SECOND ANNUAL
            United States
   Environmental Protection Agency
              SYMPOSIUM

                  on

         SOLID WASTE TESTING
                 AND
          QUALITY ASSURANCE
             PROCEEDINGS

           July 15-18, 1986
      Vista International Hotel
           Washington, D.C.
         Symposium Managed By
The American Public Works Association
        1313 East 60th Street
      Chicago, Illinois 60637

-------
SYMPOSIUM PROGRAM COMMITTEE
David Friedman (Chairman)
Manager, Methods Program
Office of Solid Waste
U.S. Environmental Protection Agency
Paul Friedman
Chemist
Office of Solid Waste
U.S. Environmental Protection Agency
Duane Geuder
Chemist
Office of Emergency and Remedial Response
U.S. Environmental Protection Agency
Gail A. Hansen
Chemist
Office of Solid Waste
U.S. Environmental Protection Agency
Kenneth Jennings
Environmental Scientist
Office of Waste Programs Enforcement
U.S. Environmental Protection Agency
Ronald Mitchum
Director
Quality Assurance Division, EMSL-LV
Office of Research and Development
U.S. Environmental Protection Agency
Agnes Ortiz
Chemical Engineer
Office of Solid Waste
U.S. Environmental Protection Agency
Florence Richardson
Quality Assurance Officer/Chemist
Office of Solid Waste
U.S. Environmental Protection Agency

-------
PROCEEDINGS INTRODUCTION
One of the major environmental problems facing the United
States, as well as other nations, is the need for safe
handling and disposal of hazardous waste. A fundamental
component of all programs relating to waste management is
the need to perform measurements. These measurements
include waste composition and properties; effectiveness of
management processes; engineering properties of materials
used in constructing management units; and, last but not
least, long term performance of such management units. Thus
the pivotal roles played by the measurement methodology and,
its attendent, quality assurance.
The analysis of complex waste matrices presents the
environmental community with demanding analytical problems
for which solutions are being developed at a rapid rate.
This annual symposium series, presented by the EPA’s Office
of Solid Waste, is designed to focus on recent developments
in testing methods and quality assurance of importance to
both the RCRA and CERCLA programs.
The symposium highlights developing requirements for quality
assurance as well as new analytical procedures intedned to
be used in EPA’s national RCRA and CERCLA hazardous waste
management programs. Our purpose in holding these symposia
is several fold. First, is as a means of communicatina what
EPA is doing regarding the activities EPA has already
initiated to upgrade the state—of—the—art as reflected in the
regulations and in SW—846. Second, to describe the
direction EPA’s program is taking with respect to testinc
and quality assurance issues. Third, as a forum for
discussion between Agency personnel and representatives from
public and private laboratories involved in waste sampling
and evaluation.
The presentations describe work in progress. Current plans
are that ASTM will publish the complete proceedings in the
near future and that prior to publication, the material
presented during the symposium wil be updated.
DAVID FRIEDMAN
CHIEF, METHODS SECTION
OFFICE OF SOLID WASTE

-------
Second Annual
United States Environmental Protection Agency
SYMPOSIUM
SOLID WASTE TESTING AND QUALITY ASSURANCE
Vista International Hotel Washington, D.C.
July 15-18, 1986
PROGRAM
Tuesday, July 15, 1986
11:00 am - 1:00 pm REGISTRATION
1:00 pm - 1:20 pm OPENING SESSION
Opening Remarks :
Marcia Williams, Director, Office of Solid Waste, USEPA,
Washington, D.C.
Henry Longest, Director, Office of Emergency and Remedial
Response, USEPA, Washington, D.C.
1:20 pm - 1:45 pm CONFERENCE OVERVIEW
Overview presented by David Friedman, Manager, Methods
Program, Office of Solid Waste, USEPA, Washington, D.C.
1:45 pm - 5:00 pm SESSION I,
CHAIRPERSON:
Agnes M. Ortiz, Chemical Engineer, Office of Solid Waste,
USEPA, Washington, D.C. 20460
TESTING THE COMPATIBILITY OF SOIL LINERS AND WASTE LEACHATE
Robert S. Truesdale, Research Triangle Institute, P. 0. Box
12194, Research Triangle Park, NC 27709
ROUND-ROBIN STUDY OF LEACHING METHODS AS APPLIED TO SOLID
WASTES FROM COAL-FIRED POWER PLANTS
Ishwar P. Muraka, Electric Power Research Institute, 3412
Hillview Avenue, Palo Alto, CA 94303
COFFEE BREAK
COLLABORATIVE STUDY OF THE TOXICITY CHARACTERISTIC LEACHING
PROCEDURE (TCLP) FOR METALS, PESTICIDES, AND SEMI-VOLATILE
ORGANIC COMPOUNDS
W. Burton Blackburn, S—Cubed, P.O. Box 1620, La Jolla, CA
92038
COMPARISON STUDY OF PREPARATIVE AND ANALYTICAL TECHNIQUES FOR
THE DETERMINATION OF SELENIUM IN WATER, SEDEMENT AND
VEGETATION MATRICES
Milad Iskander, California State Department of Health
Services, 2151 Berkeley Way, Berkeley, CA 94704

-------
Wednesday, July 16, 1986
7:00 am - 8:00 am CONTINENTAL BREAKFAST
8:00 am - 12:00 pm SESSION II
CHAIRPERSON:
Ronald Mitchum, Director, Quality Assurance Division,
Environmental Monitoring and Support Laboratory, USEPA, Las
Vegas, NV 89114
DEVELOPMENT AND VALIDATION OF RCRA METHOD 8280 FOR DIOXINS
AND FURANS
Steve Billets, Environmental Monitoring and Support
Laboratory, USEPA, Las Vegas, NV 89114
SINGLE LABORATORY EVALUATION OF METHOD 8080 FOR
ORGANOCHLORINE PESTICIDES AND PCB’s
Werner Beckert, Environmental Monitoring and Support
Laboratory, USEPA, Las Vegas, NV 89114
INNOVATIVE TECHNOLOGIES FOR HAZARDOUS WASTE ANALYSIS
Ronald Mitchum, Environmental Monitoring and Support
Laboratory, USEPA, Las Vegas, NV 89114
COFFEE BREAK
APPENDIX VIII ANALYSES IN GROUNDWATER
Robert W. April, Office of Solid Wastes, USEPA, 401 M Street
S.W., Washington, D.C. 10460
COMPARISON OF THE lOX (EPA) AND AOX (DIN) METHODS FOR THE
DETERMINATION OF ORGANIC HALOGEN COMPOUNDS IN WATER AND SOLID
WASTE
Raimund Roehi, California State Department of Health
Services, 2151 Berkeley Way, Berkeley, CA 94704
EVALUATION OF GAS CHROMATOGRAPH/MASS SPECTROMETER (GC/MS)
METHOD 8240 AND 8270 FOR APPLICATION TO APPENDIX VIII
COMPOU NDS
James Longbottom, Environmental Monitoring and Support
Laboratory, USEPA, Cincinnati, OH 45268

-------
Noon - 1:30 pm LUNCH
1:30 pm - 5:00 pm SESSION III
CHAIRPERSON:
Gail A. Hansen, Chemist, Office of Solid Wastes, USEPA,
Washington, D.C. 20460
HYDRIDE GENERATION METHODS FOR DETERMINATION OF ARSENIC AND
SELEN IUM
Steve Callio, Region VIII, USEPA, 999 18th Street, Denver, CO
80202- 2413
THE RATIONALE FOR FILTRATION OF GROUNDWATER SAMPLES
Olin C. Braidst, Geraghty and Miller, Inc., 6800 Jericho
Turnpike, Syosset, NY 11791
METHODS FOR THE ANALYSIS OF ORGANOMETALLIC COMPOUNDS IN
WASTES
Frederick Brinckman, National Bureau of Standards, Building
223, Gaithersburg, MD 20899
COFFEE BREAK
MICROWAVE PROCEDURE FOR THE DETERMINATION OF METALS IN OILY
WASTE
Thomas Copeland, ERCO, 205 Alewife Brook Parkway, Cambridge,
MA 02138
INTER-LABORATORY EVALUATION OF ICP METHOD 6010
Thomas Hinners, Environmental Monitoring and Support
Laboratory, USEPA, Las Vegas, NV 84114
METHODS FOR EVALUATING SOLIDIFIED WASTE
Albert Liem, Alberta Environmental Centre, Bag 4000,
Vegreville, AB, Canada
6:00 pm - 7:30 pm RECEPTION FOR SPEAKERS AND ATTENDEES

-------
Thursday, July 17, 1986
7:00 am - 8:00 am CONTINENTAL BREAKFAST
8:00 am - Noon SESSION IV
CHAIRPERSON:
Duane Geuder, Chemist, Office of Emergency Response and
Remediation, USEPA, Washington, D.C. 20460
DYNAMIC VALIDATION OF SUPERFUND/RCRA ANALYTICAL METHODS
Gareth Pearson, Environmentla Monitoring and Support
Laboratory, USEPA, P.O. Box 15027, Las Vegas, NV 84114
THE SUFFICIENCY, REDUNDANCY, AND APPROPRIATENESS OF SURROGATE
AND MATRIX SPIKE COMPOUNDS IN ORGANIC ANALYSIS
Forrest 0. Gardner, Lockheed Corporation, 1050 E. Flamingo
Road, Las Vegas, NV 89119
STATISTICAL APPROACH TO MULTI-ANALYTIC DATA QUALITY
Ivan T. Show, S-Cubed P.O. Box 1620, La Jolla, CA 92038
COFFEE BREAK
QUALITY ASSURANCE IN THE GROUNDWATER MONITORING TASK FORCE
FACILITY ASSESSMENT PROGRAM
Michael Kangas, ICAIR Life Systems, Inc., 24755 Highpoint
Road, Cleveland, OH 44122
ANALYSIS FOR CHLORINATED DIBENZO-P-DIOXINS AND DIBENZOFURANS
IN ENVIRONMENTAL SAMPLES AND EMISSIONS FROM CONBUSTION AND
INCINERATION PROCESSES
Robert Harless, Environmental Monitoring and Support
Laboratory, USEPA, Research Triangle Park, NC 27711
SAMPLING AND ANALYSIS FOR DELISTING DATA
VERIFICATION/DELISTING SPOT CHECKS
Myles Morse, Office of Solid Wastes, USEPA, Washington, D.C.
20460
Noon - 1:30 pm LUNCH

-------
1:30 pm - 5:00 pm SESSION V
CHAIRPERSON:
Denise Zabinski, Chemist, Office of Solid Wastes, USEPA,
Washington, D.C.
DEVELOPMENT OF PERFORMANCE EVALUATION SAMPLE SERIES FOR THE
GROUNDWATER MONITORING SURVEY AT DUMP SITES
Edward Berg, Environmental Monitoring and Support Laboratory,
USEPA, Cincinnati, OH 45268
OSWER LABORATORY EVALUATION PROGRAM - A PROGRESS REPORT
Florence M. Richardson, Office of Solid Wastes, USEPA,
Washington, D.C. 20460
IMPACTS AND INTERFACE OF CERCLA MONITORING REQUIREMENTS WITH
OTHER STATE AND FEDERAL PROGRAMS
Dennis M. Stainken, New Jersey Department of Environmental
Pollution, Trenton, NJ 08625
COFFEE BREAK
ANALYSIS OF NON-HOMOGENOUS MIXTURES
Robert L. Fisher, New Jersey Department of Environmental
Pollution, Trenton, NJ 08625
HAZARDOUS WASTE ANALYSIS USING GAS CHROMATOGRAPHY/MASS
SPECTROMETRY (GC/MS) AND LIQUID CHROMATOGRAPHY/MASS
SPECTROMETRY (LC/MS)
James A. Poppiti, Finnegan Corporation, 1383 Piccard Drive,
Rockville, MD 20850
UPDATE ON COOPERATIVE INVESTIGATION OF TEST METHODS FOR
SOLIDIFIED WASTE CHARACTERIZATION
Julia Stegemann, Waste Water Technology Centre, Environment
Canada, 867 Lake Shore Road, Burlington, ON, Canada
7:30 pm - 9:30 pm SESSION VI
(For USEPA and State Agency Attendees Only)
CHAIRPERSON:
Kenneth Jennings, Environmental Scientist, Office of Waste
Program Enfircement, USEPA, Washington, D.C. 20460
CORRECTIVE ACTIONS UNDER RCRA, INTERIM MEASURES
Jackeline Moya and Staff, Office of Waste Program
Enforcement, USEPA, Washington, D.C.

-------
Friday, July 18, 1986
7:00 am - 8:00 am CONTINENTAL BREAKFAST
8:00 am - Noon SESSION VII
CHAIRPERSON:
J. Howard Beard, Chief, Physical Science Section, Office of
Waste Program Enforcement, USEPA, Washington, D.C. 20460
ISSUES AND PROBLEMS RELATED TO CERTIFICATION AND PERFORMANCE
EVALUATION OF HAZARDOUS WASTE LABORATORIES
Robert L. Stephens, California State Department of Health
Services, 2151 Berkeley Way, Berkeley, CA 94704
SAMPLING HOLDING TIME, A PROGRESS REPORT
Michael Mackarinec, Oak Ridge National Laboratory, P.O. Box
10, Oak Ridge, TN 37830
HAZARDOUS WASTE REGULATIONS IN CANADA
Nancy P. Cathcart, Environment Canada, 351 Saint Joseph
Boulevard, Hull, PQ, Canada K1A 2C8
COFFEE BREAK
OVERVIEW OF RCRA ENFORCEMENT
Kenneth Jennings, Office of Waste Program Enforcement,
Washington, D.C. 20460
SW-846 UPDATE
Paul Friedman, Office of Solid Wastes, USEPA, Washington,
D.C. 20460
TECHNICAL ISSUES RELATED TO THE TOXICITY CHARACTERISTICS AND
LAND DISPOSAL RESTRICTIONS
David Friedman, Office of Solid Wastes, USEPA, Washington,
D.C. 20460

-------
POSTER SESSION
Liquid Releast Test (LRT)
for Liquid Loaded Sorbents
Evaluation of Methods for
Determining Chlorine in
Waste Oils
Evaluation of Selected Digestion
Methods for Determining
Hexavalent Chromium in Solid
Waste Matrices
Demonstration of the Solid
Ignitabi 1 ity Procedures
Benjamin
Research
P.O. Box
Research
Carpenter
Triangle Institute
12194
Triangle Park, NC 27709
Alvia Gaskill, Jr.
Research Triangle
P.O. Box 12194
Research Triangle
Jerry Messan
Battel le
Columbus Laboratories
505 King Avenue
COlumbus, OH 43201
Video Tapes
David Binstock
Research Triangle
P.O. Box 12194
Research Triangle, NC
Institute
Park, NC 27709
Development of Field Test
for Monitoring Organic
Hallides
RCRA Laboratory Certification
USEPA Reference Standard and
Quality Assurance Materials
for Analysis of Environmental
Con tam i nan t s
The Quantitation of PCB’s, PCDF’s,
and PCDD’s by Electron
Impact GC/MS Using Response
Factor Estimation
Demonstration of the
Toxicity Characteristic
Leaching Procedure
Ray Tarrer
College of Engineering
Auburn University
230 Ross Hall
Auburn, AL 36849-3500
Dennis Stainken
Department of Environmental Protection
Office of Science and Research CN 402
Trenton, NJ 08625
R. E. Thompson/P. A. Wylie
Northrop Services, Inc.
P.O. Box 12313
Research Triangle Park, NC 27709
Andrew 0. Sauter
A. D. Sauter Consulting
2356 Aqua Vista
Henderson, NV 89015
Nancy Rothman
Energy Resources, Inc.
ENSCO
185 Alewife Parkway
Cambridge, MA 02138
Institute
27709

-------
FIRST SESSION
Tuesday, July 15, 1986
1:45 p.m. - 5:00 p.m.
Chairperson:
Agnes M. Ortiz
Chemical Engineer
Office of Solid Waste
U.S. Environmental
Protection Agency
Washington, D.C.

-------
TESTING THE COMPATIBILITY OF SOIL
LINERS AND WASTE LEACHATE
ROBERT TRUESDALE, L. J. GOLDMAN, RESEARCH TRIANGLE
INSTITUTE, RESEARCH TRIANGLE PARK, NORTH CAROLINA
ABSTRACT
In order to satisfy RCRA requirements, a soil liner must
have a permeability of 1OE—07 cm/s or less and must be
compatible with the wastes or waste leachate it is meant to
contain. The objective of this research effort is to
develop and evaluate SW—846 test procedures for measuring
the permeability and chemical compatibility of soil liner
materials so that adherance to this requirement may be
demonstrated. Procedures were developed for fixed—wall and
flexible—wall permeameters. Two test devices were chosen
because each device has its advantages and disadvantages and
best area of application. Fixed—wall tests may be best for
chemical compatibility testing because the confining pres-
sure in flexible—wall may prevent desiccation cracks from
forming. Flexible—wall tests are better for determining the
permeability of core samples from the field because of
reduced probability of side—wall leakage. Compatibility
tests are to be conducted in tandem on identical laboratory—
compacted soil samples, with baseline permeability measure-
ments conducted with construction—site water on one sample
and chemical permeability measurements conducted on the
other sample.
Ruggedness testing results showed that for both test
methods, molding water content and the application of back—
pressure were the most important variables influencing
permeability measurements. Evaluation of both test methods
involved a collaborative testing program with sixteen
laboratories, one chemical, and one soil. Triplicate test-
ing of each test method by each laboratory enabled inter—and
intralaboratory precision to be determined.
INTRODUCTION
The current Resource Conservation and Recovery Act (RCRA)
double liner guidance, developed in response to the
Hazardous and Solid Waste Amendments (HSWA) of 1984,
recommends the use of a liner composed of inorganic
materials (e.g., compacted soil) as the lowest component in
the liner system of a hazardous waste storage or disposal
facility. To satisfy RCR.A requirements, this liner “must be
constructed of materials that have appropriate chemical
properties...to prevent failure (due to) physical contact
with the waste or leachate to which they are exposed” (40
CFR Part 264, 221, 264, 251, 264, 301). It has been recog-
nized for many years that the permeability of clay soils may
be altered by the presence of certain chemicals in the
permeating fluid. The use of these materials as hazardous

-------
waste facility liners creates the dilemma of how to predict
and/or measure any permeability changes that might occur in
them from contact with waste chemicals; i.e., to test the
compatibility of liner materials with the chemicals to be
contained by the liner.
Clay—chemical compatibility testing is necessary to select
suitable liner material during facility design. In addition
to meeting chemical compatibility requirements, RCRA
requires that the inorganic liner must have a compacted
hydraulic conductivity (permeability) less than or equal to
1 x io—7 cm/sec. It is necessary, therefore, to measure the
hydraulic conductivity of laboratory—compacted samples of
liner material to select a suitable liner material. The
hydraulic conductivity of undisturbed samples of field com-
pacted liner materials also must be measured to demonstrate
that the hydraulic conductivity achieved in laboratory—
compacted samples can be achieved in the field and for con-
struction quality assurance to ensure that the constructed
liner will perform as designed.
Currently, there are several tests for determing the
hydraulic conductivity of inorganic materials used to line
waste facilities (e.g., see SW—846 Method 9100). Most of
these tests also have been used to measure the chemical com-
patibility of liner materials. However, none of these tests
have been adopted as standard procedures for low—
permeability materials. Selection of the test method and
procedure can have a profound effect on the test results.
In addition, specific test methods may be suitable for
determining hydraulic conductivity but unsuitable for com-
patibility tests. For these reasons, the Research Triangle
Institute (RTI), under contract to EPA’S Office of Solid
Waste (OSW), has developed and evaluated standard test pro-
cedures for determining the hydraulic conductivity of in-
organic liner materials and for determining the effects of
liquid wastes or waste leachates on the fluid conductivity
of these materials.
The objective of this research effort was to develop and
evalute procedures for measuring hydraulic conductivity and
chemical compatibility. To accomplish this objective, the
following approach was used:
o Review available literature on permeability and compa-
tibility test methods.
o Select the most suitable test devices.
o Develop procedures for the selected test devices.
o Evaluate these procedures for ruggedness and vari-
ability through a collaborative testing program.
This paper summarizes the results of this research effort,
except for results of the colloborative testing program
which were not available at the time of its preparation.

-------
DEFINITIONS
The following definitions are used in this paper:
o Hydraulic conductivity or permeability testing, is a
method for determining volumetric flux, per unit cross—
sectional area, of water through a sample of inorganic
liner material.
o Compatibility testing is method for determining the
difference in permeability of a sample of inorganic
liner material to water versus a liquid waste or waste
leachate.
BACKGROUND
The procedure most commonly used for measuring the hydraulic
conductivity of a compacted soil is to enclose the sample
tightly in a cylinder (permeameter), saturate the sample,
and then pass a liquid (permeant) through the sample. The
pressure differential across the sample is expressed in
terms of hydraulic gradient ( a dimensionless quantity),
which is the change in pressure head across the sample
divided by the height of the sample. The gradient can be
controlled by superimposing air pressure above the permeant
supplied to the influent end of the sample and by regulating
the backpressure applied at the effluent end of the sample.
The flow of water through the sample is measured, and the
hydraulic conductivity is calculated using Darcy’s law:
K0/A (1)
dh/dl
where:
K = hydraulic conductivity (permeability) (cm/s)
Q = volumetric flow rate (cm 3 /s)
A = cross—sectional area of flow (cm 3 )
dh/dl = hydraulic gradient (dimensionless)
In clay—chemical compatibility testing, the permeability (K)
of a clay soil permeated by a certain chemical is measured.
A change or lack of change in the volume of K (when compared
to K for water) may be due to a combination of two factors:
o Difference in the permearit fluid viscosity and density
(compared to water or other baseline permeant fluid).
o Change in porous medium characteristics as a result of
clay—chemical interactions.
To separate these effects, it is necessary to report the
results of compatability tests in terms of intrinsic

-------
permeability (k), which is a property of the porous medium
alone, both for the tests with the baseline permeant fluid
and for the tests with the chemical permeant fluid in ques-
tion.
Intrinsic permeability has units of length squared (e.g.,
cm 2 ) and is related to hydraulic conductivity, K, by
Equation 2:
K = k g or k = Lu
I’ pg
where:
p = density of the fluid
= dynamic viscosity of the fluid
g = acceleration due to gravity
In practice, most researchers report and discuss their test
results in terms of hydraulic conductivity or permeability
(K) rather than intrinsic permeability (k). Provided that
the density and viscosity of the test fluid (at the test
temperature) are known, one can calculate the k value to
correspond to each K value reported.
Permeameter Types
Three types of permeameters have been used to measure
hydraulic conductivity of fine-grained soils. The fixed-
wall permeameter (Figure 1) consists of a rigid cylinder of
plastic or metal, often 4 inches in diameter, which has been
modified to contain a soil sample and to allow a permeant to
flow through it. A soil sample is usually compacted
directly in the cylinder.
In a flexible-wall permeameter, (Figure 2) a cylindrical
column of soil is encased laterally in a flexible membrane
(often latex rubber) and enclosed at the ends with porous
stones. The enclosed soil sample is placed in a fluid-
filled cell that is pressurized to provide a confining pres-
sure on the sides of the sample (Figure 2-3). The confining
fluid and the permeant are contained in two entirely
separate systems which do not allow the two fluids to mix.
The sample to be tested may be prepared in a compaction mold
and extruded for testing in the flexible-wall cell or may be
samples taken from the field using a shelby tube or other
coring device.
Consolidation cells are commonly used in the field of geo-
technical engineering to determine the compressibility and
rate of settlement of soils. Consolidation occurs when
water is squeezed out of the soil and is therefore a
function of permeability. With proper modification, a
fixed-ring consolidation cell can be used to measure permea-

-------
Permeant Fluid Intel
Pressure Outlet
Permeant
1— •.: 1
Effluent Outlet
Po us Stone
Figure 1. Fixed-Wall Pertneaxneter
Cell Pressure
Inlet
Pressure Outlet
flexible
Membrane
Effluent Outlets
Figure 2. Flexible-Wall Permeameter

-------
bility and is commonly used for applications such as earth
dams, retaining walls, and slurry trenches. This device has
not been used widely in the evaluation of the chemical com-
patibility of clay liner materials or for the measurement of
hydraulic conductivity. For this reason, test procedures
using this device were not evaluated.
Boynton and Daniel (1985) used compaction-mold,
consolidation—cell, and flexible—wall permeameters to
measure the hydraulic conductivity of kaolinite and fire
clay. While there were differences in conductivities
measured with the three devices, the type of permeameter did
not seem to have a large effect on the results. Differences
in measured permeabilities between devices were
substantially less than one order of magnitude.
In a series of permeability tests performed on three types
of clay, Peirce (1984a,b) found that fixed—wall, flexible—
wall, and consolidation permeameters gave essentially the
same results.
Factors Influencing Testing Results
Important factors that influence permeability measurements
include sample characteristics and preparation, permeant
properties, design of the test apparatus, and selection and
control of variables during performance of the test. Some
of the more important factors are discussed below.
Maintenance of Field Moisture Content-—
In preparing soils for permeability tests, some laboratories
air dry soil, while others maintain the soil at or near the
field moisture content during sample preparation. Drying
facilitates breaking up clods, sieving the soil, and obtain-
ing a homogeneous soil mass for testing. With some soils,
however, rehydrated dried soil has different properties than
soil maintained at field moisture. Sangrey et al. (1976)
found that drying and rewetting significantly altered the
liquid limits of several clays from field conditions. In
spite of rehydration times of several weeks (a 24 hour
curing period is commonly used for laboratory tests), the
properties of some clays were irreversibly altered by
drying. However, Daniel and Liljestrant (1984) tested the
permeability of a Gulf Coast clay prepared with and without
air drying. This clay did not show any appreciable
difference in permeability as a result of the different
sample preparation regimes. This suggests that although the
properties of some clays may be irreversibly changed by
drying, some clays are not appreciably affected.
Clod Size Control—-
“Clod” is the term used to described lumps of clay. Daniel
(1981) demonstrated that clod size can significantly affect
laboratory permeability measurements. For a single clay,

-------
samples compacted from 3/8 inch clods had a permeability of
2.5 x i — cm/sec; samples compacted from 3/16 inch clods
had a permeability of 1.7 x io8 cm/sec; and samples
conpacted from 1/16 inch clods had a permeability of 8.5 x
10—9 cm/sec. This illustrates the importance of controlling
clod size to obtaining consistent laboratory permeability
measurements.
Compaction Methods--
Several techniques have been developed for preparing labora-
tory compacted clay samples for permeability testing. For a
fix—wall permeability test the clay is compacted directly in
the permeameter, which in this case serves as the compaction
mold. For flexible wall tests, the clay is compacted in a
separate compaction mold, extruded and then trimmed to fit
the permeameter. When samples are trimmed to length, care
must be exercised to avoid forming smear zones from the cut-
ting tool sliding against the clay. These smear zones can
decrease the permeability of the sample by as much as 20
percent (Carpenter, 1982).
There are three methods that are commonly used to compact
test samples in the laboratory. These are:
o static compaction using a hydraulic or mechanical
press;
o impact compaction using a drop hammer; and
o kneading compaction using the Harvard Miniature
Compactor.
Recently Dunn and Mitchell (1984) reported that when other-
wise identical samples were compacted by different methods
to 90 and 95 percent of their maximum dry density, there
were notable differences in their hydraulic conductivities.
At both dry densities, static compaction produced samples
with the highest hydraulic conductivities, impact was second
highest, and kneading the lowest.
Sample Size—-
Boynton and Daniel (1985) compacted test samples with
various diameters ranging between 1.5 and 6 inches. The
measured hydraulic conductivities showed an increase with
sample diameter; the smallest diameter having the lowest
conductivity. However, the highest and lowest conductivi—
ties differed by only a factor of 2, which was not
considered by the authors to be of any practical
significance.
Carpenter and Stephenson (in press) used a flexible—wall
permeameter to determine the influence of sample length—to—
diameter ratio on permeability. All the samples were tested
at a gradient of 200. For samples with 2.8—inch and 4—inch

-------
diameters, they observed a slight decrease in permeability
as the length—to—diameter ratio increased.
Anderson and Bouma (1973) experimented with a series of
undisturbed core samples of different lengths to determine
the effect of sample size on permeability. They found that
permeabilities of cores 17 cm in length were lower by half
an order of magnitude than were the permeabilities of 5—cm
length cores.
Sample Saturation——
A soil sample, even when compacted, has some degree of
porosity. The pores are filled with either gas (usually
air) or liquid. Because water cannot flow through a gas
bubble, entrapped air within the interconnected pores blocks
flow channels causing a reduction in permeant flow and a
correspondingly low apparent permeability. Soaking the
sample from the bottom with the top open to the atmosphere
(a technique used by some in an attempt to saturate the
samples) may not result in complete saturation. Smith and
Browning (1942) found that in 200 specimens soaked from the
bottom, the degree of saturation averaged 91 percent with
the lowest value at 78 percent.
The extent of the error in permeability measurement attribu-
table to entrapped gas bubbles has been studied by several
investigators. Decreases in permeability by factors ranging
from 2 to 5 have been reported by Johnson (1954). Olsen and
Daniel (1979) presented evidence that the coefficient of
permeability may increase by about four orders of magnitude
as the degree of saturation increases from around 20 to 100
percent.
The application of backpressure to a soil sample is the most
effective method for achieving saturation (Matyas, 1967;
Daniel et al., 1984). Backpressure is pressure applied to
the pore fluid within a sample for the purpose of compress-
ing entrapped air bubbles and causing the gas to go into
solution in the permeant. This is accomplished by
simultaneously applying pressure to both the inflow and out-
flow ends of the soil sample.
Hydraulic Gradient——
To measure the permeability of compacted clay within a
reasonably short period of time, it is necessary to conduct
the tests at hydraulic gradients that are substantially
greater than those encountered in the field. The two
implied conditions in the Darcy equation are: the flow rate
is directly poportional to the hydraulic gradient and the
relationship between flow rate and hydraulic gradient is
linear through the origin. There is no single accepted
hydraulic gradient for use in permeability testing.

