIFIC
REPAIR INC.
THERMO ENV.
8610
AID 511
ECD
PID
ECD
PID
NO SPLIT- CAP
LESS
NO SPLIT- CAP
LESS
NO SPLIT- *CAP
LESS
NO SPLIT- CAP
LESS
NO SPLIT- PACKED
LESS
MOBILE LAB GAS CHROMATOGRAPH
HEWLETT-PACKARD 5880 PID.ECD YES SPLIT- CAP
LESS
FIXED LABORATORY GAS CHROMATOGRAPH
VARIAN 3400 PID,ECD YES CAP
(IN SERIES)
FIXED LABORATORY GC/MS
FINNIGAN 4021D MASS SPEC YES PURGE CAP
&TRAP
GC INTEGRATORS
HEWLETT-PACKARD 3396A
ANALYTICAL COLUMNS
MANUFACTURER TYPE LENGTH I.D. FILM THICKNESS
J&W SCIENTIFIC DB-624 30m 0.53mm 3.0 micron
-PHOTOVAC CPSIL-5 10m 0.53mm 2.0 micron
460
-------
TABLE II
3-LEVEL DATA QUALITY STRAW MAN-VGA
LEVEL I
RSD < 15%
BIAS < 10%
DL < CRQL
LEVEL IT
15 < RSD < 30%
10 <|BIAS|< 30%
CRQL< DL < 3*CRQL
LEVEL Iff
30 < RSD < 50%
30 < BIAS|< 50%
3 CRQL
-------
TABLE IV
COMPARISON DATA STANDARD GC ANALYSIS BY
PURGE AND TRAP VERSUS HEADSPACE TECHNIQUES
N % Bias Precision Estimate of Sensitivity
% BSD Sample area/Standard area
Purge and Trap 25 -1 9.2
Headspace 25 -1.4 3.7
.96
1.01
TABLE V
PRELIMINARY FIELD DATA
GC/ECD 1,1,2,2-TCA RESULTS FROM
THE ANALYSIS OF PERFORMANCE EVALUATION MATERIALS
INSTRU. I.D.
INJ. VOL.
SENSITIVITY
% BIAS
% RSD
1
2
3
4
5
6
100
25
100
25
100
500
ul
ul
ul
ul
ul
ul
.221
.119
.295
.227
.513
.226
+ 14
+21
+83
-11
-16
-38
.0
.3
.2
.4
.3
.8
18.1
33.9
13.6
35.4
11.4
10.0
LESC CONTRIBUTORS
E. Neil Amick
Marilew H. Bartling
Kevin A. Cappo
John W. Curtis
Betty A. Deason
Vicki A. Ecker
Forest C. Garner
Clare L. Gerlach
Michael T. Homsher
Henry B. Kerfoot
Tim E. Lewis
William D. Munslow
Ramon A. Olivero
Eric A. Steindl
Larry D. Woods
John H. Zimmerman
462
-------
COST ANALYSIS FOR USING MOBILE LABORATORIES
VERSUS FIXED-BASE LABORATORIES
FOR SITE CHARACTERIZATION AT FUSRAP SITES
Gomes Ganapathi, Ph.D. and David G. Adler
Bechtel National, Inc.
800 Oak Ridge Turnpike
Oak Ridge, Tennessee 37830
Mark Carkhuff
Weston Analytical Laboratory
208 Welsh Pool Road
Lionville, Pennsylvania 19353
ABSTRACT
This report outlines the potential cost
savings from using mobile analytical
laboratories as compared to fixed-base
laboratories. The costs of using a
mobile laboratory for characterizing
sites for chemical contaminants are
compared to the costs of employing
fixed-base laboratories. Cost estimates
were based on discussions with
commercial analytical laboratories,
Bechtel National, Inc. experience with
site characterization activities, and
analyses conducted in-house. Results
from the comparative analysis for three
hazardous waste sites are presented.
Unlike fixed-base laboratory analyses,
costs for mobile laboratory work are
generally not calculated on a cost-
per-sample basis. Rather, costs are
determined by the number of days the
laboratory is rented for on-site use.
Although site characterization using
on-site mobile laboratories is generally
less expensive, this cost advantage can
easily be lost due to poor planning or
unexpected delays in sampling activity
while the laboratory is on site.
Strategies for coordinating sampling and
analysis activities that minimize the
time frame for employing a mobile
laboratory are presented.
1.0 INTRODUCTION
Bechtel National, Inc. (BNI) is
conducting Remedial Investigation/
Feasibility Studies (RI/FS) as part of
the Department of Energy's (DOE)
Formerly Utilized Sites Remedial Action
Program (FUSRAP). FUSRAP is a program
managed by DOE to identify and clean up
or otherwise control sites where
residual radioactive contamination
(exceeding current guidelines) remains
from activities carried out under
contract to the Manhattan Engineer
District and the Atomic Energy
Commission during the early years of the
nation's atomic energy program, or from
commercial operations causing conditions
that Congress has mandated DOE to
remedy.
This report outlines the potential cost
savings from using mobile analytical
laboratories for site chemical
characterization. The costs of using a
mobile laboratory for characterizing
sites for chemical contaminants are
compared to the cost of employing a
fixed-base laboratory. Cost estimates
are based on BNI experience with site
characterization activities, discussions
with commercial analytical laboratories,
and analyses conducted in-house.
Results from the comparative analysis
for three FUSRAP sites are presented.
2.0 OVERVIEW OF SAMPLING AND ANALYSIS
STRATEGIES FOR FUSRAP SITES
As with all waste site characterization
strategies, the objectives of sampling
and analysis tasks for the FUSRAP sites
are
o to determine the nature and
extent of contamination
(radiological characterization
being the major effort)
o to quantitatively and
qualitatively characterize the
contamination in all media of
concern
o to identify contaminant transport
pathways
o to assist in the planning of site
remedial actions
2.1 Contract Laboratory Program and
Mobile Laboratory Use
Compliance with Contract Laboratory
Program (CLP) requirements by fixed-base
laboratories ensures that all analytical
data generated are subject to rigorous
463
-------
Quality Assurance/Quality Control
(QA/QC) protocols and extensive
evidential documentation. As such, data
generated using these protocols are
"litigation quality" data. The
acceptability of mobile laboratory data
and results from other relatively new
field analysis techniques have not been
tested in courts.
Recent drafts of the Environmental
Protection Agency (EPA) RI/FS guidance
document (Ref. 1) and Data Quality
Objectives document (Ref. 2) recommend
that tasks such as risk assessment.
Potentially Responsible Parties
determination, and remediation
validation activity be supported by CLP
analytical results. An increasingly
common practice is to rely on field
techniques for most of these activities
and screening, and then to use fixed-
base laboratories to confirm field
results. EPA generally promotes these
techniques as a means of saving time and
money during site characterization
activities (Refs. 2 and 3).
The methods expected to be applicable to
FUSRAP site characterization activities
include the use of mobile laboratories
fitted with gas chromatograph-flame
ionization detectors (GC-FID) for
on-site soil-gas/liquid organics
analysis, and atomic absorption (AA)
techniques for inorganics analysis. It
is, however, important to recognize that
field analytical techniques will never
replace the role of fixed-base
laboratories.
3.0 COST ANALYSIS
This section provides estimates of the
relative cost-effectiveness of employing
a mobile laboratory versus a fixed-base
analytical laboratory for site charac-
terization purposes. The analysis used
site characterization plans for the
FUSRAP Colonie site in New York, the
St. Louis Airport Site (SLAPS) in
Missouri, and the Tonawanda site in New
York as a basis for cost comparisons.
Cost estimates were calculated for two
separate characterization options.
Under Option A, sites were characterized
using an on-site mobile laboratory.
Option A costs include the cost of
sending 10 percent of all samples (as
duplicates) to a fixed-base CLP
laboratory for confirmation. Under
Option B, all samples are analyzed by a
fixed-base CLP laboratory.
For each site, three Option A costs are
provided. These three estimates reflect
the costs that would accrue for three
different time periods (7 days, 14 days,
and 30 days) that the mobile labora-
tories would be on site. The results
from the comparative cost analysis are
provided in Table I.
Table I
MOBILE LABORATORIES VERSUS FIXED-BASE LABORATORIES: A COST COMPARISON FOR SLAPS AND THE COLONIE SITE
__ SITE
SAMPLES " — __________^
Groundwater Samples
Surface Water Samples
Soil/Sediment Samples
TOTAL SAMPLES
~~--— -_ TURN AROUND
---^TIME
OPTION ^~~-~-~~^____^
A. Mobile Lab plus confirmation
at CLP Lab (10%)
B. Fixed-Base, CLP Lab
SAVINGS DUE TO OPTION A
VERSUS OPTION B
ST. LOUIS
AIRPORT SITE
51
4
46
101
7 days
$57,100
NA
$44,900
44%
14 days
$76,000
NA
$26,000
25%
30 days
$107,000
$102,000
(-$5,000)
(-5%)
COLONIE
SITE
24
5
126
155
7 days
$28,800
NA
$38,400
57%
14 days
$44,700
NA
$22,500
33%
30 days
$51,700
$67,200
$15,500
23%
Table Notes:
1. All samples analyzed for volatiles, semivolatiles, metals, and indicator ions as required by field
sampling plans.
2. Identical instrumentation (GC/FID and AA) used by both the mobile and fixed-base laboratories.
3. Costs for the mobile laboratory include transportation, technician per diem, and all required
reagents and other support items.
4. Fixed-base laboratory costs include an assumed packaging and shipping cost of $50/sample.
464
-------
After discussing these results with
appropriate FUSRAP personnel, it was
determined that significant cost savings
are possible through the use of mobile
laboratories. However, cost savings are
tied to the length of time the mobile
laboratory must be on site. This is
generally determined by the rate at
which samples can be generated, not the
rate at which samples can be analyzed by
the mobile laboratory.
3.1 SLAPS, Missouri
Borehole sampling at the SLAPS took
almost 60 days, due to difficulties in
mobilizing drilling rigs. This was
substantially longer than anticipated.
If a mobile laboratory had been employed
for this site, it would have been under
utilized and the costs of site charac-
terization would have been much higher
than if a fixed-base laboratory were
employed.
The number of days that BNI rents a
mobile laboratory could be reduced by
beginning drilling operations a few days
before the mobile laboratory arrives on
site. However, the benefits of this
phased approach would be limited by the
allowable holding time for volatile
organics samples (7 days).
3.2 Colonie Site, New York
Figure 1 provides a proposed phased
schedule for sampling and analysis at
the Colonie site using a mobile
laboratory. With this schedule, the
mobile laboratory is never "overloaded"
with samples, the sample generation rate
is reasonably achievable, and no samples
would be held in excess of allowable
holding times. Using this approach and
schedule, a savings of approximately
33 percent should be possible for
chemical analysis, even assuming that
10 percent of all samples are duplicated
for analysis in a fixed-base laboratory.
3.3 Tonawanda Site, New York
A third cost comparison between the use
of mobile laboratories versus fixed-base
laboratories for site characterization
was analyzed for the Tonawanda site in
New York. Although radiological and
chemical contamination was suspected at
the site, no information on the location
or magnitude of contamination was
available to guide placement of the
monitoring wells.
Project geologists were presented with
the site background and asked to
estimate the number and type of sampling
locations that would be necessary to
adequately characterize the site. The
geologists predicted that the number of
ACTIVITY
MOBILIZE
DRILL-RIG
MOBILIZE
MOBILE LAB
DEMOBILIZE
DRILL-RIG
DEMOBILIZE
MOBILE LAB
COLLECT
SAMPLES
5 SK+5 SO SAMPLES
II GW SAMPLES
12 SL SAMPLES
12 SL SAMPLES
12 SL SAMPLES
12 SL SAMPLES
12 SL SAMPLES
12 SL SAMPLES
12 SL SAMPLES
12 SL SAMPLES
12 SL SAMPLES
13 SL SAMPLES
13 NEW GK "SAMPLES
DAY
1
2
3
4
5
6
1
8
9
10
II
12
13
14
15
16
17
18
ANALYZE
SAMPLES
5 S»+5 SO SAMPLES
IIGW SAMPLES
IISL SAMPLES
II SL SAMPLES
IISL SAMPLES
IISL SAMPLES
IISL SAMPLES
II SL SAMPLES
IISL SAMPLES
IISL SAMPLES
IISL SAMPLES
II SL SAMPLES
II SL SAMPLES
13 NEW GW SAMPLES
(TOTAL = 155 SAMPLES ANALYZED IN 14 DAYS) KOTEi
G» - GROJtHATER
SO - SEDIIOT
SL - SOIL
S« - SURFACE WATER
FIGURE 1
PROPOSED SAMPLING AND ANALYSIS SCHEDULE
USING A MOBILE LABORATORY
FOR COLONIE FUSRAP SITE
soil borings required under both options
would be identical, since radiological
characterization would not be assisted
by the mobile chemical analysis
laboratory. However, the availability
of quick turnaround, on-site chemical
analysis was predicted to reduce the
number of groundwater monitoring wells
required by 10 percent.
The relative costs of these two
approaches are shown in Table II. As
with the Colonie and SLAPS analyses,
significant potential cost savings are
apparent. The magnitude of the savings
is explained primarily by the cost
savings associated with analyzing a very
large number of samples (a total of 120)
on site.
465
-------
TABLE II
CHEMICAL SAMPLING AND ANALYSIS COST BREAKDOWN DETAILS FOR OPTIONS A AND B
FOR TONAWANDA, NEW YORK SITE
Task Details
Soil Gas Sampling,
Analyzing
Soil Boring, Sampling,
Packing, Shipping
Surface Water + Sediment
Sampling, Packing, Shipping
Per diem, Travel, Stay, Etc.
Analysis Samples
QC Samples
Metals
VOCs
BNAEs
IONS
Analytical Cost
TOTAL
SAVINGS USING MOBILE LAB
OPTION A
Mobile
Quantity Costs
30 Samples
15 LF/30 Samples
20 SW + 20 SD
100
20
AA
GC/FID
GC/FID
YES
Lab -t-10% CLP
For 20 Days
$ 6,100
$ 45,100
$ 2,700
$ 3,000
$ 36,400
$ 93,300
$ 80,000
OPTION B
Fixed-Base Lab
Quantity Costs
30 Samples
15 LF/30 Samples
20 SW + 20 SD
100
20
AA
GC/MS
GC/MS
YES
N/A
$ 45,900
$ 3,600
$ 2,200
$ 121,500
$ 173,200
A summary of typical analytical levels
and qualitative differences between a
fixed-base laboratory and a mobile
laboratory is presented in Table III.
This summary is in reference to analysis
only.
4.0 CONCLUSION
When properly equipped, mobile
laboratories can offer essentially the
same list of services provided by
fixed-base laboratories. On-site
TABLE III
FIXED-BASE LABORATORY VERSUS MOBILE LABORATORY - A SUMMARY OF QUALITATIVE DIFFERENCES
Factors
Fixed-Base Lab Equipped With All
CLP Instruments (GC/MS/AA/ICP)
Field Peplovable Analytical Instruments
/Mb
Type of Analysis
Resolution (ppm/ppb)
Data Quality
Duali tati ve/Quanti tative
Results
Cost of Equipment
(in a scale of 0 to 10)
Cost of Analysis
(per sample cost)
Time Frame (to get
validated analysis)
Prioritized Data Uses
Limitations
Technician/Analyst
Requ i rements
Hazardous Substance List (HSL) Volatile; and a few semivolatiles HSL organics
organics/inorganics
AA - Very high (ppl
XRF - Medium (ppm)
HSL inorganics
nics
Very high (ppb) High/need MS for finer resolution Very high (ppb)
(low ppm to high ppb)
Dependent on QA/QC steps employed/required
Quantitative Quantitative Quantitative Quantitative
10 (including support facilities) 6 87
- Stringent QA/QC
- Standard Methods
$1,000/Sample (Organics)
J200/Sample (Metals)
(because of strict QA/QC)
Contractually 30 to 40 days
confirmation
toxicology/risk assessment
all other program activities
e.g., ROD, rem-design, etc.
PRP determination
litigation
$100/Sample or $l,500-$2,000/day rental charges for 14 days, if lab is rented.
Real-time to several hours
- presence or absence of contaminants
- preliminary site characterization and screening
- engineering design - temporary alternatives
- removal action (resulting in alternate water supply, fencing, etc.)
- health and safety
tentative ID of non HSL parameters
sufficient time required for
package validation
high cost
- tentative 10 of all parameters
- techniques/instruments not very sophisticated
- loss of instrument sensitivity during mobilization
- results may not support legal issues
BS chemist/well trained
- preferably BS chemist with sufficient training
- should be capable of interpreting field data
466
-------
chemical analyses can be done on
matrices such as soil, liquid, sludge,
water, and air samples. Typically,
results from analysis can be made
available within 4 to 24 hours after
sampling. The advantages of employing
on-site mobile laboratories (over
traditional off-site, fixed-base
laboratories) include
o Potential cost savings
o Greatly reduced analytical time
o An ability to modify sampling
strategies on an almost
instantaneous basis
o A potential decrease in the number
of samples being sent to fixed-
base laboratories
o Assistance to the geological
decision-making process for
determining the location of
boreholes/monitoring wells and the
most effective depth of monitoring
wells
o Rapid determination of the
presence or absence of contaminants
o Avoidance of concerns such as
holding times and packing and
shipping of samples
Potential disadvantages of employing
on-site mobile laboratories include the
following:
o If samples are not ready, the
costs are still incurred.
o The utility of the data for
litigation purposes is
questionable.
The cost of employing a mobile
laboratory is dependent upon several
factors including the time that a
laboratory must remain on site, the
distance a laboratory must be
transported to a site, and the
equipment, services, and personnel
required for the characterization
activity. The QA/QC protocols employed
and the level of documentation provided
by mobile laboratory services are
typically negotiable and defined
contractually.
Unlike fixed-base laboratory analyses,
costs for mobile laboratory work are not
generally calculated on a cost-per-
sample basis. Costs are determined by
the number of days the laboratory is
employed plus a fixed fee for trans-
portation of the laboratory. These
costs decrease as the duration of use
increases and the rental rate per day is
higher if the mobile laboratory is hired
for a shorter duration. Accordingly,
the cost-effectiveness of employing
mobile laboratories (over fixed-base) is
directly linked to the rate at which
site samples can be generated for
analysis. Although site character-
ization using mobile laboratories is
generally less expensive, this cost
advantage can easily be lost due to poor
planning or unexpected delays in
sampling activity while the mobile
laboratory is on site.
REFERENCES
1. "Guidance for Conducting Remedial
Investigations and Feasibility Studies
Under CERCLA," OSWER Directive
9335-3-01, March 1988.
2. "Data Quality Objectives for Remedial
Response Activities,"
EPA/540/G-87/003, March 1987.
3. "Proceedings of the Third Annual
Symposium on Solid Waste Testing and
Quality Assurance," Environmental
Protection Agency, Washington, D.C.,
July 1987.
467
-------
ENZYME IMMUNOASSAY FOR THE QUANTITATION OF
AN ALKALINE PROTEASE IN AIRBORNE SAMPLES
Larry S. Miller, Victor Moore, Abbie Wardwell,
Michelle Buchwalter, and Laurence A. Smith+
Battelle, Biotechnology Section, 505 King Avenue, Columbus, OH 43201
+The Procter & Gamble Company, Ivorydale Technical Center,
5299 Spring Grove Avenue, Cincinnati, OH 45217
ABSTRACT
Alkaline proteases are used as additives in
detergent products. Because they are allergenic
to workers in manufacturing areas, the work
environment is monitored for airborne levels of
these enzymes. A sensitive and specific inhibi-
tion enzyme immunoassay (IEIA) was developed for
the quantitation of an alkaline protease,
subtilisin. This IEIA had a detection limit of 1
ng/ml (36.7 pM), a response range of 1-50 ng/ml,
no crossreactivity with other microbial proteases,
and excellent inter- and intra- assay reproduci-
bility. The presence of detergent product or
environmental materials did not interfere with the
method. Air filter samples from manufacturing
areas producing subtilisin-containing detergent
product were analyzed for antigenic and enzymatic
activities of the alkaline protease. A high
correlation coefficient of 0.893 was obtained
between these methods. By the IEIA, the air
samples were found to contain 1-18 ng/ml of
subtilisin with a mean of 7.2 ng/ml and it was
possible to detect 0.15 ng subtilisin protein per
cubic meter of air. In addition to sensitivity
and specificity, the IEIA also circumvented
problems associated with the reference method such
as interference by other environmental proteases
and sample color.