-------
Gradients of 5 to 20 are recommended by some (Zimmie, 1981)
and gradients as high as 362 have been used by others (Brown
and Anderson, 1982)
Over the past several decades, there have been several
studies aimed at evaluating the validity of Darcy’s law by
measuring the dependence of permeability on hydraulic
gradient. Oakes (1960), Mitchell and Younger (1967), and
others have published data that indicate a departure from
linearity at low hydraulic gradients. Bowles (1979)
observes that in clays a threshold gradient of 2 to 4 may be
necessary to produce any flow.
Since Darcy’s law indicates a linear relationship between
flow rate and hydraulic gradient, many workers have used
elevated hydraulic gradients to reduce testing time.
However, if hydraulic gradients are excessive, piping
(opening flow channels and increasing hydraulic
conductivity) or particle migration (blocking flow channels
and reducing hydraulic conductivity) may occur, and this can
significantly influence permeability measurement. Although
such effects can occur and have been reported (Daniel, 1981;
Mitchell and Younger, 1967), studies have been conducted at
elevated gradients with no evidence of piping or particle
migration (e.g., Brown and Anderson, 1982).
Zimmie et al. (1981) have recommended use of hydraulic
gradients between 5 and 20 for laboratory studies. Research
performed at Louisiana State University concluded that tests
conducted under hydraulic gradients as high as 100 are
“acceptable for testing, reducing testing times to realistic
and practical duration” (Acar and Field, 1983). Dunn and
Mitchell (1984) reported that increasing the gradient in
steps from 20 to 200 caused an irreversible decrease in
hydraulic conductivity.
Confining Pressure——
In flexible—wall cells the membrane—encased sample is
subjected to a confining pressure that is greater than the
pore pressure within the sample. There is an ongoing
controversy about whether flexible—wall cells are
appropriate for compatibility testing of permeant fluids
that may cause the soil sample to shrink and crack. Because
of the confining cell pressure, shrinkage cracks that might
occur in the field and that might appear in fixed—wall
devices may not form in the flexible—wall cell. Not all
researchers are in agreement on this issue, since the effect
has not been demonstrated clearly in comparative tests.
However, Boynton and Daniel (1985) allowed desiccation
cracks to form in soil samples which were then tested with
confining pressures of 2, 4, 8, and 15 psi. They found that
the permeability decreased markedly as the effective stress

-------
increased. Confining pressures in the range of 4 to 8 psi
were sufficient to begin closing the cracks, and confining
pressures in excess of 8 psi closed the cracks and greatly
decreased permeability.
Current Practices in Commercial Laboratories
Peirce (l984a) , examined the permeability testing procedures
of three commercial laboratories. These procedures are
presented in Table 1. For this study, seven other
laboratories were contacted and questioned about various
aspects of their flexible—wall permeameter testing
protocols. The information gathered is presented in Table
2.
In reviewing both Peirce’s and our findings, it becomes
evident that no two laboratories are using the same test
protocol for flexible—wall perrneameter permeability testing.
The influence on permeability of factors such as material
handling, sample preparation, and sample size, makes it
questionable whether permeability test results reported by
differenct laboratories are comparable.
TEST PROCEDURES DEVELOPMENT
Test procedures for the test cells are detailed in Truesdale
et al. (1985). Two test devices were selected for test
procedure development and evaluation: fixed—wall permea-
meters and flexible—wall permeameters. Advantages and
disadvantages of these devices are summarized in Table 3.
Detailed test procedures for each of these devices were
developed (Truesdale et. al., 1985). The test procedures
developed for fixed— and flexible—wall test methods are very
similar; they differ only in sample dimensions and in the
application of confining pressures to the sample in the
flexible—wall tests. Rationale for the selection of several
key test parameters common to both test devices is presented
in this section.
Sample Preparation
The physical properties of some clay change irreversibly
upon drying. For this reason it is important to maintain as
much moisture in the sample as is practicable. However, it
is necessary to dry the clay to the point that clods can be
reduced easily to the specified size and homogenized ( 4
mesh); poor preparation will introduce variability in the
test results. The procedures described in ASTM —D 698
(Section 4.1, Method A) were selected for preparing samples
for laboratory compaction, because the method is proven and
is designed for this purpose.
Sample Compaction
Impact compaction, a modified form of ASTM—D 698, was
selected for situations where laboratory compacted samples

-------
TABLE 1. COMMERCIAL LABORATORY PREPARATION OF CLAY
SAMPLES FOR PERMEABILITY TESTING
Clay
Commercial Lab #1
Commercial Lab #2
Commercial Lab 43
Clay
Type
Field clays, mostly
kaolinite and illite
Field clays
Field clays (CHI
Excavation
Procedure
Test pits with backhoe
—______________________
Obtained from client
Backhoe-bulk sample from
stockpile
Hanoling
and Storage
Double wrap of plasti<’ and
burlap
Airlight plastic bags
Bulk samples as received
Processing
Size
Control
Hand pick or break lumps
>¾’
No 4 sieve
Hand pick
Organic
Control
Hano pick
Hand pick weeds and
roots
Hand pick
Moisture
Control
Immediately test for
moisture content
Prevent drying
Prevent drying
Homogenity
Visual inspection—cut and
quarter large samples
Use representative sample
from bulk
Representative sample
from bulk
Additives
None
None
None
Preparation
Wetting
Substance
By priority
1 Site water
2 Tap water
3 Distilled water
AS received
Deionized water
Wetting
Procedure
Mist bottle and mixing
As received
Added in small increments—
Allow to absorb over night
Percent
Moisture
Optimum moisture plus I.20.o
As received
Optimum
Size
Control
¼-inch sieve
As received
No 4 sieve
Compaction
Test Vessel
Dimensions
2 8 diameter 3-5 height
4’ mold
Shelby tube. 28 diameter
Compaction
Procedures
Standard Modified
Proctor Proctor
Standard or modified
Proctor
Standard or modified Proctor
Lift Depth
Varies Varies
Approximately 1 inch
2”i inches
No of Lifts
3 5
3 or 5
3
Scarily Lift
Yes
Yes
Yes
Testing
Preflush
Substance
Depends on project site tap
distilled water
Demineralized water
100 psi
Deionized water
Head
50-100 gradient
Variable
Preflush
Time (Pore
Volumes)
1 pore volume
No set number
3 runs—until constant
K achieved
Chemical
Permeants
Tested
Organic anc inorganic
1 5 pH H 2 S0 4
Brine boiler wastewater
N A
From: Peirce et al. (198 Ia)

-------
T 8LE 2. SX ASPWTS F1 1B1E-WLL T61 P XWJ S AT S d9 IE UL
LA3ORAT U6 (A-G)
int 3 ce of
Cl Size
Ycsiture Cate t
jctfa
Yo
iture Mditia Carp ctia f t
A. Air dry
A. —
A.
Tap I*3ter
A. Static coipxtim—5 lifts-scarify
tve lifts.
B. intain close to
field sosture
8. lly beak
up to pass 14
sieve
B. Tap watw—

are fa 24 h
B. Ib,d-t rp with 0.75-irchdi teta’ rvd—
l.5 lifts—scarify betwee lifts to depth3
of/B inch.
C. jtstbel
plastic limit if
clur are lar
thm $4 sieve
C. Use jbber
mallet to
large cloth
C. Tap mter-
sprey o —
bead-
are fa 24 h
C. Hx d-twp with 0.75-inch -di ret ’ ta er
ioved in diaga al patt across sarple-
5 lifts—scarify betw lifts.
D. intain close
to field ,rclsture
0. —
0. Tap aotw-
hx bead
0. K,rvard Miniature Cbrpacta—l0 lifts—
scarify t twee lifts
E. intain field
zois re
E. lly beak
up clcth—-
ti face
tPv’ojgh 14
sieve
E. Siteaat
tap t ’
E. Hmid-thrp lexiw, ight of soil to
.olure
F. *lntatn field
soisture
F. getable mixer-
prothxss fire
str. ,th of clay
(1110 ii lilb)
in c h a ss
sectia
F. 1 ia ,ized i ter—
sp’ay a-
F. in ight of soil to
rEdBt 1Riflad % lu1Q
6. Ovei-dry t
clabc
6. iially break
up cloth
6. Tap v tw ‘ site
ter if roq ircd
6. rvard Miniature Carpactor with ‘40 -lb’
spririg— act in 6 lifts to s 1f1ad
lure-scarify ta m lifts

-------
TABLE 2. (Cmtfr iod)
8. 2.9-Inch d1ai ter
1mgth-to d i3reter
ratio of 1—1.5
8. BacI r jre 70-100 i—
iaintain ii tI1 95%
saturatia is hie d
Perireablilty to i. te’—
gradie t of 20-30 f 1 day.
Perircability to 1each,t —
gr ia t of 70-100 for 3 re vo1ut
B. Gr. ieit of 10
Tet Is r .nti1 appr iciin,te1y 1 , ek
of w lstø t re in are obtained.
C. 2-inch diareter
C. Backpressure— maintain
offect1 e stress of 1 1 d
increase pr ires uitll 95%
saturatia Is achieved
C. Perireability to lexhate3—
of 45 i-c vary greatly
depmding a Job req.nrareits.
0.2-inch diareter
1e gth-to-diareter
ratio of I
0. 8ackpre sure 90 i—
gradiait uider 10
0. ‘ iø,t w der 10ff p ible
te’ t ts r i u tl1 stable f1 obtained.
LaacF ,te t ts ru for 2 pore o1Lr .
E. 21-Inch diareter
6-7-Inch t ight
E. Bac e re In &1 lncre, ts
(maintain eff ti e stres of 5 i)
u ti1 95% saturatia Is achieved.
E. P ieabi1 tytomter—
maxiiiui gradlmt of 25. T ts ru u ti1
inf1 o tfl are eqjal
Panrcability to leath,t —
higt ’ gr ieits. Te3ts r’.s LI tl1 W1o i
f1a a 1afte2porevokm .
F. 2.8-inch dixeter
4-inch P ight
F. 8ackpr jre
F. G’adie,ts bet a 6.5 . d 9.
6. 2-inch direeter
6. Backpr ire 1 uir to ev ight
at 9.5 psi pore p re
(0.5 psi differa,tlal top
d bottce) aid 10 psi cell pr xe—
increase to 15-20 psi cell aid
14.5-19.5 pore pr jre aid c ck
saturati i. Increase pr eir by
0.5 psI u,tll 95% saturatiai Is achieved.
6. Pr& ire drop acrces s i maintained
be m 0.5 15 psI dep ing
sreple charxtB’istice.
A.
2.85-inch diareter
lmgthto-direeter
ratio of 2
S ç1e Size Sarple Ssbiratiai Test Ca iditia
A.
A. 8ackpre s ire 70-100 psi—
maintain Lntil 95%
saturatiai Is achieved

-------
TABLE 3. ADVANTAGES AND DISADVANTAGES OF SELECTED TEST METHODS
FIXED-WALL PERMEAMETERS
Advantages
o ShrInkage cracks can form (rigid walls, no confining pressure)
o Simplest device
o Easily made compatible with most chemicals
Disadvantages
o Sidewall leakage can occur
o Cannot measure degree of saturation
o Cannot easily measure sample consolidation or swelling
o More difficult to use undisturbed samples
o Sample diameter is fixed
FLEXIBLE—WALL PERMEAMETERS
AdvantaEes
o Eliminates sidewall leakage
o Can simulate field stresses on sample
o Can use a variety of sample sizes
o Easy to test undisturbed samples
o Can measure degree of saturation
o Can measure vertical and volumetric deformations
Disadvantages
o ConfIning pressure may prevent shrinkage cracks from forming
o Permeant fluid may be incompatible with membrane
o Cells are more expensive than fixed-wall cells

-------
are to be tested, because it is a reporoducible method that
has proven successful in a variety of permeability and com-
patibility tests. This method was not selected on the basis
of what is most representative of field conditions; there is
insufficient information currently available to indicate
which laboratory compaction method most closely generates
permeabilities that are achieved in the field with full—
scale compaction equipment.
Sample Thickness
Peirce (1984a,b) proved that a 2—in, thick sample is
adequate for reproducible testing in a reasonable amount of
time.
Sample Diameter
Boynton and Daniel (1985) demonstrated that sample diameters
from 1.5 to 6 in. do not affect testing results
significantly. A 4—in, fixed—wall cell was chosen because
it is dimensionally the same as a standard compaction mold
(ASTM—D 698). A 2.8—in, sample was chosen for flexible—wall
tests to accommodate the most common sampling tube (Shelby
tube) size. Other sample diameters (from 1.5 to 6 in.)
probably can be substituted without a significant change in
test results.
Sample Saturation
It is necessary to saturate the sample fully to ensure
accurate test results. Backpressure saturation has been
proven effective by a variety of researchers and is speci-
fied in this procedure.
Gradient
Gradients of 2 to 200 have proven acceptable to researchers.
High gradients can promote consolidation of certain types of
soil. Low gradients result in excessively long testing
times. A gradient of 100 was selected as a compromise bet-
ween the two extremes.
Termination Criteria
As there is no standard protocol for performing a permea-
bility test, there also is no standard method for
determining when a test should be ended. The problem is not
so great when water is the permeant since eventually
(usually within several days or weeks) a steady state will
be reached as evidenced by a constant value of K measured
over a period of several days. With chemicals or leachates
as permeants, the situation may be quite different, since
chemical interactions between the permeant and the clay that
could affect the permeability could continue over a long
period of time.

-------
In several laboratories the criterion for ending a test is
based on the number of pore volumes of permeant that have
passed throught the sample. Sometimes the criterion is that
a steady—state flow has been established. Often , these two
requirements are combined with termination being based on a
steady—state flow in conjunction with a certain minimum
number of pore volumes having been passed through the
sample. Pierce (1984a) has refined this procedure as
follows:
Readings of column level and time are taken at certain
intervals throughout the test. The hydraulic
conductivity is computed for each time interval. From
this set of data, the first ten points are taken and a
linear regression analysis is performed to determine
the slope of the hydraulic conductivity vs. time curve.
The first point is then dropped and another value is
added on the other end. The slope is calculated again,
and so on. In the beginning of the test, this slope
will be fairly large, but as the test progresses it
will decrease and approach zero when steady-state is
obtained.
Two criteria must be met for a test to be terminated for
these procedures.
o the slope of the curve does not significantly vary from
zero at the 95 percent confidence level; and
o at least one pore—volume of the liquid has passed
through the sample.
APPLICATIONS OF TEST PROCEDURES
This section briefly discusses important aspects of the
application of the fixed—wall test and flexible-wall proce-
dures.
Permeability Testing--Compacted Samples
Boynton and Daniel (1985) and Peirce (l984a and b) has dem-
onstrated that for permeability measurements on laboratory—
compacted samples, the two devices are practically equiva-
lent. For situations requiring this kind of permeability
measurement (e.g., preliminary liner material selection),
selection of the test device type probably will not
influence test results if the test procedures are equiva-
lent. Other than the test cells and a few aspects of their
operation (e.g., a cell pressure system for the flexible—
wall cell) the procedures for both test devices developed
during this study are identical.
Permeability Testing——Undisturbed Samples
Permeability testing of undisturbed samples from the field
is necessary for confirming laboratory—derived permeabi—

-------
lity/density/moisture relationships in a field—compacted
test fill and for quality control. of liner construction.
The samples are collected in sampling tubes (e.g., ASTM D
1587) and placed directly into the permeameters. For this
loading the sample into the test cell and because of the
reduced probability of sidewall leakage. Because of their
reliability, control of sidewall leakage, and ease of teting
undisturbed samples, flexible—wall cells were preferred by
most of the cornmerical laboratories contacted during this
study.
Compatibility Testing——Compacted Samples
The purpose of compatibility testing of inorganic liner
materials is to quantify the change, or lack of change, in
permeability of the materials when they are exposed to a
liquid waste or waste and leachate, and thus determine the
suitability of a liner material for use for hazardous waste
containment. Although tests such as Atterberg limits and
hydrometer particle size analysis have been suggested as
useful and simple index tests that may be used to determine
compatibility, permeameter tests are the only methods that
are adequately developed for use as standards.
Compatibility testing involves measuring the permeability of
a representative sample of liner material to a baseline test
fluid and comparing this result with the permeability of the
same material to the waste fluid in question. An increase
in permeability, after correcting for density and viscosity
differences between the same fluids, may be defined as in-
compatibility.
The test method presented in this document requires baseline
and chemical testing to be conducted on separate representa-
tive samples of the liner material. Although many research-
ers follow baseline testing with chemical testing on the
same sample, for some low—strength soils, the gradient
applied during baseline testing could consolidate the sample
so that the sample is changed for the chemical test. En
addition, testing separate samples allows the soil to be
saturated with the chemical to be tested, ensuring better
contact of the chemical with the soil. However, with side—
by—side testing it is critical that the samples be the same;
they should be taken from the same homogeneous sample and
compacted in exactly the same manner.
The choice of test cell device may influence test results
more for compatibility testing than for regular permeability
testing, although no controlled studies have been conducted
that postively demonstrate this. Fixed—wall tests may pro-
duce more variable results, because sample shrinkage from
chemical attack may cause sidewall leakage or blowout in
some samples and not in others. On the other hand, fixed—
wall tests may provide more conservative measures of permea-
bility, because they allow the sample to shrink and crack.
Boynton and Daniel (1985) have demonstrated that confining

-------
pressures of 4 to 8 psi in flexible—wall cells can begin to
close previously formed desiccation cracks and that this
effect becomes more pronounced at higher cell pressures.
The collaborative testing program will demonstrate the dif-
ferences between the two methods because tests will be
conducted with a clay—chemical combination that will
definitely produce desiccation cracks in the clay.
Construction—site water is specified as baseline testing
fluid in the test procedures. This allows the baseline
tests to be used to aid in liner soil selection, and ensures
that any interactions of the soil with the baseline fluid
are the same as those that may occur in the field. Tap
water may be used only if it can be shown that results will
be comparable to those with site water. Atterberg limit
tests and hydrometer grain size distribution tests without
dispersing agents may be rapid and reliable means for estab-
lishing comparability.
All materials to be used for constructing compatibility test
devices (e.g., valves, tubing, and cell walls) must be com-
patible with the chemicals tested. Chemically resistant
materials such as Teflon and stainless steel are commonly
used. If there is doubt about chemical/material compati-
bility, a sample of the material in question can be soaked
in the chemical and then checked for any sign of distortion,
swelling, dissolution, corrosion, or other degradation. It
is especially important that the tubing and buretts used in
the permeant measuring system are not subject to chemical
attack or creep, because tubing creeps under pressure and
distortion can affect the calibration of flow measurement
devices. Some guideines for tubing material selection are:
o Nylon tubing is the stiffest and least susceptible to
expansion under pressure or creep. It may, however, be
affected adversely by some chemicals (e.g. carbon
tetrachioride).
o Teflon tubing is more resistant; however, it may creep
under sustained pressure, and it is hard to see
through.
o Glass is the best material for flow measuring devices
because of its stiffness and resistance to most chemi-
cals. However, it must be thick enough to withstand
testing pressures and handled carefully to avoid
breakage.
TEST METHOD EVALUATION
As discussed above, two test methods, fixed—wall and
flexible—wall, were selected for evaluation. Each appears
to satisfy the required measurement objectives and is
compatible with existing equipment in commerical laborator-
ies. Current literature on labratory testing of
soil/chemical compatibility does not provide sufficient in—

-------
formation to allow the selection of the best technique or to
demonstate the equivalency of the two procedures. Our own
testing of the procedures suggests that they may be compar-
able; however, further data are needed for a definitive
conclusion.
The criteria suggested for comparison of the tests are
precision, sensitivity, and ruggedness. Precision can be
measured if defined in terms of the random error associated
with the measurement process (i.e., the reproducibility of
the measurement). Precision is commonly measured by the
standard deviation associated with repeated measurements of
the same quantity.
In addition to being precise, the test procedure must also
be sensitive to differences in hydraulic conductivity. As
an extreme example, a test procedure that always gives the
measurement tI2 I would be 100 percent precise but completely
insensitive to variations in hydraulic conductivity.
Sensitivity can be determined by testing various clay—
permeant combinations with the two test procedures and
observing the magnitude of the differences in the compati-
bility measurements. Each measurement would involve running
parallel permeability measurements for tap water and
chemical and defining compatibility as the increase in per-
meability of the chemical vs. tap water.
Finally, the test procedures should be rugged, that is,
relatively insensitive to minor variations in the
operations, equipment, and materials used in testing. The
ruggedness of a test procedure is generally assessed by
varying in a controlled manner the conditions under which
testing occurs.
A two—phase approach to test evaluation was followed: first,
a ruggedness test was conducted to refine both procedures
for standard use in clay—permeant compatibility measurement
and then collaborative testing was carried out to determine
the between— and within—laboratory variability of the
procedures.
Ruggedness Testing
Ruggedness testing of a procedure should always precede
determination of its precision in a collaborative testing
program (Youden and Steiner, 1975; ASTM, 1964). The purpose
of ruggedness testing is to determine a procedure’s sensi-
tivity to minor reasonable variations in the method’s
variables by deliberately changing these variables in a
planned group of tests. This is necessary so that test pro-
cedures may be specified and closely controlled as necessary
to avoid excessive variation among the collaborative test
cooperators, as this may lead to misleading or inconclusive
results.

-------
Ruggedness testing of the two test procedures involved
repeated measurements of a “standard’ t clay—permeant combina-
tion made by a single laboratory with procedures and
parameters varied systematically. The test design that was
used for ruggedness testing is a multifactorial design
developed by Plackett and Burmari (1946) and used for
ruggedness testing by Youden and Steiner (1975). The design
determines the effect of varying seven factors (test
variables) in eight tests for each procedure. The test
matrix is pictured in Table 4. Table 5 is a list of the
conditions altered and their assigned values.
Table 6 and Table 7 present ruggedness testing results for
flexible—wall and fixed—wall devices, respectively. These
tables present the differences for each condition; these
differences are calculated for each condition from the
average of the four high value test runs (capital letters)
minus the average of the four low value test runs (lower
case letters). The conditions are ranked in each table
according to the size of the differences.
Several observations may be made from these tables. First,
the flexible—wall test procedure is more rugged than the
fixed—wall procedure. This is not surprising because of the
confining pressure in the flexible—wall device reduces the
probability of sidewall leakage. Inspection of the signs in
the tables also is instructive. A negative sign indicates
that the high value condition results in a lower permea-
bility than the low value condition. These signs are
consistent between fixed—wall and flexible—wall tests for
the four highest—ranking conditions, suggesting that changes
in these conditions have real effects on the results. The
change in signs between devices for the latter three
differences suggests that these effects are not significant
(Youden and Steiner, 1975).
Collaborative Testing
Following ruggedness testing and refinement of the test
procedure, a collaborative study was carried out to
determine the variability —— both between and within labora-
tories —— that can be expected from the normal useage of the
test procedures and to compare the two test procedures to
see which is more precise. Although it would have been
desirable to test multiple clay—permeant combinations, this
was much too costly and time—consuming to implement.
Instead, for each test procedure, each laboratory made
repeated permeability measurements in parallel on a single
clay soil using a standard water (0.005 N CaSO 4 ) and a
chemical (methanol) previously shown to cause a measurable
change in the permeability of the soil. The average
compatibility measurements for the tests are compared to
determine if the difference between the two average compati-
bility measurements is explained by the measurement error of

-------
TABLE 4. RUGGEDNESS TEST DESIGN
Experimental
Conditions
Combination or Determination
No.
1 2 3 4 5 6
7 8
1
A A A A a a
a a
2
B B b b B B
b b
3
C c C c C c
C c
4
D D d d d d
D D
5
E e E e e E
e E
6
F f f F F f
f F
7
6 g g 6 g G
6 g
TABLE
5.
LIST
OF
OF
ALTERED CONDITIONS AND VALUES
PERMEABILITY/COMPATIBILITY TEST
FOR RUGGEDNESS
METHODS
TESTING
Condition
Letter
High value
(capital letter)
Low value
(lower case letter)
1.
Sample preparation
A,a
Air drying and
grinding
Limited drying and
manual size
reduction
2.
Scarification
B.b
Yes
No
3.
MoIsture content
C.c
2-3% above optimum
2—3% below optimum
4.
Lift thickness
D,d
1.5 inch
0.75 inch
5.
Compactive energy
E,e
110% std. proctor
90% std. proctor
6.
Hydraulic gradient
F,f
200
100
7.
Backpressure
G.g
Yes
No

-------
TABLE 6. TABLE OF DIFFERENCES--FLEXIBLE-WALL
Parameter
Difference (cm/sec)
Moisture Content
-2.30 E—7
Lift Height
—1.81 E—7
Back Pressure
1.78 E—7
Compactive Energy
-9.23 E-8
Drying Procedure
8.37 E-8
Scarification
—4.66 E—8
Hydraulic Gradient
4.54 E—8
Mean (cm/sec)
1.28
E-7
Std. Dev. (cm/sec)
1.98
E—7
TABLE
7.
TABLE
OF
DIFFERENCES--FIXED-WALL
Parameter
Difference (cm/sec)
Compactive Energy
-5.36 E-7
Moisture Content
-4.90 E-7
Lift Height
-4.31 E—7
Back Pressure
4.20 E-7
Hydraulic Gradient
-9.94 E—8
Drying Procedure
-9.19 E-8
Scarification
8.69 E-8
Mean (cm/sec)
2.61
E-7
Std. dev. (cm/sec)
4.27
E—7

-------
the two procedures. This is a crude measure of sensitivity.
More accurate measures would require much more testing than
was possible in this study.
To measure the precision of the compatibility tests, each
laboratory in the collaborative study made repeated measure-
ments with each test procedure of the change in permeability
of the standard clay between the standard water and the
standard chemical permeant. The measurements with standard
water also will be used to determine the precision of the
test procedures for measuring permeability.
To initiate the study, a sufficient quantity of the standard
clay soil was obtained for all repetitions of the test
procedures. The clay soil was thoroughly homgenized and
then split into samples, with each sample randomly assigned
to a laboratory/test procedure/repetition combination. The
homogenization ensured that the samples of clay are were
similar as possible. The randomization of samples provided
additional protection should systematic differences in the
clay remain after homogenization.
To determine the number of measurements necessary to
adequately determine the precision of the tests, it was
decided to set , the probability of determining that there
is a difference in the two test procedures when there is
none, at 0.05, and , the probability of determining that
the tests are the same when they are different, at 0.05, for
a standard error ratio of 2 (i.e., a standard error ratio of
2 must be detected without error 95 percent of the time).
To meet these requirements it was determined that 30 repeti-
tions of each test procedure were needed. This is in basic
agreement with ASTM D2777 (Determination of Precision and
Bias of Methods of Committee D—19 on Water) which suggests
that a total of 31 replicates be used in testing.
In this testing program, sixteen laboratories agreed to
conduct each procedure in triplicate, for a total of 48
replications for each test procedure. This ensures that the
minimum requirements discussed above will be met and that
interlaboratory precision and total variability can be
determined with an adequate degree of confidence even if all
laboratories do not complete the full testing program. The
long testing times and expense of conducting these tests
makes incomplete results a real possibility.
CONCLUS ION
1. The most commonly used laboratory methods for determin-
ing the permeability of soil liner materials to water
or waste leachates involves the use of permeameter
cells in which a permeant fluid is forced through a
saturated soil layer at an elevated gradient.
2. No standard test procedures for determining the chemi-
cal compatibility of soil liner materials currently

-------
exist. The literature reveals large variability in the
test methods and procedures used by various
researchers. However, two types of permeability cells
predominate: fixed—wall, in which the sample is
contained by a rigid—walled container, and flexible—
wall, in which the sample is contained by flexible
membrane. Both methods have distinct advantages and
disadvantages. However, several studies have demon-
strated that these two test methods produce comparable
results when used to test a clay soil’s permeability to
water.
3. Surveys of commercial laboratories reveal considerable
variation in test procedures. The most commonly used
permeameter cell is the flexible—wall device, with ease
of testing undisturbed samples and no sidewall leakage
being cited as the reasons for this preference. Fixed—
wall cells are also used is some laboratories.
4. Review of the literature has revealed that permeability
and compatibility test results are influenced by test
methods and procedures. The variability in methods and
procedures that prevails among testing organizations
thus makes it difficult to compare test results from
laboratory to laboratory.
5. Two test cells, fixed—wall and flexible—wall, were
selected to be evaluated for ruggedness and
interlaboratory precision. These test cells were
selected based on their technical suitability and
widespread acceptance and use. Test procedures have
been developed for each cell type, based on information
gathered from the literature and investigators’
experience with permeability testing.
6. A ruggedness test was conducted to determine the
methods’ sensitivity to minor changes in procedures.
Test procedures and parameters evaluated included
sample preparation technique, clod size, moisture
content, compactive energy, sample thickness, gradient,
and saturation method. Results of the ruggedness
testing were used to refine the test procedures prior
to collaborative testing.
7. A collaborative testing program was designed to deter-
mine the interlaboratory precision of the two methods.
A total of 60 tests, 30 for each method, were conducted
to estimate and compare the variability of the two test
methods. The tests were conducted using a single clay
soil, a standard baseline permeant (0.005 N CaSo4), and
a waste permeant (methanol) that will change the clay
soil’s permeability.

-------
REFERENCES
Acar, Y. B. and S. D. Field. 1982. Organic Leachate
Effects to Hydraulic Conductivity in Fine—Grained Soils.
Report GE—82—0l, Hazardous Waste Research Center, Louisiana
State University. Baton Rouge, LA.
Anderson, J. L., and J. Bouma. 1973. Relationships Between
Saturated Hydraulic Conductivity and Morphometric Data of an
Argillic Horizon. Soil Science Society American
Proceedings. 37:408—413.
Bowles, J. E. 1979. Physical and Geotechnical Properties
of Soils. McGraw-Hill Book Company. New York. 478 pp.
Boynton, S. S., and D. E. Daniel. 1985. Hydraulic
Conductivity Tests on Compacted Clay. J. of Geotechnical
Engineering . 111(4) :465—478.
Brown, K. W. and D. Anderson. 1982. Effects of Organic
Solvents on the Permeability of Clay Soils. Draft Report.
Texas A&M University. Texas Agricultural Experiment
Station. College Station, TX. 153 pp.
Carpenter, G. W. 1982. Assessment of the Triaxial Falling
Head Permeability Testing Technique. PhD. Dissertation.
University of Missouri. Rolla, MO.
Daniel, D. E. 1981. Problems in Predicting the
Permeability of Compacted Clay Liners. In: Symposium on
Uranium Mill Tailings Management, Fort Collins, CO. pp.
665—67 5.
Daniel, D. E. and H. M. Liljestrand. 1984. Effects of
Landfill Leachates on Natural Systems. A Report to Chemical
Manufacturers Association.
Dunn, R. J., and J. K. Mitchell. 1984. Fluid Conductivity
Testing of Fine Grained Soils. J. of Geotechnical
Engineering . 110 (11) :1648—1665.
Johnson, A. I. 1954. Symposium on Soil Permeability. ASTM
STP 163, American Society for Testing and Materials.
Philadelphia, PA. pp. 98—114.
Mitchell, J. K., and J. S. Younger. 1967. Abnormalities in
Hydraulic Flow through Fine—Grained Soil. ASTM STP 417.
pp. 106—139.
Oakes, D. 3. 1960. Solids Concentration on Effects in
Beritonite Drilling Fluids. Clay and Clay Minerals . 8:252—
273.