Key words: immunoassay, environmental monitoring,
air samples, allergens
INTRODUCTION
Environmental monitoring for low levels of toxic
or hazardous compounds in air, water, and soil
samples has been performed using sophisticated
analytical chemistry techniques. Although these
methods can detect substances in the ppm to ppb
range, they usually require complex equipment and
need extensive sample preparation. In addition,
they are limited in number or type of samples and
are costly to perform for routine analysis.
In some cases, immunoassays may offer an alterna-
tive to conventional analytical chemistry
techniques that are now used for environmental
analysis. Immunoassays employ specific antibodies
that can detect compounds in the ppb to ppt range.
Because of the antibody specificity, it is
possible to analyze a complex mixture with only
minimal sample preparation, to detect several
substances in one sample, and to use small sample
volumes (0.5 to 1.0 ml).
Subtilisin Carlsberg (EC 3.4.4.16) is an alkaline
protease from Bacillus licheniformis which has
properties useful for detergent products (1). It
is necessary to control airborne levels of this
enzyme in manufacturing sites because of its
allergenicity as reported by Pepys et al (2). An
enzyme activity assay has been developed for this
purpose (3). However, this procedure is affected
by interfering substances, the stability of
subtilisin in the detergent product, and the
presence of other proteases in the air samples
from environmental microorganisms.
An immunoassay approach offers potential ad-
vantages over enzyme assays for evaluation of air
samples because of their specificity and sen-
sitivity. This study reports the development and
validation of an IEIA for subtilisin and its
application to the analysis of air samples for
this allergen.
DEVELOPMENT OF AN INHIBITION ENZYME IMMUNOASSAY
FOR SUBTILISIN
Rabbit anti-subtilisin sera were prepared and used
to develop an inhibition enzyme immunoassay
(IEIA). The IEIA is outlined in Figure 1 and is a
modification of the indirect enzyme immunoassay
described by Voller et al (4). Diluted antiserum
was mixed with an equal volume of the sample or
subtilisin standards. After an overnight
incubation to allow subtilisin to react with the
antibody binding sites, this solution was added to
microwells containing subtilisin adsorbed to the
plastic surface. Free antibody binding sites
would bind to the immobilized subtilisin while
subtilisin-antibody complexes would be inhibited.
The level of rabbit anti-cellulase antibody that
was bound to the solid-phase subtilisin was
469
-------
determined by adding goat anti-rabbit IgG,
alkaline phosphatase conjugate to the microwell.
After an incubation step, the excess conjugate was
removed by washing the microwell. Para-nitro-
phenylphosphate was added and this chromogenic
substrate was converted to a yellow product by the
alkaline phosphatase bound to the microwell. The
absorbance at 405 nm was determined using a
colorimeter that was designed to read 96 microwell
plates (HicroELISA Autoreader, Dynatech). The
concentration of subtilisin in a sample was
determined from an inhibition curve. For the
IEIA, absorbance values for the microwell solution
was inversely proportional to the concentration of
subtilisin in the sample or standard solution.
Incubate
flntigen-Rb
Mixture
Coat Well
with Rntigen
An inhibition curve for the IEIA was developed
using a 1:20,000 dilution of the anti-subtilisin
sera and concentrations of subtilisin from 1-100
ng/ml. This assay was found to have a limit of
detection of 1 ng subtilisin protein/ml and a
response range of 1-50 ng/ml (figure 2). The
total assay time was approximately 20 hrs which
included 16 hrs for the initial incubation of
sample and diluted antisera. This schedule
allowed for the preparation of multiple samples on
the first day and the' completion of the test and
data analysis on the following day.
100
SUBTILISIN CONCENTRATION (ng/ml)
Incubate with
Rntibody
Incubate with
Conjugate
Rdd Substrate-
Chromogen
Read DD 405nm
Figure 2. Inhibition curve for the subtilisin
IEIA. Each point represents the mean of tripli-
cate samples and the standard deviation by error
bars.
The specificity of the anti-subtilisin sera was
evaluated using other microbial proteases (1-1000
ng protein/ml) in the IEIA. They included
thermolysin, proteinase K, pronase, and newlase.
The concentration of antigen required for 50%
inhibition (ID50) was determined and only
subtilisin was inhibitory (ID50 42 ng/ml). The
other microbial proteases did not demonstrate
inhibitory activity at any concentration (ID50
>1000 ng/ml). This demonstrated that the
antibodies used in the IEIA were highly specific
for subtilisin.
Rntigen
J^Con jugate
* Chromogen
Figure 1. Inhibition enzyme immunoassay for
subtilisin.
IEIA VALIDATION STUDIES
These studies were performed to assess the effects
of various conditions on the inhibition curve for
the subtilisin IEIA. Any factors found to affect
the IEIA were assessed and measures to remove or
control their effect were developed.
Anionic detergents have been shown to have an
inhibitory effect on immunoassays (5). Detergent
products containing anionic and other surfactants
were evaluated at 0.001%, 0.01%, or 0.1% in the
IEIA using 1-100 ng/ml concentrations of sub-
tilisin. These conditions represent a 100 to
10,000 fold excess of detergent product to
470
-------
subtilisin by weight. The inhibition curve was
not affected by the granular detergent product and
was similar to the inhibition curve in figure 2.
The surfactants in the detergent product may have
had no effect on the IEIA due to the presence of
bovine serum albumin which has binding sites for
long alkyl chain molecules including sodium
dodecyl sulfate (6).
Recovery studies were performed to determine if
there were any interactions between air sample
components and subtilisin. Material was eluted
from 12 air filters and subtilisin was added to an
approximate final concentration of 10 ng/ml. The
antigenic activity of subtilisin added to the air
sample solutions was compared to a buffer control.
A 0.7 to 1.1 fold difference in antigen activity
between the air sample and the control was
obtained with an average ratio of 0.8.
The source of this variation for the recovery
studies was examined. The material that was
eluted from the air samples were diluted to
determine if an interfering substance in the
sample matrix would lose its effect. The
antigenic activity demonstrated excellent
proportionality for each two fold dilution and so
this effect did not appear to be due to components
in the air sample solution. In addition, the
effect of the subtilisin protease activity on the
IEIA was evaluated and did not affect the
inhibition curve. Next, the samples were treated
with a 0.45 m membrane filter to evaluate if
particles in the air samples may be the source of
this difference. The removal of particles by
filtering the solution demonstrated a 20-30% loss
in activity and appeared to be the source of this
difference. This situation was corrected by
increasing the salt concentration of the elution
buffer to 0.5 M to facilitate desorption of
subtilisin from the air particles.
The reproducibility of the subtilisin IEIA was
determined from inter- and intra- assay studies.
Interassay reproducibility was evaluated by
repeating the inhibition curve for the IEIA on
three consecutive days while the intraassay
reproducibility was determined with triplicate
inhibition curves on the same day. The IEIA was
highly reproducible for both inter- and intra-
assay studies with <10% coefficient of variation.
The inhibition curves were similar to figure 2.
ANALYSIS OF AIR SAMPLES FOR SUBTISIN BY THE IEIA
Air samples from manufacturing sites were
collected and evaluated for subtilisin content by
the IEIA and a reference method for subtilisin
enzyme activity (7). Air samples were collected
on glass fiber filters for 120 minutes with a
total air volume ranging from 88 to 100 m3.
Material was eluted from the glass fiber filter by
rapidly mixing with 20 ml of buffer containing 0.5
M sodium chloride, 0.02 M Tris, 0.01% Tween 20,
0.1% bovine serum albumin, 0.02% sodium azide,
0.001 M calcium chloride, 0.001 M phenylboronic
acid, and 0.001 M sodium thiosulfate at pH 8.0.
The fragmented filter particles were separated
from the solution by centrifugation at 2000g for
10 min at 4°C. The supernatant was stored at 4°C
and used for subtilisin analysis. The enzyme
assay (7) was performed by adding sample or
standards to a solution of 20 mM Tris buffer, pH
8.0 containing 2mM succinyl-ala-ala-pro-phe-
paranitroanalide. Subtilisin or other serine
protease would convert this chromogenic substrate
to a yellow product that was detected spectro-
photometrically at 405 nm.
The subtilisin content of 20 air samples was
determined for antigenic activity by the IEIA and
for enzyme activity by the reference method. The
values ranged from 4.8-18.4 ng/ml for the enzyme
assay and 0-18.0 ng/ml for the IEIA with means of
10.7 ng/ml and 7.2 ng/ml, respectively. For the
IEIA data, these concentrations were equivalent to
14 to 360 ng subtilisin per filter pad or 0.15 to
3.79 ng/m3 air. A high correlation coefficient
was obtained between these methods (r=0.893).
However, the linear regression line slope (0.8)
indicated generally higher values for the
antigenic than the enzymatic activity (figure 3A).
Furthermore, the y-intercept for the enzyme
activity assay at 7 ng/ml indicated a bias in this
method. Therefore, the enzyme assays was
evaluated for sources of these differences.
20-,
X
10
IEIA (ng/ml)
201 r=0.852
10
IEIA (ng/ml)
Figure 3. Correlative studies for the subtilisin
IEIA and enzyme assay. Initial data is in panel
A. Enzyme assay results corrected for sample
color and endogenous sample protease activity in
panel B.
471
-------
Some of the air samples had a color ranging from
yellow to dark gray which produced absorbance
values as high as 0.117. Since the sample color
interfered with the enzyme assay, these values
were corrected. This correction was not
necessary for the IEIA because the air sample
solution was washed from the reaction before
determining the absorbance.
The air samples were found to contain proteolytic
enzymes that was not from the detergent product.
Air samples from a manufacturing site not
containing subtilisin were collected and analyzed
for ambient protease activity. The mean enzyme
activity of these samples was 4.5 ng/ml with a
range of 3.9 to 5.0 ng/ml. The enzyme activity
assay uses a chromogenic substrate reagent which
is reactive with subtilisin and other proteases.
Analysis of these samples by the subtilisin IEIA
revealed that there was no detectable subtilisin
antigenic activity. The source of these proteases
may be environmental microorganisms that occur in
the air.
Correcting the enzyme assay data for sample color
and background enzyme activity, it was possible to
decrease the variation in the data between the
enzyme assay and IEIA methods. The correlation
coefficient remained high (r=0.852) and the
intercept for the regression line was moved to
near the origin (figure 3B). The line slope
(0.67) suggests that there is more detectable
antigenic activity in the samples than enzymatic
activity. This bias may be attributed to these
assays detecting two different properties of the
subtilisin molecule, its antigenic and enzymatic
activities. The IEIA can detect degradation
products as well as intact subtilisin while the
enzyme activity assay will detect only enzymati-
cally active forms of subtilisin.
SUMMARY AND CONCLUSIONS
Subtilisin, a microbial protease, is used in
detergent products (1) and can result in allergic
reactions upon exposure of workers in manufactur-
ing areas (2). Although this has been controlled
by reducing dust levels, it is necessary to
monitor the air for this enzyme. An IEIA for
subtilisin was developed for the analysis of air
samples. This highly reproducible assay has a
sensitivity of 1 ng/ml and does not detect other
microbial proteases. Validation studies of this
method demonstrated that it detected subtilisin in
detergent product and the air sample matrix.
The utility of the IEIA for subtilisin quantita-
tion was determined from comparative studies with
air samples from detergent manufacturing sites. A
high correlation coefficient (r=0.852) was
obtained between the IEIA and a reference method
for subtilisin enzyme activity. It was found that
the IEIA could detect subtilisin at levels as low
as 0.15 ng per cubic meter of air. This level of
detection and other characteristics of this assay
are comparable to those reported for radioim-
munoassays of other protein allergens, esperase
(8) and papain (9). However, the enzyme immunoas-
say offers advantages over radioimmunoassays in
that special handling and reagent stability are
not problems. The subtilisin IEIA was easy to
use, required only 300 ul of sample, used simple
sample preparation to transfer the dust to
solution, and analyzed simultaneously 23 samples
and standards in triplicate.
There are many potential applications of immunoas-
says for environmental monitoring. For subtilisin
and other airborne allergenic proteins, immunoas-
says provide an approach for their detection in
the manufacturing environment which is not
feasible by standard chemical methods. This
approach is also applicable to small organic
compounds such as pesticides, herbicides,
insecticides, and fungicides (10). Other
immunoassays formats can be used to automate the
analysis of large numbers of samples, to evaluate
one sample for several compounds, and to perform
near real-time analysis at a disposal or storage
sites.
ACKNOWLEDGEMENT
This study was supported by The Procter & Gamble
Company. Secretarial assistance was provided by
Shan Wolfington and Sandra Jennings. We wish to
thank Dr. Jeanette van Emon and Dr. P. Albro for
their comments.
REFERENCES
(1) Daubman, C. and Aunstrup, K., "The variety of
serine proteases and their industrial
significance", Proteinases and Their
Inhibitors, edited by V. Turk and L. J.
Vitals, Permagon, New York, New York, 1981,
pp. 231-244.
(2) Pepys, J., Hargreave, F. E., Longbottom, J.
L., and Faux, J., "Allergic reactions of the
lungs to enzymes of Bacillus subtilis",
Lancet. 1969, pp. 11-81-1184.
(3) Friedman, S. D. and Barkin, S. M., "Enzymatic
activity of proteases in detergent systems
comparison of assay methods and the role of
interfering substances", J. Amer. Oil Chem.
Soc.. Volume 46, 1969, pp. 81-84.
(4) Voller, A., Bartlett, A., and Bidwell, D. E.,
"Enzyme immunoassays with special reference
to ELISA techniques", J. Clin. Path.. Volume
31, 1978, pp. 507-520.
(5) Halfman, C. J., Dowe, R., Jay, D. W., and
Schneider, A. S., "The effect of dodecyl
sulfate on immunoglobulin hapten binding",
Molec. Imm.. Volume 23, 1986, pp. 943-949.
(6) Reynolds, J. A., Herbert,'S., Polet, H., and
Steinhardt, J., "The binding of diverse
detergent anions to bovine serum albumin",
Biochem.. Volume 6, 1967, pp. 937-947.
(7) Rothgeb, T. M., Goodlander, B. D., Garrison,
P. H., and Smith, L. A. "The raw material,
finished products, and dust pad analysis of
detergent proteases using a small synthetic
substrate", J. Amer. Oil Chem. Soc.. Volume
65, 1988, pp. 806-810.
472
-------
(8) Agarwal, M. K., Ingram, J. W., Dunnette, S.,
and Gleich, G. J., "Immunochemical
quantitation of an.airborne proteolytic
enzyme, Esperase, in a consumer products
factory", Am. Ind. Hyg. Assoc. J.. Volume 47,
1986, pp. 138-143.
(9) Wells, I. D., Allan, R. E., Novey, H. S., and
Culver, B. D., "Detection of airborne
industrial papain by a radioimmunoassay", Am.
Ind. Hyg. Assoc. J., Volume 42, 1981,
pp. 321-322.
(10) Vanderlaan, M., Watkins, B. E., and Stanker,
L., "Environmental monitoring by
immunoassay", Envir. Sci. Technol.. Volume
22, 1988, pp. 247-254.
473
-------
GAS CHHCHATOGRAPHIC AND MASS SPECTRCMEnKnc ANALYSIS
OF TARGET AIR TOXICS AT REMEDIAL HAZARDOOS WASTE SITES
ABSTRACT
D.W. Hodgson3, B.C. Miller, R.A. Ross and T.S.Viswanathanb
NSI Technology Services Corporation
400 State Ave., Gateway Center Tower II, Suite 311
Kansas City, KS 66101
R.D. KLeopfer3 and W.W. Bunn
U.S. EPA laboratory
25 Funston Rd.
Kansas City, KS 66115
remediation of known hazardous waste sites. This
type of pollution is often a local event of short
duration that differs significantly from such
events as the Bhqpal, India incident or the
Seveso, Italy incident in that the emissions can
be controlled. Cleanup activities at the
Superfund sites managed by the Environmental
Protection Agency (EPA) are accompanied by well-
planned air-monitoring programs designed to
safeguard the short and long-term health of
workers, scientists and residents in the vicinity
of the waste-sites. An example of such a program
is the monitoring of air-samples from dioxin
sites in eastern Missouri. During cleanup,
ambient air samples are collected on polyurethane
foam cartridges and analyzed for 2,3,7,8-
tetrachlorodibenzodioxin (2). In this case, the
pollutant is not volatile, and the air pollution
is, for the most part, caused by construction
machinery which releases dust particles
containing the bound pollutant into the air.
GC and GC/MS analytical methods were used for the
analysis of air-toxics at two Superfund waste
sites. Data was utilized for ambient air
monitoring and for supporting remedial actions.
Ambient air samples fron the waste site,
collected on Tenax cartridges, were analyzed by
using the modified Gas Chromatography-Mass
Spectrometry (GC/MS) method, TO-1 (1). The
compounds of interest, chloroform and carbon
tetrachloride, were detected at the 1 ng level in
roost samples. To support remedial actions, Air
Stripper water samples and Vapor Extraction
System (VES) vapor samples from the site were
analyzed by direct injection on an electron
capture gas chrcmatograph at a laboratory close
to the remedial site. Tedlar bags and gas
syringes were used as vapor sample collection
containers. Analytical results for these samples
ranged from the detection limit (.3 mg/m3) to 19
mg/m3 for chloroform and 1.2 to 910 mg/m3 for
carbon tetrachloride. Problems encountered in
sampling and analysis, data quality assessment
procedures, and logistics in providing fast
turnaround analytical results are discussed along
with other related issues.
INTRCOOCITCN
Air quality, both indoor and outdoor is
increasingly becoming an important issue for the
general public. To combat outdoor pollution,
governments around the world have adopted laws
and regulations that control emissions from a
number of static and dynamic sources such as
those from coal-fired power plants, manufacturing
plants, autoncbiles and even fireplaces from
individual homes. A different but increasingly
common source of air pollution is one that
results from cleanup efforts aimed at the
Current Address:
Hall-Kimbrell Environmental Services
4848 W. 15th Street
Lawrence, KS 66044-0307
Address correspondence to this author
One of the remediation projects, supported
analytically by the recently formed Environmental
Services Assistance Team (ESAT - An EPA
contractor providing technical and analytical
support for Superfund activities), was concerned
with the restoration of clean ground water to two
communities in central Nebraska. The ground
water, contaminated with the fumigants chloroform
and carbon tetrachloride, was cleaned-up by air
stripping. Soil gas containing the contaminants
was removed by using a Vapor Extraction System
(VES). The concentrations of the target toxics
in the groundwater before, during and after
treatment, and the VES emissions were monitored
by ESAT chemists in cooperation with EPA and
other EPA-contractors. This paper summarizes the
analytical methodology, quality assurance
procedures, logistics and other considerations
involved in providing the required analytical
support for this complex remediation effort.
HOJECT DESCRIPTICN
The project at the Waverly, Nebraska site
consisted of analyzing VES vapor samples and air-
stripper process water samples for chloroform and
carbon tetrachloride using a gas
chromatography/electron capture (GC/ECD)
475
-------
technique. To minimize blank-contamination and
to provide data in a timely manner the analysis
was not performed at the waste-site but in a
nearby laboratory. Samples were collected in the
field and delivered to the laboratory by an EPA-
contractor. Results were reviewed and reported
to the designated project managers by telephone.