-------
Olson, R. E., and D. E. Daniel. 1979. Field and Laboratory
Measurements of the Permeability of Saturated and Partially
Saturated Fine—Grained Soils. In: ASTM Symposium on
Permeability and Groundwater Contaminant Transport.
Philadelphia, PA. 67 pp.
Peirce, 3. J. 1984a. Effects of Inorganic Leachates on
Clay Liner Permeability. Report of U. S. Environmental
Protection Agency. MERL—SHWRD. Cincinnati, OH.
Peirce, J. J. 1984b. Soil Preparation and Permeability
Testing with Consolidation Cell Apparatus. Final Report.
EPA Contract No. 68—03—3149—24—2. Cincinnati, OH.
Plackett, R. L. and J. P. Burman. 1946. The Design of
Optimum Multifactorial Experiments. Biometrika . 33:305.
Sangrey, D. A., D. K. Noonan, and G. S. Webb. 1976.
Variation in Atterberg Limits of Soil Due to Hydration
History and Specimen Preparation. Soil Specimen Preparation
for Laboratory Testing. ASTM STP 599. American Society for
Testing and Materials. pp. 158—168.
Smith, R. M., and D. R. Browning. 1942. Persistent Water—
tinsaturation of Natural Soil in Relation to Various Soil and
Plant Factors. Soil Science Society of American
Proceedings . 4:114—119.
Truesdale, R. S., et al. 1985. Laboratory Methods for
Testing the Permeability and Chemical Compatibility of
Inorganic Liner Materials. Final Report. EPA Contract 68—
03—3149—10—6. U. S. Environmental Protection Agency.
Washington, D.C.
Youden, W. S. and E. H. Steiner. 1975. Statistical Manual
of the Association of Official Analytical Chemists. AOAC.
Washington, D.C.
Zimmie, T. F., 3. S. Doynow, and 3. T. Wardell. 1981.
Permeability Testing of Soils for Hazardous Waste Disposal
Sites. In: Proceeding of the Tenth International
Conference on Soil Mechanics and Foundation Engineering,
Vol. 2. Stockholm, Sweden. pp. 403—406.

-------
COLLABORATIVE STUDY OF THE TOXICITY CHARACTERISTIC
LEACHING PROCEDURE (TCLP) FOR METALS, PESTICIDES
AND SEMI-VOLATILE ORGANIC COMPOUNDS
W.B. BLACKBURN, DR. TAYLOR, S-CUBED, LA JOLLA, CALIFORNIA;
L. R. WILLIAMS, U.S. ENVIRONMENTAL PROTECTION AGENCY LAS
VEGAS, NEVADA; AND l.A. KIMMEL, U.S. ENVIRONMENTAL
PROTECTION AGENCY, WASHINGTON, D.C.
ABSTRACT
The US EPA has developed a new procedure, the Toxicity
Characteristic Leaching Procedure (TCLP), to more
effectively simulate the leaching of hazardous waste in a
landfill environment. The procedure involves an 18-hour
extraction of a sample with either an acetic acid or sodium
acetate solution and subsequent analysis of the leachate for
metals, pesticides, and semi-volatile organic compounds. To
validate the method, three waste samples were sent to 24
different volunteer government and commercial laboratories
for extraction and analysis Eighteen of the laboratories
participated in the organic and pesticide analysis, and all
participated in the metals analysis. The results and
statistical analysis of this collaborative test will be pre-
sented. Implications for the use of the test by the
analytical community will also be discussed.

-------
ROUND-ROBIN STUDY OF LEACHING METHODS AS APPLIED
TO SOLID WASTES FROM COAL FIRED POWER PLANTS
ISHWAR P. MURARKA, Ph.D, LAND AND WATER QUALITY STUDIES,
ELECTRIC POWER RESEARCH INSTITUTE, PALO ALTO, CALIFORNIA
ABSTRACT
In 1979, the Electric Power Research Institute (EPRI)
sponsored a multiple laboratory investigation of the
variability and reproducibility of the U.S. EPA Extraction
Procedure (EP) used for classifying wastes. Four
laboratories extracted five utility solid wastes (two fly
ashes, two bottom ashes, and one scrubber sludge) and
analyzed the concentrations of eight elements (As, Ba, Cd,
Cr, Pb, Hg, Ag and Se) in the extracts. In late 1985 and
early 1986, EPRI sponsored another round-robin study to
compare the results of EPA ’s new Toxicity Characteristics
Leaching Procedure (TCLP) with the EP method. In this
latest study, three laboratories were used to extract and
analyze the concentrations of fourteen constituents (Ag, As,
B, Ba, Cd, Cr, F, Hg, Mn, Pb, Se, SO4, V and Zn) from seven
utility wastes (3 fly ashes, 2 bottom ashes, and 2 scrubber
sludges). The results were evaluated to determine the
reproducibility of the two methods, the factors contributing
to the variability of the EP and TCLP extracts, and the dif-
ferences in mean concentrations between the two extract
types for the selected wastes. The results show that
reproducibility differs by constituent, waste type, and
between the two extraction methods. Generally, the concen-
trations measured in the TCLP extracts are higher than those
obtained by the EP method. It also appears that the
reproducibility of the TCLP results (as measured by the
coefficient of variation) are equal to or better than the EP
method (e.g., for As, B, Cd, Cr, Mn and V). Differences in
the extractions between laboratories accounted for at least
25 percent of the total variability more frequently for the
EP method than the TCLP method. For the TCLP, the
analytical variability components (i.e., differences in
analyses between laboratories and differences in analyses of
duplicate splits by the same laboratory) accounted for at
least 25 percent of the total variability more frequently.
INTRODUCTION
In 1978, the U.S. Environmental Protection Agency (U.S.
EPA) proposed a laboratory method to be used in the classi-
fication of solid waste as hazardous or non-hazardous. The
method, which came to be called the EP method, was based
upon a dilute acetic extraction of the waste, followed by
analysis of the extraction liquid for the eight inorganic
chemicals included in the National Interim Primary Drinking
Water Standards. The EP test was subsequently promulgated
as the method for use iii regulations stemming from the

-------
Resource Conservation and Recovery Act (RCRA). The Electric
Power Research Institute (EPRI) sponsored a round-robin
study in 1979 to evaluate the reproducibility and
variability of extract concentrations obtained using the EP
test on utility industry solid wastes. The results of this
study are summarized in this paper. Detailed discussions of
the study and conclusions are presented in two EPRI reports,
Rose et al, 1981 and Eynon and Switzer, 1983.
In late 1985, the U.S. EPA issued a draft description of a
new laboratory method, the Toxicity Characteristics Leaching
Procedure (TCLP), as a candidate for replacing the EP method
for the classification of solid wastes as hazardous or non-
hazardous. It is expected that the TCLP will soon be
published in the Federal Register for public comment prior
to its final promulgation (slated for mid-1986). With the
cooperation of U.S. EPA staff, EPRI sponsored a round-robin
study in late 1985 and early 1986 to evaluate the reproduc-
ibility of the new TCLP, to identify factors contributing to
variability in the TCLP and EP results, and, to quantify the
differences in extract concentrations obtained by the two
methods. The results of both the 1979 and 1985-1986 EPRI
studies are presented below. A detailed discussion of the
latter study will be presented in an EPRI report by Mason
and Carlile, to be published in the fall of 1986.
1979 EP ROUND-ROBIN STUDIES
In 1979, EPRI selected five wastes (two fly ashes, two
bottom ashes, and one scrubber sludge) and four laboratories
(Radian Corporation, Camp Dresser McKee, Acurex Corpora-
tion, and Systems, Science Software) to evaluate the EP
test. Each of the five waste samples was homogenized and
split into 16 subsamples. Each laboratory then received
four of these subsamples on which quadruplicate EP extracts
were produced. In turn, each extract liquid was split into
eight aliquots. Each laboratory retained two of the result-
ing aliquots and exchanged two aliqouts with each of the
other three participating laboratories. Each laboratory
analyzed all the aliquots for eight elements (As, Ba, Cd,
Cr, Pb, Hg, Se, and Ag), using both flame and graphite
furnace atomic absorption spectroscopy.
The measured concentrations were then analyzed statistically
using a nested analysis of variance (ANOVA) model. This
type of model allows total variability to be determined and
quantifies the relative contribution of individual sources
to this variability. This method relies on the grouping of
components and is sometimes referred to as a hierarchic
ANOVA. This type of approach can be used to analyze both
fixed effects (e.g., use of two different extraction
procedures) and random effects (e.g., analytical variability
of duplicate samples). In the 1979 and the 1985-1986

-------
studies, all the components were considered to be random
effects. The variance components model used for the 1979
study is given below:
Xjj = log ijk]. = + a + + ck + djJk + e Jkl
where
i = extraction laboratory 1,2,3,4
j = extraction replicate 1,2,3,4
k = analysis laboratory 1,2,3,4
1 = analysis replicate 1,2
ijk1 = measured concentration
= the overall mean
a = The variance in the average amount extracted
at different extraction laboratories
(between-laboratory extraction variability)
b 1 = The variance in different extracts prepared
at the same laboratory (within-laboratory
extraction variability)
ck = The variance in the average analytical
results at each laboratory. (between-labora-
tory analysis variability)
djJ = The variance in the average analytical
results for a given extract at each labora-
tory. (within-extract, between-laboratory
analysis variability)
et kl = The variance in replicate analysis of the
same extract at the same laboratory
(within-laboratory analysis variability)
Table 1 summarizes the results of the variance components
analyses for each chemical, waste, and analytical method for
which above detection limit concentrations were found in the
extracts. The largest component of the variability in most
of the waste samples for Ba, Cd, Pb and Se was due to
between-laboratory analysis (component ck in Table 1). The
largest contributing component for Hg and for As and Cr in
about a third of the wastes was between-laboratory extrac-
tion variability (component a. in Table 1). Variability in
chromium concentrations appea s to be an exception, as dif-

-------
Table 1
RELATIVE SIZE OF VARIANCE COMPONENTS — 1979 ROUND ROBIN STUDY
Percent of Total Variance
Element Analysis Prep Waste* - a 1 b 1 Ck djjk el kl
Arsenic FURNACE AA EP OFA—AC 32 2 51 9 6
DFA-ALK 4 44 31 7 13
SS 47 12 37 2 2
FLAME AA EP DFA—AC 35 2 14 17 32
DFA-ALK 10 33 4 36 17
SS 29 1 10 23 38
Barium FURNACE AA EP DFA—AC 0 0 82 5 12
DFA—ALK 0 3 73 18 6
WA—AC 0 7 65 12 15
WA—ALK 2 0 88 0 11
SS 1 0 90 0 10
FLAME AA EP WA—ALK 6 0 63 16 15
Cadmium FURNACE AA EP DFA—AC 28 1 45 7 20
DFA—ALK 0 29 41 2 28
WA—AC 12 12 56 2 17
WA-ALK 7 6 65 11 10
SS 13 5 70 4 8
FLAME AA EP DFA—AC 29 0 32 13 26
Chromium FURNACE AA EP DFA—AC 0 28 4 19 50
OFA—ALK 11 39 13 13 24
WA—AC 0 71 9 10 10
WA—ALK 42 18 2 22 16
SS 70 24 3 0 3
FLAME A.A EP OFA—ALK 3 65 15 13 4
Mercury FLAME AA EP SS 53 4 0 24 19
Lead FURNACE AA EP DFA—AC 12 0 54 7 27
SS 6 0 32 25 38
Selenium FURNACE AA EP OFA—AC 23 2 34 6 36
OFA—ALK 16 34 11 22 17
SS 0 15 13 23 49
FLAME AA EP OFA-AC 4 0 86 3 6
OFA—ALK 6 15 58 14 7
SS 2 0 72 4 21
* DFA—AC = Acidic dry fly ash;
DFA—ALK = Alkaline dry fly ash;
WA—AC = Acidic t bottom ash;
WA-ALK = Alkaline wet bottom ash;
SS = Flue gas desulfurization sludge

-------
ferences between the results of multiple extracts prepared
at the same laboratory accounted for more of the variability
(component b . in Table 1), with a relatively small contri-
bution from 3 between-laboratory analysis variability
(component ck in Table 1).
These results suggested that both the EP test and common
laboratory and chemical analysis procedures at the various
laboratories contributed to the variability in the re-
producibility of extract concentrations. One method to
reduce this variability would be to carry out multiple
extractions and average the results. This would yield a
statistically more precise estimate of chemical concentra-
tions.
1985-86 EP AND TCLP ROUND-ROBIN TESTS
In 1985-86, EPRI selected three laboratories (Battelle,
Pacific Northwest Laboratories; Radian Corporation; and Oak
Ridge National Laboratory) and seven wastes (three fly
ashes, two bottom ashes, arid two scrubber sludges) for use
in evaluating the TCLP test and for comparing its results
with those of the EP test. In this round-robin test, each
waste sample was freeze-dried after collection from a power
plant, homogenized, and split into 16 subsamples of about
100 g each. The acid-base characteristics of each waste (as
required by TCLP) were tested in triplicate by Battelle, PNL
to determine which extraction medium (i.e., acetic acid or
sodium acetate buffer) should be used. The results are
given in Table 2. Sodium acetate buffer was required for
all wastes except the two FGD sludge samples and the alka-
line fly ash sample.
Two of the participating laboratories received four
subsainples of each of the seven waste types, while the third
laboratory received the remaining eight subsamples of each
waste. The laboratories receiving four subsamples carried
out the EP and TCLP extractions using two subsamples for
each extraction method and each waste. The third laboratory
carried out the EP and TCLP extractions using four
subsamples of each waste. Duplicate extracts at each
laboratory were split into six aliquots each. Then the
laboratories exchanged two aliquots with each of the other
two participating laboratories, while retaining two of the
aliquots. The laboratory receiving four subsamples carried
out quadruplicate extractions and duplicate aliquot analyses
on those extracts. Each laboratory analyzed the sample
aliquots using graphite furnace atomic absorption (GFAA),
flame atomic absorption (FAA), cold vapor atomic absorption,
inductively coupled argon plasma spectroscopy (ICAP), ion
chromatography (IC), or ion specific electrode (ISE) as
appropriate for the fourteen constituents of concern (As, B,
Ba, Cd, Cr, Mn, Pb, Hg, Ag, Se, V, Zn, F, SO 4 ).

-------
Table 2
MEASURED pH VALUES AND THE CORRESPONDING EXTRACTION MEDIUM
Selected
Waste*
Initial pH
Final pH
Medium**
Wi
12.0 ± 0.3
11.1
± 0.2
M2
W2
3.9±
.8
———
Mi
W3
11.0 ± 0.4
2.05
± 0.03
Mi
W4
8.0 ±
0.7
1.46
± 0.01
Mi
W5
8.0 ±
0.4
5.9
± 0.2
M2
W6
10.1 ± 0.6
6.1
± 0.5
M2
W7
7.7 ±
0.5
1.7
± 0.1
Mi
* Wi — alkaline fly ash
W2 — acidic fly ash
W3 — alkaline bottom ash
W4 — neutral bottom ash
W5 — forced oxidized FGD sludge
W6 — FGD sludge
W7 — neutral fly ash
**M1 is a sodium acetate buffer solution
M2 is an acetic acid solution

-------
The data from this study were analyzed statistically to
estimate the variability in extract concentrations and the
factors which might account for the estimated variability.
Data were analyzed separately for each chemical, each waste,
and each extraction method using the following variance com-
ponents model:
ijkl =,a + a 1 + b 1 + Ck + djjk + e kl
where
i = extraction laboratory 1,2,3
j = extraction replicate 1,2 for laboratories 2
and 3 and 1,2,3,4 for laboratory I
k = analysis laboratory 1,2,3
1 = analysis replicate 1,2
= the overall mean
a = The variance due to the difference in the
average amounts extracted at different
extraction laboratories (between-laboratory
extraction variability)
b 1 = The variance due to the difference in
different extracts prepared at the same
laboratory (within-laboratory extraction
variability)
Ck = The variance due to the difference in average
analytical results at each laboratory
(between-laboratory analysis variability)
djjk = The variance due to the difference in average
analytical results for a given extract at
each laboratory (within-extract, between-
laboratory analysis variability)
eijk l = The variance due to replicate analysis of the
same extract at the same laboratory (within-
laboratory analysis variability)
This model is similar to the type used in the 1979 study.
In the 1985-86 study, the measured concentrations were not
log-transformed. The variance components were calculated
using a maximum likelihood estimation procedure. Statisti-
cal analyses were not conducted if more than 25 percent of
the data for a given constituent and waste type were below
detection.
Table 3 lists the percent of the total variance assigned to
each component by chemical constituent, analytical
technique, extraction method, and waste type. These results
show that the variance components can be quite different
from one waste type to another and for the different
constituents. For example, the largest source of

-------
Table 3
RELATIVE SIZE OF VARIANCE COMPONEUTS FOR THE
1985—86 ROUND ROBIN STUDY
Components of Variance as
Analysis Extraction Waste % of Total Variance
Element Method Method No . a 1 Ck djjk el kl
As GFAA EP W7 50 19 0 28 3
TCLP W7 4 0 0 57 39
B ICAP TCLP Wi 2 9 47 29 13
W2 0 12 30 31 27
W3 56 22 0 11 ii
W5 0 0 17 77 6
W6 0 1 19 74 6
W7 0 0 17 56 27
EP Wi 50 22 0 22 6
W2 29 0 37 22 12
W3 6 52 0 0 42
W5 0 0 0 56 44
W6 0 0 0 82 18
W7 0 10 29 30 3i
Ba ICAP TCLP Wi 2 0 67 9 22
W3 33 1 1 7 58
W7 0 82 1 0 17
EP Wi 50 5 41 3 1
W3 90 7 1 1 1
W7 80 14 4 1 i
Cd GFAA TCLP Wi i4 4 43 0 40
W2 0 9 21 32 38
W4 0 0 0 0 100
W5 0 0 21 22 57
W6 0 6 24 15 55
W7 0 4 20 2i 55
EP Wi 61 7 17 0 15
W2 0 13 5 27 55
W4 15 5 12 0 68
W5 5 2 0 54 40
W6 69 6 0 21 4
W7 0 10 0 34 56

-------
Table 3
continued
Components of Variance as
Analysis Extraction Waste % of Total Variance
Element Method Method No . a 1 Ck djjk elJkl
Cd ICAP TCLP W2 0 0 0 0 100
W5 0 0 34 28 38
EP W2 0 2 18 74 6
W5 0 0 35 28 37
Cr GFAA TCLP Wi 22 1 0 71 6
W2 71 14 0 14 1
W3 0 5 0 84 11
W5 0 0 0 80 20
W6 0 3 14 81 2
EP Wi 34 5 3 56 2
W2 0 1 43 50 6
W3 13 16 33 26 12
W5 0 1 46 52 1
W6 0 0 14 78 8
ICAP TCLP Wi 0 18 41 37 4
EP Wi 33 18 29 15 5
F ISE TCLP Wi 5 1 11 67 16
W2 1 0 71 0 28
W5 0 0 44 19 37
W6 7 4 43 26 20
W7 0 6 42 8 44
EP Wi 34 0 18 37 11
W2 15 4 18 0 63
W5 5 2 0 89 4
W6 48 2 2 47 1
W7 54 0 11 15 20
Mn FAA TCLP Wi 25 25 0 30 20
W2 9 1 0 78 12
W3 0 71 9 0 20
W5 0 1 29 54 16
W6 0 7 3 58 32
EP Wi 46 40 1 12 1
W2 74 1 0 15 10
W3 0 94 2 2 2
W5 39 51 4 0 6
W6 98 1 1 0 0

-------
Table 3
continued
Components of Variance as
Analysis Extraction Waste % of Total Variance
Element Method Method No . a Ck djjk el kl
Mn ICAP TCLP Wi 69 24 0 4 3
W2 11 39 15 9 26
W3 21 59 16 0 4
W4 26 47 16 3 8
W5 8 0 0 66 26
W6 1 0 6 1 92
W7 10 0 32 11 47
EP Wi 58 38 0 3 1
W2 60 12 3 12 13
W3 0 79 0 0 21
W4 0 2 0 3 95
W5 20 0 0 28 52
W6 98 1 0 1 0
W7 37 16 34 6 7
Se GFAA TCLP W7 18 1 0 75 6
EP W7 0 13 0 76 11
SO IC TCLP Wi 0 0 43 57 0
W2 8 0 69 22 1
W3 0 0 37 61 2
W5 4 0 74 22 0
W6 0 3 0 96 1
W7 0 0 33 67 0
EP Wi 9 1 41 50 0
W2 0 1 59 38 2
W3 0 1 49 44 6
W5 0 0 78 21 1
W6 0 5 0 92 3
Wi 0 0 29 71 0
V ICAP TCLP Wi 0 4 59 8 29
W2 83 15 0 1 1
W3 0 0 10 17 73
W4 0 13 0 0 87
W5 6 0 58 21 15
W6 0 2 22 70 6
Wi 0 13 9 43 35

-------
Table 3
continued
Components of Variance as
Analysis Extraction Waste % of Total Variance
Element Method Method No . a. C dijk el kl
V ICAP EP Wi 14 8 35 19 24
W2 0 0 56 0 44
W3 0 0 15 56 29
W4 1 6 42 16 35
W5 0 2 57 14 27
W6 2 0 24 60 12
W7 0 7 19 49 25
GEM TCLP Wi 72 6 0 16 5
W7 0 9 0 78 13
EP Wi 0 59 0 0 41
W7 7 19 0 60 14
Zn FAA TCLP Wi 89 1 0 4 6
W2 38 6 0 32 24
W4 74 1 0 10 14
W5 77 3 0 12 8
W6 9 1 0 7 83
W7 49 10 3 2 36
EP Wi 83 13 0 1 3
W2 50 4 27 0 19
W4 83 12 0 4 1
W5 30 37 0 23 10
W6 91 5 0 4 0
W7 83 7 4 6 0
ICAP TCLP Wi 83 2 3 5 7
W2 42 ii 7 23 17
W3 80 0 0 0 20
W4 70 0 0 12 18
W5 0 1 0 51 48
W6 55 7 6 6 26
W7 51 11 0 0 38
EP Wi 76 15 0 6 3
W2 36 0 47 0 17
W3 69 17 2 0 12
W4 80 ii 0 1 8
W5 12 0 0 38 50
W6 43 4 0 50 1
W7 84 ii 0 3 2

-------
variability in the Zn concentration is associated with the
extraction by different laboratories (component aj) in all
but two of the wastes. However, the largest source of
variability in the SO 4 concentration is associated with dif-
ferences between analysis laboratories (component ck and
d Jk).
To focus on the overall differences between the EP and TCLP,
a summary table has been prepared listing the number of
times a particular variance component accounted for more
than 25 percent of the total variance (Table 4A) and for
more than 75 percent of the total variance (Table 4B).
Different analytical techniques for a given constituent and
the different waste types have been combined. The following
discusses chemical constituents for which the source of
variability were the same in three or more cases. The sum-
mary table (Table 4) indicates that differences in the
amounts extracted by the laboratories (component a )
accounted for more than 25 percent of the total variability
more often in the EP extracts than the TCLP extracts. In
the EP extracts, this is true in 3 or more cases for Ba, F,
Mn and Zn. In the TCLP extracts, this is true in three or
more cases only for Zn. In the case of Zn, the extraction
step (at) account for over 75 percent of the variability in
seven cases in EP extracts and four cases in TCLP extracts.
This high variability may be due in part to the presence of
higher concentrations of Zn in the blanks from extraction
laboratory three (Table 5). The variability in duplicate
extracts by the same laboratory (component b ) exceeded 25
percent in five cases for Mn in both the EP and TCLP
extracts. Only once at most did the extraction variability
associated with the same laboratory exceed 25 percent for
either the EP or the TCLP for any other constituent. This
implies that each laboratory was consistent in its extrac-
tion method although there must be some actual differences
in the implementation of the procedures between labs.
The between-laboratory analytical variability (component ck)
accounted for 25 percent and 75 percent of the total vari-
ability approximately the same number of times for the El’
and TCLP extracts. This component of variability accounted
for at least 25 percent of the total variability in three or
more cases for Cr, SO 4 and V in the EP extracts and F and
S04 in the TCLP extracts. In only one case, for S04 in an
EP extract, did this component account for 75 percent or
more of the total variability. These results imply that
there was nothing inherent in either the EP or TCLP extracts
which caused higher variability in chemical analyses.
The analytical variability can be further separated into
variability associated with analysis of duplicate extracts
by the three different laboratories (compoent djjk) and the
variability associated with analysis of duplicate splits by
the same laboratory (component ejjkl). Analysis of

-------
Table 4
FREQUENCY DISTRIBUTION BASED ON VARIANCE COMPONENT
ACCOUNTING FOR AT LEAST (A) 25% (B) 75% OF TOTAL VARIABILITY
1 2 3 2 2 3 8
O 1 0 0 0 0 5
O 2 1 1 4 0 1
1 3 0 5 5 3 1
O 3 0 5 0 1 2
2 11 4 13 11 7 17
Se SO 4 V
0 12
1 1
4 2
4 2
6 1
1 10 15 18
33
8
20
30
18
TCLP Method
O 1 1 0 1
O 0 1 0 0
O 2 1 2 1
1 5 0 2 5
1 2 1 8 0
O 2 0 0 2 11 18
o 5 0 0 0 0 6
4 2 0 5 2 0 19
2 5 1 4 3 2 30
3 5 0 0 4 5 29
b 1 0 0 0 0 0 2 0 0 0 0
0 0 0 0 0 0 0 1 0 0
1 0 0 1 1 0 1 1 0 0
0 0 0 0 0 1 0 0 0 0
Total 1 2 0 1 1 5 1 2 0 7 20
TCLP Method
O 0 0 0 0 0 0
O 1 0 0 0 0 0
O 0 0 0 0 0 0
1 0 0 3 0 1 1
O 0 2 0 0 1 0
B Ba Cd Cr F Mn
Variance
Component As
(A) EP Method
a 1
b 1
Ck
djjk
el kl
Total
Total
O 0
0 0
0 5
1 5
0 0
109
a 1
b 1
d.
e 1 kl
Total
(B) EP Method
2 10 4 12 7 9 19 1 9 11 18 102
0 2 0 0 0 2 0 0 0 7 11
Ck
djjk
e 1 kl
2
1
5
1
a 1
b 1
Ck
djjk
elJkl
o 1 4
O 0 0
O 0 0
1 1 0
0 1 1
5
I
0
8
S
Total
1 1 2 3 0 2 1 1 3 5 19

-------
Table 5
MEASURED MEAN CONCEr TRATI0NS IN PROCEDURE BLANKS
mg/i
Extraction Lab Analysis Lab
BaICAPEP 1 2 3
1 .013 .02 <.02
2 <.004 .005 <.02
3 .447 .14 .43
Ba ICAP TCLP
1 .032 .07 .045
2 .125 .045 .067
3 .574 .53 .59
Zn FAA EP
1 <.02 .01 .01
2 .04 <.01 .01
3 .19 .10 .23
Zn FAA TCLP
1 .025 .04 .031
2 .086 .03 .037
3 .355 .35 .412
Zn ICAP EP
1 .017 .01 <.02
2 .045 .015 <.02
3 .244 .125 .218
Zn ICAP TCLP
1 .025 .04 .031
2 .111 .03 .037
3 .433 .35 U/A

-------
duplicate extracts by different laboratories contributed
more than 25 percent of the variability in at least three
cases for B, Cd, Cr, F, SO4, and V in HP extracts and for B,
Cr, Mn, SO 4 and V in TCLP extracts. For Cr, the variability
exceeded 75 percent in three of the TCLP extracts. The
total number of times when this component (djjk) exceeded 25
percent of the total variability is the same ror both EP and
TCLP extracts (i.e., 30 times). Analysis of duplicate
splits by the same laboratory (component ejjk l) contributed
more than 25 percent of the variability in at least three
cases for B, Cd and V in EP extracts and for Cd, F, Mn, V
and Zn in TCLP extracts. The total number of cases with
more than 25 percent variability was higher for the TCLP
extracts (29 times) than for the HP extracts (18 times).
Overall, components associated with the extraction step (ai
and b ) account for at least 25 percent of the total
variance more frequently for the HP method than for the TCLP
method. For the TCLP, the analysis step (components, ck,
d 1 jk, and ejjk l) more frequently account for 25 percent of
the total variance.
Table 6 gives the mean concentrations in the EP and TCLP
extracts of each waste by extraction laboratory and by
analysis laboratory. Values for concentrations near or
below detection are not listed in this table. The mean con-
centrations in the TCLP extracts when compared to EP
extracts were higher in most wastes for As, Ba, Cr, F, Mn,
Ph, Se, V and Zn. For As, Ba, Cr, Mn, and V differences
were large for many of the wastes. Differences were small
in those cases where the concentrations in the wastes were
high (e.g., B).
The differences in the two extraction methods can also be
summarized by waste type. Table 7 lists the mean concentra-
tion and coefficient of variation for each extraction method
by constituent. These results show that Ag and Hg con-
centrations are below detection in both the EP and the TCLP
extracts for all wastes extracted. Se is present in measur-
able quantities in only the acidic and neutral fly ash
samples. Pb is detectable in only the acidic fly ash
sample. V is not detectable in the two bottom ash samples
or in one of the scrubber sludge samples.
The coefficient of variation was typically low (i.e., less
than 30 percent) for B and Mn (Table 7). Based on the
results shown in this table, the reproducibility of the TCLP
test is equal to or better than the HP method for As, B, Cd,
Cr, Mn and V. Because of the many values below the detec-
tion limit, differences between EP and TCLP could not be
determined for Ag, Hg, Pb or Se.
Table 8 shows the relative frequency distribution of the
ratio of mean TCLP to HP concentrations. Only about 13