Support to EPA for this project was provided by
NSI from 1-20-88 to 2-19-88.
The ambient air-monitoring project at the
Murdoch, NE and Waverly, NE sites consisted of
analyzing Tenax sorbent cartridges through which
a known volume of ambient air had been drawn.
The pollutants, CHC13 and CC14, were desorbed
thermally and analyzed using a gas
chromatography/mass spectrometry (GC/MS)
technique. Support to EPA for this project was
completed in July '88.
The steps involved in ESAT's analytical support
consisted of the following:
1. Selection of appropriate analytical methods
and the development of a Quality Assurance
Project Plan for the Waverly, NE project and a
Standard Operating Procedure for analytical
support of the Murdoch, NE project.
2. Performance of laboratory analysis using
methods identified in step 1.
3. Review of analytical data by analysts and
reporting by telephone for rapid turnaround.
4. Detailed follow-up review of analytical data
to assess precision, accuracy, completeness and
other performance criteria associated with the
analytical methods.
5. Recommendations for future field or on-site
analytical support.
EXPERIMENTAL
The GC/ECD analyses were performed using a
Hewlett Packard (HP) 5890 gas chromatograph
equipped with a HP 10 MByte hard disk drive, HP
model 3393A integrator, and a HP model 7634A
autosampler. The EPA Method no. 501.2, "Analysis
of Trihalomethanes in Drinking Water by
liquid/Liquid Extraction" was used with the
following modifications:
1. A 30 m DB-Wax (J&W Scientific) megabore (0.53
mm I.D.) capillary column of 1.0 micrometer film
thickness was used for chromatographic
separation. Chromatography was done isothermally
at 40°C using argon /5%-methane as the carrier
gas. The GC run times were typically 3.5 minutes
in length.
2. A separate, four point calibration curve was
developed for each class of water samples tested.
The calibration ranges consisted of the following
concentrations (micrograms/liter): For effluent
water, CHC13 - 0.2 to 0.7; CC14 - 0.4 to 1.4; for
midprocess water, CHC13 - 1.0 to 3.4; CC14 - 10
to 34; for influent water, CHC13 - 18 to 60; CC14
- 900 to 3000.
3. For vapor samples, 15 to 40 microliters of 1
ppm CHC13 and 1 ppm CC14 gas-phase standards were
injected to construct the 4 point calibration
curve. The calibration range was 80 pg to 260 pg
for CHCL3 and 100 pg to 320 pg for CCL^. Static
dilution bottles for diluting either vapor
samples or gaseous standards were unavailable to
the analysts at the field laboratory.
4. Dilution of water samples with organic-free,
distilled, deionized water was employed to
overcome the constraints of a limited calibration
range for samples that exceeded the upper
calibration range.
All the GC/ECD measurements reported in this
article were performed at the Nebraska Department
of Environmental Control (NDEC) laboratory in
Lincoln, NE. To support this project, the NDEC
laboratory kindly provided counter and hood
space, electricity and access to their facility
outside the normal working hours. The equipment,
and all other supplies including gas cylinders,
syringes, glassware, etc. were transported from
the EPA regional laboratory in Kansas City, KS to
the Lincoln, NE location.
ANAISnCAL STANDARDS
Analytical standards of chloroform (10
micrograms/liter) and carbon tetrachloride (5
micrograms/liter) in methanol were obtained from
the NSI-operated EPA Analytical Standards
Repository in Research Triangle Park, NC. A
vapor phase analytical standard mixture of
chloroform and carbon tetrachloride (1 ppm v/v
each) in nitrogen was obtained from Scott
Specialty Gases.
SAMPLING AND ANALYSIS OF VAPOR EXTRACTION SYSTEM
VAPOR SAMPLES
Vapor phase samples collected in Tedlar bags or
50 cc Luerlock glass gas-syringes with a
stainless steel ball valve were employed for
sampling. The syringe was used in the later part
of the project since the VES sampling ports were
at or below atmospheric pressure and the samples
collected in collapsible Tedlar bags did not
yield analytically reproducible results. Either
of these methods were preferred over sorbent
absorption techniques due to the high
concentration of the analytes and the speed of
sampling with Tedlar bags and/or syringes. Also,
charcoal, Tenax and XAD sorbent cartridges were
not used since the desorption of analytes from
these tubes and subsequent analysis requires a
relatively long period of time. In addition,
some protocols require three separate analyses;
one each for the front particulate filter, the
front sorbent and the back sorbent. This would
have tripled the analysis time.
Vapor samples collected in Tedlar bags or gas
syringes were transported to the lab at ambient
temperature and analyzed by injecting an
appropriate volume of sample (< 1 microliter to
500 microliters) directly into the gas
chromatograph. The injection volume was adjusted
476
-------
effectively to bring the analybe GC/ECD response
within the calibration range. Response factors
(RF) in units of picograra of analyte standard
injected per unit GC peak area were defined;
these were used to calculate the sample
concentration using the following equation :
* RF ) / Vs
where:
= concentration of analyte in
pg/microliter or mg/cu.m. units
Ag = GC peak area for sample
Vs = Volume of air injected (microliters)
RF = Response factor for analyte (pg/A)
SAMPLING AND ANALYSIS OP WATER SAMPLES
Water samples were collected in 40 ml vials with
teflon-lined screw caps and shipped to the
laboratory in ice filled coolers. Ten
milliliters of the sample were extracted with two
milliliters of pentane; one microliter of the
extract was injected into the gas chromatograph
for analysis.
SAMPLING AND ANALYSIS OF AMKLENT AIR
Ambient air samples were collected by drawing a
known volume of air through clean Tenax
cartridges. The thermal desorption and mass
spectral analyses were performed by using a
Finnigan 4000 GC/MS/DS system interfaced to a
Tekmar 5010 thermal desorber.
Tenax cartridges were fortified with the internal
standard d4-l,2-dichloroethane (for use in
quantitative analysis) by injection into the
desorber gas stream. The pollutants and the
internal standards were prepurged to remove water
from the sorbent at 40°C using a helium flow of
15 ml/min. The pollutants and the internal
standards were desorbed thermally by heating the
cartridges to 190°C using a helium flow of 15
ml/min. and trapped using a liquid-nitrogen
cryogenic trap at -125°C inside the Tekmar 5010
desorber. The analytes were subsequently
desorbed thermally at a lower capillary column-
compatible helium flow rate of 1-2 ml/minute and
were trapped again on a cryogenic trap at -125°C
at the capillary interface unit near the head of
the capillary column. A quick injection of the
trapped analytes was made into the GC column by
heating the trap at the interface unit rapidly to
250°C. The analytes were separated on a 50 m DB-
5 ( J&W Scientific) GC capillary column (0.32 mm
I.D. and 1 micrometer film thickness) using the
following conditions: Initial temperature: 30°C
for four minutes; temperature programming:
6°C/minute ramp to 90°C followed by a second ramp
at 25°C/minute to 200°C to prepare the column for
the next run.
All calibration and internal standards were
prepared by vaporizing a known amount of high
purity, liquid phase standard material in a 2
liter gas dilution bottle equipped with a mini-
inert valve. The concentration of each compound
was calculated from the following equation :
C = ( d * I ) / V (2)
where:
C = compound concentration (ng/microliter)
d = density of compound (g/ml)
I = amount injected (microliters) into the
Static Dilution Bottle, and
V = Volume of Static Dilution Bottle
(liters)
The calibrations were performed by absorbing the
appropriate vapor phase analytical standards and
the internal standard onto a blank Tenax
cartridge in the same manner that the internal
standard was added to the sample cartridges prior
to analysis. Calibration curves consisting of a
blank and two concentration levels of analytes
were employed for initial calibration. Each day
samples were analyzed, the calibration factors
were verified at one concentration level.
RESDiaS
A. GC Analysis
A total of 123 samples consisting of 56 vapor
samples and 67 water samples were analyzed for
chloroform and carbon tetrachloride during a four
week period from 1-22-88 to 2-19-88 in support of
remedial efforts at the Waverly, NE Superfund
site. The analysis of 123 samples required 1057
separate injections including those needed for
the multi-point calibration at multiple
calibration ranges for each of the two matrices,
laboratory method blank for the two matrices,
matrix spike, matrix spike duplicates, field
duplicates, laboratory duplicates, sample
dilutions, etc. The analytical results for the
samples varied from none detected (the lowest
detection limits that were reported are shown
below in parentheses) to very high concentration
levels:
For water:
CHC13 from: (0.2) to 170 micrograms/liter
CC14 from: (0.1) to 6400 micrograms/liter
For vapor:
CHC13: from (0.1 to 0.4) to 23 mg/m3
CC14: from (0.2) to 910 mg/m3
The following performance factors were assessed
for determining data quality for the project:
1. GC Retention Time Stability
This factor affects the ability to identify the
analyte qualitatively. From the calibration
standard analysis runs, the following retention
time (RT) data were collected. The mean RT,
standard deviation (s) and the percent relative
standard deviation (RSD) from 16 data points
collected between 1-23-88 and 2-18-88 for air
analyses were:
477
-------
For CHC13:
RT = 0.858 min.; s = 0.004 min.; RSD = 0.46%
For CC14:
RT = 1.971 min.; s = 0.009 min.; RSD = 0.47%
2. Method Blank Results
Twelve of fourteen water method blanks, which
consisted of distilled deionized water, contained
low levels of CHC13 (0.1 to 0.4 micrograms/liter)
and 00.4 (0.1 micrograms/liter). The remaining
two water method blanks showed elevated levels of
CHC13 (3.3 to 3.6 micrograms/liter) and CC14 (1.2
to 1.6 micrograms/liter). The laboratory method
blanks for vapor samples, which consisted of 50
to 500 microliters of air obtained from outside
the laboratory, were mostly free of analytes.
Small amounts of chloroform (0.1 & 0.2 mg/cu.m.)
and CC14 (0.1 & 0.7 mg/cu.m.) were observed in 2
of the nine blanks. At least one matrix specific
method blank was analyzed for each day water or
air samples were tested. Detection limit
thresholds for the analysis were calculated as
twice the level found in the blank. Positive
values for analytes below the threshold were
reported as not detected. Sample results above
the threshold were reported without correction
for blank concentrations.
3. Reproducibility of Response Factors
Instrument responses for CHC13 and CC14 were
linear over the short ranges of concentrations
described in an earlier section. The response
was non-linear over wider concentration ranges.
The response factor stability over the project
period was excellent. For all the calibration
runs performed during the project, the percent
relative standard deviation of the response
factors varied from 19% to 22% for CHC13 and from
23% to 26% for CC14.
4. Reproducibilitv of Analysis
Duplicate determinations were performed on 16
water and 15 vapor samples, a minimum of one
duplicate analysis for each day water or air
samples were tested. This represents a frequency
of more than 20%. The duplicate results for
CHC13 and CC14 in water samples are shown in
Tables I and II respectively. The numbers in
parentheses are detection limits. The percent
difference (%D) for duplicate results was
calculated using the following equation:
(%D) = 1200 * (S1-S2)/(S1+S2)| ( 3 )
where S^ and S2 are the results of replicate
analysis. The %D is generally lower at
concentrations >5 micrograms/liter. The mean %D
for vapor analysis, based on 13 determinations,
was 17% for chloroform and 38% for carbon
tetrachloride.
TABLE I
REPRODUCIBILITY OF CHCL3
RESULTS IN WATER SAMPLES
(micrograms/liter)
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Sl
16
19
0.3
29
1.9
0.3
21
(0.3)
1.9
23
0.6
19
20
(0.3)
44
44
S2
19
12
0.7
22
1.0
0.5
21
(0.3)
2.0
20
(0.1)
19
3.9
(0.3)
41
43
%D
16
37
130
24
49
67
0
0
5
13
—
0
81
0
7
2
n = 15
mean %D = 29
S.D. = 38
TABLE II
REPRODUCIBILITY OF
RESULTS IN WATER SAMPLES
(micrograms/liter)
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
sl
44
1260
(3)
2970
26
0.5
2240
(1.8)
37
2540
0.9
2450
3140
1.8
3020
6380
S2
60
1190
(3)
3300
33
0.2
2360
(1.8)
42
2240
0.9
2340
395
2.0
4240
6290
%D
36
6
0
11
27
60
5
0
14
12
0
5
87
11
41
1
n = 16
mean %D = 20
S.D. = 25
478
-------
5. Spite
A total of eleven water and eight vapor
samples, representing a frequency of 16% and 14%,
respectively, were spiked and analyzed. The
amcxmt spiked into water sanples varied from 25
to 563 micrtxjrams/liter for chloroform and from
20 to 2000 micxograms/liter for OC14. The eleven
water samples contained from 1 to 44
nucrograms/liter of chloroform and from 1 to 6380
micrograms/liter of OC14 before spiking. The
recovery of the spiked analytes, in valid
experiments, varied from 0% to 150% for
chloroform and 0% to 145% for OC14 with the
following statistics:
CHC13: n = 9; mean = 83%; S.D. = 36%
CC14: n = 10; mean = 90%; S.D. = 49%
Ihe eight vapor samples contained chloroform from
0.1 to 23 mg/m3 and CC14 from 0.8 to 448 mg/m3
before spiking. The amount spiked into the
sanples varied from 38 to 500 mg/m3 for
chloroform and from 60 to 2000 rag/m3 for CC14.
Since static dilution bottles were unavailable,
matrix spikes were performed by injecting small
volumes (<25 microliters) of methanolic CHC13
and CC14 standards directly into the sample
syringe. Incomplete vaporization of spiked
analytes cannot be ruled out under these
conditions. The recovery of spiked analytes
varied from 16% to 121% for CHC13 and 7% to 329%
for CC14 with the following statistics:
CHC13: n = 7; mean = 60%; S.D. = 33%
CC14: n = 7; mean = 94%; S.D. = 112%
6. Performance
Audit
Two water audits were performed during this
project. The true value of OC14 in both audit
samples was 20.4 micrograms/liter. The observed
values were 24.7 and 22.2 micrograms/liter, which
translates to % recoveries of 121% and 109%,
respectively. The true value of chloroform in
both audit samples was 20.2 micrograms/liter .
The observed values, 73 and 162 micrograms/liter,
were much higher. A co-eluting interference
which was partially resolved on the column, as
evidenced by GC peak broadening, led to higher
analytical results for chloroform. Pour other
chlorinated hydrocarbons were also present in the
audit sanple. No reference standards were
available on-site to determine the identity of
the co-eluting peak which was not observed in the
analysis of actual samples.
7. overall Assessment of GC Data Quality and
Analytical Rig-port
The detection limits of 5 micrograms/liter for
water samples and 1 micrograms/m3 in vapor
samples for CHC13 and OC14 were easily
acccnplished. The required analytical turnaround
times of 8 hours for vapor sanples and 48 hours
for water sanples were also met. The objective
of >90% ccnpleteness was met as well. With
regard to precision, the data was quite variable
for both water and vapor sanples. The highest
of variability was observed for sanple
concentrations near the detection limits, as
would be expected. The accuracy of the GC
analytical results was estimated at +/~ 150%,
which was within the target range of +/~ 200%. A
number of factors inherent in the method created
the significant variation observed in the data.
Some of these are listed below:
1. lack of stability of water samples.
2. large dilution factors for some water
samples.
3. Wide range of injection volumes for vapor
sanples required to have the instrument
response within the calibration range.
4. Difficulty in spiking and withdrawing vapor
samples from the 50 ml glass syringe
containers.
B. GC/MS ANALYSIS
Over 125 ambient air sanples collected in Tenax
cartridges have been analyzed to date for
chloroform and carbon tetrachloride using the
modified method TO-1 in support of ongoing
remedial projects. The following performance
characteristics were observed for this method
over a one week period:
1. Calibration Factors
Relative response factors for analytes with
respect to the internal standard, d4-l,2-
dichloroethane were defined as follows:
) * (
where:
FRF = relative response factor for CHC13
or CC14
Aa = area for the analyte quantitation
ion; in/z 83 for CHC13 and m/z 117
for CC14
= area for the internal standard
quantitation ion of m/z 65
= amount (ng) of internal standard, and
Wa = amount (ng) of analyte
The relative response factors determined daily
over a one week period varied substantially. The
following statistics were observed for eleven
such determinations:
Analyte
CHCU
Mean RRF : 1.38 1.24
Standard Deviation: 0.67 0.98
Percent Relative
Standard Deviation: 49% 79%
The mean value and standard deviation of the
absolute response for 50 ng of the internal
standard were 16803 area units and 12703 area
units, respectively. Sixteen determinations were
made.
479
-------
2. Retention Time Stability
The retention time (RT) stability for the
internal standard and the analytes was
substantially worse for the GC/MS method than for
the GC method as shown below:
Compound
: Internal CHC13
Standard
ecu
No. of analyses :
Mean RT (s) :
Standard deviation:
Percent Relative
standard deviation:
16
235
21
8.9
12
200
14
7.0
12
246
18
7.3
The manual trigger mechanism employed for
coordinating three events, sample injection by
the Tekmar 5010, the start of the GC run, and the
start of the MS data acquisition contributed
somewhat to the observed scatter in the retention
times. The major contributor, however, was the
poorly designed GC oven-sub-ambient temperature
controller. Since automated data reduction
procedures required a narrow RT window, analyte
identification using mass spectral match and
quantitations were done manually.
FKXSLfUS
The analytical results for chloroform and carbon
tetrachloride for vapor samples collected in
Tedlar bags were not reproducible. These
compounds had a tendency to adhere strongly to
the walls of the bag. Agitation of the bags led
to laboratory sub-samples with higher values.
Variations up to 700% were observed for seven
replicate determinations with higher values
corresponding to higher degrees of agitation.
Higher results were also observed when the
samples collected outdoors in the cold winter
months were allowed to warm up to room
temperatures. Additional problems were
encountered in performing matrix spite recovery
determinations. Consequently, Tedlar bags were
discontinued in favor of glass syringes as vapor
sampling devices. One drawback with the
syringes, however, was the collection of an
oil/water emulsion in the syringe ball valve,
which occasionally contaminated the analytical
syringes.
A number of administrative problems needed to be
resolved for providing the field analytical
support. For example, the transportation of the
GC equipped with a radioactive Nickel-63 source
in the electron capture detector across the state
lines required prior permission from the state
government agencies. The use of versatile
automated GC equipment enabled the one-person
field analytical crew to perform a number of
tasks simultaneously. Projects of this type
require experienced, adaptable and versatile
chemists who can operate without much supervision
and are also capable of reviewing their own data
for quality and usability within the limited
analytical turnaround time.
A number of analytical problems were observed in
the analysis of Tenax samples. The clogging of
capillary columns by ice (due to incomplete
removal of moisture during the prepurge step)
and/or particulates from the cartridge was a
serious problem. Using the system described
above, one large tank of liquid nitrogen is
required for the analysis of approximately 30
samples.
CONCHJSION
From our experience, we can conclude that
providing sampling and analytical support for air
monitoring is far more complex than it may
appear. The lack of availability of gaseous
standards, inapplicability of laboratory methods
to field-like conditions, the inadequacy of
sampling devices and other factors make
determination of organic pollutants in air a
challenge to the scientific community. Highly
automated equipment with multiple detectors
exhibiting linearity over wide concentration
ranges is mandatory for rapid and dependable
analytical determinations. The availability of a
laboratory near the waste site, rather than
setting up a laboratory on-site, is desirable
since this minimizes background and cross
contamination in the analytical determinations.
ACKNOWLEDGMENTS
The authors would lite to acknowledge the
generous support provided by the Nebraska
Department of Environmental Control laboratory
personnel. The projects were managed by the
Superfund section of the Region VII EPA with
support from the contractors: Woodward-Clyde
Consultants and Tetratech, Inc. The assistance
of Mr. Lee Taylor in preparation of this
manuscript is also acknowledged.
REFERENCES
1. USEPA, "Compendium of Methods for the
Determination of Toxic Organic Compounds in
Ambient Air", EPA-600/4-64-041, April 1984.
2. Fairless, B.J., Hudson, J., KLeopfer, R.D.,
Holloway, T.T., Morey, D.A., and Babb T.,
"Procedures Used To Measure the Amount of
2,3,7,8-Tetrachlorodibenzo-p-dioxin in the
Ambient Air near a Superfund Site Cleanup
Operation", Env. Sci. Tech., Vol. 21, No.6, 1987,
pp. 550-555.
480
-------
ON-SITE SOIL GAS ANALYSIS OF GASOLINE COMPONENTS
USING A FIELD-DESIGNED GAS CHROMATOGRAPH-MASS SPECTROMETER
Albert Robbat, Jr. and George Xyrafas
Tufts University, Chemistry Department
Trace Analytical Measurement Laboratory
Medford, Massachusetts 02155
ABSTRACT
A Gas Chromatograph-Mass Spectrometer (GC-MS)
specifically designed for hazardous waste site field
investigations has been evaluated for the identification
and quantification of volatile gasoline components.
Response factors for benzene, toluene, ethylbenzene and
isomeric xylenes have been determined. The field GC-
MS responded linearly over two orders of magnitude
with minimum detectable quantities of Ippb for these
compounds. The GC-MS was brought to the site for
field investigation of gasoline contaminated soil.
Comparative soil gas measurements are presented.
Key words: field gas chromatograph-mass spectrometer,
gasoline contamination, on-site field
investigation.
INTRODUCTION
The current inability of a field usable, chemical
method of analysis, to unambiguously identify the wide
diversity of organic compounds found on EPA's
Hazardous Substance List (HSL) is increasingly becoming
the critical, limiting step for hazardous waste site
investigations. State and private sector environmental
laboratories are overburdened with work, with data
turnaround time typically exceeding several months.
Clearly, an urgent need exists for analytical
instrumentation capable of performing on-site chemical
analysis. The obvious goal is to provide site managers
with the necessary information required to make
immediate decisions. A field usable gas chromatograph-
mass spectrometer (GC-MS) provides the most reliable
means for immediate compound identification and
quantification for the diverse compounds found on the
hazardous substance list. Development of reliable
analytical instruments that are easily transportable and
able to withstand changing climactic conditions should
move the analyst from the laboratory into the field.
In another paper presented at this symposium (1), a
GC-MS (Bruker Instruments) specifically designed for
on-site investigations was evaluated for its ability to
separate and quantify HSL volatile organic compounds
(VOCs). It was found that under well-defined operating
conditions the GC-MS, employing a 30m x 0.32mm i.d.
fused silica capillary column with 1.8/im film of DB624,
separated 31 of the 35 HSL compounds. Compounds
that coeluted were differentiated based on their mass
spectrum. Dynamic ranges were determined for these
compounds under the same operating conditions used to
separate all of the compounds simultaneously. It was
found that the minimum detectable quantity was
between 10 and 40 ng of compound injected (S/N=3)
and that the linear dynamic range extended over two
orders of magnitude.
In the present study, the applicability of field GC-
MS toward VOCs will be discussed for soil gas analysis.
Massachusetts Department of Environmental Quality
Engineering (DEQE) allowed the authors (as part of an
ongoing program, initiated in 1986) to evaluate the GC-
MS for chemical analysis of a gasoline contaminated site.
The gasoline vapors presently pass from groundwater
through the basement floor of a church building. Initial
work determined that groundwater flow is above
bedrock, that the depth to bedrock is not uniform, and
that the groundwater flow velocitites are low. DEQE's
goal is to remove the volatile gasoline constituents from
the contaminated soil beneath and around the church.
The church has been closed for over a year due to the
immediate health and safety risks associated with the
vapors. DEQE with its contractors performed a soil gas
extraction pilot study. The objective was to investigate
the effect of vapor extraction from different well points;
determine whether soil under the basement floor was
porous (soil gas transfer); and finally, determine the
effect of the water table on soil gas vapors.
EXPERIMENTAL SECTION
The experimental work reported in this paper was
performed by DEQE contractors and the authors. A
Foxboro century 128 organic vapor analyzer (OVA),
HNU photoionization meter, and a Photovac model
10A10 GC were used by the contractors to perform their
analyses. Before measurements were made by the
contractors, the soil gas probes and wells were uncapped
and air drawn from them to purge the dead volume.
The samples were collected in an air tight syringe and
analyzed by GC. Ambient air samples were collected on
Tenax/Ambersorb using a low flow pump. The trapped
material was analyzed after extraction by methanol using
GC-MS.
481
-------
A detailed description of the Brucker GC-MS and
the development of well-defined operating conditions
established specifically for on-site analysis of HSL VOCs
under varying climatic conditions have been provided in
other papers at this meeting, (Trainor and Laukien as
well as Robbat and Xyrafas). The field GC-MS and the
accompanying HNU (response/non-response) data were
produced by the authors. Soil gas probes and wells were
uncapped for two hours before VOC air sampling was
performed. Samples were collected and trapped in a
tube containing Tenax from the probes, wells and
ambient air with a Gillian (model 513) air sampling
pump The tube was lowered about six inches into the
probes and wells. The building was ventilated by
opening doors and windows as allowable.
RESULTS AND DISCUSSION SECTION
Response factors, RF, were determined for the
gasoline components benzene, toluene, ethylbenzene and
isomeric xylenes (BTEX) as prescribed by EPA
procedures. Known BTEX concentrations between the
range of 10 ppb and 200 ppb and an internal standard,
1,4-difluorobenzene (50 ppb), were injected into a tube
containing Tenax. The tube was placed into a GC
desorption oven and held at 220 °C for 45 sec. The oven
was directly interfaced to the GC-MS. The average RF
percent standard deviation value for each compound
based on fifteen GC-MS experiments was: benzene,
11.0%; toluene, 6.6%; ethylbenzene, 20.6%; meta plus p-
xylene's, 21.0%; and o-xylene, 22.4%. A GC-MS total
ion current chromatogram for the BTEX compounds at
20 ppb is shown in Figure 1. Figure 2 is representative
of the typical dynamic range plots obtained using the
field GC-MS for these compounds. The data presented
above were obtained with the GC-MS housed in a
Chevrolet Blazer, in a parking lot. Blank experiments
were performed to ensure that the carrier gas, ambient
air, did not contaminate the sample.
Figure 3 depicts soil gas probes and wells located
within and around the church. Table 1 lists outside soil
gas probes (J), their corresponding depths, and total
BTEX concentrations. The analyses shown in the table
were performed five months apart.
Since the basement floor is about five feet below the
exterior ground level, the J-probes are generally at or
above the floor. The HNU probe (authors) was inserted
directly into the well to determine relative response
before the samples were collected for GC-MS analysis.
Figure 4 reveals a typical field GC-MS chromatogram.
The profile is of a VOC sample collected from soil gas
probe J-9. RTEX concentrations were: benzene, 13 ppb;
tolune, 36 ppb; ethylbenzene, 38 ppb; m- & p-xylene,
88 ppb; and o-xylene 51 ppb. It appears from the table
that the contamination is moderately localized on the
southeast side. The absence of BTEX in the outside
ambient air sample indicated that the sample collection
procedure and GC-MS experiments were free of
contamination from the site.
The height of the interior wells, I-wells, were
measured from the basement floor to the water level and
found to be on average 1.35 feet. The sample collection
tube was placed six inches into the well. The VOC
samples were collected onto Tenax tubes and thermally
desorbed into the GC-MS. Figure 5 is representative of
the BTEX GC-MS total ion current chromatograms
obtained from the interior monitoring wells (e.g., 1-2):
•IN
1.
I.
3.
*.
9.
t.
1,4-dlflaoroWura*
telMM
•Ujrlbmnw
Figure 1. GC-MS chromatogram of benzene, toluene,
ethylbenzene, m-,p- and O-xylene (BTEX), compounds
Injected onto Tenax and then thermally desorbed
directly onto the GC column.
benzene
Figure 2.
benzene.
Linear dynamic range plot of
benzene, 1.6ppm; tolune, 4.4ppm; ethylbenzene, ND; m-
& p-xylene, 98ppm; and o-xylene 48.5ppm. Figure 6 is
the GC-MS of the inside ambient air at the center of the
basement: benzene, 60 ppb; tolune, 204 ppb;
ethylbenzene, 83.5 ppb; m- & p-xylene, 444 ppb; and o-
xylene 181 ppb. Table 2 summarizes the results from
three separate 1988 sampling dates. As evident from the
tables the GC-MS and HNU results are in excellent
agreement.
482
-------
The results of the soil gas study suggest that the
highest concentrations of VOCs are in fact on the
southeast side of the church. The highest BTEX
readings were found in J-9 and 10 as well as 1-1 and 2.
This is the side closest to the gas station. Much lower
concentrations were detected at all other wells. The
apparent dramatic decrease between results of February
and June may be due to the changing hydrogeologic
conditions (water table), the groundwater gasoline
recovery system put into operation between the
respective sampling dates, and/or better vapor venting.
We have demonstrated that the field GC-MS is
stable, reliable and performs under various climatic
conditions (1). For example, it rained hard during the
gasoline study while the RF and groundwater
experiments were accomplished in 90°C temperatures
and high humidity. The quality of data produced by the
field GC-MS (e.g., RF, dynamic range, and field study)
was shown to be statistically equivalent to commercial
laboratories. The instrument provides for easy
identification and accurate quantitation of the many
HSL compounds normally requested of laboratory
methods during comprehensive site investigations.
ACKNOWLEDGEMENT
The authors thank Bruker Instruments for use of the
GC-MS instrument and Chevrolet Blazer as well as the
many technical discussions related to this work. The
authors also thank Jack Duggan, DEQE, for coordinating
the site investigation. The authors greatly appreciate the
help and interest of many other DEQE personnel for
allowing us to participate in this and other
environmentally important site investigations. Without
their help it would have been nearly impossible to gain
access to appropriate sites.
REFERENCES
1. Robbat, A., Jr. and Xyrafas, G. "Evaluation of
a Field-Based, Mobile, Gas Chromatograph-
Mass Spectrometer for the Identification and
Quantification of Volatile Organic Compounds
on the EPA Hazardous Substance List", Field
Screening Methods for Hazardous Mobile Waste
Site Investigations, First International
Symposium, October 11-13, 1988, Las Vegas,
Nevada.
2. Trainer, T.M. and Laukien, F.D. "Design and
Performance of a Mobile Mass Spectrometer for
Environmental Field Investigations", Field
Screening Methods for Hazardous Waste Site
Investigations, First International Symposium,
October 11-13, 1988, Las Vegas, Nevada.
©*.
O t
rw—i
KOMFLAM
Figure 3. Interior/exterior placement of church
monitoring solid gas probes and wells.
Table 1. Soil Gas Results from Probes outside of Church.
Total BTEX Concentration (ppm)
Location (J)
9
10
11
12
13
14
15
Outside
Ambient Air
GC
9.0
5.0
5.0
5.0
2.5
5.0
2.5
4.5 refusal
3.5
ND
ND
960
50
ND
ND
ND
ND
368.4
GC-MS
ND
0.314
0.458
ND
0.178
0.167
ND
HNU
NR
NR
R
R
NR
NR
NR
R
R
NR
The GC analysis was performed on-site by a DEQE
contractor, February 1 and 2, 1988.
The GC-MS analysis was performed on-site by the
authors, June 29, 1988
The HNU meter was calibrated to toluene and was used
to determine the relative BTEX response before samples
were collected for GC-MS analysis.
Indicates samples were not collected.
483
-------
Jin
I.
2. 1.4-dlflwirobraMM
3.
4.
3.
I. O-zylnc
».« l.t I.* !.• 4.* I.*
FiguiE 4. GC/MS of VOCs collected from J-9.
nut
•.*
*.t
Figure 6. GC/MS of the VOC ambient air collected
from within the church.
Table 2. Soil Gas Results from Interior of Church
Total BTEX Concentration (ppm)
Location
-1
-2
-3
-4
-5
1-6
SUMP
J-l
GC
16,271
9,641
1,217
440
27.4
44
HNU
160
155
4
12
30
20
2
GC-MS
199.9
152.5
0.4
0.3
16.7
1.6
5.1
Inside Ambient Air 2.4
1.0
1.0
GC analysis performed on-site by DEQE contractors, February
1 and 2, 1988.
HNU analysis performed by DEQE contractors April 7, 1988.
GC-MS analysis performed on-site by authors, June 29, 1988
All analytical instruments calibrated for benzene, toluene,
ethylbenzene and isomeric xylenes.
Indicates samples were not collected.
Figure 5. GC/MS of VOCs collected from (1-2).
484
-------
REFLECTANCE SPECTROSCOPY (0.2 to 20 /m) AS AN ANALYTICAL METHOD FOR THE
DETECTION OF ORGANICS ON SOILS
Trude V.V. King and Roger N. Clark, U.S. Geological Survey,
P.O. Box 25046 MS 964, Denver, Colorado, 80225, USA
ABSTRACT
Reflectance and transmission spectroscopy (2 to
20/jm) has been used as method for the
identification of organic solvents on a soil.
Fundamental C-H stretch absorptions (~3.4>jm) and
the overtones of these absorptions, which occur
at —1.7/im, can be used to uniquely characterize
the organic solvents. Spectral differences exist
between toluene, benzene, trichloroethylene and
gasoline in transmission spectra and in
reflectance spectra when they are mixed with a
Na-rich clay. Key words: reflectance
spectroscopy; organic solvents; transmission
spectroscopy.
INTRODUCTION
Reflectance spectroscopy is a quick and
inexpensive method for identifying organic
contaminants in both surface and sub-surface
environments, for example:
1. Mapping the aerial distribution of
potentially hazardous spills.
2. Identifying the potential sources of
hazardous and/or toxic materials.
3. Monitoring toxic sources as a function of
time, e.g., during and after a water
rejuvenation program.
Laboratory experiments using mixtures of clay
standards and organic contaminants have shown
that it is possible to identify both the clay and
the associated organic contaminant by reflectance
spectroscopy methods. Samples of neat clays and
organic solvents have been measured in both
transmission and reflectance modes (0.2-20 /jm) to
characterize fundamental and overtone vibrational
absorptions as well as crystal field absorptions.
These spectra are entered in the U.S.G.S
reference sample data base and used as comparison
standards for unknown mixtures. Attempts are
underway to estimate abundances of organics in a
clay mixture using a radiative transfer model by
deriving absorption coefficients and index of
refraction for the materials of Interest. These
techniques could help evaluate environmental
problems by in-situ measurements, laboratory
measurements, and potentially with airborne
imaging spectrometers.
DISCUSSION
Most minerals, organic and chlorinated solvents
show diagnostic absorption features in the
reflected solar radiation part of the spectrum
(0.3 to 2. 5/im) and in the infrared part of the
spectrum (3 to 200 /im) . These absorption
features can be used to determine the mineralogy
of the surface, and in a number of cases, the
trace elements present in those materials. In
addition, spectroscopy can identify organic and
chlorinated solvents which may be associated with
a specific mineral or mineral assemblage.
Laboratory calibration investigations included
clay/organic mixtures in which the organic phases
were characterized by transmission measurements
and then mixed with a host-clay and characterized
by reflectance measurements. Transmission
spectra of benzene, methanol, toluene,
trichloroethylene, and gasoline show numerous
absorptions in the near infrared (Figure 1).
These absorptions are different in position and
shape from mineral and water absorptions, making
their detection easy. Because the intensity of
absorption bands vary with wavelength (i.e.
fundamental absorptions are stronger than
overtones) detectability levels are a function of
the absorption feature selected for analysis.
Thus, detectability limits can range from parts-
per-million (ppm) to a few percent.
Simulations of likely environmental problems
involved mixing the solvents with
montmorillonite. There was no additional sample
preparation and spectral measurements took less
than fifty minutes per sample. Spectrometers
with detector arrays could reduce the measurement
time to less than one second. Figure 2 shows the
near-infrared reflectance spectra for the neat
Na-montmorillonite sample (SWy-1) and the
mixtures of montmorillonite and solvents. In each
example the clay was mixed with large quantities
of the solvent for illustrative purposes, however
detection levels in this spectral region are on
the order of 1 wt%, when computer analysis is
employed. In samples of montmorillonite mixed
with gasoline, gasoline bands are prominent
between 1.65-1.8 /im, and near 2.3 /jm. The benzene
bands are very strong in the 1.7, 2.16 and 2.46
/jm regions when mixed with montmorillonite at the
33 and 14 wt% levels. The toluene absorptions
are very strong, at 1.15, 1.68-1.7, 2.2, and
485
-------
2.35-2.45 pm, in mixtures of 28, 14 and 7 wt.%.
The absorptions from trichloroethylene are very
strong in the montmorillonite at 24 and 39 wt.%.
The absorptions in the 1-2.6 pm region are
combination overtones. The fundamentals, some of
which occur in the 3-5^m region, are 30 to 100
times more intense.
Figure 3 shows the mid-infrared spectra of
gasoline (super-unleaded), measured in
transmission mode, and the reflectance spectra of
montmorillonite (SWy-1) and mixtures of
montmorillonite and gasoline. The prominent C-H
stretch vibration between approximately 3000-2800
cm wavenumber (-3.33-3.5 /jm) is used, in this
example, to illustrate the capabilities of
spectroscopy in the detection and identification
of organic/clay mixtures. As the amount of
organic contaminant decreases the strength of the
fundamental band also decreases. However, based
on these measurements in the 3.3-3.5-pm
wavelength region, detection limits might be as
low as 10 to 30 ppm.
APPLICATION
Mapping the aerial distribution of potentially
hazardous spills using airborne reflectance
spectroscopy techniques will require a timely
response and instrumentation with high spectral
resolution and signal to noise. Slow rates of
penetration of the organic material into the
underlying soil should in most instances, hold
the organic for a period of time sufficient for
remote detection. However in some instances, a
quick response to assess spills would be
necessary before the organic material evaporates.
Advantages of mapping the aerial extent of a
spill using airborne remote spill as well as
providing a synoptic view which is difficult too
obtain by ground-based observations.
In-situ methods of spectral characterization can
use the remote, airborne-type of spectroscopic
study mentioned above either independently or in
conjunction with laboratory measurements.
Laboratory measurements, similar to the results
presented in Figures 1-3, of samples collected
from specific sites could aid in determining the
direction of movement of a toxic plume. By
collecting samples in the field and returning
them to the laboratory in sealed sample holders,
the integrity of the sample would not be
compromised. Laboratory measurements using
currently available technology can more
accurately determine low concentrations of
organic contaminants in soils than instruments
now available for synoptic field measurements
because of greater spectral resolution. Maps of
the concentrations of organic contaminants in
soils would provide information on the source
(type of organic or chlorinated solvent) and the
direction of movement. An added benefit of the
spectroscopic method of mapping organic/solid
interactions, is the possibility of detecting and
characterizing secondary changes in the
contaminant as a result of chemical or physical
reactions with the surrounding material.
Down-hole spectroscopic measurements would prove
beneficial for the characterization and
monitoring of organic/solid interactions below
the Earth's surface. Although instrumentation
does not currently exist for these types of
measurements, infrared fiber optics could be
used. Down-hole spectroscopy will provide in-
situ measurements for non-recoverable, subsurface
samples. This method of measurement would be
valuable for monitoring the concentration of
organics contaminants in soils as a function of
water rejuvenation. Because much of any spill is
caught and held by the soil, it would be
beneficial to monitor the changes in organic
contaminant concentrations as a function of time.
1.5 2.0 2.5
WAVELENGTH (urn)
Figure 1. Transmlttance spectra (O.lmm thick) of super
unleaded gasoline,benzene, toluene,and trichloroethylene In
the 0.9-3.0-fin wavelength region show the characteristic
overtone and combinations features of the solvents. Of
particular Interest is the 1.7-^m wavelength region whlh Is
the first overtone of the fundamental C-H stretch which
occurs near 3.4/JB.