-------
Waste
No.
Wi
Element
As
Ba
B
Analysis
Method
GFAA
I CAP
I CAP
Table 6
MEAN CONCENTRATIONS
LABORATORIES AND
DETERMINED BY
BY THE ANALYSIS
ThE EXTRACTING
LABORATORIES
Extraction
Method
EP
TCLP
Extraction
1 2
(mg/i)
Lab Analysis Lab
3 1 2
(mg/i)
3
0.006 0.0232
0.0108 0.0053
0.0052 0.0167 0.0074
0.0096 0.01 0.008
0.0067
0.0079
W2 EP
TCLP
0.0062 0.0053
0.5030 0.1451
0.0047 0.01 0.0026
0.2403 0.2952 0.4267
0.0023
0.2358
Wi EP
TCLP
0.0401 0.0238
0.1579 0.1476
0.092 0.0467 0.047
0.137 0.1389 0.1542
0.0603
0.1558
Wi EP
TCLP
0.2962 0.5950
0.2989 0.3604
0.3623 0.4285 0.5267
0.3318 0.4049 0.4075
0.2542
0.1433
W2 EP
TCLP
0.0658 0.0855
0.0783 0.1154
0.0836 0.0769 0.1125
0.106 0.1025 0.1333
0.0420
0.0558
W3 EP
TCLP
0.2708 0.3243
0.6816 0.6785
0.7468 0.3692 0.4833
1.1418 0.7074 0.9292
0.4561
0.8567
W4 EP
TCLP
0.0574 0.0219
0.696 0.2031
0.4713 0.1317 0.2092
0.6597 0.2725 0.2867
0.1850
0.3056
W5 EP
TCLP
0.0678 0.0705
0.0896 0.1090
0.1106 0.0948 0.105
0.1759 0.1464 0.1292
0.0400
0.0800
W6 EP
TCLP
0.1827 0.1518
0.1959 0.2286
0.1498 0.1852 0.205
0.206 0.2613 0.2675
0.0933
0.0800
W7 EP
TCLP
0.1284 0.1593
0.3742 0.4453
0.2582 0.1597 0.1975
0.5513 0.4276 0.4375
0.1783
0.447
Wi EP
TCLP
17.3 14.6
17.8 18.2
19.4 17.4 17.4
17.2 18.9 17.2
16.5
16.5
W2 EP
TCLP
43.6 47.1
45.2 43.9
46.2 46.3 46.7
45.2 46.1 45.7
43.1
42.3
W3 EP
TCLP
1.31 1.39
1.50 1.42
1.61 1.45 1.46
1.88 1.66 1.56
1.37
1.53
W4 EP
TCLP
0.22 0.15
0.31 0.19
0.29 0.16 0.32
0.30 0.23 0.30
0.20
0.30
W5 EP
TCLP
0.76 0.78
0.99 0.79
1.01 0.96 0.85
0.95 1.02 1.05
0.67
0.66

-------
Table 6 continued
Extraction Lab
3
Analysis Lab
1 2
3
Element
B
Analysis
Method
ICAP
Waste
No.
W6
Extraction
Method_
EP
TCLP
1
2
(mg/i)
(mg/i)
1.82
2.19
1.63
1.70
1.57
1.68
1.73
1.95
2.05
2.32
1.27
1.38
W7
EP
TCLP
1.19
1.24
1.22
1.24
1.18
1.22
1.16
1.21
1.42
1.36
1.04
1.14
Cd
GFAA
Wi
EP
TCLP
0.0149
0.185
0.0083
0.014
0.0157
0.0159
0.0152
0.0198
0.0109
0.0132
0.0128
0.0152
W2
EP
TCLP
0.2237
0.2369
0.2592
0.2460
0.2606
0.2526
0.2398
0.2459
0.275
0.2725
0.2233
0.2142
W4
EP
TCLP
0.0014
0.0017
0.0022
0.0062
0.0014
0.0011
0.0014
0.0018
0.0022
0.0074
0.0014
0.0018
W5
EP
TCLP
0.0371
0.0281
0.0320
0.0271
0.0295
0.0288
0.0306
0.0312
0.0375
0.0255
0.0327
0.0262
W6
EP
TCLP
0.0063
0.0044
0.0020
0.0045
0.0073
0.0040
0.0055
0.0053
0.0046
0.0080
0.0058
0.0041
W7
EP
TCLP
0.0050
0.0062
0.0058
0.0058
0.0060
0.0062
0.0051
0.0069
0.0059
0.0044
0.0058
0.0064
ICAP
Wi
EP
TCLP
0.0238
0.0224
0.0226
0.0208
0.0209
0.0219
0.0256
0.0297
0.0112
0.0104
0.03
0.0226
W2
EP
TCLP
0.2313
0.2298
0.2102
0.242
0.2288
0.2279
0.248
0.2366
0.1858
0.2367
0.2308
0.2242
W5
EP
TCLP
0.0289
0.0346
0.0277
0.0256
0.0323
0.0245
0.0388
0.0345
0.0195
0.0177
0.0273
0.0326
Cr
GFAA
Wi
EP
TCLP
0.38
0.42
0.26
0.32
0.42
0.29
0.35
0.37
0.41
0.34
0.31
0.34
W2
EP
TCLP
0.014
1.33
0.008
0.53
0.025
0.56
0.0046
0.94
0.0425
0.89
0.0038
0.73
W3
EP
TCLP
0.005
0.0123
0.0078
0.0081
0.013
0.0087
0.0056
0.0119
0.0142
0.0078
0.0058
0.0096
W4
EP
TCLP
0.0015
0.0064
0.0015
0.0027
0.0022
0.0030
0.0015
0.0042
0.0017
0.0044
0.002
0.0043
W5
EP
TCLP
0.0355
0.0373
0.0239
0.0634
0.0302
0.0251
0.0187
0.0337
0.0617
0.0675
0.0148
0.0258

-------
Table 6 continued
Extraction Lab
Analysis Lab
1 2
(mg/l)
3
Element
Cr
F
Mn
Analysis
Method
GFAA
Waste
No.
W6
Extraction
Method
EP
TCLP
1 2
Jmg/lJ
3
0.0511
0.011
0.0023
0.0226
0.0246
0.0056
0.0024
0.0087
0.0894
0.0267
0.0024
0.0047
W7
EP
TCLP
0.0027
0.0666
0.0022
0.0578
0.0025
0.0502
0.0016
0.0604
0.0041
0.0575
0.002
0.0588
ICAP
Wi
EP
TCLP
0.4419
0.5089
0.3592
0.4304
0.4743
0.5564
0.4827
0.529
0.3725
0.405
0.4067
0.455
W2
EP
TCLP
0.0671
1.3513
0.07
0.5705
0.0609
0.6967
0.765
1.1341
0.0383
0.7992
0.08
0.7575
W3
EP
TCLP
0.0325
0.0388
0.0389
0.0417
0.0379
0.0406
0.0183
0.0287
0.0158
0.0158
0.08
0.08
W5
EP
TCLP
0.0524
0.1250
0.0764
0.0729
0.0627
0.0927
0.0586
0.0810
0.0508
0.0642
0.08
0.16
W6
EP
TCLP
0.1097
0.1207
0.1032
0.0539
0.0483
0.0762
0.0414
0.0445
0.0558
0.0717
0.1867
0.160
W7
EP
TCLP
0.0419
0.1136
0.0337
0.0654
0.0333
0.075
0.0084
0.0725
0.0317
0.0758
0.08
0.1192
ISE
Wi
EP
TCLP
0.1697
0.3636
0.4683
0.1833
0.2182
0.3817
0.145
0.1443
0.3758
0.4492
0.3375
0.4083
W2
EP
TCLP
1.675
1.432
1.6958
1.4367
1.3525
1.6492
1.39
0.82
1.76
2.07
1.67
1.87
W3
EP
TCLP
0.0425
0.0488
0.0631
0.092
0.094
0.150
0.0297
0.0371
0.1183
0.1283
0.0558
0.1292
W4
EP
TCLP
0.0438
0.0425
0.504
0.0828
0.0634
0.1702
0.0235
0.0429
0.105
0.1125
0.0358
0.14
W5
EP
TCLP
5.85
7.94
9.09
9.34
7.53
7.84
6.11
6.59
8.75
11.29
7.52
7.69
W6
EP
TCLP
1.45
1.94
3.18
2.24
1.93
1.91
1.86
1.83
2.60
2.38
1.82
1.89
W7
EP
TCL.P
0.47
1.44
1.26
1.54
0.97
1.48
0.89
1.44
1.02
1.68
0.66
1.33
FAA
Wi
EP
TCLP
3.84
4.62
2.78
4.45
3.54
5.40
3.72
4.83
3.20
4.82
3.28
4.75

-------
Table 6 continued
Extraction Lab Analysis Lab
Analysis Waste Extraction 1 2 3 1 2 3
Element Method No. Method ( mg/i) ( mg/i )
Mn FAA W2 EP 3.26 3.81 3.61 3.46 3.59 3.56
TCLP 3.75 4.24 4.09 3.81 4.16 4.10
W3 EP 0.41 0.40 0.56 0.45 0.45 0.47
TCLP 0.61 0.59 0.67 0.61 0.61 0.65
W5 EP 1.2777 1.6188 1.3707 1.3811 1.3833 1.4683
TCLP 1.5046 1.5188 1.4989 1.5328 1.425 1.555
W6 EP 1.81 0.76 0.18 1.16 0.85 0.96
TCLP 1.93 1.98 1.90 1.99 1.85 1.94
W7 EP 0.14 0.15 0.13 0.15 0.09 0.16
TCLP 0.17 0.18 0.12 0.19 0.10 0.19
ICAP Wi EP 3.78 2.52 3.22 3.51 3.05 3.04
TCLP 4.49 4.00 5.08 4.60 4.42 4.52
W2 EP 3.24 3.55 3.42 3.39 3.43 3.34
TCLP 3.69 3.91 3.82 3.80 3.87 3.71
W3 EP 0.45 0.41 0.58 0.50 0.47 0.46
TCLP 0.57 0.61 0.67 0.63 0.62 0.59
W4 EP 0.0478 0.0255 0.0273 0.0267 0.0583 0.0226
TCLP 0.0374 0.0293 0.0288 0.0344 0.0333 0.0288
W5 EP 1.13 1.57 1.34 1.39 1.22 1.34
TCLP 1.45 1.50 1.62 1.56 1.49 1.48
W6 EP 1.87 0.78 0.20 1.18 0.97 0.92
TCLP 1.98 1.90 1.88 1.97 1.84 1.95
W7 EP 0.1403 0.1603 0.1616 0.1592 0.1558 0.1408
TCLP 0.1674 0.1747 0.1674 0.1764 0.1683 0.1625
Pb GFAA W2 EP 0.0193 0.0111 0.0222 0.0219 0.0082 0.0217
TCLP 0.3143 0.0787 0.1066 0.2282 0.1417 0.1583
Se GFAA W7 EP 0.0623 0.047 0.0727 0.0566 0.0517 0.0756
TCLP 0.1524 0.1484 0.0991 0.1361 0.1283 0.1408
S0 IC Wi EP 996 354 1144 502 1710 447
TCLP 933 529 1139 490 1759 500
W2 EP 2794 2552 3169 2189 4052 2475
TCLP 3304 2970 3945 2691 4898 2833

-------
Table 6 continued
Extraction Lab
3
Analysis Lab
1 2
(ma/i)
3
Element
S0
Analysis
Metnod
IC
Waste
No.
W3
Extraction
Method
EP
TCLP
1
2
(mg/i)
370
404
114
50
502
543
52
52
992
1017
49
45
W4
EP
TCLP
51
181
15
16
150
195
12
16
208
416
10
15
W5
EP
TCLP
1415
1458
1278
1169
1661
1722
996
971
2444
2574
1053
967
W6
EP
TCLP
1627
1671
1626
1198
1241
1204
1292
1205
1856
1789
1459
1233
W7
EP
TCLP
561
589
191
228
454
411
186
197
972
970
174
191
V
ICAP
Wi
EP
TCLP
.0529
.0687
.1123
.0709
.0412
.0764
.0212
.0373
.1158
.1492
.08
.04
W2
EP
TCLP
.0344
1.0495
.0325
.1135
.0358
.3023
.02
.6638
.0633
.51
.0242
.42
W3
EP
TCLP
.0154
.0136
.0129
.0156
.0098
.0145
.0069
.0104
.0225
.0197
.02
.02
W4
EP
TCLP
.0118
.0364
.0102
.0078
.0093
.0087
.0061
.0223
.0135
.0053
.02
.03
W5
EP
TCLP
.0203
.0428
.0392
.0317
.0417
.0342
.0091
.0097
.0758
.07
.02
.04
W6
EP
TCLP
.1129
.1683
.0613
.1183
.0537
.0811
.0398
.0916
.16
.235
.525
.0667
W7
EP
TCLP
.0874
.2142
.0444
.2153
.0783
.1981
.0563
.2033
.1217
.2258
.0425
.2025
GFAA
Wi
EP
TCLP
.0356
.0842
.0561
.0424
.0369
.0764
.0482
.0751
.0293
.0550
.0470
.0620
W2
EP
TCLP
.0115
1.1006
.0087
.2130
.0087
.3220
.0200
.7369
.0020
.4542
.0040
.5658
W3
EP
TCLP
.0142
.0151
.0092
.0116
.0087
.0126
.0200
.0206
.0038
.0069
.0063
.0098
W4
EP
TCLP
.0015
.0091
.0088
.0087
.0087
.0087
.0200
.0175
.0021
.0021
.0040
.0040

-------
Analysis
Method
GF AA
Extraction
______ Method
EP
TCLP
W6 EP
TCLP
W7 EP
TCLP
Wi EP
TCLP
W2 EP
TCLP
W3 EP
IC L P
W4 EP
TCLP
W5 EP
TCL P
W6 EP
TCL P
W7 EP
IC L P
Wi EP
IC L P
W2 EP
IC L P
W3 EP
IC L P
W4 EP
IC L P
W5 EP
TCLP
W6 EP
IC L P
Table 6 continued
Extraction Lab
1 2 3
(mg/i)
.0015 .0090
.0116 .0053
.0612 .0487
.1971 .2135
0.0928 0.0743
0.1298 0.1351
4.86 5.35
5.06 5.48
0.022 0.017
0.037 0.100
0.035 0.023
0.55 0.112
1.38 1.56
1.39 1.62
0.13 BDL
0.17 0.35
0.078 0.107
0.134 0.216
0.109 0.076
0.129 0.109
5.14 5.49
5.16 5.25
0.027 0.040
0.045 0.111
0.048 0.034
0.079 0.137
1.175 1.501
1.438 1.555
0.127 0.048
0.181 0.184
0.165 0.153
0.252 0.201
5.60 5.32
5.45 5.35
0.091 0.108
0.166 0.162
0.071 0.072
0.210 0.184
1.484 1.252
1.365 1.533
0.112 0.127
0.271 0.282
.0 040
.0044
.0 651
.1183
.0713
.2150
0.185
0.2408
5.25
5.32
0.103
0.202
0.068
0.183
1.50
1 . 64
0.13
0.26
0. 149
0.284
0.179
0.223
5.08
5.26
0.123
0.204
0.082
0.206
1.342
1.583
0. 126
0.206
Waste
No.
W5
Analysis Lab
1 2
3
Element
V
Zn FAA
I CAP
.00d7 .0200
.0091 .0175
.0 549
.1082
.05 59
.1996
0.16
0.2349
4.95
5.32
0.082
0. 135
0.060
0.165
1.45
1.57
0.12
0.26
0. 109
0.192
( mg/i )
.0023
.0021
.0198
0446
.0605
.20 58
0.17
0. 2 383
5.32
5.48
0.097
0.166
0.063
0.182
1.51
1.63
0.12
0.41
0.148
0.239
.0707 .0311 .0327
.1027 .1039 .0663
.07 62
.2 108
0.372
0. 4843
5.34
5.67
0.262
0.398
0.141
0.399
1.54
1.88
0.22
0.44
0.231
0.385
0. 332
0.479
5.52
5.73
0.281
0.415
0.152
0.427
1.505
1.465
0.186
0.425

-------
Table 6 continued
Extractior Lab Analysis Lab
Analysis Waste Extraction 1 2 3 1 2 3
Element Method No. Method ( mg/i) ( mg/i )
Zn ICAP W7 EP 0.096 0.123 0.253 0.143 0.165 0.149
TCLP 0.156 0.240 0.438 0.250 0.246 0.292
BDL = below detection limit.

-------
Table 7
MEAN CONCENTRATIONS AND COEFFICIENT OF
VARIATION (Cv) MEASURED IN THE EP-TCLP EXTRACTS
ALKALINE ACIDIC SCRUBBER SCRUBBER NEUTRAL
ELEMENT/ FLY ASH FLY ASH BOTTOM ASH BOTTOM ASH SLUDGE SLUDGE FLY ASH
CHEMICAL EXTRACT, mean CV mean CV mean CV mean CV mean CV mean CV mean CV
TECH. METHOD (mg/i) (5) (mg/i) (5) (mg/i) (5) (mg/i) (5) (mg/i) (5) (mg/i) (5) (mg/i) (5)
Ag
GFAA
EP
TCLP
BOL
BDL
BDL
BDL
BDL
BIlL
BDL
BDL
BDL
BDL
BI lL
BDL
BDL
BIlL
As
GFAA
EP
TCLP
0.0109
0.0088
101
38
0.0054
0.3168
72
72
0.0053
0.0053
74 0.0054
75 0.0053
72
74
0.0066
O.0D56
88
69
0.0083
0.0104
39
20
0.0509
0.1486
67
14
Ba
ICAP
EP
TCLP
0.4057
0.3272
44
45
0.0771
0.0977
42
43
0.4297
0.8187
51 0.1709
41 0.2867
121
105
0.0814
0.1213
53
57
0.1636
0.2088
34
54
0.1766
0.4460
34
31
B
ICAP
EP
TCLP
17.14
17.72
14
8
45.45
44.85
6
6
1.43
1.59
20 0.22
18 0.27
96
45
0.84
0.92
34
29
1.69
1.89
24
29
1.20
1.23
20
12
Cd
GFAA
EP
TCLP
0.013]
0.0195
31
25
0.2454
0.2444
21
16
0.0020
0.0007
194 0.0017
35 0.0035
52
358
0.0333
0.0280
25
17
0.0053
0.0043
48
38
0.0055
0.0060
18
32
Cd
ICAP
EP
TCLP
0.0226
0.0218
41
46
0.2242
0.2329
17
43
0.0081
0.0076
63 0.0076
49 0.0091
49
77
0.0296
0.0289
39
41
0.018
0.015
64
84
0.0096
0.0099
67
73
Cr
GFAA
EP
TCLP
0.3572
0.3518
26
27
0.0157
0.8602
145
50
0.0082
0.01
75 0.0017
46 0.0043
38
55
0.0304
0.0415
85
86
0.0285
0.0129
219
116
0.0025
0.0590
54
30
Cr
ICAP
EP
TCLP
0.4268
0.4696
18
15
0.0661
0.9207
30
44
0.0361
0.0402
82 0.0294
68 0.0435
115
146
0.0627
0.0997
35
61
0.0893
0.0873
91
87
0.0369
0.0879
97
39
F
ISE
EP
TCLP
0.28
0.32
69
72
1.58
1.50
24
45
0.06
0.09
67 0.05
92 0.09
72
131
7.33
8.33
30
33
2.07
2.02
41
17
0.86
1.48
51
14
Hg
Cold
Vapor
EP
TCLP
BDL
BDL
--
--
801
BIlL
--
--
BDI
801
-- BD I
—- BDL
--
——
BDI
BDL
--
-—
BDL
BIlL
-—
BDL
BIlL
Pb
GFAA
EP
TCLP
BDL
BDL
0.018
0.18
54
67
BDL
BDL
BDL
BDL
BDL
BDL
BDL
BIlL
BDL
BDL
Mn
FAA
EP
TCLP
3.43
4.8
17
12
3.53
4
7
8
0.46
0.62
25 0.03
12 0.03
33
47
1.41
1.51
12
5
1.01
1.93
70
6
0.14
0.16
32
32
Mn
ICAP
EP
TCLP
3.23
4.52
20
10
3.39
3.79
5
4
0.48
0.61
24 0.035
11 0.032
157
17
1.33
1.52
26
8
1.04
1.93
70
8
0.15
0.17
10
6
Se
GFAA
EP
TCLP
BDL
BDL
BIlL
0.0249
65
BDL
BDL
BDL
801
BDL
BDL
BDL
BDL
0.06
0.14
39
27
504
IC
EP
TCLP
848
874
90
84
2834
3396
34
33
333
340
163 70
173 136
202
233
1448
1451
49
57
1511
1389
33
36
418
427
120
113
V
GFAA
(P
TCLP
0.04
0.07
47
28
0.01
0.60
86
74
801
801
BD I
BDL
BDL
BDL
0.05
0.09
61
43
0.06
0.21
25
11
V
ICAP
EP
TCLP
0.0672
0.0717
92
91
0.0343
0.5445
74
81
0.0131
0.0144
86 0.0106
68 0.0203
76
275
0.0324
0.0369
113
83
0.08
0.127
100
84
0.07
0.21
78
10
Zn
FAA
EP
TCLP
0.17
0.24
83
72
5.15
5.37
6
7
0.09
0.16
140 0.06
106 0.18
86
96
1.48
1.61
8
14
0.12
0.31
66
91
0.13
0.23
53
60
Zn
ICAP
EP
TCLP
0.17
0.23
74
79
5.36
5.36
6
6
0.11
0.18
125 0.07
101 0.?
74
87
1.37
1.48
25
23
0.12
0.26
57
56
0.15
0.26
48
59
BDL = below detection limit

-------
Table 8
RELATIVE FREQUENCY DISTRIBUTION OF RATIO OF
MEAN TCLP CONCENTRATIONS TO MEAN EP CONCENTRATIONS
Re 1 at I v e
Frequency
Ratio Range Frequency* ( % value )
<0.5 2 1.5
0.8— <1.0 16 12.0
1.0— <1.20 62 46.6
1.20—<1.50 17 12.8
1.50—<2.OO 16 12.1
2.00— 5.00 12 9.0
>5.00 8 6.0
TOTAL 133 100%
* When both TCLP and EP concentrations are below detection limit, the ratio
has been assigned a value of 1.0.

-------
percent of the ratio values are less than 1, with 12 percent
falling in the range of 0.8 to 1. Forty-seven percent of
the ratios fell between 1.0 and 1.2, another 25 percent were
between 1.2 and 2.0, and the remaining 15 percent exceeded
2.0. Overall, mean TCLP concentrations exceeded the EP con-
centrations 86 percent of the time.
SUMMARY AND CONCLUSIONS
The two EPRI sponsored round-robin studies have produced
data on the reproducibility and variability of extract con-
centrations produced by the EP and the TCLP tests when
applied to electric utility industry solid wastes. Both
these studies show that reproducibility differs for the
various constituents, waste types, extraction methods,
laboratories carrying out the extractions, and laboratories
analyzing the extracts. The TCLP and EP tests do differ
with respect to the relative size of the components of
variance. It appears that generally the reproducibility of
the TCLP is equal to or better than the EP method for As, B,
Cd, Cr, Mn and V. In the 1979 study the largest variability
in measured concentrations in extracts from the EP method
appeared to be due to differences in between-laboratory
analyses (the Ck component), while in the 1985-86 study of
the EP test it appears to be mostly due to differences in
the amount extracted at different extraction laboratories
(component a ) and to a lesser extent the analytical varia-
bility associated with analysis of duplicate extracts by
different laboratories (component dljk). The most
frequently encountered source of variability in the TCLP
extracts appears to be due to this same component (djjk).
In general, higher concentrations are measured in the TCLP
extracts compared to the EP extracts for a given constituent
and waste type. This is thought to be due to differences in
the extraction fluids (e.g., acetic acid for the EP test and
either acetic acid or sodium acetate buffer for the TCLP
test), and lack of pH adjustment during extraction in the
TCLP test.
REFERENC ES
1. Rose, S.J., J. Dane, B. Bynon and P. Switzer.
Extraction Procedure and Utility Industry Solid
Waste . Electric Power Research Institute, Report No.
EA-1667, 1981.
2. Eynon, B. and P. Switzer. A Statistical Comparison
of Two Studies on Trace Element Composition of Coal
Ash Leachates . Electric Power Research Institute,
Report No. EA-3181, 1983.
3. Mason, B. and D. Carlile. Round Robin Evaluation for
Selected Elements and Anionic Species from TCLP and
HP Extractions. Electric Power Research Institute ,
Draft Report, 1986.

-------
COLLABORATIVE STUDY OF THE TOXICITY CHARACTERISTICS
LEACHING PROCEDURE (TCLP) FOR METALS, PESTICIDES, AND
SEMIVOLATILE ORGANIC COMNXJNDS
D.R. TAYLOR AND W.B. BLACKBURN, S—CUBED, LA JOLlA, CA;
L.R. WILLIAMS, ENVIRONMENTAL MONITORING ANS SUPPORT
LABORATORIES, U.S. ENVIRONMENTAL PROTECTION AGENCY, LAS
VEGAS, NV; T.A. KIMMELL, OFFICE OF SOLID WASTE, U.S.
ENVIRONMENTAL PROTECTION AGENCY, WASHINGTON, DC
ABSTRACT
The United States Environmental Protection Agency has
developed a new procedure, the Toxicity Characteristic
Leaching Procedure (TCLP), to more effectively simulate
the leachinci of hazardous waste in a landfill
vironmertt. Thu proc: 1iir involves an 18—hour
extraction of a s n ple with either an acid or sodium
acetate solution, and u u u. -uL analysis of the
leachate for metals, pesticides, and semi—volatile
organic compounds. To vali.date the method, three waste
samples at two different pH levels were sent to 23
differ. nt volunteer government and commercial
laboratories for extraction and analysis. The results
and statistical analysis of this collaborative test are
presented and discussed. The results of the
collaborative study indicate that. the TCLP can be
applied consistently by a diverse group of
Ørganizations for the analyses cc.nsidered in this
study.
DISCLAIMER
Although the research described in this report has been
funded wholly or in part by the United States
Environmental Protection Agency through Contract No.
68—03—1958, it does not necessarily reflect the views
of the Agency and no official endorsement should be
inferred.