486
-------
- SW»-1 t SUPER UNLCADED
1. 0
1.5
2.0
2. 5
3.0
WAVELENGTH (urn)
Figure 2. Reflectance spectra of SWy-1 (Na-montmorillonlte)
and SWy-1 mixed with super unleaded gasoline, benzene,
toluene and trlchloroethylene. In the mixed spectra, it is
possible to Identify spectral characteristics of the clay
and the mixed contaminant. Although all the mixed spectra
have adsorptions at similar wavelength positions( eg.
1.7/im), absorptions features related to contaminants can be
distinguished from one another based on exact position and
intensity. Absorption features which are correlated, to some
degree, with the presence of organics in the clay are marked
with bars.
487
-------
WAVENUMBER (cm~1)
Figure 3. Transmittance spectra of super-unleaded gasoline
(top spectra) shows the vibratlonal absorptions of gasoline
from 2 to 25 pm. The strongest adsorptions In the 3.6-fjm
wavelength region is attributed to the C-H stretch in the
gasoline (shown by bar). The bottom spectra is the
reflectance spectra of SWy-1 (Na-montmorillonite) without
added gasoline and shows the OH-absorption region (-2.7/im)
characteristic of the clay. The two intermediate spectra
illustrate the mixture of -12 wt% and 1 wt% gasoline in the
clay. The fundamental C-H adsorption is detectables in both
Intermediate spectra and illustrates the usefulness of
reflectance spectroscopy in identifying clay/organic
mixtures.
488
-------
FIELD USE OF A MICROCHIP GAS CHROMATOGRAPH
.R.W. Sherman, T.H. McKinney, Institute for Environmental Studies LSU Baton
Rouge, LA,
M.F. Solecki, National Oceanic and Atmospheric Administration Seattle WA
R.B. Games, U.S. Coast Guard R&D Center Groton CT
B. Shipley, U.S. EPA Region 9, San Francisco, CA '
When responding to chemical emergencies and
evaluating waste sites it is important to have accurate
information about the identities and quantities of the
chemicals involved. It is advantageous to have real
time information so that response personnel and
project managers may make timely decisions to
accelerate mitigative actions, cleanup, and to
adequately protect workers and the public.
A field deployable microchip gas chromatograph
linked to an external personal computer has been
used to evaluate the nature and extent of
contamination at a number of hazardous waste sites
and chemical emergencies. The system has been
used to analyze air and water samples as well as
identify the contents of drums containing volatile
liquids.
The instrument used for these field deployable
analyses was a MichromonitorSOO (Microsensor
Technologies, Inc., Fremont, CA) interfaced with a
Macintosh personal computer (Apple Computers,
Cupertino, CA). The MichromonitorSOO (M500)
contains four high resolution microchip gas
chromatographs (u.GC) with capillary columns and
thermal conductivity detectors. Vapor samples are
introduced with an internal vacuum pump, and the
sample is routed to the appropriate p.GC. Each (J.GC
has a different length capillary column and is used to
analyze volatile compounds with a predefined range
of volatility. The M500 is designed to be portable
and has the ability to identify several dozen
compounds which have data stored in its internal
ROM library (1). The device as it comes from the
manufacturer will detect only those vapors which
were preselected. Other compounds could be
present but would go undetected if data are not
contained in the library. This limitation is overcome
by the use of user-friendly software which has been
developed to run on a Macintosh computer. The
connection between the M500 and the Macintosh
may be a direct link with wires, telephone lines, radio
frequency modems, or any combination of these.
The Macintosh software contains three modes of
operation. The first mode allows access to all of the
MSOO's standard features by having the Macintosh
emulate the control panel of the M500. This mode is
particularly useful for troubleshooting and
debugging.
The second mode of operation has the M500 dump
raw digital data to the Macintosh which converts the
data into high resolution gas chromatograms. This
allows an evaluation of peak shapes, elution profiles,
and instrumental parameters such as drift.
The third mode of operation is the most complex, but
it provides more useful anlytical data. The M500
analyzes a sample and stores the temperature, data
on the retention times, and peak areas for all peaks
in the chromatogram. These data are transmitted to
the Macintosh. The Macintosh uses the temperature
and the retention time of the air peak to calculate
expected retention times for normal hydrocarbon
standards as if they had been analyzed with the
sample. Using these calculated hydrocarbon
retention times the software then calculates modified
Kovats retention indices for all peaks in the
chromatogram (2). Identification is made by
comparison of the retention index of the unkown
compound with library values which are a part of the
Macintosh software.
Since the M500 has a small sample loop and uses a
thermal conductivity detector, the detection limits for
most compounds are about 10ppm. Since this
detection limit is too high for most environmental
applications, a sample preconcentrator has been
developed to lower detection limits by factors of 100
(3). This device is used to concentrate large samples
of volatile organic compounds (200-500 ml) on
ambient temperature sorption traps and desorb the
organics at 250°C into smaller volumes (2 mL). The
traps use a combination of Tenax GC and
Spherocarb as the sorbents. In addition to
concentrating air samples this device may be used to
concentrate volatile compounds from water samples
using a purge and trap technique.
489
-------
This analytical system has been used at a number of
waste sites, emergency removals, and fires involving
hazardous materials. It has been used by LSD
personnel, Members of the Coast Guard Atlantic
Strike Team, and Pacific Strike Team, EPA Region 9,
and the EPA-Environmental Response Team. Brief
descriptions of some of these field uses are given
below.
In April 1986 the M500-Macintosh system was used
to analyze drums at a Superfund site in Utah. By
using the M500 and a HAZCATT procedure the
drums were prepared for disposal with a reduced
need for laboratory analysis.
In November 1986 the system was used to sample
the contents of underground storage tanks at an
abandoned latex manufacturing plant in New Jersey.
The results given by the M500 in real time were
consistent with the results of GCMS analyses which
were performed off site and had a turnaround time of
several days.
In February 1987 the system was used at a chemical
storage facility in Maryland. The M500 was used to
verify that drums were labelled correctly. It was not
necessary to send samples for laboratory analyses,
and an estimated $50,000 was saved.
In the spring of 1987 the M500 was used at a fire
located at a rubber manufacuring facility in Virginia.
Some firefighters were using the runoff water to wash
themselves and their clothing. The M500 detected
benzene, toluene, and other hydrocarbons in the
headspace of the runoff water. The operator of the
M500 informed the fire chief, and recommended that
the fireman stop rinsing in the runoff, and any firemen
who had been in the runoff water go to the hospital
for observation. In the ambulance on the way to the
hospital several firemen broke out with a rash.
In August 1987 the M500 with a concentrator was
used at a Superfund site in Washington. Although
primarily used for air monitoring, the M500 was also
able to identify contents of unlabelled drums which
were uncovered during the excavation process.
In the fall of 1987 the M500 was used at a drum
removal action in California. The M500 yielded data
which was consistent with labels and HAZCATT
information. The drums were able to be disposed of
with no further analysis, resulting in a considerable
savings.
In winter of 1987 the M500 was used for air
monitoring at two municpal landfill fires in Hawaii.
The M500 was used along with other field
instruments, including another field GC. The MSOO's
results were consistent with all of the other
instruments, but it provided data much quicker and
was found to be easier to set up and maintain.
The previous examples, and others, have
demonstrated the value and usefulness of these high
resolution field deployable analytical devices for a
wide variety of field applications.
REFERENCES
1. Kovats. E.. Helv. Chem. Acta41.1915.. 1958.
2. Wohltjen, H., "Chemical Microsensors and
Microinstrumentation", Anal. Chem. 56, 87A-
103A, 1984.
3. Overton, E.B., McKinney, T.H., Steele, C.F.,
Collard, E.S., Sherman, R.W., "Rapid Field
Analysis of Volatile Organic Compounds in
Environmental Samples", Proceedings of the
Third Annual Symposium on Solid Waste
Testing and Quality Assurance. USEPA, July
1987, Washington, D.C.
490
-------
RAPID ASSESSMENT OF PCB CONTAMINATION AT FIELD SITES USING
A SPECIALIZED SAMPLING, ANALYSIS AND DATA REVIEW PROCEDURE
William W. Freeman
Principal Scientist
Roy F. Weston, Inc.
West Chester, PA
Joel Karmazyn
Senior Project Scientist
Roy F. Weston, Inc.
West Chester, PA
ABSTRACT
An operational procedure was needed to
accurately assess the extent of PCB con-
tamination at field locations in as short
a time and as economically as possible.
The procedure which was developed includes
a sampling grid lay-out and expansion
scheme, a rapid extract ion/analysis pro-
cedure and a data validation step. Most
results are "turned around", including the
grid expansion-decision making process in
48 hours. Key factors include the rapid
extraction procedure and the use of a
laboratory dedicated to the analysis of
just these samples.
The objective was not only to obtain
accurate test results in a timely manner,
but also to make efficient, economical use
of the field team's time. The entire
"turn-around" process is described here,
from the original identification of
potential PCB sources at the site, through
sampling procedures, transfer to the
dedicated lab, analytical procedures and
data review.
When site assessment begins, a standard
sampling grid pattern is used. Grid
expansion procedures are described in
which each PCB detection point will
generate 2 to 4 new sampling points, de-
pending on their location in the grids.
After reviewing and validating the lab
data on a daily basis, a decision is made
whether to extend the grids, stop sampling
in a particular direction, resample
specific areas, etc.
While evaluating analytical procedures,
several proven and accepted extraction/
analysis techniques were examined before
choosing an orbital shaker extraction
method and gas chromatography technique
which would provide the best combination
of speed and accuracy. The procedure is
described wherein 10 gram samples of soil
are extracted for 30 minutes and then
tested for PCB content. Standard QC and
QA procedures are employed, including
field blanks, matrix spikes, duplicate
samples, etc. This procedure can be used
for specific action levels such as 10 or
25 ppm etc., as well as general site
characterization.
Sample transport time, analysis, and data
review are all coordinated between the
field team and home office so that no time
is lost. For example, while waiting for
analytical data from a particular on-site
grid area, the field team can be sampling
other on-site or off-site areas, extending
grids from areas previously found con-
taminated, etc. One of the chief ad-
vantages of this sampling and analysis
procedure is that it facilitates the rapid
characterization of plumes of surface
contamination.
INTRODUCTION
Once a site of known or potential PCB
contamination has been identified at a
field location, it is desirable to obtain
analytical results for soil samples as
soon as possible. This aids in making
preliminary decisions regarding the extent
of contamination and the type of remedial
action which may be required. Project
management personnel need information fast
in order to make decisions on depths for
soil borings, directions of sampling grids,
e t c .
A procedure was developed which not only
enables project personnel to obtain
accurate test results in a timely manner,
but also makes economical, efficient use
of the field teams' time.
The effectiveness of this procedure is
based on the coordination of many aspects
of the operation. These include the
sampling team, the delivery personnel
(sample "runners"), laboratory receipt
custodians, lab analysts and the data
review personnel. A laboratory with
personnel and equipment dedicated to the
analysis of just these samples is used.
491
-------
When properly coordinated, this unique
operational procedure allows the field
team sufficient time to sample other areas
of interest while waiting for results irom
previous samples. Management can quickly
acquire and assess the results and relay
instructions on items such as grid ex-
pansions to the field team. The entire
procedure, including the decision making
process, can be accomplished within 48
hours (or less) from the time that a set
of samples is taken at the field site.
Procedures such as pre-labeling bottles,
pre-printing Chain-of-Custody forms and
the use of a dedicated delivery person to
transport samples between the site and the
laboratory help in the rapid and efficient
assessment of the extent of the PCB con-
tamination .
PROCEDURE SUMMARY
Once the potentially contaminated area has
been identified and the field sampling
team with it's equipment is set up at the
site, the following general sequence of
events takes place.
1. Survey site; lay out initial sampling
grids.
2. Take initial sets of samples.
3. Preserve and package samples for same
day or overnight delivery to the
laboratory.
4. Laboratory personnel perform rapid
extraction procedure and gas chro-
matography analysis at the laboratory
dedicated to the analysis of just
these samples.
5. Field personnel continue sampling
activities in other areas while
waiting for initial test results.
6. Data review personnel perform
validation and "reasonableness"
checks on lab data.
7. As validated test results become
available (usually within 48 hours),
project management personnel make
decisions regarding grid expansions or
other additional sampling procedures.
8. This sequence should continue until
the site assessment procedure is
complete.
SPECIFIC PROCEDURES
1. Soil Sampling Grids
The soil sampling grid system should be
established by surveying selected
sampling locations, using fixed objects
s reference points if possible. Other
sampling location/grid nodes should be
surveyed so that the overall grid system
is well defined. Survey stakes should be
set up at the four corner grid locations
that surround a potential source of
contamination and then a standard 25-foot
square grid system should be put in place.
Once the grid has been established, surface
soil samples should be collected from
each corner of the 25-foot square and
(optional) a composite sample from
within the square, approximated by the
center of the square.
2. Sampling Activities
Standard sampling equipment and procedures
should be used. The equipment includes
stainless steel augers, scoops and
trowels, mixing bowls for homogenizing
composites, etc. Clean sample containers
of the proper size and composition should
be used. Pins or wire/plastic flags
should mark all sample locations.
The actual sampling procedure, including
type, number and depth of samples should
be delineated on a site specific basis.
All QC samples such as duplicates, blanks,
etc., should be included in the sampling
plan and strictly adhered to by the field
personnel.
3 . Packaging/Shipment of Samples
This is one of the more important phases
of the rapid turn-around site assessment
procedure. It is crucial that samples
are identified, packaged, preserved and
shipped to the dedicated laboratory in
the proper fashion. Complete identi-
fication information must be recorded on
the sample container label as well as on
the Chain-of-Custody sheets (see example
in Figure #4), including exact location
of the sample as well as the tests that
are to be performed at the laboratory.
All standard practices involving sample
integrity, safety, D.O.T. regulations,
etc., should be followed. These include
items such as custody seals on jars,
placing ice in the shipping container for
sample preservation and using the correct
labels for shipping any hazardous (PCB)
materials.
In order to insure that the samples arrive
at the dedicated lab as soon as possible,
a "runner" should be utilized whenever
possible both to deliver samples to the
lab and bring new materials to the field
site. This procedure, using a dedicated
van and driver, is practical and efficient
even if the field site is up to 100 miles
from the lab. It enables samples to be
delivered to the laboratory on the same
day that they are taken, so that the
analytical process can begin as soon as
492
-------
possible. At least one round trip per
day can be made and possibly more,
depending on the distance involved.
In lieu of a dedicated "runner", an over-
night delivery service can be utilized.
This is not as efficient as a dedicated
driver and van, but will get the samples
to the lab late on the morning following
the sampling events of the previous day.
4. Continuation of Sampling Activities
While waiting (up to 48 hours) for results
on a particular set of field samples,
other activities such as additional
surface soil sampling, soil borings,
groundwater or surface water sampling,
etc., can be carried out. If these types
of events are scheduled at least 2 days
in advance, then continuous, efficient
use can be made of field personnel and
sampling teams.
5. Data Review
When results become available from the
dedicated laboratory, they should
immediately be delivered to the project
personnel who perform the data review
process. This review of the "quick-turn-
around" data cannot be as complete as, for
example, the validation of a complete
Tier II data package. However, since the
data will be used for grid expansions and
other field related activities, it is
important to have a review step in place.
The review should include items such as
matching the Chain-of-Custody sample IDs
with the data summary IDs, checking
dilution factors, reviewing QC data (e.g.
matrix spike/matrix spike duplicate), etc.
If discrepancies are discovered, a more
detailed review of the lab data including
the chromatograms can be conducted.
This review step should also include a
"reasonableness" check by someone
familiar with the specific site and the
sampling locations. This check is as
important as the analytical data review
even though it is only a quick check of
test results-vs-sample locations. Since
field sites may often yield dozens of
samples, depending on grid expansions,
it is necessary to perform this "reason-
ableness" check in order to help assure
that test results reflect the true
location of PCB contamination areas.
The analytical review and "reasonableness"
check should take no more than one hour
even for a batch of up to 20 samples. It
is convenient to use a checklist such as
the one shown in Figure #5.
6. Decision Making
As soon as validated test results become
available, they should be delivered
simultaneously to the site supervisor and
the project manager. "Delivered" means
at least by telephone and preferably in
hard copy form. This can be accomplished
by use of a telefax machine in a mobile
office at the site or, when practical,
delivery of reports to the site by the
dedicated runner during a trip to the site.
The data should then be examined by the
site supervisor and project manager and
decisions made as soon as possible re-
garding grid expansions. Site maps or
prints should be available at both the
home office and field site to aid in
identifying exact field locations.
This sequence of sampling analysis -
decision making grid expansion - sampling.
etc., should be continued until the
sampling plan is completed or terminated
due to grid overlap.
7. Material Preparation
In a site assessment/samp 1 ing operation,
one of the most time consuming activities
involves the filling out of documents such
as Chain-of-Custody forms and properly
identifying the numerous sample containers.
A key to conducting a rapid, efficient,
economical operation such as this is the
use of pre-printed Chain-of-Custody forms
(see sample, Figure //2) and pre-labeled
sample j ars.
Once the initial sampling grids are
defined, forms can be pre-printed and jar
labels filled out for these locations.
Any minor variations in the sampling
scheme can simply be crossed off or
corrected on the forms, as necessary. As
analytical data is received from the lab,
validated, and grid extension decisions
are made, the next batch of Chain-of-
Custody forms and jar labels can be
prepared. The dedicated runner can also
assist in bringing forms and labeled jars
from the home office.
THE DEDICATED LABORATORY
During this type of remedial investigation
or site assessment, one of the most
expedient ways to generate rapid, accurate
analytical data is to make use of a
"dedicated laboratory". This type of
facility can typically analyze twenty (20)
or more soil or sediment samples per day
for PCB content, and provide results in
approximately forty-eight (48) hours, or
less, from the time the samples are taken.
1. Analytical Procedures
Analyses of PCBs for screening purposes
can be conducted at prescribed char-
acterization levels such as 5, 10, or
493
-------
25 ppm in soil samples. The laboratory
can use a rapid extraction technique1
similar to the one developed by the U.S.
EPA Releases Control Branch. Other
extraction procedures including the
soxhlet method were evaluated at the
WESTON Analytics Laboratory in Lionville,
PA, but the referenced rapid extraction
procedure provided the best combination
of speed, low cost and accuracy at
characterization levels of as low as 5 ppm
for PCBs in soil. The chosen procedure
can include an acid clean-up of the
extract and analysis by gas chromatography.
2. Method Summary
Ten (10.0) grams of sample is placed in a
250 ml screw top Erlenmeyer flask or glass
jar. Eighty (80) ml of hexane and twenty
(20) ml of methanol is added directly to
the flask or jar. The container is placed
on a gyrotory shaker unit positioned
behind a safety shield in a fume hood and
then is shaken at 400 rpm for thirty (30)
minutes. The flask is then removed to a
stationary area, where the suspended
solids and particulate matter is allowed
to settle for approximately thirty (30)
minutes.
The hexane extract is removed using Pasteur
pipets while avoiding agitation of the
so 11/sediment layer or particulates in the
bottom of the flask. The sulfuric acid
clean-up is performed at this point. The
analysis for PCB content is then conducted
by electron capture gas chromatography
under conditions similar to those shown in
Appendix A.
A separate test for percent moisture
(drying at 105°C) can be run concurrently.
Results (PCB content) should be expressed
on a dry weight basis.
& Quality Control
3 . Quality Assurance
Checks
The daily quality of the analytical data
generated by the dedicated laboratory is
monitored by the analysis of method blanks,
fortified method blanks, etc. External
(field) quality control samples typically
include field and trip blanks and field
duplicates .
The purpose, description and frequency of
laboratory QC samples typically used in
the dedicated laboratory facility at the
WESTON Analytics Laboratory are shown in
Append ix B.