-------
1.0 INTRODUCTION
As part of its effort to improve the Extraction Procedure (EP) protocol and to expand its
applicability to all toxic constituents that could leach from hazardous wastes in landfills,
the U.S. Environmental Protection Agency (EPA) recently proposed the Toxicity
Characteristic Leaching Procedure (TCLP). The TCLP was published as a draft
protocol on December 20, 1985. It was officially published in the Federal Register on
January 14, 1986, as part of the Land Disposal Restrictions Rule and again for public
comment on June 13, 1986, as part of EPA’s efforts to expand its Toxicity
Characteristic. The protocol calls for the extraction of semi-volatile organic
compounds, metals and pesticides using a bottle or jar similar to the EP procedure.
For the extraction of volatile organic compounds (VOCs), a new device known as a
Zero Headspace Extractor (ZI-lE) is used.
The purpose of this study was to assess the interlaboratory precision of the TCLP for
the determination of metals, pesticides, base-neutral/acid extractable organic
compounds, and volatile organic compounds. As the volatile organic compounds
portion of this study is ongoing, this paper reports on the results of the study for
metals, pesticides and base-neutral/acid extractable organic compounds. In addition,
while this paper incorporates the majority of the data from the participating
laboratories, we anticipate receiving additional data at a later date which will be
incorporated into a final analysis.
TABLE 1.1. Participating Laboratories
ALBERTA ENVIRONMENTAL CENTRE LEMSCO
CHEMICAL WASTE MANAGEMENT MICROBAC LABS
COMPUCHEM LABORATORIES NUS
EG&G IDAHO OAK RIDGE NATIONAL LABORATORY
ENSECO PEI ASSOCIATES
ENVIRODYNE RADIAN CORPORATION
ENVIRONMENTAL SCIENCE & ENGINEERING ROCKY MOUNTAIN ANALYTICAL
ENVIRONMENTAL TESTING & CERTIFICATION S-CUBED
INDUSTRIAL & ENVIRONMENTAL ANALYSTS THERMAL-ANALYTICAL LABS
IT CORPORATION U.S. DEPT. OF ENERGY-MORGANTOWN
LANGSTON LABS WESTERN RESEARCH INSTITUTE
WILSON LABS
1

-------
2.0 METHODS
2.1 Collaborative Study Design
A total of 23 laboratories participated in the analytical aspects of the study. These
laboratories are shown in Table 1.1.
Each of the laboratories agreeing to participate were invited to a meeting at S-CUBED
in San Diego, California. This meeting was held to discuss with the participants the
details of the collaborative study and to familiarize them with the TCLP procedure.
Familiarization samples were distributed to the participants for return to their
laboratories and subsequent TCLP analysis for semi-volatile organic compounds and
metals.
A flowchart outlining the structure of the collaborative study is given in Figure 2.1. After
analysis of familiarization samples (consisting of spiked sludges), the participating
laboratories showing outlier values were contacted to identify and rectify possible
analytical problems before proceeding with the collaborative study samples. In
addition, statistical methods were refined after evaluation of familiarization sample data.
The collaborative study samples consisted of three different matrices; (1) an ammonia
lime still bottom, (2) an API separator sludge, and (3) a fossil fuel fly ash. Each of these
wastes was identified throughout the study as Waste A, Waste B, and Waste C,
respectively. The pH was adjusted on split portions of each waste so that one portion
would require the use of Extraction Fluid 1; the other portion requiring the use of
Extraction Fluid 2 as defined in TCLP protocol. A total of six different samples were
created: three different sample matrices each with two different pH’s.
The samples were distributed among the participating laboratories such that each
laboratory received two portions from a sample waste type, each portion having a
different pH, thus requiring a different extraction fluid. In addition, a single sample at
one pH level of another waste type was received by each laboratory, for a total of three
samples for leaching and analysis. Samples were shipped to the participating
laboratories via overnight courier in glass containers with Teflon-lined caps. Samples
were packaged with ice packs to ensure the samples remained cold during transit.
2

-------
2.2 Analytical Methods
Toxicity Characteristic Leaching Procedure
The TCLP is a second generation extraction procedure which improves upon the
existing EP technique and allows for the inclusion of an expanded list of volatile and
semi-volatile organic compounds. A new ZHE has been developed for a separate
leaching of wastes to be analyzed for volatile organic compounds. Existing EP
apparatus, with few modifications, may be employed for leaching of wastes to be
analyzed for metals and semi-volatile organic compounds (including pesticides).
An outline of the TCLP procedure is shown in Figure 2.2. As with the EP, an initial
liquid/solid separation is made. Glass fiber filters (0.6-0.8 pm pore size) are used for all
filtrations in the TCLP rather than the membrane filters used in EP. After the initial
separation, a pH measurement is made of the solid phase for determination of the
appropriate extraction fluid to be used in the leaching procedure. The waste and
extraction fluid are loaded into the extraction vessel and rotated at 30 a 2 RPM for 18
hours.
No pH adjustments are made during this leaching period. A final liquid/solid separation
is made at the conclusion of the 18-hour leaching. The solid phase is discarded and
the final filtrate is either combined physically with the initial filtrate and analyzed, or the
initial.and final leachates are analyzed separately and the results are mathematically
combined.
Chemical Analysis Procedure
For the purpose of comparability, all participating laboratories were requested to
adhere to SW-846 1 analytical protocol. A summary of these methods is given in Figure
2.3.
3

-------
SAMPLE (LEACHATE) PREPARATION METHODS
Analyte Group Technique Reference 1
Semi-Volatile Organic Compounds Solvent Extraction 3510
Metals Acid Digestion Series 3Oxx
Pesticides Solvent Extraction 3510
ANALYTICAL PROCEDURES
Analyte Group Technique Reference
Semi-Volatile Organic Compounds Gas Chromatography/Mass Spectroscopy 8270
Metal&- Atomic Absorption Series 7Oxx
I Inductively Coupled Plasma 6010
Pesticides Gas Chromatography/Electron Capture Detection 8270
Figure 2.3. Chemical Analysis Methods
4

-------
3.0 STATISTICAL DESIGN
From a statistical point of view, the purpose of the study was to determine the efficacy
of the TCLP. Efficacy was judged by the precision of the method measured among
replicates conducted under the same conditions by a group of collaborating
laboratories. Precision was determined by appropriate variance estimates derived from
a designed statistical experiment and the subsequent analysis of variance (ANOVA) 2 .
Since there was no direct knowledge of the exact contents of the test samples, bias
was not measured directly. However, the relative distribution of bias among
collaborating laboratories was considered.
The following discussion presents a step-by-step outline of the statistical aspects of the
program and details the statistical procedures used. In several instances, hypothetical
numbers are used to demonstrate various aspects of the analyses. Relevant tables
and graphs are presented, where appropriate, in idealized form.
Raw data were submitted by each collaborating laboratory in the following form (Table
3.1). A separate submission was made for each type of sample (semi-volatile organic
compounds, metals, and pesticides). For all submittals, a careful distinction was made
between zero results and missing values.
TABLE 3.1. Raw Data Submittal Form
Laboratory Test Sample No.
Familiarization Sample One Two Three
List of Materials Rep 1 Rep 2 Rep 1 Rep 2 Rep 1 Rep 2 Rep I Rep 2
1 * * * * § § *
2 § ii $1 * ii
3 $1 * § § § § §
S S S S S S S
S S S S S S S S
S S S S S S S S S
m § § § $1 § § § §
5

-------
However, the statistical analysis began using procedures similar to those outlined in the
AOAC Manual 3 (Pages 74 and 75). First, the data were tabularized as shown in Table
3.2.
TABLE 3.2. Tabularized and Ranked Data
Sample 1 Sample 2 Sample 3
Total
Lab Sum Rank p Sum Rank p Sum Rank Rank
I I I $1 * § § § § § *
2 $1 II § *
3 $1 § *Q It It ItIt II It *
• . . • S S • S S • S S S S
• . S S S S • S S •S • S •
• . S S S S S S S • S • • S
C §It § It It § §* It * It
The ir formation in Table 3.2 was used for two separate, but related, purposes:
(1) The identification of outlying laboratories, and
(2) The identification of outlying individual samples results.
The Rep column in Table 3.2 represents the sum of all concentrations of the
appropriate analytes, herein called total recover)’. This particular univariate treatment of
the data was somewhat artificial. However, there was at least some justification based
on the following: at the time that the study was designed, it was felt that if a
laboratory’s measurements were consistently high (or low) on a given sample, then the
measurements for all or most of the materials list would also be consistently high (or
low). Therefore, dependencies among materials would tend to emphasize
measurement differences among laboratories, but would not greatly influence the
ranking of the laboratories.
Computational procedures for outlying laboratories in Table 3.2 were as follows. The
Sum column was calculated separately for each sample as the sum of the two
corresponding Rep values. The Rank column for each sample was then calculated
6

-------
based on the Sum column. If two or more laboratories tied for a given rank, each was
given the mean rank of all tied ranks. For example, if three laboratories tied for rank k,
then each was given the rank of [ k + (k + 1) + (k + 2)]/3. Finally, the three Rank
columns were summed to obtain the Total Rank column.
Each Total Rank was compared with the upper and lower 95 percent rank limits given
in Appendix Table B (AOAC Manual) for the appropriate number of collaborating
laboratories. The rank limits are based on the fact that any laboratory that is
consistently high or low on all samples will have a Total Rank that is inordinately high or
low, respectively. Laboratories having a Total Rank outside the tabulated limits were
identified as outliers.
Individual outliers were identified using data from all laboratories, including those
identified as outliers in Table 3.2. Each sample was considered independently. First,
Sum values for the sample were arranged in ascending order and designated as S(1)
through S(n), where S(1) had the lowest rank and S(n) the highest. The following ratios
were computed for the Sum column values designated S(1) and S(n) (i.e., the two with
the most extreme ranks):
S(n) highest: ( S)n — S(n—2 )
S(n) — S(3)
S(1) lowest: S(3) — S(1 )
S(n-2) — S(1)
If either ratio exceeded the tabulated value (Appendix Table C.1, AOAC Manual) for n
measurements, then that value was identified as an outlier. After the removal of
outliers, the data were reranked and the entire procedure repeated.
The experiment design was defined by the following independent variables and
relevant characteristics:
C - Collaborating Laboratories c = 23
S - Samples (test samples of Table 2.1) s = 3
B - Batches (sample lots) b = 2
R - Replicates (duplicate analyses) r = 2
Each independent variable was considered as a random variable, meaning that it can
be interpreted as a sample of a larger population. The dependent variable was total
recovery.
7

-------
The resultant ANOVA is shown in Table 3.3. In the source of variation (SV) column, the
notation CS indicates an interaction (i.e., the synergistic effect of C and S); the notation
B/CS indicates that B is nested within CS (i.e., B might be different in each C and S
combination). The numbers in parenthesis in the degrees of freedom (DF) column are
based on the above definitions. A significant F-ratio indicates a failure of the extraction
method on the indicated independent variable. For example, the significant F-ratio on
C would indicate that the variability among collaborating laboratories was unacceptably
high, or alternatively, that the potential for systematic error between laboratories was
too great. The exception is that a significant Sample (S) main effect is allowable
because the samples were in fact different.
TABLE 3.3. Analysis of Variance (ANOVA)
SV DF SS MS F-Ratio
Total csbr (600) SS(T) -
Mean 1 ( 1) SS(mean) — -
C c-i (14) SS(C) MS(C) MS(C) / MS(CS)
S s-i ( 1) SS(S) MS(S) MS(S) / MS(CS)
CS (c-1)(s-i) (14) SS(CS) MS(CS) MS(CS) / MS(B/CS)
B/CS cs(b-1) (270) SS(B/CS) MS(B/CS) MS(B/CS) / MS(R/CSB)
R/CSB csb(r-1) (300) SS(R/CSB) MS(R/CSB)
Following the ANOVA, the following variance and standard deviation estimates were
calculated:
Variance Estimates:
Between Replicates: S = MS(R/CSB)
Batches (Lots): = (MS(B/CS) — S 0 ]/r
Cs Interaction: C = (MS(CS) — SB - rS 0 ]/rb
Among Laboratories: S = (MS(C) - SB — rS 0 — rbS 5 ]/rbs
8

-------
Standard Deviation Estimates:
Repeatability: S 0 = 4 MS(R/CSB)
Reproducibility: SR = + + S
A note is in order on the calculation of sums of squares (SS). Because of the
prevalence of missing values, a method for unbalanced ANOVA was used. By this
method, only the values actually present contribute to the sums of squares; there is no
attempt to estimate missing values. The degrees of freedom (DF) reflect the number of
valid cases making up each sum of squares. In the Results section of this paper
(Section 4.0), it will be noted that the degrees of freedom are generally very low in
relation to comparable values if all cases were present.
9

-------
4.0 RESULTS AND DISCUSSION
Tables 4.1 to 4.18 give the results of the collaborative study. Tables 1 to 6 present
semi-volatile organic compound results for the 22 target compounds used in the study.
Summarized values are in pg/L, the standard deviation, and relative standard deviation
(RSD) are given. The RSD generally varied in the range of 20 to 120 percent for these
compounds. No discernible trend was observed with respect to either the base/neutral
or acid fractions from one sample batch to another.
Tabtes 4.7 to 4.12 give results for the metals extraction and analyses. The RSD for
these results generally vary between 20 and 110 percent, with most values being below
90 percent. The variations of the different metals were not consistent from waste to
waste. Thus, the TCLP does not appear to create particular extraction/analysis
problems for any particular element. The use of the lower pH extraction system did
result in generally higher metal results and less variable data.
Tables 4.13 to 4.18 present pesticide data. Only a limited number of pesticides were
spiked in this study, so much of the data reported were detection limits which varied by
organization. The RSD values for most compounds, especially dieldrin, 4,4-DDE, and
endrin, do not reflect actual analytical results, but only detection limit variability.
Tabl s 4.19 to 4.27 present the results of the rank order test for the laboratories for
semi-volatile organic compounds, metals, and pesticides, respectively. For both the
semi-volatile organic compounds and metals, the results were good. No laboratories
showed unacceptable total recoveries for organic compounds, and only two of the
laboratories had unacceptable metals data. The pesticides data (Table 4.21) again
reflect the variable detection limft problem for unspiked pesticides and a number of
laboratories were found to be statistically unacceptable.
Tables 4.22, 4.23, and 4.24 report the analysis of variance results for the study. The
key factor in these results is found in the last column of each table. The first factor
gives the probability that the total recoveries for the different laboratories were
statistically different. The values of 0.54, 0.47, and 0.85 for semi-volatile organic
compounds, metals and pesticides, respectively, indicate that the performance of the
laboratories was essentially the same. This was less true of the pesticides than for the
other two types of analyses. The next value, 0.96, 0.93, and 0.65 for the three types of
analyses, reflects whether the samples differed significantly from each other. In the
case of semi-volatile organic compounds and metals, this was clearly the case. For
pesticides, the results suggest this was not true. None of the pesticide samples
10

-------
contained background levels of pesticides, unlike the other materials. The smaller
number of variables (six) and lack of background material was undoubtedly
contributory to the statistical similarity of the results.
The third probability value of 0.01, 0.06, and 0.02 reflects the probability that the total
recovery varies among laboratory sample combinations. These results reflect well on
the overall consistency of the study, since they indicate little probability that there was a
significant difference in the overall analyte concentration patterns observed among
laboratories for different samples.
Finally, the last results from Tables 4.22, 4.23, and 4.24 indicate that the total recoveries
did not vary within a batch. Thus, Samples A-I and A-2, B-l and B-2, and Samples C-l
and C-2, which were designed to require different TCLP leaching pH solutions, all yield
comparable recoveries. As would be expected for the semi-volatile organic
compounds and pesticides where water solubility is a critical factor, these probabilities
are very low, 0.000 and 0.09, respectively. For metals, not surprisingly, the probability
is much higher, 0.69, but it is still not significant at the 95 percent confidence level.
Overall, the statistical design used in this study was successful in providing the answers
required. If the design and analysis were in error, they were in error on the
conservative side as far as the result required to call the TCLP successful. The sums of
squares (SS) and mean square error (MSE) in the ANOVAs were quite large. However,
they vere artificially so because of the use of total recovery as the dependent variable.
The major weakness in the statistics was the use of total recovery in outlier detection
and as the dependent variable in the ANOVA. Its use assumes that the semi-volatile
organic compounds, pesticides, and metal samples are independent of one another.
This assumption is not necessarily true for all anatytes. Multivariate methods to deal
with this problem are under development at the present time (see paper by Show,
Williams and Taylor in these proceedings). Whether these methods will yield similar
results remains to be seen. When the multivariate methods become available, they will
be used in addition to the univariate treatment used in this study.
11

-------
TCLP COLLOBORATIVE STUDY RESULTS
Semi-volatile Organic Compounds
Table 4.1
SAMPLE A-i
Compound i SD RSD
Phenol 6950 1650 24
1 ,2-Dichlorobenzene 117 90.3 77
2-Methylphenol 933 538 58
4-Methy lphenol 3540 1910 54
Nitrobenzene 361 223 62
2-Nitrophenol 8.20 6.40 78
1,2,4-Trichlorobenzene 7.31 4.60 63
Naphthalene 774 618 80
Hexachlorobutadiene 8.72 2.41 28
2-Methylnaphthalene 136 79.7 59
2 1 4,6-Trichlorophenol 363 139 38
2,4,5-Trichlorophenol 1800 1120 62
2-Chloronaphthalene 78.1 58.5 75
4-Nitrophenol 20.0 21.6 108
Hexaôhlorobenzene 7.14 4.88 68
Pentach orophpenol 1230 737 60
Phenanthrene 126 70.7 56
Di-N-Butylphthalate 8.74 7.05 81
Fluoranthene 25.6 14.9 58
Bis(2-ethylhexyl) Phthalate 9.21 3.71 40
Chrysene 5.97 5.06 85
Di-N-Octylphthalate 7.31 3.91 54
12

-------
TCLP COLLOBORATIVE STUDY RESULTS
Semi-volatile Organic Compounds
Table 4.2
SAMPLE A-2
Compound SD RSD
Phenol 15900 8410 53
1,2-Dichlorobenzene ND ND —
2-Methylphenol 1880 856 46
4-Methy lphenol 10300 5150 50
Nitrobenzene 6.49 4.88 75
2-Nitrophenol 16.2 15.5 96
1,2,4-Trichlorobenzene 7.11 4.42 62
Naphthalene 1300 1200 92
Hexachlorobutadiene ND ND —
2-Methylnaphthalene 134 106 79
2,4 ,6-Trich lorophenol 7.10 4.42 62
2,4 ,5-Trichlorophenol 37.7 31.7 84
2-Chloronaphthalene ND ND —
4-Nitrophenol 28.4 31.0 109
Hexaàhlorobenzene ND ND —
Pentachlorophpenol 36.3 29.6 82
Phenanthrene 130 82 63
Di-N-Butylphthalate 8.80 6.57 75
Fluoranthene 13.9 13.5 97
Bis(2.ethylhexyl) Phthalate ND ND —
Chrysene 6.49 4.88 75
Di-N-Octylphthalate ND ND ND
13

-------
TCLP COLLOBORATIVE STUDY RESULTS
Semi-volatile Organic Compounds
Table 4.3
SAMPLE B-I
Compound i SD RSD
Phenol 113 81.0 72
1,2-Dichlorobenzene 19.0 13.8 73
2-Methylphenol 136 92.6 68
4-Methylphenol 158 98 62
Nitrobenzene 164 108 66
2-Nitrophenol 5.92 4.89 83
1 ,2,4-Trichlorobenzene 6.52 4.76 73
Naphthalene 133 79.5 60
Hexachlorobutadiene 6.98 4.02 58
2-Methylnaphthalene 70.4 45.6 65
2,4,6-Trichlorophenol 88.8 43.9 49
2,4,5-Trichlorophenol 405 251 62
2-Chloronaphthalene 5.89 4.18 71
4-Nitrophenol 20.2 26.0 128
Hexaèhlorobenzene 5.43 5.02 92
Pentachlorophpenol 174 175 101
Phenanthrene 5.28 4.08 77
Di-N-Butylphthalate 6.13 5.35 87
Fluoranthene 5.50 4.71 85
Bis(2-ethylhexyl)Phthalate 16.4 21.4 130
Chrysene 5.08 4.86 96
Di-N-Octylphthalate 5.71 5.10 89
14

-------
TCLP COLLOBORATIVE STUDY RESULTS
Semi-volatile Organic Compounds
Table 4.4
SAMPLE B-2
Compound I SD RSD
Phenol 74.2 60.1 81
1,2-Dichlorobenzene 20.1 13.5 67
2-Methylphenol 34.0 29.5 87
4-Methylphenol 45.0 29.4 65
Nitrobenzene 110 67.1 61
2-Nitrophenol 18.8 26.3 140
1,2,4-Trichlorobenzene 3.6 4.98 138
Naphthalene 165 32.9 20
Hexachlorobutadiene 3.6 4.98 138
2-Methylnaphthalene 65.3 46.3 71
2 ,4 ,6-Trichlorophenol 276 264 96
2,4,5-Trichlorophenol 346 301 87
2-Chloronaphthalene 6.66 4.44 67
4-Nitrophenol 4.90 7.35 150
Hexathlorobenzene 2.92 4.54 156
Pentachlorophpenol 308 173 56
Phenanthrene 3.93 4.32 110
Di-N-Butylphthalate 6.47 4.88 75
Fluoranthene 3.20 4.38 137
Bis(2-ethylhexyl)Phthalate 25.3 31.6 125
Chrysene 3.6 4.98 136
Di-N-Octylphthalate 3.1 4.25 137
15

-------
TCLP COLLOBORATIVE STUDY RESULTS
Semi-volatile Organic Compounds
Table 4.5
SAMPLE C-i
Compound SD RSD
Phenol 24.7 21.6 87
1 ,2-Dichlorobenzene 22.4 21.8 97
2-Methylphenol 51.9 55.4 107
4-Methylphenol 26.8 25.1 94
Nitrobenzene 103 68.1 66
2-Nitrophenol 5.33 7.14 134
1,2,4-Trichlorobenzene 4.22 5.04 119
Naphthalene 35.2 34.2 97
Hexachlorobutadiene 4.45 4.83 108
2-Methylnaphthalene 4.22 5.04 119
2,4,6-Trich lorophenol 29.8 28.8 96
2,4,5-Trichlorophenol 176 235 134
2-Chloronaphthalene 39.2 49.6 126
4-Nitrophenol 16.1 23.1 144
Hexaôhlorobenzene 3.83 4.08 106
Pentachlorophpenol 50.1 53.2 106
Phenanthrene 4.60 4.81 105
Di-N-Butytphthalate 7.47 6.47 87
Fluoranthene 22.0 23.0 104
Bis(2-ethylhexyl)Phthalate 259 400 154
Chrysene 4.75 5.12 108
Di-N-Octylphthalate 11.2 15.2 136
16

-------
TCLP COLLOBORATIVE STUDY RESULTS
Semi-volatile Organic Compounds
Table 4.6
SAMPLE C-2
Compound SD RSD
Phenol 23.3 8.4 36
I ,2-Dichlorobenzene 13.3 13.9 104
2-Methylphenol 31.9 23.1 72
4-Methy lphenol 22.0 16.8 76
Nitrobenzene 54.0 30.4 56
2-Nitrophenol 6.3 9.5 150
1 ,2,4-Trichlorobenzene ND ND —
Naphthalene 36.6 31.6 86
Hexachlorobutadiene 7.38 3.33 45
2-Methylnaphthalene ND ND —
2,4,6-Trichlorophenol 20.3 9.5 47
2,4,5-Trichlorophenol 140 71.6 51
2-Chloronaphthalene 49.5 47.0 95
4-Nitrophenol ND ND —
Hexaôhlorobenzene 4.50 4.12 92
Pentachlorophpenol 287 51.9 18
Phenanthrene ND ND —
Di-N-Butylphthalate ND ND —
Fluoranthene 19.6 10.3 53
Bis(2-ethylhexyl) Phthalate 7.83 12.3 157
Chrysene ND ND —
Di-N-Octylphthalate 4.03 4.91 122
17

-------
TCLP COLLOBORATIVE STUDY RESULTS
Metals
Table 4.7
SAMPLE A-i
Metal SD RSD
Aluminum 5.83 3.60 62
Cadmium 52.6 31.4 60
Calcium 19600 8500 43
Chromium 1.54 1.43 93
Cobalt 13.1 7.3 56
Copper 3.10 1.71 55
Iron 2.22 1.36 61
Lead 2.97 2.67 90
Magnesium 2220 446 20
Manganese 131 58 44
Nickel 1.50 1.49 99
Thallium 7.11 5.37 76
Vanadium 0.87 1.12 129
Table 4.8
SAMPLE A-2
Metal I SD RSD
Aluminum 14.9 13.5 91
Cadmium 27.6 14.8 54
Calcium 19300 1300 7
Chromium 3.79 3.79 100
Cobalt 6.70 3.78 56
Copper 0.53 0.28 53
Iron 8.3 6.9 83
Lead 3.89 2.59 67
Magnesium 2330 598 26
Manganese g.gi 0.32 3
Nickel 1.93 1.82 94
Thallium 2.691 2.90 108
Vanadium 0.98 1.04 106
18

-------
TCLP COLLOBORATIVE STUDY RESULTS
Metals
Table 4.9
SAMPLE B-i
Metal i SD RSD
Aluminum 2.60 1.79 69
Cadmium 4.59 2.82 61
Calcium 12000 5590 47
Chromium 56.1 22.7 40
Cobalt 6.84 7.88 115
Copper 2.15 0.58 27
Iron 1.45 1.62 112
Lead 3.12 3.13 100
Magnesium 421 78.3 19
Manganese 1.39 0.50 36
Nickel 0.95 1.00 105
Thallium 41.7 47.3 113
Vanadium 0.76 0.27 35
Table 4.10
SAMPLE 6-2
Metal I SD RSD
Aluminum 20.8 18.1 87
Cadmium 0.48 0.37 77
Calcium 105 41.4 39
Chromium 105 18.0 17
Cobalt 0.50 0.15 30
Copper 8.35 2.33 28
Iron 0.93 0.63 68
Lead 12.4 13.6 110
Magnesium 72.1 61.9 86
Manganese 0.13 0.06 46
Nickel 0.39 0.08 20
Thallium 74.1 40.4 54
Vanadium 11.3 7.03 62
19

-------
TCLP COLL.QBORATIVE STUDY RESULTS
Metals
Table 4.11
SAMPLE C-I
Metal SD RSD
Aluminum 521 429 82
Cadmium 87.1 67.4 77
Calcium 10900 2960 27
Chromium 18.5 14.0 76
Cobalt 51.7 40.2 78
Copper 0.35 0.36 103
Iron 31.1 25.8 83
Lead 8.69 7.40 85
Magnesium 1380 960 70
Manganese 19.0 15.4 81
Nickel 0.69 0.42 61
Thallium 40.9 22.3 54
Vanadium 0.52 0.37 71
Table 4.12
SAMPLE C-2
Metal SD RSD
Aluminum 2690 937 35
Cadmium 86.7 17.7 26
Calcium 11900 5740 48
Chromium 84.1 23.7 28
Cobalt 89.0 9.23 10
Copper 4.67 1.48 32
Iron 532 294 55
Lead 45.7 8.3 18
Magnesium 2360 164 7
Manganese 65.8 23.1 35
Nickel 1.17 0.66 56
Thallium 41.3 13.2 32
Vanadium 5.99 2.62 44
20

-------
TCLP COLLOBORATIVE STUDY RESULTS
Pesticides
Table 4.13
SAMPLE A-i
Compound i SD RSD
p-BHC 27.6 27.8 99
r-BHC (Lindane) 58.9 36.6 161
Aldrin 15.1 28.0 54
Dieldrin 4.16 6.69 62
4,4-DDE 13.2 35.1 38
Endrin 3.05 6.89 44
Table 4.14
SAMPLE A-2
Compound SD RSD
p-BHC 17.4 19.5 71
r-BHC (Lindane) 17.8 24.6 72
Aldrin 2.00 1.87 107
Dieldrin 2.63 2.36 111
4,4-DDE 1.50 0.71 211
Endrin ND ND
21

-------
TCL.P COLLOBORATIVE STUDY RESULTS
Pesticides
Table 4.15
SAMPLE B-I
Compound SD RSD
p-BHC 3.90 2.22 176
r-BHC (Lindane) 6.61 4.08 162
Aldrin 2.29 2.16 106
Dieldrin 1.88 2.22 85
4,4-DDE 1.71 1.86 92
Endrin 0.70 0.45 156
Table 4.16
SAMPLE B-2
Compound i SD RSD
p-BHC 3.08 3.01 102
r-BHC (Lindane) 15.64 20.18 76
Aldrin 0.98 0.90 109
Dieldrin 2.17 3.33 65
4,4-DDE 0.43 0.40 107
Endrin 0.63 0.48 131
22

-------
TCLP COLLOBORATIVE STUDY RESULTS
Pesticides
Table 4.17
SAMPLE C-i
Compound SD RSD
fl-BHC 56.7 48.0 118
F-BHC (Lindane) ND ND
Aldrin ND ND
Dieldrin ND ND
4,4-DDE ND ND
Endrin ND ND
Table 4.18
SAMPLE C-2
Compound SD RSD
p-BHC ND ND
r-BHC (Lindane) ND ND
Aidrin ND ND
Dieldrin ND ND
4,4-DDE ND ND
Endrin ND ND
23

-------
SUMMARY RESULTS
SEMI-VOLATILE ANALYSIS
Table 4.19
SAMPLE A
LOT1 LOT2 MEAN
Lab No. Total Rank Total Rank Total Rank a = 0.05
1 22200.50 6.00 16472.50 2.00 19336.50 4.00 Acceptable
4 12306.25 3.00 12117.45 1.00 12211.85 2.00 Acceptable
5 25888.00 7.00 36529.50 3.00 31208.75 5.00 Acceptable
7 19185.25 5.00 42998.95 4.00 31092.10 4.50 Acceptable
9
10
12
14
15
17
18 11094.90 2.00 - 11094.90 2.00 Acceptable
19
20 10044.65 1.00 - - 10044.65 1.00 Acceptable
23 - - 149001.41 5.00 149001.41 5.00 Acceptable
24 18266.00 4.00 - - 18266.00 4.00 Acceptable
Notes 1. Totals represent total recovery for the lab summed over all chemicals and all lots.
2. Ranks are based on total recovery - Rank 1 is given to the lowest total recovery, etc.
3 “Not Acceptable” indicates that total recovery is a statistical outlier.
24

-------
SUMMARY RESULTS
SEMI-VOLATILE ANALYSIS
Table 4.20
SAMPLE B
LOTI LOT2 MEAN
Lab No. Total Rank Total Rank Total Rank a = 0.05
1 1524.00 4.00 - 1524.00 4.00 Acceptable
4 739.35 2.00 - 739.35 2.00 Acceptable
5 2798.00 9.00 - 2798.00 9.00 Acceptable
7 1679.05 5.00 - - 1679.05 5.00 Acceptable
9 5376.00 10.00 2480.00 6.00 3928.00 8.00 Acceptable
10 1817.20 6.00 1792.95 4.00 1805.07 5.00 Acceptable
12 2688.50 7.00 1984.00 5.00 2336.25 6.00 Acceptable
14 1223.30 3.00 1714.35 3.00 1468.82 3.00 Acceptable
15 97.00 1.00 505.00 1.00 301.00 1.00 Acceptable
17 2797.00 8.00 1266.50 2.00 2031.75 5.00 Acceptable
18 -
19.
20 - - - - - -
23 39292.24 11.00 - 39292.34 11.00 Not Acceptable
24 - -
Notes : 1. Totals represent total recovery for the lab summed over all chemicals and all lots.
2. Ranks are based on total recovery - Rank 1 is given to the lowest total recovery, etc.
3. “Not Acceptable” indicates that total recovery is a statistical outlier.
25

-------
SUMMARY RESULTS
SEMI-VOLATILE ANALYSIS
Table 4.21
SAMPLE C
LOT1 LOT2 MEAN
Lab No. Total Rank Total Rank Total Rank a = 0.05
1 —
4 -
5 - -
7 -
9 1448.00 8.00 - 1448.00 8.00 Not Acceptable
10 742.85 5.00 - 742.85 5.00 Acceptable
12 964.00 7.00 - 964.00 7.00 Acceptable
14 926.60 6.00 - - 926.60 6.00 Acceptable
15 -
17 332.00 1.00 - - 332.00 1.00 Acceptable
18 572.90 2.00 512.35 1.00 542.63 1.50 Acceptable
19. 688.00 4.00 742.00 2.00 715.00 3.00 Acceptable
20 2808.00 8.00 1453.00 4.00 2130.50 6.00 Not Acceptable
23 - - - - - - -
24 601.00 3.00 819.00 3.00 710.00 3.00 Acceptable
Notes . 1. Totals represent total recovery for the lab summed over all chemicals and all lots.
2. Ranks are based on total recovery - Rank 1 is given to the lowest total recovery, etc.
3. “Not Acceptable” indicates that total recovery is a statistical outlier.
26

-------
SUMMARY RESULTS
PESTICIDES ANALYSIS
Table 4.22
SAMPLE A
LOll LOT2 MEAN
Lab No. Total Rank Total Rank Total Rank a = 0.05
1 133.00 5.00 42.50 4.00 87.50 4.50 Acceptable
2 157.70 7.00 11.30 1.00 84.50 4.00 Acceptable
4 131.00 4.00 82.50 5.00 106.75 4.50 Acceptable
5 146.75 6.00 29.55 3.00 88.15 4.50 Acceptable
7 427.50 8.00 450.00 6.00 438.75 7.00 Not Acceptable
9 -
10 - -
12 - -
14 -
17
18 1153.00 9.00 - - 1153.00 9.00 Not Acceptable
20. 15.74 9.00 18.80 2.00 17.27 2.00 Not Acceptable
21’ 53.00 3.00 - - 53.00 3.00 Acceptable
24 0.15 1.00 0.15 1.00 Not Acceptable
Notes : 1. Totals represent total recovery for the lab summed over all chemicals and all lots.
2. Ranks are based on total recovery - Rank 1 is given to the lowest total recovery, etc.
3. “Not Acceptable” indicates that total recovery is a statistical outlier.
27

-------
SUMMARY REPORTS
PESTICIDES ANALYSIS
Table 4.23
SAMPLE B
LOTI LOT2 MEAN
Lab No. Total Rank Total Rank Total Rank a = 0.05
1 72.25 6.00 72.25 6.00 Not Acceptable
2 23.85 4.00 23.85 4.00 Acceptable
4 13.85 1.00 - 13.85 1.00 Acceptable
5 18.55 2.00 16.55 2.00 Acceptable
7 450.00 9.00 - - 450.00 9.00 Not Acceptable
9 22.03 3.00 45.33 2.00 33.68 2.50 Acceptable
10 199.00 8.00 199.00 5.00 199.00 6.50 Not Acceptable
12 120.50 6.00 182.25 4.00 151.38 5.00 NotAcceptable
14 0.00 1.00 0.00 1.00 0.00 1.00 NotAcceptable
17 - 168.50 3.00 168.50 3.00 Acceptable
18
20 - -
21
24 -
Notes : 1. Totals represent total recovery for the lab summed over all chemicals and all lots.
2. Ranks are based on total recovery - Rank 1 is given to the lowest total recovery, etc.
3. “Not Acceptable” indicates that total recovery is a statistical outlier.
28

-------
SUMMARY RESULTS
PESTICIDES ANALYSIS
Table 4.24
SAMPLE C
LOT1 LOT2 MEAN
Lab No. Total Rank Total Rank Total Rank a = 0.05
1 -
2 - -
4 -
5 -
7 -
9 53.29 3.00 53.29 3.00 Acceptable
10
12 135.80 6.00 - - 135.80 6.00 Acceptable
14 78.00 4.00 - 78.00 4.00 Acceptable
17 -
18 1129.00 7.00 532.50 4.00 830.75 5.50 Acceptable
20 6.70 2.00 18.80 3.00 12.75 2.50 Not Acceptable
21 96.00 5.00 0.00 1.00 48.00 3.00 Acceptable
24 0.15 1.00 0.15 2.00 0.15 1.50 Not Acceptable
Notes 1. Totals represent total recovery for the lab summed over all chemicals and all lots.
2. Ranks are based on total recovery - Rank 1 is given to the lowest total recovery, etc.
3. “Not Acceptable” indicates that total recovery is a statistical outlier.
29

-------
SUMMARY RESULTS
METALS ANALYSIS
Table 4.25
SAMPLE A
LOll LOT2 MEAN
Lab No. Total Rank Total Rank Total Rank a = 0.05
1 24143.55 5.00 22643.27 4.00 23392.41 4.50 Acceptable
2 29946.54 7.00 29191.73 5.00 29569.14 6.00 Acceptable
4 23380.59 4.00 21899.73 3.00 22640.16 3.50 Acceptable
5 20874.81 2.00 20540.91 1.00 20707.86 1.50 Acceptable
7 21567.46 3.00 20792.69 2.00 21180.07 2.50 Acceptable
8 37458.75 9.00 - - 37458.75 9.00 Not Acceptable
9
10 - - -
12
14
16 - -
17 - - - - - -
I 8 . 4706.57 1.00 - - 4706.57 1.00 Not Acceptable
20 24283.10 6.00 - 24283.10 6.00 Acceptable
21
24 18733.10 1.00 - - 18733.10 1.00 Acceptable
Notes : 1. Totals represent total recovery for the lab summed over all chemicals and all lots.
2. Ranks are based on total recovery - Rank 1 is given to the lowest total recovery, etc.
3. “Not Acceptable” indicates that total recovery is a statistical outlier.
30

-------
SUMMARY RESULTS
METALS ANALYSIS
Table 4.26
SAMPLE B
LOT1 LOT2 MEAN
Lab No. Total Rank Total Rank Total Rank a = 0.05
1 23020.36 10.00 - 23020.36 10.00 Acceptable
2 25501.81 11.00 - 25501.81 11.00 Acceptable
4 13000.52 7.00 - - 13000.52 7.00 Acceptable
5 12775.09 6.00 - 12775.09 6.00 Acceptable
7 12100.93 4.00 - 12100.93 4.00 Acceptable
8 - - - - - -
9 10570.35 3.00 447.07 2.00 5508.71 2.50 Acceptable
10 8032.94 2.00 328.36 1.00 4180.65 1.50 Acceptable
12 12524.85 5.00 867.67 4.00 6696.26 4.50 Acceptable
14 14369.60 7.00 493.95 3.00 7431.78 5.00 Acceptable
16 13109.94 6.00 3060.35 5.00 8085.14 5.50 Acceptable
17 3956.20 1.00 21178.55 6.00 12567.38 3.50 Acceptable
18 -
20 -
21
24
Notes : 1. Totals represent total recovery for the lab summed over all chemicals and all lots.
2. Ranks are based on total recovery- Rank 1 is given to the lowest total recovery, etc.
3. “Not Acceptable” indicates that total recovery is a statistical outlier.
31

-------
SUMMARY RESULTS
METALS ANALYSIS
Table 4.27
SAMPLE C
LOTI LOT2 MEAN
Lab No. Total Rank Total Rank Total Rank a = 0.05
1 - —
2
4 -
5 -
7 - - - - - -
8 29768.75 11.00 28049.75 5.00 28909.25 8.00 Acceptable
9 15447.70 7.00 - - 15447.70 7.00 Acceptable
10 17285.06 8.00 - 17385.06 8.00 Acceptable
12 13543.60 5.00 - 13543.60 5.00 Acceptable
14 9684.43 3.00 - 9684.43 3.00 Acceptable
16 9163.10 2.00 - 9163.10 2.00 Acceptable
17 18762.40 10.00 - - 18762.40 10.00 Acceptable
18. 5771.71 1.00 9058.73 1.00 7415.22 1.00 Acceptable
20 10114.00 4.00 17322.40 3.00 13718.20 3.50 Acceptable
21 17894.81 9.00 20534.51 4.00 19214.66 6.50 Acceptable
24 14988.95 6.00 13382.80 2.00 14185.88 4.00 Acceptable
Notes 1. Totals represent total recovery for the lab summed over all chemicals and all lots.
2. Ranks are based on total recovery - Rank us given to the lowest total recovery, etc.
3 “Not Acceptable” indicates that total recovery is a statistical outlier.
32

-------
Table 4.28 ANOVA Semi-Volatile Organic Compounds
SV DF SS MSE F-Ratio Prob(F>f)
Total 75 38.3x10 9 -
Mean 1 9.4x10 9
C 14 3.1x10 9 2.2x10 8 1.02 0.540
S 2 1.6x10 9 8.0x10 8 3.68 0.093*
CS 28 6.1x10 9 2.2x10 8 0.43 0.014
B/CS 28 14.2x10 9 5.1x10 8 0.08 0.0003
R/CSB 2 12.3x10 9 61.7x10 8
Definitions : C = Collaborating Labs Main Effect
S = Samples Main Effect
CS = Lab-Sample Interaction
B/CS = Batches (Lots) nested within Lab-Sample Interaction
R/BCS = Replicates nesled within B/CS nested Effect
SV = Source of Variation
DF = Degrees of Freedom
SS = Sums of Squares
MSE = Mean Squared Error
Significance : * - Indicates significant effect
** - Indicates highly significant effect
33

-------
Table 4.30 ANOVA Pesticides
SV DF SS MSE F-Ratio Prob(F>f)
Total 71 63.4x10 5
Mean 1 2.1x10 5
C 13 10.1x10 5 7.8x10 4 1.61 0.856
S 2 1.1x10 5 5.3x10 4 1.09 0.649
CS 26 12.5x10 5 4.8x10 4 0.43 0.017
B/CS 26 29.3x10 5 11.2x10 4 0.41 0.086
R/CSB 3 8.2x10 5 27.4x10 4
Definitions C = Collaborating Labs Main Effect
S = Samples Main Effect
CS = Lab-Sample Interaction
B/CS = Batches (Lots) nested within Lab-Sample Interaction
R/BCS = Replicates nested within B/CS nested Effect
SV = Source of Variation
DF = Degrees of Freedom
SS = Sums of Squares
MSE = Mean Squared Error
Significance . * - Indicates significant effect
** - Indicates highly significant effect
34

-------
Table 4.29 ANOVA Metals
SV DF SS MSE F-Ratio Prob(F>f)
Total 84 28.7x10 9
Mean 1 3.1x10 9
C 15 3.3x10 9 21.8x10 7 0.94 0.469
S 2 1.3x10 9 65.4x10 7 2.83 0.927
CS 30 6.9x10 9 23.1x10 7 0.56 0.056
B/CS 31 12.9x10 9 41.5x10 7 1.63 0.693
R/CSB 5 1 .3x1 0 25.4x1
Definitions : C = Collaborating Labs Main Effect
S = Samples Main Effect
CS = Lab-Sample Interaction
B/CS = Batches (Lots) nested within Lab-Sample Interaction
R/BCS = Replicates nested within B/CS nested Effect
SV Source of Variation
DF = Degrees of Freedom
SS = Sums of Squares
MSE = Mean Squared Error
Significance * - lnthcates significant effect
** Indicates highly significant effect
35

-------
Table 4.31 ANOVA Semi-Volatiles Familiarization
SV DF SS MSE F-Ratio Prob(F>f)
Total 38 90.2x1
Mean 1 60.6x10 5
C 16 16.2x10 5 10.1x10 4 1.23 0.656
B/C 16 13.2x10 5 8.3x10 4 25.98 0.998
R/BC 5 O.2x10 5 0.3x10 4
Definitions . C = Collaborating Labs Main Effect
S = Samples Main Effect
CS = Lab-Sample Interaction
B/CS = Batches (Lots) nested within Lab-Sample Interaction
R/BCS = Replicates nested within B/CS nested Effect
SV = Source of Variation
DF = Degrees of Freedom
SS = Sums of Squares
MSE = Mean Squared Error
Significance . * - Indicates significant effect
** - Indicates highly significant effect
36

-------
Table 4.32 ANOVA Pesticides Familiarization
SV DF SS MSE F-Ratio Prob(F>f)
Total 43 19.6x10 5
Mean 1 8.1x10 5
C 17 8O.8x10 5 47.5x10 3 2.41 0.961*
B/C 17 33.5x10 5 19.7x10 3 189.51 0.999**
R/BC 6 0.1x10 5 0.1x10 3
Definitions . C = Collaborating Labs Main Effect
S = Samples Main Effect
CS = Lab-Sample Interaction
B/CS = Batches (Lots) nested within Lab-Sample Interaction
R/BCS = Replicates nested within B/CS nested Effect
SV = Source of Variation
DF = Degrees of Freedom
SS = Sums of Squares
MSE = Mean Squared Error
Significance : * - Indicates significant effect
** Indicates highly significant effect
37

-------
Table 4.33 ANOVA Metals Familiarization
SV DF SS MSE F-Ratio Prob(F>f)
Total 63 11.8x10 7
Mean 1 9.7x10 7
C 22 12.9x10 7 58.5x10 5 1.85 0.921
B/C 22 7.0x10 7 31 .6x1 7.07 0.999**
R/BC 18 0.8x10 7 4.5x10 5
Definitions C = Collaborating Labs Main Effect
S = Samples Main Effect
CS = Lab-Sample Interaction
B/CS = Batches (Lots) nested within Lab-Sample Interaction
R/BCS = Replicates nested within B/CS nested Effect
SV = Source of Variation
DF = Degrees of Freedom
SS = Sums of Squares
MSE = Mean Squared Error
Significance . * - Indicates significant effect
** - Indicates highly significant effect
38

-------
5.0 CONCLUSIONS
The results of the collaborative study indicate that the TCLP can be applied consistently
by a diverse group of organizations. Total recoveries for both semi-volatile organic
compounds and metals were sufficiently good to indicate that variability is within
reasonable statistical limits (at a 95 percent confidence level) for these two groups of
parameters. The pesticide results were distorted by the limited amount of data and the
wide range of detection limits reported by the participating laboratories, It is felt that a
larger data set would yield comparable results to that for other organic compounds.
39

-------
6.0 REFERENCES
(1) Test Methods for Evaluating Solid Wastes: U.S. Environmental Protection
Agency. Office of Solid Waste. Washington, DC, 1982.
(2) Kempthorne, 0. The Design and Analysis of Experiments; Robert E. Krieger
Publishing Company. Huntington, NY, 1973; pp. 631.
(3) Youden, W. J.; Steiner, E.H. Statistical Manual of the Association of Official
Analytical Chemists; The Association of Official Analytical Chemists: Arlington,
VA, 1975.
40
S-CUBED, A Division of Maxwell Laboratories, Inc.

-------
COMPARISON STUDY OF PREPARATIVE AND ANALYTICAL TECHNIQUES
FOR THE DETERMINATION OF SELENIUM IN
WATER, SEDIMENT AND VEGITATION MATRICES
MILAD S. ISKANDER, NABIL L. YACOUB, ARTHUR HOLDEN, CHARLES
SMITH AND ROBERT D. STEPHENS CALIFORNIA PUBLIC HEALTH
FOUNDATION 2151 BERKELEY WAY, ROOM 520 P.O. BOX 520
BERKELEY, CALIFORNIA
ABSTRACT
In an effort to develop precise and accurate methodologies
for the determination of Se, a potentially toxic and dif-
ficult to determine element, in various sample matrices,
different approaches of sample preparation and analysis were
evaluated. Modification of methods was applied when needed
and modified methodologies were assessed.
Nitric acid digestion, with and without H 2 0 2 and HC1, was
considered. Selenium (total) in various digestates was
determined by Inductively-Coupled Plasma-Atomic Emission
Spectroscopy (ICP-AES), Heated Graphite Atomizer-Atomic
Absorption (HGA-AA), and Hydride Generation-Atomic Absorp-
tion (HG-AA) Spectrometry.
The data generated by this work showed that there is a
reasonable agreement among the Se results obtained by ICP-
AES and HGA-AA in the sediment and vegetation samples. The
HGA-AA technique, however, is preferred for the determina-
tion of low levels of Se in water (sub ppm) due to its
higher sensitivity relative to that of ICP-AES. Several
problems were encountered with the hydride generation
technique (HG-AA) and its use required difficult manipula-
tions and extreme care. Generally, Se recoveries of 77% to
108% were attained at levels of 2 ppm and 30 ppm total Se.
This paper describes the various methods of samples prepara-
tion considered, and the operating parameters of the
different instrumental techniques used. A comparative dis-
cussion of these methods and techniques is also reported
along with recommendations pertinent to the use of each.

-------
Table of Contents
I. Abstract
II. Introduction
III. Methods and Materials
a. Sample Preparation
b. Digestion Techniques
c. Instrumental Analysis (ICP-AES, HGA-AA and HG-AA)
d. Operating Parameters
e. Matrix Effect and Modification
IV. Comparative Results
a. Treatments
b. Instruments
V. Discussion and Recommendations
VI. References

-------
SECOND SESSION
Wednesday, July 16, 1986
8:00 a.m. - 12:00 p.m.
Chairperson:
Ronald Mitchum
Director
Quality Assurance Division
Environmental Monitoring and
Support Laboratory, USEPA,
Las Vegas, NV 89114

-------
DEVELOPMENT AND VAlIDATION OF RCRA METHOD
8280 FOR DIOXINS AND FURANS
STEPHEN BILr ETS, U.S. ENVIRONMENTAL MONITORING SYSTEMS
LABORATORY, U.S. ENVIRONMENTAL PROTECTION AGENCY, LAS VEGAS,
NEVADA; AND JOHN M. BALLARD, TIMOTHY L. VONNAHME, NATHAN J.
NUNN AND DAVID R. YOUNCMAN, LOCKHEED ENGINEERING AND
MANAGEMENT SERVICES COMPANY, LAS VEGAS, NEVADA
ABSTRACT
RCRA Method 8280 for the analysis of chlorinated dibenzo—2—
dioxins and dibenzofurans, as published in the Federal
Register in April 1983, revealed the need for several
modifications to allow for the determination of the target
analytes in complex matrices, such as industrial sludge and
still—bottom samples. Details of these modifications and of
the subsequent applicaton of the revised method to a limited
number of samples which were analyzed in the course of a
single laboratory evaluation will be reported.
Further evaluation of RCRA Method 8280 for the analysis of
polychiorinated dibenzo—p—dioxins and dibenzofurans has been
performed. The Method has been modified to provide for the
quantitation of total tetra—through octa—chiorinated doixins
and dibenzofurans and has been applied to sample matrices
derived from industrial polychlorophenal sources as well as
to fly—ash, still—bottom, and Missouri soil samples. As an
additional test of Method performance, an inter—laboratory
validation study was conducted in two parts. A two—part
study was used becasue the Method had been extensively
revised since its publication in the Federal Register, and
it was felt that participating laboratories would be
unfamiliar with some of the proposed procedures. The first
phase was intended to allow the participants to acquire
familiarization with the Method by analyzing relatively
simple matrices for a few specified analytes which had been
spiked into the samples. The second phase required the
total quantitation of tetra—through octa—CDD’s and CDF t s in
complex samples containing the arialytes at both low (ppt)
and extremely high (ppm) levels; no spiking was used for
these samples. A method detection limit study using all
available 13 C 12 — labeled PCDD and PCDF isomers spiked into
seven different sample matrices was also performed and the
results indicated both matrix and homolog specific
differences.
The revised Method 8280 has undergone a period of continual
development, new documentation which will be reported
includes Method performance data on complex samples from
polychlorophenol use processes, results from an inter—
laboratory study of the revised method, and method detection
limits of selected PCDD’s and PCDF’s in a variety of
environmental and hazardous waste matrices.

-------
INTRODUCT ION
On a molecular basis, 2,3,7,8—tetrachlorodibenzo—p dioxin
(2,3,7,8—TCDD) is one of the most poisonous ynthetic
chemicals known. The compound has been shown in animals to
possess teratogenic, embryotoxic, carcinogenic, and co—
carcinogenic properties in addition to acute toxicity.
Because of its chemical stability, lipophilic character, and
extreme toxicity, it presents potentially severe health
hazards to the human population. Although 2,3,7,8—TCDD is
the most toxic of the 75 chlorinated dibenzo—p dioxins
(PCDD’s), many of the others (as well as the 135 clTlorinated
dibenzofurans (PCDF’s) which have similar genesis,
structures, and properties) are known to possess relatively
high toxicity to humans and animals. For this reason, the
entire spectrum of PCDD’s and PCDF’s is of environmental
concern.
Based on this information, it was concluded that samples
containing tetra—, penta—, hexa—, hepta—, and octa—CDD’s and
—CDF’s are likely to exhibit increased toxicity. A method
to analyze hazardous wastes for the relevant PCDD’s and
PCDF’s was included in the Resource Conservation and
Recovery Act (RCRA) requirements for hazardous waste
monitoring. A single—laboratory evaluation of RCRA Method
8280 for the analysis of PCDD’s and PCDF’s in hazardous
waste has been the subject of a previous report prepared for
the Office of Solid Waste. That report presented results
obtained with sample matrices including pottery clay, a
Missouri soil, a fly—ash, a still—bottom from a
chiorophenol—based herbicide production process and an
industrial process sludge. Major revisions to the Method
which was first published in 1983 were necessary to accom-
modate the analysis of complex samples such as sludge and
still—bottom.
The revised Method 8280 has undergone a period of continual
development, and this report presents results obtained
during the further evolution of the Method. The Method has
now been modified to enable the quantitation of total tetra—
through octa—chiorinated dioxins and dibenzofurans and has
been applied to six different sample matrices derived from
industrial polychiorophenol sources arid also to fly—ash,
still—bottom, and Missouri soil samples. An interlaboratory
validation of the Method was conducted in two phases: Phase
I required the analysis of spiked and unspiked clay and
sludge samples for certain specified PCDD/PCDF analytes, and
Phase II required the analysis of soil, sludge, fly—ash, and
still—bottom samples for total tetra— through octa—chiori—
nated dioxins and dibenzofurans. Method detection limits of
13 C 12 —labeled polychiorinated dioxins and dibenzofuraris in
seven matrices have also been determined.

-------
RCRA Method 8280 for the analysis of chlorinated dibenzo—R—
dioxins and chlorinated dibenzofurans, as published in the
Federal Register in April 1983, revealed the need for
several modifications to allow for the determination of the
target analytes in complex hazardous waste matrices, such as
industrial sludge and stilibottom samples. Subsequently,
the Method has been further refined in several important
ways as needed for the characterization and assessment
aspects of RCRA. A summary of these changes is as follows:
In order to improve the accuracy of quantitation of the
hepta— and octa—CDD’s and —CDF’s, a second internal standard
(l C12—OCDD) is added together with 13 C 12 —2,3,7,8—TCDD prior
to sample workup. Some of the ions specified in the
multiple ion detection (MID) descriptors have been changed
so as to increase sensitivity by monitoring the most intense
ion in the isotopic cluster. To ensure that coeluting
polychiorinated diphenyl ethers (PCDE’s) are not con-
tributing to the signal response clue to PCDF’s, the
molecular ion of the appropriate PCDE was included in each
MID descriptor. In addition, the criteria for the positive
identification of PCDD and PCDF isomers were made more
explicit. Instrument tune criteria employing PCDD standard
reference materials were substituted for those based on the
use of decafluorotripheriylphosphine (DFTPP). The section on
the calculation of concentrations of the target analytes was
expanded to include a procedure for measuring unknown PCDD
and PCDF isomers.
This report presents data on the performance of the Method
as it was applied to the analysis of a variety of wastes
derived from the use of polychiorophenols in the wood—
preserving industry. As an additional test of Method
performance, an interlaboratory validation study was con-
ducted. This study was divided into two phases because the
Method had been extensively revised since its first publica-
tion in the Federal Register, and it was felt that partici-
pating laboratories would be unfamiliar with some of the
proposed procedures. The first phase was intended to allow
the participants to acquire familiarization with the Method
by analyzing relatively simple matrices for a few specified
analytes which had been spiked into the samples. The second
phase required the total quantitation of tetra—through octa—
CDD’s and -CDF’s in complex samples containing the analytes
at both low and extremely high concentration levels; no
spiking was used for these samples. A method detection
limit study using all available 1 3 Cl2_labeled PCDD and PCDF
isomers spiked into seven different sample matrices was also
performed and will be reported.
Data obtained from Phase I of the interlaboratory study
indicate that the Method is biased high and that the bias
appears to decrease as the concentrations of the analytes
increase. Data from the method detection limit (MDL) study
can also be used as an indicator of intralaboratory

-------
precision. For seven replicate determinations of a TCDF and
a PeCDD in fly-ash with each at a measured concentration of
2.6 times their final calculated MDL’S, the relative
standard deviations (RSD’s) were 12.3 percent and 12.2
percent, respectively. Similar determinations for a PeCDF
and a TCDD which were measured at a level 6.0 and 4.4 times
their MDL’S gave RSDtS of 5.2 percent and 7.2 percent,
respectively.
Encouraging results were obtained from Phase I of the inter—
laboratory study in which specific analytes spiked into clay
and sludge samples were quantitated. The good overall re-
covery (greater than 50 percent) of the internal standard
and the small differences between the spiked concentrations
and the mean measured values both indicate that the Method
can provide acceptable data in a multi—laboratory evalua-
tion. Phase II of the interlaboratory study which required
the quantitation of total tetra— through octa—CDD’s and —
CDF’s in 10 aliquots of 4 sample types also provided
generally satisfactory results. The internal standards
( 13 C 1 2—2,3,7,8—TCDD and - 3 Cl2—OCDD) were recovered in
overall acceptable yields ranging from 51 to 82 percent.
However, quantitation of the analytes was less precise than
in Phase I. Two major, probable reasons for this are as
follows:
o the complex samples themselves, some of which con-
tained endogenous amounts of the target analytes
at low and at extremely high levels. This
required a large dilution effect which minimized
the value of the internal standard, and
o the analysis required the identification, co-
nfirmation, and guantitation of unknown peaks for
each congener without an authentic reference
material which could be used to confirm the
identification.
Statistical analysis of the Phase II data revealed
that:
o the recovery of the - 3 C 12 —2,3,7,8—TCDD internal
standard was a function of sample type whereas
that of the ‘ 3 C 12 —OCDD internal standard was not;
o the laboratories were equivalent in accuracy for
all analytes except OCDD; and
o the laboratories were equivalent in precision for
31 of the 40 possible matrix/analyte combinations.
As a result of the experience gained during the single—and
multi—laboratory testing of the Method with a variety of

-------
environmental samples, several modifications and areas of
further study are recommended as follows:
o The Method should allow for the use of disposable,
open carbon columns as an option to the HPLC
carbon column cleanup. This would allow for an
increase in the rate of sample throughput and
would also reduce solvent consumption.
o Gas chromatography (GC) conditions should be
modified to improve the resolution between the
internal standard ( 1 - 3 C 12 —2,3,7,8—TCDD) and the
recovery standard ( 13 C 12 —l,2,3,4—TCDD). If this
cannot be readily achieved, then use of an alter-
native recovery standard should be considered.
o The elution windows (defined by first and last
eluting isomers) of the tetra— through octa—CDD
and —CDF congeners should be established for the
GC conditions used in the Method.
o Method 8280 should be written to require as many
GC/MS analyses as necessary by using the appro-
priate MID descriptors whenever an elution overlap
is noted in a sample. The descriptors should in-
clude at least one ion for each overlapping
homologue.
o Kovats Indices should be determined for available
PCDD’s and PCDF’s. This would aid laboratories in
the identification of isomers not known or
available and would be useful in a CC screening
program.
o The need to monitor for polychiorinated diphenyl
ethers (PCDE’s) in the final sample extract should
be investigated.
o A source of a well—defined CC performance standard
should be identified. Column performance guide-
lines should be established for a variety of
columns.
These changes have been incorporated into the final version
of the Method.
In order to assess method performance, 10 waste samples
derived from the industrial use of pentachiorophenol (PCP)
were provided to the EPA. These samples together represent
4 different matrix types, viz , sludge, fuel oil, alcohol
fuel oil, and soil. The range of matrix types encompassed
is expected to be representative of those to be analyzed
under RCRA regulations and is expected to provide varying
degrees of sample complexity.

-------
Each sample was analyzed in duplicate for the quantitation
of total tetra— through octa—CDD’s and —CDF’s. Two criteria
were applied to confirm that peaks in the extracted ion cur-
rent profiles (EICP’s) of the quantitation ions were due to
the targeted analytes and were not due to either interfer-
ences or spurious noise signals. These were the following:
o single—to—noise ratio greater than 3 to 1
o the presence of the confirmation ion such that the
relative intensity of the quantitation ion and the
confirmation ion was within the limits specified
in the Method.
Signal responses that did not meet these criteria are
reported as “ND” (not detected). Quantitation was usually
performed against 13 C 12 —2,3,7,8—TCDD as the internal
standard, and values were corrected for the recovery usina
this compound as an isotopic diluent. However, due to the
extremely high levels of hexa—, hepta—, or octa-CDD’s/CDF’s
present in some samples, these analytes were quantitateci
against 13 C 12 —l,2,3,4—TCDD or - 3 C 12 —OCDD added to the
extracts after dilution. It is a disadvantage of the
quantitation method that multiple GC/MS analyses are there-
fore required for samples containing both low and high
analyte concentration level. To spike the sample with the
appropriate L 3 Cl2 standard at the levels found would have
caused an unnecessarily large expense. Whenever possible,
sample re—run requirements were imposed to lessen the need
for quantitation by anything other than the isotopic
diluent. Several characteristics are evident in the data
presented in Table 1. First is the total absence of detect-
able levels of TCDD in all of the 10 samples and the occur-
rence of PeCDD in only 3 samples, second is the very high
levels of hepta— and octa—CDD present in the PCP process
samples.
For purposes of this study, the method detection limit (MDL)
is defined as the minimum concentration of a substance that
can be identified, measured, and reported with 99 percent
confidence that the analyte concentration is greater than
zero and is determined from analysis of a given matrix con-
taining the target analyte. The data for this study was
obtained using all available ‘ 3 C 12 —labeled analytes spiked
into seven different sample types at a concentration of
twice the estimated MDL of each analyte. This experimental
design was used in order to obtain MDL values in each matrix
without spiking with unlabeled PCDD’s and PCDF’s and without
changing the integrity of the sample. In order to establish
an appropriate spiking level, the MDL was estimated as that
concentration at which the response of the appropriate
quantitation ion gave a signal/noise ratio of 3 to ].
Statistical considerations required that a minimum of seven
replicates of each sample type should be processed through

-------
TABLE 1. ANALYSISa OF PCP PROCESS SAMPLES USING METHOD 8280
= ====—====——=——=== =——===——========= ——==================——=========== ==============
Fuel Alcohol
Sludge oil Sludge Sludge Fuel oil fuel oil Sludge Soil Soil Soil
PCDDI B-6d B-7b B-8b B-12h A-2g A-3g A-4g A-5g A-6.lg A—6.2g
PCUF (ppb) (ppb) (ppb) (ppb) (ppb) (ppb) (ppb) (ppb) (ppb) (ppb)
TCDD NDb ND ND ND ND ND ND ND ND ND
PeCDD ND ND ND ND ND ND ND ND 27 ND
HxCDD 2150 2186 ND ND 2079 762 726 283 730 396
HpCDD 51520C 67176C 2166C 978C 38195C 17956C 59600C 12945C 24700C 12300C
OCDD 72300C 154000C 2670C 2550C 59100C 24500C 106000C 16500C 26300C 15000C
TCDF ND ND ND ND ND ND ND ND ND ND
PeCDF ND 154 ND ND 246 ND ND ND 61 ND
HxCDF 68 2933 ND ND 2852 76 1568 65 252 56
HpCE)F 343 1342 ND ND 1913 1118 1948 533 1695 434
OCDF 4100C 75QQC ND 76 447 741 3200C 900C 3 080 C 1690C
13
C 12 _ 66.8 69.0 64.3 67.8 69.2 60.0 62.9 77.0 75.4 74.8
2,3,7,8—
TCDD per-
cent recovery
8 Mean of duplicates, concentrations shown are for the total of all isomers within a given
homologous series.
bND is below the detection limit for the sample matrix. Detection limits are estimated as 5 ppb
for the tetra- through hexa-isomers, and 10 ppb for the hepta— and octa—isomers.
CDue to the extremely high levels of HpCDD, OCDD, and OCDF detected in the GC/MS analysis, the
extracts were diluted after normal quantitation of the tetra-, çnta-, hexa-CDD/CDF and
hepta-CDF. HpCDD, OCDD, and OCDF were en quantitated versus 1 C 12 —1,2,3,4-TCDD added after
dilution; the values are corrected for C 12 —2,3,7,8—TCDD recovery.