Quality Control samples should represent
a minimum of 5% of the field samples.
Fortified samples are spiked at WESTON
Analytics with a commercially available
Aroclor 1254 PCB standard. The blank
matrix consists of a 10 gram sample of
sodium sulfate. Table I represents a
typical statistical evaluation of lab QA
data. The average percent recovery for
matrix spikes and matrix spike duplicates
was 100.5%. As an additional QA step, and
to help confirm the accuracy of the rapid
extraction method, approximately 10% of
the samples can also be analyzed for PCBs
using more conventional methods. WESTON
Analytics routinely performs this type of
program for PCB contaminated soils. A
very good correlation has been obtained
using both the rapid extraction method and
the SW-846 method. A comparison of
typical PCB analysis results obtained
by both methods is shown in Table II.
SUMMARY
Site assessment sampling activities for
known or potentially contaminated soils
and sediments can be conducted in a rapid,
efficient and accurate manner if all
aspects of the operation are coordinated
properly. The time (and dollars) spent
by having a sampling team and mobile
equipment in the field can be minimized
by use of:
o A field team that has pre-planned
their activities.
o A rapid delivery system (dedicated
"runner") for samples and
materials.
o The use of pre-planned paperwork
and sample containers.
o A dedicated laboratory for 48-hour
(or less) turnaround of analytical
result s.
o Coordination between the field
supervisors and project management
personnel regarding data review
and decision making.
The major advantage of this type of pro-
gram is the ability to make timely
decisions regarding items such as the
direction of a contamination plume,
construction and extension of sampling
grids and managing the field team's time.
If the sample delivery system, dedicated
laboratory and decision making process all
function as designed, the field team
should have no "down" time until they are
waiting for what is presumed to be the
last batch of lab results. Even then, the
time can be used for activities such as
decontamination of equipment, updating
maps, etc., while waiting for the "final"
set of analytical data.
REFERENCES
•*- Emergency Response Analytical Methods
for Use On Board Mobile Laboratories.
Draft Document, June 1987.
494
-------
Grid Layout
u.
PJ
\+2S Ft-M
A Round ! Sample Location
FIGURE #1
Identification of Grid Sample Locations
r
25'
\ /
Interior
/ \
Perimeter
Corner
Perimeter
Inner
-25'
FIGURE it2
495
-------
Expansion of Grid Based on Round 1 Test Results
I
I
I
-cfc-
I
I
I
I
-+•
A Round 1 Sample Location
® Round 1 "Hit" for PCB
0 Proposed Round 2 Sample Location
Received By .
Date_
FIGURE #3
Custody Transfer Record/Lai? Work Request
Assigned to _
Client Contact-
Phone
RFW Contact
Date Due
Project Number .
SAMPLE IDENTIFICATION
ANALYSES REQUESTED
Sovlllto.
0010
0020
0030
CHwHIONo.
ABC Co.
ABC Co.
ABC Co.
D~cnpto.
Grid Poinc 13-x
Grid Point 14-y
Grid Point 15-z
UMrll
S
S
S
MiCoMcM
22 Aug. 88
22 Aui. 88
22 AUK. 88
C«Atafew/PTM«m«*«
Glass Jar/Ice
Glass Jar/Ice
Glass Jar/Ice
PCB
X
X
X
1. Please run by Rapid Fjttraction/G.C. Method
2. Identify PCB Isoner - 1248, 1254, etc.
3. Please phone and Fax data to Field Site & Home Office A.S.A.P.
FIGURE #4
496
-------
DATA REVIEW CHECKLIST
Site Name:_
Batch ID f_
Task
Delivery Due Date:.
Reviewer:
Date:
Chain of Custodv Matches Samole IDs
Samole Holdina Tines Met *
Case Narrative Flagged Samples
Exceeding Holdina Time *
Data Summary Sheet Matches Chronoloov
OC Data Included
Matrix Blanks Clean
Matrix Spike Recovery Acceptable
ADDrooriate Detection Limits
J - Values Correct/Missina
B - Values Correct/Missina
O - Values Correct/Missina
MS and MSD Recoveries Reported
In Percentaae
Case Narrative Describes Analytical
Difficulties *
Typos
Other Comments/Reasonableness Check:
*Where applicable
Date Submitted To Project Manager:.
Submitted For Revisions:
Response Due Date
Returned To Reviewer:
Date Of Package Approval:.
Initials:
FIGURE #5
497
-------
TABLE I
STATISTICAL EVALUATION OF TYPICAL OA DATA
PC Parameter No. of Samples Results
Matrix spike, matrix 40 Avg. % Recovery = 101%
spike duplicate (Std. Dev. = 16.5)
Blank spike/blank 25 Avg. % Recovery = 103%
spike duplicate (Std. Dev. = 12.9)
Duplicate samples 40 Relative % Difference =
24%
NOTE: The relative percent difference (RPD) is a measure of the
analytical precision of the laboratory procedure. The
Contract Laboratory Program has established a guideline of
50 or less RPD for the analysis of PCBs in soil The RPD
of 24 for this rapid extraction method is well within the
guideline.
TABLE II
Comparison of Analytical Results for Samples Extracted by the
Rapid Extraction Method and EPA SW - 846 Method
PCS Content (mg/kg)
Sample No. Rapid Extraction SW-846
1 3.0 2.5 J
2 2.1 5.3
3 23 27 J
4 27 13
5 220 37
6 .61 J .66 J
7 .56 J 1.8 J
8 36 22
9 6.8 273
10 ND 3.5 J
11 2.3 2.0 J
12 3.4 2.8
13 18 17
14 4.0 110
15 5.0 7.7
16 120 120
17 14 9.3
18 ND .17 J
19 17 12
20 1400 980
ND - Not Detected
J - Present below detection limit
498
-------
APPENDIX A
TYPICAL GAS CHROMATOGRAPHIC CONDITIONS
Analytical Columrun: *
Carrier Gas:
Carrier Flow Rate:
Detector:
Detector Temperature:
Injector Temperature:
Oven Temperature:
Program Rate:
Time of Analysis:
1.5% SP 2250/1.95% SP 2401 on 100/120
Supelcoport 6 ft x 4 mm glass column
(mixed phase support)
Argon:Methane (95%:5%)
40 ml/min
Electron Capture (BCD)
260«>C
225°C
180°C (Dependent on PCB Aroclor)
Isothermal
Variable (Dependent on PCB Aroclor}
* The mixed phase column recommended for PCB analyses should be
packed and conditioned in accordance with Method 7.3
(Instructions for Packing and Conditioning Glass Analytical
Columns for Gas Chromatography Instruments) and performance
tested according to the procedure and specifications described
in Appendix C of Method 7.4 (Protocol for Performance Testing
Packed, Conditioned, Glass Analytical Columns for Gas
Chromatography Instruments).
APPENDIX B
Purpose and Frequency of
Laboratory Quality Control Samples
(Typical for PCB in Soil Analyses)
Sample Type
Purpose
Frequency
Method Blank
Monitor Laboratory
Contamination
Fortified Method Blank Laboratory Accuracy
Field Blank
Trip Blank
Monitor Sample
Collection
Monitor Contamination
during Field
Operations
Daily with each
preparation batch
of 20 or fewer
samples
Daily with each
preparation batch
of 20 or fewer
samples
Daily or once
for each sample
collection method
At least once per
sampling episode
Fortified Sample
Duplicate Sample
Monitor Bias due to
Sample Matrix
Monitor Precision
As required by
analytical
protocol per
batch of 20 or
fewer samples
As required by
analytical
protocol per
batch of 20 or
fewer samples
499
-------
First International Symposium
Field Screening Methods for
Hazardous Waste Site Investigations
Participants' List
Prabha Acharya
Arizona State Laboratory
1520 W. Adams
Phoenix, AZ 85007
(602) 255-1188
James H. Adams, Jr.
USEPA, Region V
536 South Clark Street
Chicago, IL 60504
(312) 353-9317
Albert H. Adelman
ENSR Consulting and Engineering
696 Virginia Road
Concord, MA 01742
David B. Agus
University of Pennsylvania School of Medicine
4418 Spruce Street, Apt. E-l
Philadelphia, PA 19104
(215) 222-2324
E. N. Amick
Lockheed ESC
1050 E. Flamingo Road, Suite 120
Las Vegas, NV 89199
(702) 734-3287
William C. Anderson
IT Corporation
312 Directors Drive
Knoxville, TN 37923
(615) 690-3211
J. Andrade
College of Engineering MEB
University of Utah
Salt Lake City, UT 84112
(801) 581-4379
W. F. Arendale
Chemistry Department
University of Alabama
Huntsville, AL 35899
(205) 895-6473
Neil Arnold
Center for Micro Analysis
University of Utah
214 EMRL Building 61
Salt Lake City, UT 84112
(801) 581-8431
Russell Arnold
Food and Drug Administration
5600 Fishers Lane
Rockville, MD 20857
(301) 443-2872
Tad Bacon
Environ. Analytical Systems, Inc.
1400 Taylor Avenue
P.O. Box 9840
Baltimore, MD 21284-9840
(301) 321-5133
Harold Balbach
U.S. Army CERL
P.O. Box 4005
Champaign, IL 61820-1305
(217) 373-7251
Karen Bankert
USEPA Region IX (P-3-2)
215 Fremont Street
San Francisco, CA 94105
(415) 974-8856
John Barich
USEPA, Region X
1200 6th Avenue
Seattle, WA 98101
(206) 442-8562
M. H. Bartling
Lockheed EMSC
1050 E. Flamingo Road, Suite 120
Las Vegas, NV 89119
(702) 734-3322
Raymond J. Bath
NUS Corporation
1090 King Georges Post Highway
Edison, NJ 08837
(201) 225-6160
Tom Baugh
USEPA
401 M Street SW
Washington, DC 20460
(202) 382-5798
G. R. Bear
Shell Development Company
P.O. Box 1380
Houston, TX 77001
(713) 493-8024
Werner Beckert
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2137
Marian Bedinger
USEPA Environmental Monitoring Systems Lab
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2372
501
-------
Participants' List continued
Joseph V. Behar
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2203
Bob Beimer
S- CUBED
P.O. Box 1620
La Jolla, CA 92038
(619) 453-0060
Henry Beiro
Martin Marietta Energy Systems
P.O. Box 2003, M.S. 7440
Oak Ridge, TN 37831
(615) 576-1568
Walter Berger
Enseco, Inc.
7440 Lincoln Way
Garden Grove, CA 92641
(213) 598-0458
Richard E. Berkley
USEPA MD-44
Research Triangle Park, NC 27711
(919) 541-2439
Bob Berkshire
U.S. Army Environmental Hygiene Agency
Aberdeen Proving Ground, MD
21010-5422
(301) 671-2208/3739
Bernie B. Bernard
O.I. Corporation
P.O. Box 2980
College Station, TX 77841-2980
(409) 690-1711
John Berthold
Babcock & Wilcox Company
1562 Beeson Street
Alliance, OH 44601
(216) 821-9110, ext. 271
Don Betowski
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2116
Stephen Billets
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2232
Tracie L. Billington
California Dept. of Health Services
83 Scripps Drive
Sacramento, CA 95825
(916) 924-2139
Hari Bindal
U.S. Air Force
HQ AFSC/DEV
Andrews Air Force Base, MD 20334
(301) 981-6341
Michael Birch
USEPA, Region IV
College Station Road
Athens, GA 30613
(404) 546-2447
Reginia Bochicciho
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2150
William Bokey
USEPA Hazardous Waste Section
College Station Road
Athens, GA 30613
(404) 546-3357
David Bottrell
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2142
Wayne Boyles
Hach Company
P.O. Box 389
Loveland, CO 80539
(393) 669-3050 ext. 2246
Julie L. Bozich
USEPA, Region VI (6H-EC)
1445 Ross Avenue
Dallas, XX 75202
(214) 655-6730
George Brills
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 7,98-2112
Peter Brown
Leeman Labs Inc.
600 Suffolk Street
Lowell, MA 01854
(617) 454-4442
William Brumley
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2684
Randy Bryson
Brunswick Defense
9509 International Court North
St. Petersburg, FL 33716
(813) 576-5482
502
-------
Participants' List continued
Rod Bushway
Department of Food Science
University of Maine
102 B. Holmes Hall
Orono, ME 04469
(207) 581-1626
Larry Butler
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2114
K. J. Cabbie
Lockheed-EMSC
1050 E. Flamingo Road, Suite 120
Las Vegas, NV 89119
(702) 734-3322
Roy Cameron
Lockheed-EMSC
1050 E. Flamingo Road, Suite 120
Las Vegas, NV 89119
(702) 734-3322
D. Cardenas
Lockheed-EMSC
1050 E. Flamingo Road, Suite 120
Las Vegas, NV 89119
(702) 734-3322
Chris Carlsen
Lockheed
1050 E. Flamingo Road, Suite 124
Las Vegas, NV 89119
(702) 734-3258
John M. Carlson
USEPA Environmental Services Division
60 Westview Street
Lexington, MA 02173
Thomas Carlson Environmental Systems
Corporation
200 Tech Center Drive
Knoxville, TN 37912
(615) 688-7900
Michael Carrabba
EIC Laboratories Inc.
Ill Downey Street
Norwood, MA 02062
(617) 769-9450
Hunt Chapman
Ecology & Environment
1700 N. Moore Street, Suite 1105
Arlington, VA 22209
(703) 522-6065
David Charters
USEPA Environmental Response Team
Edison, NJ 08873
(210) 906-6825
Doug Chatham
NUS Corporation
1927 Lakeside Parkway, Suite 614
Tucker, GA 30084
(404) 938-7710
P. K. Chattopadhyay
Ecology & Environment, Inc.
160 Spear Street, Suite 1400
San Francisco, CA 94105
(415) 777-2811
Rob Cherney
Hewlett-Packard
1421 S. Manhattan Avenue
Fullerton, CA 92631
(714) 758-5524
T. Lloyd Chesnut
Ohio University
306 Cutler Hall
Athens, OH 45701
(614) 593-2581
Tom Chiang
Lockheed
1050 E. Flamingo Road
Las Vegas, NV 89119
(702) 798-2145
Stan Christensen
ICF Technology
P.O. Box 280041
Lakewood, CO 80228-2213
(303) 236-7414
Wayne Chudyk
Tufts University
Civil Engineering Department
Anderson Hall
Medford, MA 02155
(617) 381-3211
William Claytor
Bruker Instruments, Inc.
19 Fortune Drive
Billerica, MA 01821
(508) 667-9580
Scott Clifford
USEPA, Region I
60 Westview Street
Lexington, MA 02173
(617) 860-4300
William Cole
Lockheed ESC
1050 E. Flamingo Road, Suite 126
Las Vegas, NV 89119
(702) 734-3226
Stuart P. Cram
Hewlett-Packard
MS-20 BAE
3000 Hanover Street
Palo Alto, CA 94304
503
-------
Participants' List continued
Alan B. Crockett
EG & G Idaho
P.O. Box 1625
Idaho Falls, ID 83415-2213
(208) 526-1574
Tom A. Cronk
Oak Ridge National Laboratory
P.O. Box 2567
Grand Junction, CO 81502
(303) 242-8621 ext. 212
Beneta Culpepper
U.S. Analytical Instruments
1511 Industrial Road
San Carlos, CA 94070
(415) 595-8200
John Curtis
Lockheed ESC
1050 E. Flamingo Road, Suite 120
Las Vegas, NV 89119
(702) 734-3257
Dileep K. Dandge
ST & E, Inc.
1214 Concannon Boulevard
Livermore, CA 94550
(415) 449-8516
Betty Anne Deason
Lockheed EMSCO
1050 E. Flamingo Road
Las Vegas, NV 89119
(702) 798-2227
Carla Dempsey
USEPA OERR
401 M Street SW, Mail Code OS230
Washington, DC 20460
(202) 382-5746
Jane E. Denne
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2655
Gabriel Dib
IT Corporation
4585 Pacheco Boulevard
Martinez, CA 94553
(415) 372-9100
Randall K. Dickinson
United Engineers
30 S. 17th Street
Philadelphia, PA 19101
(215) 422-4987
Chuck Dittmar
Groundwater Technology
24168 Haggerty Road
Farmington Hills, MI 48024
(313) 471-2031
Joseph J. DLugosz
USEPA Environmental Monitoring Systems Lab
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2598
D. Dobb
Lockheed-EMSC
1050 E. Flamingo Road, Suite 120
Las Vegas, NV 89119
(702) 734-3322
J.R. Donnelly
Lockheed
1050 E. Flamingo Road
Las Vegas, NV 89119
(702) 798-2299
Doug Dowis
Arizona Instrument
P.O. Box 336
Jerome, AZ 86331
(602) 634-4263
Brian Dozier
Reynolds Electrical & Engineering Co., Inc.
P.O. Box 98521 M/S 738
Las Vegas, NV 89193-8521
(702) 295-6879
Jo Ann Duchene
ICAIR-Life Systems Inc.
24755 Highpoint Road
Cleveland, OH 44122
(216) 464-3291
Peter H. Duquette
Bio-Metric Systems, Inc.
9924 West Seventy-Fourth Street
Eden Prairie, MN 55344
(612) 829-2714
Philip Durgin
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) '798-2623
F. F. Dyer
Martin Marietta Energy Systems
Oak Ridge National Laboratory
P.O. Box 2008 Building 4500S
MS-6128
Oak Ridge, TN 37831-6128
(615) 574-6856
David G. Easterly
USEPA
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2108
DeLyle Eastwood
Lockheed ESCO
1050 E. Flamingo Road, Suite 208
Las Vegas, NV 89119
(702) 734-3287
504
-------
Participants' List continued
Lawrence Eccles
USEPA Environmental Monitoring Systems Lab
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2385
V. A. Ecker
Lockheed EMSC
1050 E. Flamingo Road, Suite 120
Las Vegas, NV 89119
(702) 734-3322
John Edwards
Avalon Ventures III
c/o Ms. Alexis Parks
973 Fifth Avenue
Boulder, CO 80302
(303) 443-7010
Van Ekambaram
Woodward Clyde Consultants
4582 S. Ulster Street, Suite 1000
Denver, CO 80237
(303) 694-2770
Dudley Emer
Reynolds Electrical & Engineering Co., Inc.
P.O. Box 98521 M/S 738
Las Vegas, NV 89193-8521
(702) 295-6879
Ed Eschner
Lockheed
1050 E. Flamingo Road
Las Vegas, NV 89121
(702) 734-3508
Eugene Esplain
Navajo Nation Superfund Office
P.O. Box 2946
Window Rock, AZ 86515
(602) 871-5772
Joan Etheridge
Ontario Waste Management Corp.
845 Harrington Court
Burlington, Ontario
CANADA L7N3P3
(416) 637-2452
Harold Ethridge
Louisiana Dept. of Environ. Quality
625 N. 4th Street
Baton Rouge, LA 70802
(504) 342-8925
John C. Evans
Battelle Pacific Northwest Laboratories
P.O. Box 999/ MSIN; K6-81
Richland, WA 99352
(509) 376-0934
Louis Feige
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2226
I. Cecil Felkner
Technical Assessment Systems, Inc.
1000 Potomac Street NW
Washington, DC 20007
(202) 337-2625
Bruce S. Ferguson
ImmunoSystems Inc.
8 Lincoln Street
P.O. Box AY
Biddeford, ME 04005
(207) 282-4146
Mario Fernandez, Jr.
U.S. Geological Survey
4710 Elsenhower Boulevard, Suite B5
Tampa, FL 33634
(813) 228-2124
Carlos Alberto Ferreira
CETESB-CIA De Technologia De
Saneamento Ambiental
Av. Itambe, 38 Sta. Luzia
Taubate, Sao Paulo
BRAZIL CEI 12100
0122-33-4900
Thomas L. Ferrell
Oak Ridge National Laboratory
P.O. Box 2008
Oak Ridge, TN 37831-6123
(615) 574-6214
James H. Ficken
U.S. Geological Survey
Building 2101
Stennis Space Center, MS 39529
(601) 688-1548
Stephen R. Finch
Dexsil Corporation
One Hamden Park Drive
Hamden, CT 06517
(203) 288-3509
Stanley Finger
Vitreous State Lab
Catholic University
Washington, DC 20046
(202) 259-6711
Robert Finnigan
Finnigan Corporation
355 River Oaks Parkway
San Jose, CA 95134
(408) 433-4800
Timothy L. Fisher
U.S. Army Environ. Hygiene Agency
HSHB-ML-A/Analytical QA Office
Aberdeen Proving Ground, MD
21010-5422
(301) 671-3269
505
-------
Participants' List continued
Joan Fisk
USEPA OERR
401 M Street SW (OS 230)
Washington, DC 20460
(202) 382-3115
Donald A. Flory
Flory Environmental Consultants, Inc.
3214 Churchill
Pearland, TX 77581
(713) 485-3603
Mike Franz
Enseco, Inc
7440 Lincoln Way
Garden Grove, CA 92641
(213) 598-6458
Doug Frazer
USEPA, Region X
215 Fremont Street
San Francisco, CA 94105
(415) 485-1228
Scott Fredericks
USEPA
401 M Street SW
Washington, DC 20460
(202) 475-8103
William W. Freeman
Roy F. Weston, Inc.
We ston Way
West Chester, PA 19380
(215) 344-3616
David Friedman
USEPA Office of Solid Waste
401 M Street, SW
Washington, DC 20460
(202) 382-4761
David Gahr
Lockheed ESC
1050 E. Flamingo Road
Las Vegas, NV 89119
(702) 734-3292
Richard B. Gammage
Oak Ridge National Laboratory
P.O. Box 2008, Building 7509
MS 6383
Oak Ridge, TN 37831-6383
(615) 574-6256
Gomes Ganapathi
Bechtel National, Inc.
800 Oak Ridge Turnpike
Oak Ridge, TN 37837
(615) 572-7102
Steven P. Gardner
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2580
Clare Gerlach
Lockheed
1050 E. Flamingo Road
Las Vegas, NV 89119
(702) 798-2227
Jacob Gibs
U.S. Geological Survey
810 Bear Tavern Road
West Trenton, NJ 08628
(609) 771-3900
H. K. Gibson
IT Corporation
5815 Middlebrook Pike
Knoxville, TN 37921
(615) 588-6401
Greg Gillispe
Dept. of Chemistry
North Dakota State University
Fargo, ND 58105
(701) 237-8244
Richard K. Glanzman
CH2M Hill
6060 South Willow Drive
P.O. Box 22508
Denver, CO 80222
(303) 771-0900
Dick A. Glass
E-N-G Mobile Systems, Inc.
2950 Cloverdale Avenue
Concord, CA 94518
(415) 798-4060
Garth Glenn
NUS Corporation
999 West Valley Rd.
Wayne, PA 19087
(215) 687-9510
Donald E. Glowe
Texas Research Institute
9063 W_. Bee Caves Road
Austin', TX 78733
(512) 263-2101
Bruce Godfrey
Curtis & Tompkins Labs
2323 Fifth Street
Berkeley, CA 94710
(415) 486-0900
Steven C. Goheen
Battelle, Pacific Northwest Laboratories
P.O. Box 999
Richland, WA 99352
(509) 376-3286
Larry Golden
CAE Instrument Rental
207 N. Woodwork Lane
Palatine, IL 60067
(312) 991-3300
506
-------
Participants' List continued
John L. Gordon
North American Weather Consultants
3761 South 700 East
Salt Lake City, UT 84106
(801) 263-3500
Kisholoy Goswami
ST & E, Inc.
1214 Concannon Boulevard
Livermore, CA 94550
(415) 449-8516
Thomas E. Gran
ETC Findlay Laboratory
P.O. Box 1404
16406 U.S. Route 224 E.
Findlay, OH 45858
(419) 424-4925
Doug Grant
Geraghty & Miller, Inc.
P.O. Box 273630
Tampa, FL 33688-3630
(813) 961-1921
Robert Grant
V. G. Gas Analysis Systems
Aston Way, Middlewich
ENGLAND CW10 OHT
+44 60684 4731
Robin Grant
Lockheed ESC
1050 E. Flamingo Road, Suite 120
Las Vegas, NV 89117
(702) 734-3255
Daniel Greathouse
USEPA Risk Reduction Engineering
26 W. Martin Luther King Drive
Cincinnati, OH 45268
(513) 569-7885
Mark Greene
University of Pennsylvania School of Medicine
4418 Spruce Street, Apt. E-l
Philadelphia, PA 19104
(215) 222-2324
Donald Gregonis
Albion Instruments
4745 Wiley Post Way
650 Plaza 6
Salt Lake City, UT 84116
(801) 364-2021
Alan Grey
EG & G Idaho, Inc.
P.O. Box 1625
Idaho Falls, ID 83415
(208) 526-1414
Peter Grohse
Research Triangle Institute
Research Triangle Park, NC 27709
(919) 541-6897
Deborah Gustowski-Gatto
Institute for Environmental Studies
Louisiana State University
Room 42 Atkinson Hall
Baton Rouge, LA 70803
(504) 388-4290
J. W. Haas, III
Martin Marietta Energy Systems
Oak Ridge National Laboratory
P.O. Box 2008, Mail Stop 6114
Building 4500S
Oak Ridge, TN 37831-6113
(615) 576-7607
Michael C. Hadka
W. B. Satterthwaite Associates
720 N. Five Points Road
West Chester, PA 19380
(215) 692-5770
Andrew Hafferty
Ecology & Environment
101 Yesler Way
Seattle, WA 98104
(206) 624-9537
Patrick L. Hammack
U.S. Environmental Protection Agency
1445 Ross Avenue
Dallas, TX 75202
(214) 655-2270
Frank Hammer
Finn Sugar Biochemicals
1400 N. Meacham Road
Schaumburg, IL 60173-4808
(312) 843-3200
Barbara L. Hanby
Hanby Analytical Laboratories, Inc.
4400 South Wayside, Suite 107
Houston, TX 77087
(713) 649-4500
John Hanby
Hanby Analytical Laboratories Inc.
4400 South Wayside, Suite 107
Houston, TX 77087
(713) 649-4500
Robert C. Hanisch
Enseco Corporation
4955 Yarrow Street
Arvada, CO 80002
(303) 421-6611
Ken Hanks
Arizona Dept. Environ. Quality
2005 N. Central Avenue
Phoenix, AZ 85004
(602) 257-2394
507
-------
Participants' List continued
Philip Hanst
Infrared Analysis, Inc.
11 Gaming Drive
Ossining, NY 10562
(914) 762-6975
Tony Harding
Tracer Xray
305 W. Magnolia #321
Ft. Collins, CO 80526
(303) 491-7712
Rita M. Harrell
NSI Technology Services Corporation
P.O. Box 12313
2 Triangle Drive
Research Triangle Pk., NC 27709
(919) 541-5387
Keith Harris
Grundfos Pumps
2555 Clovis Avenue
Clovis, CA 93612
(209) 292-8000 ext 302
Robert 0. Harrison
Department of Entomology
University of California
Davis, CA 95616
(916) 752-5109/6571
Christopher G. Harrod
ENSR Consulting & Engineering
740 Pasquinelli Drive, Suite 124
Westmont, IL 60559
(312) 887-1700
Don Hartman
Hach Company
P.O. Box 389
Loveland, CO 80539
(303) 669-3050 ext. 2246
Scott Hazard
Lockheed ESC
1050 E. Flamingo Road
Las Vegas, NV 89119
(702) 736-8884
Edward Heithmar
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2626
John R. Helvig
USEPA, Region VII
Air Monitoring Section
Environmental Services Division
25 Funston Road
Kansas City, KS 66115
(913) 236-3884
Robert Hemeroth
Arizona Dept. of Health Services
1520 West Adams Street
Phoenix, AZ 85007
(602) 255-1188
Charles B. Henry
Institute for Environmental Studies
Louisiana State University
42 Atkinson Hall
Baton Rouge, LA 70803
(504) 388-8521
Stephen C. Hern
USEPA Environmental Monitoring Systems Lab
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2203
Nelson Herron
Lockheed ESC
1050 E. Flamingo Road, Suite 126
Las Vegas, NV 89119
(702) 798-2176
A. Judson Hill
Westinghouse Bioanalytic Systems Company
15225 Shady Grove Road, Suite 306
Rockville, MD 20850
(301) 670-0688
D. Hillman
Lockheed EMSC
1050 E. Flamingo Road, Suite 120
Las Vegas, NV 89119
(702) 734-3322
Lance Hines
U.S. Army Engineer District Omaha
1624 Douglas Street No. 320
Omaha, NE 68102
(402) 221-7868
Thomas Hinners
USEPA Environmental Monitoring Systems Lab.
P.O., Box 93478
Las Vegas, NV 89193-3478
(702) 798-2140
Raymond H. Hirshman
Arizona Dept. of Health Services
Division of State Laboratory Services
1520 W. Adams
Phoenix, AZ 85007
(602) 255-1188
James S. Ho
USEPA Environmental Monitoring and
Systems Lab.
26 W. Martin Luther King Drive
Cincinnati, OH 45268
(513) 569-7321
508
-------
Participants' List continued
Paul A. Hodakievic
Technology Applications, Inc.
26 W. Martin Luther King Drive
Cincinnati, OH 45268
(513) 569-7321
Michael T. Homsher
Lockheed ESC
1050 E. Flamingo Road, Suite 246
Las Vegas, NV 89119
(702) 734-3312
Richard Home
Ecology and Environment Inc.
1509 Main Street, Suite 1400
Dallas, TX 75201
(214) 742-6601
Sarah Horowitz
BP Research International Research Center
4440 Warrensville Center Road
Cleveland, OH 44128
(216) 581-5284
John Hosenfeld
Midwest Research Institute
425 Volker Boulevard
Kansas City, MO 64110
(816) 753-7600
Marilyn Hoyt
ENSR Consulting & Engineering
696 Virginia Road
Concord, MA 01742
(508) 369-8910
Alissa Hudson
CHEMetrics, Inc.
Route 28
Calverton, VA 22016
(703) 788-9028
Tom Huetteman
USEPA Region IX (P-3-2)
215 Fremont Street
San Francisco, CA 94105
(415) 974-0923
David Hulst
Hulst Research Farm Services
4449 Tully Road
Hughson, CA 95326
(209) 883-2164
Alan Humphrey
USEPA Environmental Response Team, Bldg 10
Woodbridge Avenue
Edison, NJ 08837
(201) 321-6748
Michael L. Hurd
USEPA Analytical Operations Branch
401 M Street SW
Washington, DC 20460
(202) 382-7906
Rick Irvin
Engineering Toxicology Division
Texas A&M University
College Station, TX 77843
(409) 845-0731
Elizabeth Jangula
Arizona State Laboratory
1520 W. Adams
Phoenix, AZ 85007
(602) 255-1188
Lynn Jarvis
Microsensor Systems, Inc.
5610 Sandy Lewis Drive
Fairfax, VA 22032
(703) 323-0034
Stephen Jensen
Curtis & Tompkins, Ltd.
2323 Fifth Street
Berkeley, CA 94710
(415) 486-0900
Janine Jessup
Idaho National Engineering Lab.
P.O. Box 1625 CFA 633
Idaho Falls, ID 83415-4123
(208) 526-4541
Roy R. Jones
USEPA, Region X
1200 6th Avenue ES 096
Seattle, WA 98101
(206) 442-7373
Tammy Jones
USEPA Environmental Monitoring Systems Lab
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2144
Bill Jow
Groundwater Technology Environmental
Laboratories
20610 Manhattan Place, Suite 108
Torrance, CA 90501
(213) 328-7959
Freia Jung
University of California
Department of Entomology 0420
Davis, CA 95616
(916) 752-5109
Edward J. Kantor
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2690
Philip A. Keith
USEPA Environmental Monitoring Systems Lab.
P.O. Box 71825
Las Vegas, NV 89170-1825
(702) 734-3207
509
-------
Participants' List continued
Johnathan Kenny
Department of Chemistry
Tufts University
62 Talbot Avenue
Medford, MA 02155
(617) 381-3397
H. B. Kerfoot
Lockheed ESC
1050 E. Flamingo Road, Suite 120
Las Vegas, NV 89119
(702) 734-3257
Suhas N. Ketkar
Extrel Corporation
240 Alpha Drive
Pittsburgh, PA 15238
(412) 963-7530
William Kilgore
California Dept. of Health Services
83 Scripps Drive
Sacramento, CA 95825
(916) 924-2599
Trude V. V. King
U.S. Geological Survey
P.O. Box 25046
Denver, CO 80225
(303) 236-1373
Stanley M. Klainer
ST & E, Inc.
1214 Concannon Boulevard
Livermore, CA 94550
(415) 449-8516
Robert D. Kleopfer
Hall-Kimbrell Environ. Services Inc.
4747 Troost
Kansas City, MO 64110
(816) 756-3162
Bonnie Koch
Navajo Superfund Office
P.O. Box 2946
Window Rock, AZ 86515
(602) 871-5772
Thomas Koch
Maryland Medical Laboratory, Inc.
1901 Sulphur Spring Road
Baltimore, MD 21227
(301) 247-9100
Eric N. Koglin
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2432
John Koutsandreas
USEPA (RD-680)
401 M Street SW
Washington, DC 20460
(202) 382-5789
Pat Kraker
ICF Technology
P.O. Box 280041
Lakewood, CO 80228-2213
(303) 236-7414
Lisa Kulju
NUS Corporation
19 Crosby Drive
Bedford, MA 01730
(617) 275-2970
Narindar Kumar
USEPA, Region IV
345 Courtland Street
Atlanta, GA 30365
(404) 347-5065
William Lacy
COM Federal Programs Center
13135 Lee Jackson Memorial Highway No. 200
Fairfax, VA 22033
(703) 968-0900
G. Laing
Lockheed EMSC
1050 E. Flamingo Road, Suite 120
Las Vegas, NV 89119
(702) 734-3322
Myron Douglas Lair
USEPA Hazardous Waste Section
College Station Road
Athens, GA 30613-7799
(404) 546-3351
Victor Lambou
USEPA
5320 Eugene Avenue
Las Vegas, NV 89108
(702) 759-2259
Frank Laukien
Bruker Instruments, Inc.
19 Fortune Drive
Billerica, MA 01821
(508) 667-9580
Tim Launius
Roy F. Weston, Inc.
5820 Lilley Road
Canton, MI 48187
Vernon J. Laurie
USEPA (RD-680)
401 M Street SW
Washington, DC 20460
(202) 382-5795
Eugene Leach
Caterpillar
100 N E Adams Street JB 3330
Peoria, IL 61629
(309) 675-1329
510
-------
Participants' List continued
Gary Lee
Microsenser Technology Inc .
41762 Christy Street
Fremont, CA 94538
(415) 490-0900
Roger W. Lee
USGS-WRD
801 Broadway
Nashville, TN 37203
(615) 736-5424
T. L. Lewis
Lockheed
1050 E. Flamingo Road, Suite 120
Las Vegas, NV 89119
(702) 734-3287
Albert A. Liabastre
U.S. Army Environ. Hygiene Agency
Building 180
Ft. McPherson, GA 30330-5000
Russell Lidberg
Lockheed ESCO
1050 E. Flamingo Road, Suite 208
Las Vegas, NV 89119
(702) 734-3287
Mark Lieber
ICF Technology
9300 Lee Highway
Fairfax, VA 22031
(703) 934-3000
Daniel Lillian
USEPA, Region II
Environmental Services Division
Woodbridge Avenue
Edison, NJ 08837
(201) 321-6707
Thomas Limero
Krug International
1290 Hercules Drive, Suite 120
Houston, TX 77058
(713) 483-8442
A. Linenberg
Sentex Sensing Technology Inc.
553 Broad Avenue
Ridgefield, NJ 07657
(201) 945-3694
Viorica Lopez-Avila
Acurex Corporation
485 Clyde Avenue
Mountain View, CA 94039
(415) 961-5700
Alec Loudon
Bruker Instruments, Inc.
19 Fortune Drive
Billerica, MA 01821
(508) 667-9580
Norman Low
Hewlett-Packard
1601 California Avenue
Palo Alto, CA 94304
(415) 857-7381
Nile Luedtke
Martin Marrietta Energy Systems
P.O. Box 2003
Oak Ridge, TN 37831
(615) 574-8752
Marianne L. Lynch
VIAR and Company
209 Madison Street
Alexandria, VA 22314
(703) 684-5678
Patricia Mack
USEPA, Region IX
944 E. Harmon
Las Vegas, NV 89119
(702) 798-2250
Philip Malley
Lockheed ESC
1050 E. Flamingo Road, Suite 124
Las Vegas, NV 89119
(702) 734-3207
Charles K. Mann
Department of Chemistry
Florida State University
Tallahassee, FL 32306
(904) 644-3747
Joanne Manygoats
Navajo Superfund Program
P.O. Box 2946
Window, AZ 86515
(602) 871-5772
Chung-Rei Mao
Corps of Engineers
12565 W. Center Road
Omaha, NE 68144-3869
(402) 221-7494
Leslie Maple
CHEMetrics, Inc.
Route 28
Calverton, VA 22016
(703) 788-9028
Mark Marcus
Chemical Waste Management, Inc.
150 W. 137th Street
Riverdale, IL 60627
(312) 841-8360
Robert Marguccio
Ecology and Environment
1509 Main Street, Suite 1500
Dallas, TX 75201
(214) 742-6601
511
-------
Participants' List continued
Paul Marsden
S-CUBED
3398 Carmel Mt. Road
San Diego, CA 92126
(619) 453-0060
John Marshall
HNU Systems, Inc.
160 Charlemont Street
Newton, MA 02161
(617) 964-6690
Joseph D. Mastone
Roy F. Weston, ESAT Division
Landmark One
One Vande de Graaff Drive
Burlington, MA 01803
(617) 229-2050
Cindy L. Mayer
Lockheed ESC
1050 E. Flamingo Road, Suite 120
Las Vegas, NV 89119
(702) 734-3257
Aldo T. Mazzella
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2254
Russell McCallister
ICF, Inc.
9300 Lee Highway
Fairfax, VA 22031-1207
(703) 934-3909
Harry B. McCarty
VIAR and Company
209 Madison Street
Alexandria, VA 22314
(703) 684-5678
William McClenny
USEPA Environmental Monitoring Systems Lab.
Mail Drop 44
Research Triangle Park, NC 27711
(919) 541-3158
Lisa McKenzie
USEPA, Region IX
944 E. Harmon
Las Vegas, NV 89119
(702) 798-2298
Jack McLaughlin
Ecology and Environment
6440 Hillcroft, Suite 402
Houston, TX 77081
(713) 771-9460
D. McNelis
Environmental Research Center
University of Nevada LV
Las Vegas, NV 89154
(702) 739-3382
Richard E. Means
NSI Technology Services Corp.
P.O. Box 12313
2 Triangle Drive
Research Triangle Pk, NC 27709
(919) 541-5387
Richard T. Medary
Army Corps of Engineers
601 E. 12th Street
Kansas City, MO 64106-2896
(816) 426-5806
Eugene Meier
USEPA Environmental Monitoring Systems Lab
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2237
Anne Melia
Ecology & Environment Inc.