-------
the entire analytical method. Two initial replicates were
tested to verify the reasonableness of the MDL , estimate for
each sample type. When a reasonable spiking level had been
achieved, five more determinations were made at the same
spike concentration. The standard deviation (S) of the mean
concentration determined for each analyte was then cal-
culated from the seven replicte measurements for each of the
seven matrix types. The MDL was then calculated from the
equation:
MDL = t(n_l, l— = 0.99) X (S)
where t(n_l, 1—ac = 0.99) is the Student’s t value appro-
priate for a 99 percent confidence level and a standard
deviation estimate with n—i degrees of freedom.
Therefore,
MDL = 3.143(S).
The concentration (Cs) of each analyte in each sample was
determined with reference to the 13 C 12 —l,2,3,4—TCDD internal
standard, which was spiked after sample workup to give a
final concentration of 40 pg /eL.
The eight 13 Cl2—labeled PCDD’s and PCDF’s used in this study
and their MDL’S in the seven sample matrices are listed in
Table 2. Several characteristics and trends are apparent in
the data: 13 C 12 —2,3,7,8—TCDD/TCDF usually had the lowest
MDL values for each sample type while - 3 C 12 -HpCDD/OCDD
usually had the highest as might be expected, the MDL
values for all analytes generally increased in passing from
the “clean” sample types (reagent water, fly—ash) to the
more complex, organics—containing matrices (still—bottom,
industrial sludge). The MDL for 13 C 12 —2,3,7,8—TCDD in
reagent water (0.44 ppt) determined in this study using
Method 8280 compares well with the value reported for
2,3,7,8—TCDD in reagent water (2 ppt) which was determined
using Method 613 (capillary column GC/MS with selected ion
monitoring). The MDL procedure, involving seven replicate
determinations of each of the eight analytes in each of
seven sample matrices, generated other data (percent
recovery, precision) which is of interest in assessing the
performance of Method 8280. These data are presented in
Tables 3 and 4. It can be seen that good recoveries were
obtained and that the precision at low spike levels was
acceptable.
Phase I of the interlaboratory study was intended to allow
participating laboratories an opportunity to become familiar
with the requirements of the revised Method 8280. This
phase of the study was considered necessary since extensive
revisions had been made to the original version of the
Method. These included (1) changes in the procedure for the

-------
13
TABLE 2. METHOD DETECTION LIMITS OF C12_LABELED PCDD’S AND PCDF’S IN REAGENT
WATER (PPT) AND ENVIRONMENTAL SAMPLES (PPB)
A lyte Watera Ashb Sludgec Bottomd 011 d Sawdustb
2,3,7,8-TCDD 0.44 0.17 0.07 0.82 1.81 0.75 0.13
1,2,3,78-PeCDD 1.27 0.70 0.25 1.34 2.46 2.09 0.18
1,2,3,6,7,8-HxCDD 2.21 1.25 0.55 2.30 6.21 5.02 0.36
1,2,3,4,6,7,8-HpCDD 2.77 1.87 1.41 4.65 4.59 8.14 0.51
OCDD 3.93 2.35 2.27 6.44 10.1 23.2 1.48
2,3,7,8-TCDF 0.63 0.11 0.06 0.46 0.26 0.48 0.40
1,23,78—PeCDF 1.64 0.33 0.16 0.92 1.61 0.80 0.43
1,2,3,4,78-HxCDF 2.53 0.83 0.30 2.17 2.27 2.09 2.22
asample size 1,000 mL
bsampIe size 10 g
CSample size 2 g
dsample size 1 g
Note: The final sample—extract volume was 100 L for all samples.

-------
=
13 Concentration (pg/pL)
C12-Labeled in Final Extract Mean Recovery
Analyte Spiked Measured RSD Percent Percent
2,3,7,8-TCDD 40 29.7 8.0 74.3
1,2,3,7,8-PeCDD 80 56.8 11.7 71.0
1,2,3,6,7,8—HxCDD 120 110.7 14.4 92.3
1,2,3,4,6,7,8—HpCDD 150 118.3 21.9 78.9
OCDD 250 232.8 31.7 93.1
2,3,7,8-TCDF 20 12.9 11.9 64.5
1,2,3,7,8-PeCDF 40 27.7 9.2 69.3
1,2,3,4,78-HxCDF 80 57.2 11.7 71.5
TABLE 4. PERCENT RECOVERY OF ‘ 3 C 12 —LABELED PCDD’S AND PCDF’S
FROM INDUSTRIAC SLUDGE
13 Concentration (pg/uL)
C12-Labeled in Final Extract Mean Recovery
Analyte Spiked Measured RSD Percent Percent
2,3,7,8-TCDD 40 25.3 10.2 63.3
1,237,8-PeCDD 80 56.7 16.5 70.9
1,2,3,6,7,8-HxCDD 80 90.6 15.8 113.3
1,2,3,46,7,8—HpCDD 120 128.6 16.7 107.2
OCDDa 175 16.2
1,2,3,7,8-TCDF 20 15.6 14.6 78.0
1,2,3,7,8-PeCDF 40 26.4 14.1 66.0
1,2,3,4,78-HxCDF 80 49.5 14.8 61.9
apeak shape was distorted by very high 1eve of interferent.

-------
extracton of analytes from the sample, (2) modification of
the open column alumina chromatography cleanup, (3) deletion
of the HPLC cleanup, (4) addition of a carbon column
cleanup, and (5) incorporation of internal and recovery
standards into the Method.
Five laboratories were selected by the Contract Laboratory
Program to participate in the study and were each provided
with samples, analytical standards, isotopically labeled
internal and recovery standard solutions, and the revised
Method, also provided were, detailed supplemental
instructions which included guidance on the sample size, the
volume of the final extract, typical MID descriptors,
typical MID and RIC chromatograms, and the reporting
requirements for data and deliverables. Each laboratory was
provided with the six samples which had been prepared by
personnel of the referee Laboratory who also analyzed these
samples.
The results obtained by the referee Laboratories’ personnel
who were familiar with the Method, although not obtained as
part of the blind study, were valuable for comparison and
are included in the relevant Tables.
A very simple, qualitative measure of how well the combined
extraction, chromatographic cleanup, and GM/MS analysis
prescribed in the Method deals with the various analytes may
be obtained by noting the number of target analytes for
which values were reported by each laboratory. As shown in
Table 5, at least one laboratory in addition to referee
detected all of the analytes, and two laboratories reported
12 of 13. Relatively lower reporting by Laboratories II and
IV is presumed to be due to lack of familiarity with the
extraction and cleanup procedures and may be expected to
improve with experience.
A more quantitative measure of the extraction and cleanup
efficiency is provided bZ monitoring the percent recovery of
the internal standard (1 3 C 12 _2,3,7,8_TCDD) which was added
to the sample immediately prior to extraction. Table 6
shows that three of the participating laboratories, apart
from the referee, obtained acceptable results (mean recovery
greater than 40 percent). The two other laboratories
reported mean recoveries of less than 30 percent.
A summary of results obtained from the sludge samples is
described in Table 7 and the results of the study are
summarized as follows:

-------
TABLE 5. INTERLABORATORY TEST OF METHOD 8280, PHASE I:
SUMMARY OF ANALYTES REPORTED BY PARTICIPATING LABORATORIES
Number of Analytes Reported
Number by Each Laboratory
of ________________________________________
Analytes Referee
Sample Spiked Lab I II III IV V
Clay Spike 13 13 13 8 12 11 12
No. 1
Clay Spike 13 13 13 8 12 4 12
No. 2
Sludge Spike 6 6 6 3 4 6 4
No. 1
Sludge Spike 6 6 6 3 4 6 4
No. 2
TABLE 6. INTERLABORATORY TEST OF METHOD 8280, PHASE 1: PERCENT RECOVERY OF
INTERNAL STANDARD 13 C 12 -2,3,7,8—TCDD
— Participating Laboratories
Referee RSD
Sample Lab I II 111 IV V Mean Percent
Clay Blank 66.3 34.3 9.1 54 21 67 42.0 57.7
Clay Spike
No. 1 62.8 64.3 13.3 53 26 63 47.1 46.7
Clay Spike
No. 2 78.9 132 10.2 60 14 58 58.9 76.4
Sludge Blank 88.4 84.5 28.7 57 34 35 54.6 48.6
Sludge Spike
No. 1 66.1 79.0 28.4 12 25 32 40.4 64.6
Sludge Spike
No. 2 74.0 96.4 32.9 15 36 44 49.7 60.2

-------
TABLE 7.
INTERLABORATORY TEST OF METHOD 8280, PHASE I:
QUANTITATION OF ANALYTES IN SPIKED SLUDGE SAMPLES (PPB)
Participating Laboratories
Spike Referee RSD
Analyte Level Lab I II III IV V Mean Percent
2,3,7,8— 125.0 132.1 62.2 156.4 270 150 163.2 152.0 39.4
TCDD 127.9 48.9 153.2 220 155 184.8
1,3,7,8— 125.0 147.3 50.7 204.2 ND 93 177.9 122.2 47.8
TCDD 134.5 29.1 185.1 ND 84 116.1
1,2,7,8— 125.0 132.2 85.7 181.6 50 180 ND 1.16.7 50.8
TCDD 128.5 38.3 173.5 37 160 ND
1,2,8,9— 125.0 136.4 60.9 ND 160 260 805.2a 145.8 47.3
TCDD 135.6 33.7 ND 140 180 205.6
1,2,3,4,7— 125.0 113.3 67.2 ND 87 110 ND 93.3 34.9
PeCDD 118.8 42.4 ND 68 140 ND
1,2,3,7,8— 125.0 115.1 48.2 ND ND 120 170.9 116.6 44.4
PeCDF 123.4 33.4 ND ND 150 171.7
aNot included in calculation of mean and standard deviation.
ND = Not detected.
The mean value for 114 determinations of 11 analytes spiked
into clay at the 5 ppb level was 6.02 ± 2.78 ppb.
The mean value for 16 determinations of 2 analytes spiked
into clay at the 2.5 ppb level was 3.56 ± 2.35 ppb.
The mean value for 57 determinations of 6 analytes spiked
into sludge at the 125 ppb level was 126.4 ± 57.9 pph.
The objective of Phase II of the interlaboratory study was
to test the applicability of the Method to the analysis of
samples which were much more difficult and complex than
those used for Phase I. Significant revisions to the Method
were made between completion of Phase I and initiation of
Phase II.

-------
These revisions were made to several critical areas,
including:
o New multiple ion detection descriptors
o Addition of a carbon column cleanup procedure
o Requirements for sample size and final extract
volume
o New analytes to be quantified
o DB—5 column requirements
o Deletion of DFTPP tune requirement
o Sample reanalysis requirements based on recovery
criteria for the internal standards.
These revisions were based on the comments/suggestions
provided by the participating laboratories at the conclusion
of Phase I. The same five laboratories which took part in
Phase I also collaborated in Phase II, and each was provided
with identical packages of samples, analytical standards and
documentation etc., similar to those provided for Phase I.
Each laboratory was requested to analyze a total of 10 sam-
ples of 4 different sample types. As in Phase I, duplicate
samples of each sample type were provided to check accuracy
and precision. In addition, a third different sample from
each of two sample types was also to be analyzed. The
greatest difference between the requirements for Phase I and
Phase II was the analytes which were to be guantitated.
Whereas Phase I required the determination of certain
selected CDD’s and CDF’s, the Method was tested under Phase
II for the quantitation of total tetra— through octa—CDD’s
and —CDF’s. This change as specified in the Method
presented some difficulties in view of the use of different
MID descriptors at different points along the GC elution
profile. It is appreciated that not all of the isomers
within a given series may be detected using this procedure.
For instance, at least five early—eluting PeCDD isomers will
probably overlap the late—eluting TCDD/TCDF isomers. Simi-
lar overlap may occur between the penta—, hexa—, and hepta—
congener groups. However, the chroniatographic windows pre-
scribed were developed within the limitations of the PCDD/
PCDF standards available to the referee laboratory during
the course of this study. It is expected that, as addi-
tional reference standards become available, chromatographic
conditions will be refined accordingly.
Data packages from the participating laboratories were
audited with an emphasis on Method performance. In those
instances where a large difference occurred among the values
reported for a particular sample, the raw data were examined
to try to determine if the difference was a Laboratory
problem or a Method problem. One Laboratory evidenced
analytical problems due to errors or not following instruc-
tions which affected their results for all of the samples.
Only one—fifth the required amount of internal standard was

-------
added; this resulted in extremely low internal standard
responses and biased quantitation. Because of these
problems, the data from this laboratory be excluded from the
statistical analysis as not being representative of Method
performance. Problems with individual data points were also
examined and were traced to a variety of causes.
Calculation and data transposition errors were found; other
discrepancies were traced to differences in instrument sen-
sitivity, to differences in retention—time windows scanned,
and to isomer peaks which met identification criteria in one
laboratory and not in others.
As an indication of method performance, recovery data of the
internal standards is presented in Tables 8 and 9.
In general, the Method performed well when the laboratories
followed the protocol . A visual examination of the data
showed that approximately 85 percent of the values reported
by the 5 laboratories and used in the statistical analysis
were consistent among the laboratories.

-------
TABLE 8. INTERLABORATORY TEST OF METHOD 8280, PHASE II: PERCENT RECOVERY
OF INTERNAL STANDARD 13 C 12 -2,3,7,8-TCDD
Excluding
Participating Laboratories All Results Lab II Results
Referee
Sample Laboratory I II III LV V Mean RSD Percent Mean RSD Percent
Fly_Ash* 89.8 103 98.1 59 64 101 85.1 23.5 82.1 25.2
81.1 98 102 60 56 109
Soil A 74.4 42.3 53.6 53.6 46 75 57.5 24.3 58.3 26.7
Soil B* 67.2 54.3 50.0 51 42 90 58.8 25.2 59.4 27.1
62.2 54.3 62 45 46 82
Sludge A 64.9 77.7 40.6 18 33 74 51.4 47.3 53.5 49.6
Sludge B* 53.8 78.7 40.9 69 72 72 63.4 26.1 67.1 23.1
41.9 84.1 48.9 51 58 90
Still- 72.6 118 61.9 74 46 104 72.9 35.2 73.5 38.1
Bottom* 84.8 23 78.3 69 53 90.5
*B1ind duplicates.

-------
TABLE 9. INTERLABORATORY TEST OF METHOD 8280, PHASE II: PERCENT RECOVERY
OF INTERNAL STANDARD ‘ 3 C 12 -OCDD
Excluding
Participating Laboratories All Results Lab II Results
Referee
Sample Laboratory I II III IV V Mean RSD Percent Mean RSD Percent
Fly_Ash* 75.5 98 90 34 64 102 74.0 33.5 74.8 36.7
78.1 119 50 38 56 83
Soil A 46.2 53 65 29 43 112 58.0 49.9 56.6 56.8
Soil B* 68.1 87 3.2 14 95 150 63.6 71.6 75.4 53.1
67.3 67 5.8 29 58 119
Sludge A 55.5 75 O 37 40 46 50.7 30.2 50.7 30.2
Sludge B* 47.3 145 23 50 75 74 61.0 65.4 68.6 57.8
39.2 132 24 28 48 47
Still- 96.4 180 32 58 55 69.5 71.3 69.5 71.3
Bottom* 125 30.8 oa 33 46 39
=== ===== == =========—============——============
dNot included in calculation of mean and standard deviation.
*Bllnd duplicates.

-------
Sample
(1) Add j ternal Standards: 13 C 12 —2,3,7,8-TCDO
and C 12 —OCDD ( 13 C 12 —2,37,8—TCDF)
(2) Perform matrix—specific extraction.
________ V ________
Sample
Extract
(1) Wash with 20% KOH
(2) Wash with 5% NaC1
(3) Wash with conc. H 2 S0 4
(4) Wash with 5% NaC1
(5) Dry extract
(6) Evaporate to near dryness and
redissolve in hexane
(7) Alumina column
60% CH 2 C1 2 /hexane
Fraction
(1) Concentrate to 400 L
(2) Carbon column cleanup
(3) Add recovery standard 13 C 12 -1,2,3,4-TCDD
‘V
Analyze by GC/MS
Method 8280 flow chart for the analysis of PCDD’s and PCDF’s.

-------
SINGLE-LABORATORY EVALUATION OF METHOD 8080 FOR
ORGANOCHLORINE PESTICIDES AND PCB’S
WERNER F. BECKERT, ENVIRONMENTAL MONITORING SYSTEMS
LABORATORY, U.S. ENVIRONMENTAL PROTECTION AGENCY, LAS VEGAS,
NEVADA; VIORICA LOPEZ-AVILA, ACUREX CORPORATION, MOUNTAIN
VIEW, CALIFORNIA
ABSTRACT
Method 8080 was developed for the determination of certain
organochiorine pesticides (OCPs) and polychiorinated
biphenyls (PCBs) in liquids and solids. Liquid samples are
extracted according to Method 3510 (separatory funnel) or
Method 3520 (continuous liquid-liquid extractor) and solid
samples according to Method 3540 (Soxhiet extraction) or
Method 3550 (sonication), the extracts are concentrated,
fractionated on Florisil and the fractions analyzed by gas
chromatography (GC) on packed columns.
The Method 8080 protocol is being evaluated in a single
laboratory on actual and simulated wastes. It was found
that the Florisil cleanup method is problematic when both
OCPs and PCBs are present; a cleanup and fractionation on
deactivated silica gel is more advantageous. Sulfur in
extracts can be removed with tetrabutylammonium sulfite.
Toxaphene and chlordane pose special problems because of
their multiple-peak responses.
The use of capillary columns instead of packed columns in
the CC analysis is advantageous because better separations
are obtained for complex samples containing combinations of
OCP5, PCBs and other organics.
The final evaluation is in progress on liquid and solid
waste samples containing OCPs and Aroclor 1016 and 1260 at
three levels.

-------
INNOVATIVE TECHNOLOGIES FOR HAZARDOUS WASTE ANALYSIS
R. K. MITCHUM, D. F. GURKA, AND L. D. BETOWSKI,
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY, U.S.
ENVIRONMENTAL PROTECTION AGENCY, LAS VEGAS, NEVADA
ABSTRACT
The analysis of complex waste matrices represents a
challenge for the modern analytical chemist. The need for
innovative technologies to address the growing needs of the
environmental community, to provide greater sensitivity,
specificity and to accomplish the desired goal of cost-
effective analysis creates a dilemma. Many analytical
chemical goals are mutually exclusive, sensitivity enhance-
ment results in a loss of specificity, cost effectiveness
will usually result in the loss of an important analytical
parameter such as sensitivity and or selectivity. The use
of combined (hyphenated) techniques provides additional
analytical vectors which increase the combined techniques
power. There are several hyphenated techniques which have
been successfully applied to complex wastes analysis. The
use of gas chromatography/infrared spectrometry/mass
spectrometry (GC/IR/MS) represents one such technique. The
ability to use the technique for the non-target analysis of
complex wastes has been demonstrated. Data base retrieval
of infrared spectra provides functional group information by
which a mass spectral data base can be correlated. The use
of another innovative technique, liquid chromatography/mass
spectrometry/mass spectrometry (LC/MS/MS) also provides a
robust organic analytical method. The ability to address
the non-volatile, non-chromatographable fractoin of complex
wastes represents an emerging Agency concern for which
conventional methodologies can not be used to provide regu-
latory data or to write regulatory criteria. LC/MS/MS
represents an on-the-fly technique capable of the ionization
of non-volatile residues followed by the mass analysis of
the resulting complex parent ions by collisional activation
dissociation. The technique has been applied successfullly
to dyes a d dye effluents from industrial wastewater treat-
ment. The use of reduced cleanup prior to analysis and on-
the-fly mixture analysis provides a powerful analytical
tool. Although not a hyphenated technique in the
conventional sense, the introduction of imniunoassay techni-
ques, utilizing both polyclonal and monoclonal assay systems
for rapid cost effective target compound analysis for
organic compounds, provide a new tool for the chemist. The
technique has been demonstrated and found to be equivalent
to the more traditional analytical chemistries. These
methods will offer rapid cost effective procedures for both
screening and quantitation of many RCRA compounds which are
highly polar and or non-volatile.

-------
APPENDIX VIII ANALYSIS IN GROUNDWATER
BOB APRIL, OFFICE OF SOLID WASTE, U.S. ENVIRONMENTAL
PROTECTION AGENCY, WASHINGTON, D.C.
ABSTRACT
This presentation will focus on developments that have taken
place concerning the Appendix VIII monitoring requirement
since July 1985.
Appendix VIII to Part 261 of the RCRA regulations is a broad
list of chemicals of concern under RCRA. In 1982, EPA
promulgated ground-water monitoring regulations that
required monitoring of all Appendix VIII constituents. This
requirement is impossible to fulfill for reasons which I
assume the audience is familiar with.
By the fall of 1985, controversy over this requirement had
clearly become a major factor delaying the issurance of RCRA
permits. Tn December 1985, a meeting of EPA and State
experts was convened to address the problem. They prepared
a list of specific chemicals, derived from Appendix VIII and
the Superfund hazardous substances list, to generally
replace Appendix VIII for ground-water monitoring.
That list was issued by EPA as guidance in February 1985,
and as a regulatory proposal in July 1986. Future work will
include the refinement of the present ground-water
monitoring list, based on comments on the guidance and the
proposal, and on ongoing work at EPA. A second phase effort
will identify those chemicals that we really want to
monitor, as opposed to those that we are able to monitor.
It will produce a list of priority chemicals for a revised
ground-water monitoring list, similar to the Clean Water Act
priority pollutant list.

-------
COMPARISON OF THE TOX (EPA) AND AOX (DIN) METHODS FOR THE
DETERMINATION OF ORGANIC HALOGEN COMPOUNDS IN
WATER AND SOLID WASTE
RAIMUND ROEHL, THOMAS S. FISHER, CALIFORNIA PUBLIC HEALTH
FOUNDATION P.O. BOX 520 2151 BERKELEY WAY, ROOM 702
BERKELEY, CALIFORNIA
ABSTRACT
EPA method 9020 and DIN method 38409-H14 are group parameter
methods designed to quantify organic halogen compounds (OHC)
in aqueous samples. The two methods, which were developed
in the U.S. and in West Germany respectively, are based on
the same analytical principles and involve the following
major steps: adsorption of OHC on granular activated carbon
(GAC), displacement of chloride ions on the GAC by nitrate
ions, pyrolysis of the carbon and the OHC, and determination
of the halide ions formed during combusion. Despite this
general similarity, the two methods differ in several
instrumental and procedural details, particularly with
respect to the determination of purgeable OHC (POX), the
handling of samples containing particulate matter, pre-
analysis tests, use of a microcolumn or batch technique
during the adsorption step, pyrolysis conditions, as well as
calibration and quality control procedures. Differences
between the methods are described and their implications on
analyses of water, waste water and solid waste leachates are
discussed.

-------
EVALUATION OF GAS CHROMATOGRAPH/MASS SPECTROMETER
(GC/MS) METHOD 8240 AND 8270 FOR APPLICATION TO
APPENDIX VIII COMPOUNDS
J.E. LJONGBOTTOM, ENVIRONMENTAL MONITORING AND SUPPORT
LABORATORY, U.S. ENVIRONMENT PROTECTION AGENCY, CINCINNATI,
OHIO
ABSTRACT
The Resource Conservation and Recovery Act (RCRA) Appendix
VIII list of organic compounds represents a formidable
challenge to method standardization. Since the last malor
Agency multi-analyte method development effort, priority
pollutants, increased GC/MS availability and advances in
capillary column technology have influenced the approach to
method standardization. Notwithstanding the problems of
dealing with the matrices involved with the RCRA
regulations, the general approach to Appendix VIII method
standardization can be contrasted with priority pollutants
in the following manner:
(1) With the exception of highly toxic materials such as
pesticides, mass spectrometry has clearly become the
method of choice.
(2) Fused—silica capillary columns have led to standardized
chromatography of semivolatiles, and wide—bore surface
coated columns appear to be the solution for purge—and—
trap.
(3) The Appendix VIII analyte list drives the development
of HPLC/MS, heated purge/trap, and broad spectrum
pesticide analyses.
The approach to Appendix VIII is contrasted with that for
priorty pollutants. The results to date of efforts at the
Enviromental Monitroing and Suport Laboratory — Cincinnati
(EMSL—Cincinnati) to standardize generic methods for these
analytes and those on the Michigan petition to amend
Appendix v i ii are presented.

-------
THIRD SESSION
Wednesday, July 16, 1986
1:30 p.m. - 5:00 p.m.
Chairperson:
Gail A. Hansen
Chemist
Office of Solid Wastes
USE PA
401 “M” Street, S.W.
Washington, D.C. 20460

-------
HYDRIDE GENERATION METHODS FOR DETERMINATION
OF ARSENIC AND SELENIUM
STEVE CALLIO, REGION VIII, U.S. ENVIRONMENTAL PROTECTION
AGENCY, DENVER, COLORADO 80202-2413
ABSTRACT
Although hydride generation methods have long been used for
the analysis of As and Se by AA, very little work has been
done on simultaneous hydride generation for the ICP environ-
ment. This paper will present an outline of the sample prep
and instrumental parameters for the determination of As, Sb
and Se by ICP/Hydride. This will be followed by a brief
discusssion of some quality assurance measurements which
demonstrate the accuracy and precision of the method. It
will be shown that the method can be used with CLP type
digestates and has great potential for producing reliable
data for As, Sb, and Se in waste samples.
INTRODUCTION
Hydride methods have been used with AA for many years for
the analysis of As and Se, however, very little work has
been done to date using the hydride approach with ICP
Emission Spectroscopy. The work in this presentation has
been generated using the method outlined in the April, 1982
Analytical Chemistry article by Nygaard and Lowry of the
NEIC. This work was done using CLP type extracts to demon-
strate the utility of the Hydride/ICP approach for EPA
programs.
Sample prep is quite elementary. 5 ml of previously
prepared digested sample extract is placed in a Teflon cen-
trifuge tube. 5 ml of concentrated HCI is then added to the
sample, the tube is capped and placed in a 95 degree hot
water bath for 1/2 hour. This is to reduce Se to the +4
oxidation state. Appropriate standards, blanks, reference
standards and QA samples (spikes, dups etc.) are also pre-
pared in an identical fashion.
When cool, the samples are analyzed using the instrumenta-
tion described in the article by Nygaard and Lowry. The
work in this paper was performed on 14 different days over a
six month interval. This was done to demonstrate the day—
to—day reliability of the method.
Analyses of EPA reference samples show in general about 5%
RSD at the l00 g/l level for all three parameters. In all
cases shown the true value lies within 1SD of the returned
average value. Lab—prepared, digested standards show
similar behavior. Analysis of low—level standards indicate
estimated detection limits of 4—5 ugh for As and Se and
about 7 ,eLg/l for Sb. Estimation of detection limits from

-------
the analysis of blank samples show slightly lower values.
To demonstrate accuracy, analyses of spiked aqueous samples
were also performed. For samples spiked at 100 g/l the
following recoveries have been determined. For As, a mean
recovery of 1O0.3, g/l with a standard deviation of 6.4 /hg/i
was determined on 25 individuals samples. For Se the
numbers are 100.0 recovery, an SD of 4.9,a.g/i and 20
individual data points. For Sb, the average recovery was
97.8,a.g/l, with an SD of 5.2 g/l and 19 data points.
Limited work has been done with solid samples to date. Not
enough data points exist to show accuracy with solid
samples. However, analysis of solid reference materials
such as NBS River Sediment and EPA’s Municipal Digested
Sludge indicate good agreement with bench—mark values esta-
blished by other labs.
In conclusion it is stated that the ICP/Hydride method out-
lined here shows good promise of being able to provide
accurate and precise As, Sb and Se data for use in EPA pro-
grams.
The author would like to acknowledge the assistance in
setting up this method which was provided by Ed Bour and Joe
Lowry of EPA’S NEIC facility in Denver.

-------
CARR!ER ARGON
TO PLASMA 9
2
3
4
K! FLOW
P1 IASE
SEPARATOR
116- OL/ 1 0, .4/16rb I/e
7
WASTE

-------
All samples run the same way
5 ML SAMPLE + 5 ML CONC HC1
25ml Teflon centrifuge tube
1/2 hour in 95 degree water bath
cool, then run thru apparatus as described used extracts as
prepared for solids and waters data collected on 14
different days
April - October

-------
INSTRUMENT CONDITIONS
coolant gas 18 1pm
sample gas 0.8 1pm
RF Forward 1.75 kw
RF Reflected 10 watts
Observation Height 14.5 mm above coil
20 sec exposure, no background correction
As 193.8 nm linear to 5 mg/].
Sb 206.8 nm linear to 800 ugh
Se 196.0 nm linear to 400 ugh

-------
REFERENCE VALUES FOR AS
Ref 481—2 Ref WS12—2
True Value 235 48.1
N data pts 14 11
Mean Value 240.7 50.8
Std Dev 8.4 2.1

-------
REFERENCE VALUES FOR SB
WP8 —4
True Value 83.2
N. data pts 12
Mean Value 82.0
Std Dev 4.2

-------
REFERENCE VALUES FOR SE
R481—2 wS12—2
True Value 50 76.3
N data pts 14 11
Mean Value 51.1 77.1
Std Dev 1.8 3.0

-------
DIGESTED STANDARDS FOR AS
250 100 25
Ndatapts 4 10 5
Mean Value 260.4 100.7 24.5
Std Dev 5.4 3.7 1.0

-------
DIGESTED STANDARDS FOR SB
250 lOU 25
Ndatapts 4 10 5
Mean Value 253.6 100.9 24.0
Std Dev 7.7 4.6 2.2

-------
DIGESTED STANDARDS FOR SE
250 100 25
Ndatapts 4 10 5
Mean Value 254.4 100.6 24.1
Std Dev 9.2 4.9 1.4

-------
BLANK DATA
AS SB SE
N data pts 13 12 12
Mean Value 0.3 —0.7 —0.3
Std Dev 0.9 1.8 1.0

-------
SHORT TERM PRECISION OF BLANK
Typical day ** 10 burns ** ugh
As Sb Se
Std Dev 0.6 1.5 0.9

-------
SPIKE RESULTS AQUEOUS bOugh
AS SB SE
N data pts 25 19 20
Mean Value 100.3 97.8 100.0
Std Dev 6.4 5.2 4.9
Highest Rec 107.5 107.0 107.5
Lowest Rec 75.2(90.9) 86.3 90.2
1 incidence Se Rec 43.0 undigested sample spike
w/digestion 101 Recovery

-------
DUPLICAPE RESULTS (all samples 156 ugh)
AS SB SE
N data pts 21 15 16
Average Range 0.89 1.91 0.65

-------
NBS SEDIMENT
AS SB SE
True Value* 66.1 26.6 2.5
Ndatapts 5 5 5
Mean Value 62.6 23.0 1.1
Std Dev 7.8 4.9 0.4
* NEIC value by Fusion

-------
EPA SLUDGE
AS SB SE
True Va lue* 4.2 11.2 3.5
Ndatapts 3 3 3
Mean Value 3.9 10.9 4.7
Std Dev 0.3 1.2 0.5
* NEIC value by Fusion

-------
THE RATIONALE FOR FILTRATION OF GROUND WATER SAMPLES
Presented at the National Water Well Association
Ground Water Monitoring Meeting, Columbus, Ohio
Olin C. Braids, Ph.D.
Geraghty & Miller, Inc.
Syosset, New York
The issue of filtration of ground water samples is integral
to discussions of proper protocol in ground water sampling.
There are those who believe that filtration is essential to the
preparation of a water sample in order to get a representative
and accurate analysis. There are others who are equally
convinced that filtration will deleteriously effect the water
sample and lead to difficulties with the data acquired from it.
Each side of this issue has merit because the reasons for
the ground—water analysis vary and the objectives in the
analytical program also differ. There are legitimate situations
in which a ground—water sample should not be filtered before it
is analyzed for its chemical constituents in the laboratory and
there are circumstances dictating that an accurate analysis is
only obtained when filtration is accomplished. This discussion
will address the situations in which filtration should be
included in sample preparation.
Filtration, in this context is filtration through a 0.45 um
pore—size membrane. The same principles also apply if the
filtration medium is glass fiber or paper. Frequently, practical
limitations of time and sample characteristics dictate the use of
glass fiber or paper as prefilters before the final membrane
step.
In dealing with the subject of filtration, one must address
the water—quality parameters that could be affected either by the
act of filtration or by the failure to filter. The following
characteristics of water are those that would be affected by
filtration er the lack thereof.
The partial pressure of dissolved gases in water withdrawn
from the aquifer could be affected by the process of filtration.
The relationship of oxygen, carbon dioxide, and perhaps other
trace gases influences the pH and the oxidation—reduction
potential of the water. In cases where the partial pressure of a
gas such as carbon dioxide is significantly affected by the added
input from decomposition or some other process, the chemical
characteristics of water can change when that relationship is
disrupted.