6405 Metcalf Avenue
Clover Leaf Bldg. 3
Overland Park, KS 66202
(913) 432-9961
Frank J. Messina
USEPA
Woodbridge Avenue
Edison, NJ 08837
(201) 906-6170
Henk Meuzelaar
Center for Micro Analysis
University of Utah
ERML Building 61, Room 214
Salt Lake City, UT 84112
(801) 581-8431
Bryan Miller
Hewlett-Packard Company
5725 West Las Positas Boulevard
Pleasanton, CA 94566
(415) 460-1644
Dennis A. Miller
Lockheed ESC
1050 E. Flamingo Road
Las Vegas, NV 89119
(702) 798-2376
Gary Miller
Jacobs Engineering
428 S. Quail Street
Lakewood, CO 80226
(303) 232-7093
Larry S. Miller
Battelle
505 King Avenue
Columbus, OH 43201
(614) 424-5316
Gayle F. Mitchell
Civil Engineering Department
Ohio University
Athens, OH 45701 (614) 593-2476
512
-------
Participants' List continued
Ronald K. Mitchum
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2103
Bruce Molholt
USEPA, Region III
841 Chestnut (3HW16)
Philadelphia, PA 19107
(215) 597-6682
Patrick Molloy
Navajo Nation Superfund Office
P.O. Box 2946
Window Rock, AZ 86515
(602) 871-5772
Jim Moore
Tracer Xray
345 E. Middlefield Road
Mountain View, CA 94043
(415) 967-0350
Leuren Moret
ST & E, Inc.
1214 Concannon Boulevard
Livermore, CA 94550
(415) 449-8516
Frank A. Morris
Brown and Caldwell Laboratories
1255 Powell Street
Emeryville, CA 94608
(415) 428-2300
W. D. Munslow
Lockheed
1050 E. Flamingo Road, Suite 120
Las Vegas, NV 89119
(702) 734-3322
Jack Murphy
IN-SITU, Inc.
P.O. Box I
Laramie, WY 82070
(307) 742-8213
Bohdan Mykijewycz
USEPA, Region III
841 Chestnut Building
Philadelphia, PA 19107
(215) 597-3153
Royal J. Nadeau
USEPA Environmental Response Team
Woodbridge Avenue
Edison, NJ 07060
(201) 321-6743
Clifford Narquis
BP America
4440 Warrensville Center Road
Cleveland, OH 44128
(216) 581-5252
Charles H. Nauman
USEPA
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2258
William Newberry
USEPA Environmental Monitoring Systems Lab
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2167
Bruce J. Nielsen
Headquarters Air Force Engineering and
Services Center
HQ AFESC/RDVW
Tyndall AFB, FL 32403-6001
(904) 283-2942
John W. Nixon
Chemical Waste Management
150 W. 137th Street
Riverdale, IL 60627
(312) 841-8360
John M. Nocerino
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2110
Nathan Nunn
Lockheed
1050 E. Flamingo Road, Suite 126
Las Vegas, NV 89119
(702) 798-2171
Jonathan E. Nyquist
Oak Ridge National Laboratory
P.O. Box 2008
Oak Ridge, TN 37831
(615) 574-4646
R. A. Olivero
Lockheed EMSC
1050 E. Flamingo Road, Suite 120
Las Vegas, NV 89119
(702) 734-3322
Khris B. Olsen
Battelle Pacific Northwest Labs.
P.O. Box 999, MSN; K6-81
Richland, WA 99352
(509) 376-4114
Maureen O'Mara
Roy F. Weston, Inc.
Ill N. Canal, Suite 855
Chicago, IL 60606
(312) 993-1067
Ed Overton
Institute for Environmental Studies
42 Atkinson Hall
Louisiana State University
Baton Rouge, LA 70803
(504) 388-8521
513
-------
Participants' List continued
Jeff Oxenford
AWWA Research Foundation
6666 W. Quincy Avenue
Denver, CO 80235
(303) 794-7711 ext. 6016
Reddy Pakanati
Ecology & Environment
1509 Main Street
Dallas, TX 75201
(214) 742-6601
Joe Paladino
Westinghouse Bioanalytic Systems Company
15225 Shady Grove Road, Suite 306
Rockville, MD 20850
(301) 670-0688
Nancy Parson
Ecology & Environment
717 W. Temple Street
Los Angeles, CA 90012
(213) 481-3870
James R. Pasmore
Columbia Scientific Industries Corp.
P.O. Box 203190
Austin, TX 78720
(512) 258-5191
Selvin Passen
Maryland Medical Laboratory, Inc.
1901 Sulphur Spring Road
Baltimore, MD 21227
(301) 536-1400
Dwight Patterson
Xitech Instruments, Inc.
2919 Burbank Dr.
Fairfield, CA 94533
(707) 425-9283
Brent Payne
Geoscience Consultants, Ltd.
1400 Quail Street, Suite 140
Newport, Beach CA 92660
(714) 724-0536
J. Gareth Pearson
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2203
Randy Perlis
Ecology & Environment, Inc.
1776 S. Jackson Street
Denver, CO 80013
(303) 757-4984
Gary Ferryman
USEPA, Region VIII
Building 53 Denver Federal Center
P.O. Box 25366
Denver, CO 80225
(303) 236-5080
Mark Peters
Environmental Research Center
University of Nevada
4505 South Maryland Parkway
Las Vegas, NV 89154-4013
(702) 739-3142
Jimmie D. Petty
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2383
David A. Phillippi
USEPA, Region VII
726 Minnesota Avenue
Kansas City, KS 66101
(913) 236-2836
Ann M. Pitchford
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2203
Albert A. Pleva
New Jersey DEP BPA
65 Prospect Street
Trenton, NJ 08618
(609) 292-7696
Russell H. Plumb, Jr.
Lockheed ESCO
1050 E. Flamingo Road, Suite 126
Las Vegas, NV 89122
(702) 734-3265
Fred Poeppel
EXTREL
240 Alpha Drive
Pittsburgh, PA 15238
(412) 963-7530
Billy Bob Potter
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-3128
George R. Prince
USEPA Environmental Response Team
Woodbridge Avenue
Edison, NJ 08837
(201) 321-6649
Thomas H. Pritchett
USEPA Environmental Response Team
GSA Raritan Depot
Edison, NJ 08829
(201) 321-6738
David Pudvah
VG Instruments, Inc.
32 Commerce Center
Cherry Hill Drive
Danvers, MA 01923
(508) 777-8034
514
-------
Participants' List continued
Steven Pyle
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2529
Greg Raab
Lockheed ESC
1050 E. Flamingo Road, Suite 126
Las Vegas, NV 89119
(702) 734-3332
R. Rajagopal
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2664
Julio Reategui
Environ. Analytical Systems, Inc.
1400 Taylor Avenue
P.O. Box 9840
Baltimore, MD 21284-9840
(301) 321-5133
Joseph Reed
ICF Technology
P 0 Box 280041
Lakewood, CO 80228-2213
(303) 236-7414
Francis Regina
Allied Signal Corporation
P.O. Box 1021-R
Morristown, NJ 07960
(201) 455-2170
Regina Prevosto Rehm
ICF Technology
P 0 Box 280041
Lakewood, CO 80228-2213
(303) 236-7414
Les Rice
01 Corporation
P.O. Box 2980
College Station, TX 77841-2980
(409) 690-1711
Albert Robbat
Tufts University
Chemistry Department
Medford, MA 02155
(617) 381-3474
G. L. Robertson
Lockheed EMSC
1050 E. Flamingo Road, Suite 120
Las Vegas, NV 89119
(702) 734-3322
Alfredo Carlos Cardoso Rocca
CETESB-CIA Technologia Saneamento Ambiental
R. Agisse 172 AP 81 Vila Madalena
Sao Paulo, Sao Paulo
BRAZIL 05439
011-210-33-2711
Joseph Roehl
Environ. Anlaytical Systems Inc.
1400 Taylor Avenue
P.O. Box 9840
Baltimore, MD 21284-9840
(301) 321-5304
Joseph F. Roesler
Cincinnati Engineers, Inc.
4030 Mt. Carmel-Tobasco Road
Suite 225
Cincinnati, OH 45255
(513) 528-1888
David Roitman
University of Pennsylvania School of Medicine
4418 Spruce Street, Apt. E-l
Philadelphia, PA 19104
(215) 222-2324
Allen Rosenberg
HNU Systems, Inc.
160 Charlemont Street
Newton, MA 01261
(617) 964-6690
Jeffrey Rosenfeld
Lockheed ESC
1050 E. Flamingo Road, Suite 120
Las Vegas, NV 89119
(702) 734-3211
Ronald A. Ross
NSI Technology Services Corporation
Gateway Center, Tower II, Suite 311
4th and State
Kansas City, KS 66101
(913) 281-0307
Mary Ryan
Clean Air Engineering
207 N. Woodwork Lane
Palatine, IL 60067
(312) 991-3300
Mahmoud A. Saleh
Environmental Research Center
University of Nevada
4505 South Maryland Parkway
Las Vegas, NV 89154-4013
(702) 739-3142
Shad M. Sargand
Ohio University
13 Grand Park Boulevard
Athens, OH 45701
(614) 593-1467
Wayne Saunders
ICF Technology, Inc.
9300 Lee Highway
Fairfax, VA 22031
(703) 934-3000
Drew Sauter
A.D. Sauter Consulting
2356 Aqua Vista Avenue
Henderson, NV 89014
515
-------
Participants' List continued
John Scalera
USEPA Central
Central Regional Laboratory
839 Bestgate Road
Annapolis, MD 21401
(301) 266-9180
Douglas T. Scarborough
U.S. Army Toxic and Hazardous Materials Agency
Building E4460 (AMXTH-TE-A)
Aberdeen Proving Ground, MD
21010-5401
(301) 676-7569/671-3348
Kenneth R. Scarbrough
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2645
Eugene Scheide
Environmetrics, Inc.
10679 Midwest Industrial Boulevard
St. Louis, Missouri 63132
(314) 427-0550
Gregory Schiefer
ICAIR-Life Systems Inc.
24755 Highpoint Road
Cleveland, OH 44122
(216) 464-3291
Hendrik Schlesing
biocontrol institut fur chemische und
blologische Untersuchungen
P.O. Box 16 30
6500 Mainz
WEST GERMANY
+6132/2000
Steve Schroedl
Lockheed ESC
1050 E. Flamingo Road, Suite 120
Las Vegas, NV 89119
(702) 734-3257
David B. Seielstad
Kevex Corporation
355 Shoreway Road
San Carlos, CA 94070
(415) 591-3600
Jerry L. Sessions
Brunswick Defense
2000 Brunswick Lane
Deland, FL 32724
(904) 736-1700 ext 4352
Mahmoud R. Shahriari
Fiber Optic Center
Rutgers University
Piscataway, NJ 08854
(201) 932-5033
Patricia A. Sheridan
USEPA
GSA Raritan Depot
Woodbridge Avenue
Edison, NJ 08837
(201) 321-6730
Brad Shipley
USEPA, Region IX
215 Fremont Street
San Francisco, CA 94105
(415) 974-8108
Robert Shokes
Science Applications International
Corporation (SAIC)
4224 Campus Pt. Ct.
San Diego, CA 92121
(619) 535-7506
Steven J. Simon
Lockheed ESC
1050 E. Flamingo Road, Suite 126
Las Vegas, NV 89119
(702) 734-3285
Diann Sims
USEPA Central Regional Laboratory
839 Bestgate Road
Annapolis, MD 21401
(301) 266-9180
Gary Skiles
Talem Inc.
306 W. Broadway Avenue
Fort Worth, TX 76104
(817) 335-1186
John Skinner
U.S. Environmental Protection Agency
401 M Street, SW
Washington, DC 20460
(202) 382-2600
Amy Smiecinski
Environmental Research Center
University of Nevada
4505 South Maryland Parkway
Las Vegas, NV 89123
(702) 798-3382
Dick Smith
Texas Research Institute
9063 Bee Caves Road
Austin, TX 78733
(512) 263-2101
William P. Smithson
U.S. Army Environmental Hygiene Agency
Aberdeen Proving Ground, MD
21010-5401
(301) 671-2024
516
-------
Participants' List continued
Robert N. Snelling
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2525
Charles Soderquist
Enseco Inc.
2544 Industrial Boulevard
West Sacramento, CA 95691
(916) 372-1393
Michael F. Solecki
USEPA Environmental Response Team
Woodbridge Avenue
Edison, NJ 08837
(201) 906-6918
Lynn Sorensen
Microbics Corporation
2223 Faraday Street, #B
Carlsbad, CA 92008
(619) 438-8282
Joseph Soroka
USEPA, Region II
Woodbridge Avenue
Edison, NJ 08837
(201) 906-6875
G. Wayne Sovocool
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-93478
(702) 798-2212
Lisa Spadini
Marketing Services XRF
Kevex Instruments
355 Shoreway Road
San Carlos, CA 94070
(415) 591-3600
Martin L. Spartz
Dept. of Chemistry, Willard Hall
Kansas State University
Manhattan, KS 66506
(913) 532-6298
Richard D. Spear
USEPA, Region II
Woodbridge Avenue
Edison, NJ 08837
(201) 321- 6685
Thomas M. Spittler
USEPA
60 West View Street
Lexington, MA 02173
(617) 860-4734
Stanford Spurlin
Midwest Research Institute
425 Volker Boulevard
Kansas City, MO 64110
(816) 753-7600
Martin Stapanian
Lockheed EMSC
1050 E. Flamingo Road, Suite 120
Las Vegas, NV 89119
(702) 734-3322
Randy St. Germain
Department of Chemistry
North Dakota State University
Fargo, ND 58105
(701) 237-8244
James Stiles
Assessment Systems
7010 Beach Dr. SW, Suite 4
Seattle, WA 98136
(206) 937-8419
Grant Stokes
Geo-Centers Inc.
7 Wells Avenue
Newton Center, MA 02159
(617) 964-7070
Tom Stolzenberg
RMT Inc.
1406 E. Washington Ave, Suite 124
Madison, WI 53711
(608) 255-2134
Michael Story
Finnigan Corporation
355 River Oaks Parkway
San Jose, CA 94134
(408) 433-4800
Hal Stuber
James P. Walsh & Associates, Inc.
P.O. Box 2003
1002 Walnut, Suite 201G
Boulder, CO 80306
(303) 443-3282
Chris Sutton
Ruska Laboratories, Inc.
3601 Dunvale
Houston, TX 77063
(713) 975-0547
Robert Suva
AgriTech Systems Inc.
100 Fore Street
Portland, ME 04101
(207) 774-4334
Melvin J. Swanson
Bio-Metric Systems, Inc.
9924 West 74th Street
Eden Prairie, MN 55344
(612) 829-2700
George Sylvester
Van, Waters, Rogers
1363 South Bonnie Beach
Los Angeles, CA 90023
(213) 265-8122
517
-------
Participants' List continued
Michael J. Szezewski
Hewlett-Packard
Rt 41 & Starr Road
Avondale, PA 19311
(215) 268-5444
Andrew P. Szilagyi
ICF Technology Inc.
9300 Lee Highway
Fairfax, VA 22031
(703) 934-3774
Doreen Y. Tai
U.S. Geological Survey
Building 2101
Stennis Space Center, MS
(601) 688-1518
39529
Yoshi Takahashi
Dohrmann Rosemount Analytical Div.
3240 Scott Boulevard
Santa Clara, CA 95054
(408) 727-6000
Charles Tanner
ICAIR-Life Systems
24755 Highpoint Road
Cleveland, OH 44122
(216) 464-3291
Victoria Taylor
TMA/Norcal
2030 Wright Avenue
Richmond, CA 94804
(415) 235-2633
Avraham Teitz
USEPA, Region II
62A Woodbridge Avenue
Highland Park, NJ 08904
(201) 572-6089
Prakash M. Temkar
U.S. Army CERL
Box 4005
Champaign, IL 61820
(217) 373-6747
Francis Thomas
USEPA, Region V
230 S. Dearborn
Chicago, IL 60604
(312) 353-9065
J. Edward Tillman
Target Environmental Services, Inc.
8940-A, Route 108
Oakland Center
Columbia, MD 21045
(301) 992-6622
Kim Titus
Lockheed ESC
1050 E. Flamingo Road, Suite 212
Las Vegas, NV 89119
(702) 734-3285
Thomas Trainor
Bruker Instruments, Inc.
19 Fortune Drive
Billerica, MA 01821
(508) 667-9580
Jerald D. Trease
U.S. Army Engineer District, Omaha
1624 Douglas Street No. 320
Omaha, NE 68102
(402) 221-7868
Robert Turner
Microsensor Technology
41762 Christy Street
Fremont, CA 94538
(415) 490-0900
J. Vail
ICF Technology Inc.
160 Spear Street No. 1380
San Francisco, CA 94105-1535
Martin Vanderlaan
Lawrence Livermore National Lab.
Biomedical Sciences Division
P.O. Box 5507, L-452
Livermore, CA 94550
(415) 422-5721
J. Jeffrey van Ee
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2203
Jeanette Van Emon
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2203
Anna Vargas
Arizona Dept. Environ. Quality
2005 N. Central Avenue
Phoenix, AZ 85004
(602) 257-6852
David B. Vener
Xontech, Inc.
6862 Hayvenhurst Avenue
Van Nuys, CA 91406
(818) 787-7380
Ort Villa
USEPA
839 Bestgate Road
Annapolis, MD 21401
(301) 266-9180
Harold Vincent
USEPA Environmental Monitoring Systems Lab
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2129
518
-------
Participants' List continued
Tuan Vo-Dinh
Oak Ridge National Laboratory
P.O. Box 2008
Oak Ridge, TN 37831
(615) 574-6249
Rick Vollweiler
Art's Manufacturing & Supply Inc.
105 Harrison
American Falls, ID 83211
(208) 226-2017
Eric A. Wachter
Oak Ridge National Laboratory
P.O. Box 2008, MS 6113
Oak Ridge, TN 37831
(615) 576-2712
Steve Walker
Geraghty & Miller, Inc.
P.O. Box 273630
Tampa, FL 33688-3630
(813) 961-1921
Jim Walsh
James P. Walsh & Associates, Inc.
P.O. Box 2003
1002 Walnut No. 201G
Boulder, CO 80306
(303) 443-3282
Amy Walton
Jet Propulsion Lab.
4800 Oak Grove Drive
Pasadena, CA 91109
Joe Wander
Environmental Services Branch
HQ AFESC/RDVS
Tyndall AFB, FL 32403-6001
(904) 283-4234
Llewellyn Williams
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2203
Dave Wineman
USEPA, Region VI
1440 Ross Avenue
Dallas, TX 75202
(214) 655-6491
Marcus B. Wise
Analytical Chemistry Division
Oak Ridge National Laboratory
P.O. Box 2008, MS 120
Oak Ridge, TN 37831-6120
(615) 574-4867
Hank Wohltjen
Microsensor Systems, Inc.
5610 Sandy Lewis Drive
Fairfax, VA 22032
(703) 323-0034
Steven Wolfe
Ecology & Environment, Inc.
160 Spear Avenue, Suite 1400
San Francisco, CA 94105
(415) 777-2811
Greydon Woolerton
Syprotec Corporation
380 Route No. 1
West Chazy, NY 12992
(800) 361-3652
Ray D. Worden
Ruska Laboratories, Inc.
3601 Dunvale
Houston, TX 77063
(713) 975-0547
John R. Worlund
USEPA Environmental Monitoring Systems Lab
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2656
Donald T. Wruble
USEPA Environmental Monitoring Systems Lab.
P.O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2530
Dwayne Wylie
University of Nebraska
319 Manter Hall
Lincoln, NE 68502
(402) 472-2628
Thomas E. Yeates
USEPA, Region V
Environmental Sciences Division
536 South Clark Street 10th Floor
Chicago, IL 60605
(312) 353-3808
Kaveh Zarrabi
University of Nevada
4505 South Maryland Parkway
Las Vegas, NV 89154
(702) 739-3142
John Zimmerman
Lockheed ESC
1050 E. Flamingo Road, Suite 120
Las Vegas, NV 89119
(702) 734-3322
Barry Zvibleman
Environmental Instruments
2170 Commerce Avenue, Unit S
Concord, CA 94520
(415) 686-4474
519
------- |