-------
The standard protocol of long standing for dissolved metals
is to perform a filtration. Any suspended matter occurring in
water is likely to have metal ions adsorbed on it. If the water
is preserved with acid prior to analysis as the standard
protocols call for, the metals are likely to be desorbed from the
solids. This would result in dissolved concentrations of metals
being higher than originally existed.
Iron is frequently found as a constituent of ground water in
concentrations which result in its precipitation when the water
is exposed to the oxygen of the atmosphere. Under these
conditions the sample of ground water should be filtered as
rapidly as possible to prevent contact with the air and to remove
any suspended material prior to the addition of acidic
preservatives. If the sample contains suspended matter and
dissolved iron, addition of the acid prior to filtration may
desorb metals from the suspended matter. If the sample is
allowed to be in contact with air for even a matter of minutes
prior to filtration, the iron may precipitate and co—precipitate
or adsorb metals that were in solution. Acidification of the
sample at this point may redissolve the iron but may also bring
into solution more of the other metals than were in solution at
the time the sample was collected.
The presence of suspended matter in water, where the water
has been in contact with or is contaminated with very slightly
soluble organic compounds poses a problem similar to that of the
metals. Slightly soluble compounds such as PCBs, polynuclear
aromatic hydrocarbons, phthalate, esters, and many pesticides are
in this class. If an unfiltered sample is extracted with organic
solvent and analyzed, the compounds will desorb and appear as if
they were in solution.
Radioactive gases such as radon could be affected by
filtration because of the pressure change across the filtration
medium. There are methods as will be discussed later that can
eliminate or minimize losses of gases or volatile compounds in
water during the filtration procedure.
Many radioisotopes that may be included in ground—water
analyses are isotopes of metals that would be associated with the
suspended solids in a water sample. The fact that these elements
are radioactive does not influence their chemical behavior. Thus
acidic preservation of the water prior to filtration would result
in their desorption from suspended solids.

-------
Volatile organic compounds may be lost in the process of
filtration if the water is exposed to the atmosphere or if the
filtration occurs with a pressure change across the membrane
caused by a vacuum. Most volatile organic compounds listed in
the volatile category of the priority pollutants have low to
moderate affinity for the solid substrate. Thus, water samples
for the volatile analysis are frequently not filtered because the
recovery by purging in the presence of suspended matter can be
shown to be quantitative. Filtration of water in this
context requires a filter placed in the water discharge line.
Less desirably, water may be filtered as soon as possible after
collection by another means of filtration. As noted, delaying
filtration may complicate the acquisition of reliable data if the
water has an appreciable iron concentration.
The issue of filtration of ground water is raised beôause
many times water collected from monitoring wells carries
suspended matter as a result of the nature of the sediments or
construction of the well. Production wells used for drinking
purposes or for other high—volume uses are usually constructed to
tap a reasonably prolific aquifer and to produce water with good
clarity. In contrast, monitoring wells are sometimes screened in
silty or clayey zones and samples may have substantial amounts of
fine sediment. The amount of suspended matter is an artifact of
the method of water collection and well construction and is not
reproducible through time. Any influence the sediment may have
on the results of the chemical analysis must be looked on as
biasing the sample.
This discussion is based on the premise that the ground
water in question is produced from an unconsolidated aquifer or a
crystalline rock aquifer. In some locations where solution
cavitied aquifers are monitored, the aquifer water may be
carrying sediment. In the former cases, the assumption is that
sediment is not being carried in suspension.

-------
METHODS FOR THE ANALYSIS OF
ORGANOMETALLIC COMPOUNDS IN WASTES*
G.J. OLSON, F.E. BRINCKMAN, SURFACE CHEMISTRY AND
BIOPROCESSES GROUP, NATIONAL BUREAU OF STANDARDS
GAITHERSBURG, MARYLAND
ABSTRACT
An integral feature of untapped and recycled resources
resides in the complexity of the host matrix and the huge
variety of morphological and molecular forms adopted by the
sought after components. Even more complicated are cases
involving waste materials where both host matrix and
hazardous components can occur in either original
anthropogenic forms or in quite unexpected chemically or
biologically modified forms. In any case, analysts,
modellers, and regulators are confronted with special
problems where heavy elements are involved in waste
management.
Thus, heavy elements, especially the metals, create
notorious problems in waste control, abatement, or recovery
because they can undergo significant transformations in
environmental media greatly altering their adverse impacts.
A glimmering of the global hazard of such chemical and
biotransformations was recognized two decades ago with the
methylmercury poisonings in Japan. Our recent work suggests
the broad scope of potential analogous involvement of
organometallic substances in waste sources via exocellular
solubilization and methylation by global algal metabolite,
idomethane.
Organometallic and organometalloidal entities in wastes
hence may stem from original depositions or form by
subsequent naturally—ocur ring processes. Nevertheless,
these are important classes of toxicants, because, depending
upon specific metal (bid) and bound organic moieties, their
persistence, degree of bioaccumulation, and intrinsic
toxicities are greatly magnified by heightened
lipophilicity. Fortunately, for analysts and regulators,
recent work has paved the way which takes advantage of the
unique persistent molecular “signatures” and potential for
unambiguous speciation of such active agents in wastes.
Generally, combinations of chromatographic molecular
separation coupled on—line with element—selective detectors
(ESD) afford reliable results though standard reference
materials and protocols are yet lacking.
* Preliminary manuscript, do not quote or cite

-------
Assured assay of the quantity and extant molecular forms of
heavy elements in environmental and waste media depend upon
complete recovery of analytes in forms both suited to the
analytical protocol yet fully diagnostic of the original
hazardous component in the matrix tested. Consequently,
analytical procedures yielding non—destructive in situ
characterization or precise derivatization enabling complete
extraction and analyte speciation are prereauisite. Since
most environmental and waste matrices under test are highly
non—uniform and bioactive, sample size and freciuency, and
sterile controls are necessary.
In our laboratory, routine ultratrace molecular speciation
simultaneously of methyl—mercury and -tin components in
mining discharge sediments was performed using a novel
purge—and-trap gas chromatograph coupled with dual flame
photometric (Sn) and AA (Hg) detectors. This system
provides one example of current ready access to needed data
with state—of—art commercial instrumentation. Additional
illustrations from our Group employing variants of such
“hyphenated” GC—ESD or HPLC—ESD have been used to speciate
organometals in marine sediments, biological sludges,
tissues, aquatic microlayers and atmospheres. The
alternative approach, relying upon NDE ultratrace speciation
in wastes is under active study employing unique metal—
specific fluorogenic ligands as means to decorate the
complex test matrix for assay by microscopic epifluorescence
spectrophotometric imaging. Microbial solubilization and
leaching of heavy metals from wastes, sediments, or ores are
examples provided as illustrations.
INTRODUCTION
An integral feature of both untapped and recycled resources
is the complexity of the host matrix and the huge variety of
morphological and molecular forms adopted by the sought
after components. Even more complicated are cases involving
waste materials where both host matrix and hazardous
components can occur in either the original anthropogenic
forms or in quite unexpected chemically or biologically
modified forms. In any case, analysts, modellers, and
regulators are confronted with special problems where heavy
elements are involved in waste management.
The disposal of metal—contaiminated wastes in the
environment selects for microbial populations resistent to
toxic metals (1). Metal resistance in bacteria is often
conferred by genes residing on plasmic3 DNA, which can be
passed from one bacterial species to another (2) . Often the
resistance mechanisms involve metabolic alteration of the
metal including enzymatic oxidation—reduction or alkylation—
dealkylat ion reactions (3). Consequently, microbially
catalyzed processes in environmentally disposed wastes may
produce chemical species not predicted on strictly chemical

-------
grounds. These processes can alter metal mobilization and
toxicity.
Thus, heavy elements, especially the metals, create
notorious problems in waste control, abatement, or recovery
because they can undergo significant transformations in
environmental media, greatly altering their adverse impacts.
A glimmering of the global hazard of such chemical and bio—
transformations was recognized two decades ago with the
methymercury poisonings of Japan (4). Our recent work
suggests the broad scope of potential analogous involvement
of orgariometallic substances in waste sources via
exocellular solubilization and methylation by the ubiquitous
algal metabolite, iodomethane (5—7).
Organometallic and organometalloidal entities in wastes
hence may stem from original depositions, or form by
subsequent naturally—occurring processes (Table 1). For
example, phenylmercury, tributyltin and methyl arsenic
species enter the environment as a result of commercial use
as biocidal agents (4,10,11). However, methylated species
of metals, for example, mercy (12), tin (13) and arsenic
(14,15), are produced by microorganisms in various
environmental compartments including waters, sediments and
sludges. Organometals are important classes of toxicants,
because, depending upon the specific metal(loid) and the
bound organic moieties, their persistence, degree of
bioaccumulation, and intrinsic toxicities are greatly
magnified by their heightened lipohilicity (16,17).
However, determination of “total” metal levels in an
environmental matrix gives no information as to metal
species present. Sensitive chemical speciation procedures
are therefore necessary for determination of levels of
specific organometallic entities. Detection limits of parts
per billion to parts per trillion are necessary since
organometals may exert physiological effects on organisms at
these levels. Fortunately for analysts and regulators,
recent work has paved the way to an accounting of
environmental hazards which takes advantage of the unique
persistent molecular “signatures” and potential for
unambigious speciation of such active agents in wastes.
Generally, combinations of chromatographic molecular
separation coupled on—line with element—selective detectors
(ESD) afford reliable data, even though standard reference
materials and protocols are lacking.
Assured assay of the quantity and the extant molecular forms
of heavy elements in environmental and waste media requires
complete recovery of analytes in forms suited to the
analytical protocol yet fully diagnostic of the original
hazardous component in the matrix tested. There is
consequent need for analytical procedures that yield
nondestructive in situ characterization or precise
derivatization, complete extraction, and analyte speciation.

-------
Table 1. Examples of organometals released into or formed
in the environment (see references 7—9).
Metal speciesa Origin Commercial use
methylmercury biogenic
phenyirnercury anthropogenic biocide
methyltins biogenic, plastics stabilizer,
anthropogenic catalysts
tributyltin anthropogenic biocide
tr icyclohexyltin
triphenyltin
methylarsenicals biogenic, biocide
anthropogenic
tetramethyllead biogenic
tetraethyllead arithropogenic gasoline additive
methylethyllead
methyls ilicones anthropogenic polymers
phenyls ii icones anthropogenic polymers
methylselenium biogenic
methyltellurium biogenic
a Parent moiety indicated, irrespective of charge or
gegenions

-------
Since most environmental and waste matrices are highly non-
uniform and bioactive, consideration of sample size and fre-
quency, and potential sample alteration by microorganisms is
necessary.
A generalized approach that we have used to speciate metals
is illustrated in Figure 1. The sample is introduced into a
chromatograhic system (gas-GC, liquid-HPLC) that is coupled
to an element—selective detector (ESD), such as a graphite
furnace atomic absorption spectrophotometer (GFAA), flame
photometric detector (FPD) or mass spectrometer (MS). The
coupling between the chromatograph and the ESD may be built
into the instrument by the manufacturer. For example, in
our gas chromatograph system a simple glass tube couples the
column to the FPD. In other instances, heated transfer
lines (GC—AA, ref 18), or Teflon sample cut interfaces
(HPLC—GFAA, ref 19) have been devised to conduct the sample
from chromatograph to the ESD.
This paper describes our work on simultaneous ultratrace
molecular speciation of methyl—mercury and —tin components
in mining discharge sediments using a novel purge—and—trap
gas chromatograph coupled with dual flame photometric (Sn)
and AA (Hg) detectors. This system provides one example of
simple modifications of commercial instrumentation to
provide metal speciation information in environmental
samples. Additional illustrations from our Group are shown
that employ variants of “hyphenated” GC—ESD or HPLC—ESD
systems used to speciate organometals in marine sediments,
biological sludges, tissues, aquatic mircolayers and
atmospheres. An alternative approach, relying upon non-
destructive ultratrace chemical speciation in wastes is
under active study. These techniques include microscopic
Fourier transform infrared spectroscopy and the use of
Unique metal—specific fluorogenic ligands to decorate the
complex test matrix for assay by microscopic epifluorescence
spectrophotometric imaging. In addition, examples of
microbial solubilization and leaching of heavy metals from
wastes, sediments, or ores are described.
HPLC-GFAA
The commonly employed UV detector for HPLC is sensitive only
to chromophores, and is not element—selective. The coupling
of liquid chromatography to graphite furnace atomic absorp-
tion spectrophotometry greatly enhances the usefulness of
HPLC. HPLC-GFAA also offers the advantage that samples,
including aqueous solutions, can often be tested directly
without derivatization or extraction. Ion—exchange, reverse
bonded phase or size—exclusion chromatography can be
employed, with single or multielement detection (20).

-------
FL JUrC 1
SAMPLE
WASTE 11
‘V hydridize
dissolve or
I alkylate
1 extract
ligate
I ____ ____
sparge FILTRATE
or l .— or DERIVATIVE1
inject CENTRIFUGATE _____ _____
I evaporate sparge
or
CONCENTRATE absorb
ec T
CHROMATOGRAPH 1
GC, HPLC, TLC J
IJIiate
SELECTIVE
DETECTOR
(on— or off-line)
(element or
functional group)

-------
We have used HPLC—GFAA to speciate butyltins in a variety of
samples including shipyard sandblasting grit leachates (21).
Such grits are used frequently to remove weathered antifoul—
ing paints containing tributyltin biocides. Increasing
private, commerical and military application of tributyltin—
based antifouling coatings will require means to treat and
dispose of toxic solids and liquids generated by such
drydock maintenance operations, and in turn to monitor
effluents from drydocks or sludges after treatment or
disposal operations. Figure 2 is a HPLC—GFAA chromatogram
showing detection or tributyltin in grit leachates as well
as the much less toxic dibutyltin species, perhaps generated
by biological or chemical degradation.
Pore waters from sediments or sludges are also amenable to
HPLC—GFAA analysis for organometal species. We have
developed a technique whereby pore waters in sediments are
driven, under pressure, through a demountable HPLIC precolumn
which traps the analyte of interest. The precolumn is then
connected to the HPLC—GFAA system and is eluted wi.th the
proper solvents. An example of this methodology is shown in
Figure 3.
The use of size exclusion LC—AA to speciate metals in shale
oils is shown in Figure 4. Here we employed a LC effluent
splitter (1:1) to detect As and Fe simultaneously using two
GFAA instruments to sample and test the split streams auto-
matically. The absolute concentrations of As and Fe in the
shale oils and their stoichiometry were determined by
appropriate calibrations, as well as the approximate
molecular weights of compounds ligating the metals (22).
For example, in the ORNL centrifuged oil, Fe has a bimodal
MW distribution with elution volumes corresponding to about
12,000 to 2,000 daltons. Inorganic arsenic and
organoarsenic compounds were also speciated in shale oil
retort waters and oil shale (23,24) . The presence of
monomeric—methyl— and phenylarsenic species in unretorted
oil shale source kerogen suggests a biogeochemical origin of
these species (25). Thus, HPLC—GFAA is useful for monitor-
ing such processing energy products as well as starting or
waste materials (26).
GC-ESD
Gas chromatography can be used to detect inherently volatile
metal species (e.g. Hg°, (CH 3 ) 2 Hg, (CH 3 ) 4 Sn, AsH 3 ) or
derivated (e.g. hydridized or alkylated) metal species such
as various methyl and butyltin species. The metal species
are introduced into the CC by injecting gas samples and
solvent extracts or by using purge and trap (P/T) samplers.
The metal species are separated on an appropriate GC column
and detected by means of ESDs such as AA, FPD or MS.

-------
Figure 2
CALlS .
+4
+ Bu 2 Sn
Bu 3 Sn
___ L 1 11 1
liii II liii. •. uIIlII IIII
GRIT
I i liii I t1t1
T
I I I I I
0 10 20 30mm

-------
Figure 3
N 2 , 60 aIm
Stainless steel
pressure vesse
Sediment
sample, Ca. 1-2 kg
-Inert filters, 0.45w
Sediment Pore Water Press —
HPLC Pre-Column Assemb’y
Pore water eluant

-------
Figure 4
W 13S SHALE OIL. 5PM 1580
FE
ASL .
PARAHO SHALE OIL
FE
As I ii 11111 1 .IIIIJ
ORHL CENTRIFUGED SHALE OIL
FE
111 1
ill
AA
AA
AA
uv
AA
AA
uv
EXP’2
As 11111! 1 lI1lJ! IlIIWiIJIIi
— mL
I I
0 10 20 30 40
MW (KO)
EXP: 4
— tJv
EXP 2
I 1 II I
I I I I
1O 10 10 1O

-------
We have used a heated stainless steel transfer line between
a GC column and the graphite tube in a GFAA spectrophoto—
meter, run at l8000C for continous sample atomization, to
speciate a variety of organometallic compounds (18) with Hg,
a simpler system is possible. The hydrogen flame in certain
gc detectors (FID, FPD) reduces organomercurials (CH3Hg ,
(CH3)2Hg) to Hg O. The detector exhaust, diluted with
ambient air, is drawn through a flameless AA detector by
gentle vacuum (27). This system provides higher sensitivity
than corwentional GC—FID for organomercurials while also
identifying Hg° in a sample, all by means of a single
chromatogram.
Flame photometry (flame emission) is extremely sensitive for
detecting certain metals and metalloids. This technique has
long been used for P and S detection. However, with appro-
priate interference filters and flame gas conditions, the
FPD can be made selective for metals such as Sn, As, Se, Pb,
and others (28).
We have measured organotin species in environmental waters,
human urine, sediment pore waters, surface microlayers, wood
piling leachates, and ship drydock wash waters using GC—FPD
(29—32). Detection limits for organot ins are in the range
of 1—100 ng/L depending on species and matrix.
Chromatograms showing tin and sulfur species in Chesapeake
Bay waters and surface microlayers are presented in Figure
5. Relatively high levels of methyl and butyltin species
were found in microlayer samples as compared to the under-
lying water. This tendency for organometals to accumulate
in hydrophobic microlayers is an important phenomenon in
relation to waste disposal involving hydrophobic or
lipophilic components.
We have also developed a combined AA and FPD detection
system for use with P/T—CC to speciate simultaneously tin
and mercury in water, sediments, and sludges (Fig. 6).
Additional of NaBH 4 in this P/T-CC-FPD—AA system creates
volatile tin hydrides (SnH 4 , Me SnH4— , n=l—3), reduces
inorganic Hg to volatile Hg°, and produces a volatile
species are retained on a trap of Tenax CC chromatographic
packing at ambient temperature.
The trap is backflushed and heated, and desorbed compounds
are carried on to the GC column for separation. Tin and
mercury species separate and are detected by tandem FPD
(tin) and AA (Hg) detectors. The FPD detector is not sensi-
tive for Hg but reduces all Hg compounds to Hg 0 as they
leave the GC column, allowing their detection in the
flameless AA detector. Detection limits for various mercury
species are approximately 0.1 to 1.0 ,s . g/L.
We have used this method to speciate Hg and Sn in samples
obtained from sites around a lake receiving Hg and Sn wastes

-------
Figure 5
BLANK
MARY’S
DEL)GHT
M
M
/
BLACK MARSH
/
COLGATE
M
C
w
CALl BRATION
6
234
5
r T
0
4 8
12 mm

-------
Vent
P/T-GC-FPD-AA
N 2
20 mL/min
N 2
T
20 mL/mln
T
T
P/T
1
Mg(C10 4 ) 2
L____
Automatic
Purge/Trap
Sampler
0
25 1600 C
10-50 ml
aqueous
sample

-------
from domestic sewage and mining effluents. The long period
of metal input into the water system raised concerns that
biornethylation of mercury, perhaps via a transmethylation
reaction (Figure 7) with methyltin species (33,34) might
have occurred to generate methylmercury species in the
environment. Lake and river water and leachates from
digester sludge, lake sediments and mine tailings were
analyzed (Figures 8,9). Methyltin species, but no methyl—
mercury were detected in the samples, suggesting that Hg was
not in a form available for methylation.
In addition, we have employed the demountable tenax trap to
preconcentrate sample gases (Figure 10), collected over the
surface of Chesapeake Bay and over microbial bioreactors in
which cinnabar (HgS) and pyrite (FeS 2 ) minerals were under-
going biological dissolution. Following sample collection,
the tenax trap was returned to the P/T—GC—FPD——AA system and
the collected volatile metabolites were eluted and
chromatographed. Figure 11 illustrates the production of
elemental Hg by the ore leaching bacterium Thiobacillus
ferrooxidans . Mercury is volatilized as a resistance
mechanism during dissolution of ores by this organism.
MOLECULAR SPECIATION ON SURFACES
The methods described above have been applied to the
chemical speciation of organometallic molecules in solution
or gas phases. However, methods for molecular speciation on
surfaces, which are often the sites of metal binding and
transformation, are needed to fully understand environmental
metals transformations. Microscopic emission and absorption
techniques offer great promise in this area.
Epifluorescence microscopy is used in our laboratory to
image Sn(IV) accumulation on cell and materials surfaces
(17). A Sn(IV)—selective ligand, 3—hydroxyflavone, forms a
blue fluorescent complex with inorganic and certain organic
Sn(IV) compounds. Efforts are under way to extend the ranqe
of ligands and metal species for imaging.
Microscopic Fourier—transform infrared absorption
spectroscopy (FTIR) is another promising area for develop-
ment of surface molecular speciation. We have begun to
analyze surfaces of fossil fuels, ores, and ceramic
materials to characterize the binding and transformation of
metallic and nonmetallic chemical species, in areas as small
as 10—20 urn. Fixation of iron—containing processing
precipitates (jarosites, goethite, hematite) in
hydrometallurgy operations is important for minimizing
environmental impacts associated with mining waste disposal
(35). Rapid, nondestructive microscopic FTIR molecular
characterization of biogenic jarosites on pyrite minerals is
illustrated in Figure 12. The diagnostic “fingerprint” of

-------
III
As
,
As’ ’
Me Sn 4 ”
,
/
‘V
Me As
In
Me 3 As 2
Sn 1
Hg 0 ç
I-
(D
+ 2e
- Me
Me Hg +

-------
GRIBBEN BASIN SLUDGE
÷ NaBH 4
Hg (AA)
Sn (FPD)
F — 1 1 I
Hg°
H
LQ
CD
C
4 8
12 mm

-------
MeSnI-1 3 0.2
Me 2 SnH 2 0.2
Hg 0 ,4
ND
LAKE
0
Hg tr
/
MINE TAILINGS
Hg° tr
SEWAGE
Hg° 0.1
sludge Hg° 2.0
H
l-1
CD

-------
Figure 10
PUMP
FLOWMETER
__ IL?
I TENAX-GC
60-80 MESH
0
o
I a
I S
I S
IS
I I 0.8 )1 MILLIPORE
-GLASS PREFILTER
AIR
2-5 L/MIN

-------
Figure 11
S1E
W2A
2000ng
200ng
standards
46
8
10 12 14 16 18 20 22mm

-------
Figure 12
T. ferroox$dans
£000
Jarosite
0. S000
0. 0000 — _______ - - — — — -
4399 360C 2*00 2 0 60O 1200 SOC £01
2. £000
sterile
1. 20 CC 1
0.6000
—— — -—--—- . - — — - - — - —-- J
4)99 36CC 20CC 2 00C 1600 1200 S00 10

-------
jarosites includes peaks in the 1060—1150 cm - region
associated with “3 sulfate bond absorbances (36). The peak
at 1430 cm’ - assigned to NH 2 deformation suggests that the
precipitate is amnioniojarosite (36). Bacterially produced
jarosites may be distinguished from other abiotically
produced iron precipitates by means of this FTIR methodology
(36). Acceptable forms of processing precipitates for
disposal could be rapidly monitored by the same (or a
similar) procedure.
Important information on molecular binding of both metals
and organometals on solid wastes certainly is possible in
future research employing these FTIR and epifluorescence
microspcopy imaging methods.
SUMMARY
Organometallic compounds occur in the environment as a
result of anthropogenic and biogenic processes. Chemical
speciation methods are available to determine concentrations
of these molecules at ultratrace levels, helping us to
understand their environmental fate, effects, occurrence and
transformation. New methods for nondestructive speciation
of metal—containing molecules on surfaces, including
epifluorescence microscopy coupled with metal—selective
fluorescent ligands and microscopic FT—IR techniques, are
promising areas for future development.
ACKNOWLEDGEMENTS
The research described in this paper was supported in part
by the Office of Naval Research. We thank our colleagues
Edwin J. Parks, JoAnne A. Jackson, Kenneth L. Jewett, Carl
S. Weiss, Cheryl L. Matthias, and Franco Baldi with whom we
collaborated with in much of the work described herein. We
also thank Frank D’Itri, Michigan State University for his
interest and involvement in our analysis of tin and mercury
in lake water, sediment, and sludge samples.

-------
REFERENCES
(1) Babich, H. and Stotzky, C. Heavy metal toxicity to
microbe—mediated ecologic processes: a review and
potential application to regulatory policies
environ. res. 36: 111—137 (1985)
(2) Summers, A. 0. Genetic adaptations involving heavy
metals. In Current Perspectives in Microbial
Ecology, eds. M. J. Kiug and C. A. Reddy, American
Soc. Microbiol., Washington, D.C., pp. 94—104
(1984)
(3) Williams, 3. W. and Silver, S. Bacterial resistance
and detoxification of heavy metals, enzyme Microb.
Technol. 6: 530—537 (1984).
(4) Gavis, J. and Ferguson, 3. E. The cycling of mercury
through the environment. Water Res. 6: 989—1008
(1972)
(5) Manders, W. F., Olson, C. J., Brinckman, F. E., and
Bellama, J. M. A novel synthesis of rnethyltin
triiodide with environmental implications. J. Chem.
Soc. Chem. Commun. 1984: 538—540 (1984)
(6) Thayer, 3. S., Olson, G. J., and Brinckman, F. E.
lodomethane as a potential metal mobilizing aaent in
nature. Environ. Sci. Technol. 18: 726—729 (1984).
(7) Brinckman, F. E., Olson, G. 3., and Thayer, 3. S.
Biological mediation of marine metal cycles: the
case of methyl iodide. In Marine And Estuarine
Geochemistry, eds. A. Sigleo and A. Hattori, Lewis
Pubi., Chelsea, MI, pp. 227—238 (1985).
(8) Brinckman, F. E., and Bellama, 3.M., (eds.)
Organometals And Organo—metalloids: Occurrence and
Fate In The Environment. ACS Symp. Ser. 82.
American Chemical Society, Washington, D.C. (1978).
(9) Craig, P. 3. Metal cycles and biological
methylation. In Handbook of Environmental Chemistry,
Vol. 1, part A, ed. 0. Hutzinger. Springer—Verlag,
Berlin, pp. 169—227 (1980)
(10) Kronstein, M. Controlled release of polymeric
organometal toxicants. md. Eng. Chem. Prod. Res.
Dev. 20: 5—12 (1981).
(11) Woolson, E. A. and Kearney, P. C. Persistence and
reactions of 14 C—cacodylic acid in soils. Environ.
Sci. Technol. 7: 47—50 (1973).

-------
(12) Wood, J. M.; Kennedy, F. S. and Rosen, C. G.
Synthesis of methyl—mercury compounds by extracts of
a methanogenic bacterium. Nature 220: 173—174,
(1968)
(13) Huey, C., Brinckman, F. E.; Grim, S. and Iverson, W.
P. In, Proc. Internat. Conf. Transport Persistent
Chem. Aquatic Ecosystems, eds., A. S. W. de Frietas,
D. 3. Kushner, and S. U. Quadri. National Research
Council of Canada, Ottawa, pp. 73—78 (1974).
(14) Challenger, F. Biological methylation. Chem. Rev.
36: 315—361 (1945).
(15) Baker, K. D.; Wong, P. T. S.; Chau, Y. K.; Mayfield,
C. I. and Inniss, W. E. Methylation of arsenic by
freshwater green algae. Can. 3. Fish. Sci. 40:
1254—1257 (1983)
(16) Thayer, J. S. Organometallic Compounds and Living
Organisms. Academic Press, New York, 273 Pp. (1984).
(17) Brinckman, F. E.; Olson, G. J.; Blair, W. R. and
Parks, E. J. Implications of molecular speciation
and topology of environmental metals: uptake
mechanisms and toxicity of organotins. In, Proc.
Tenth ASTM Symp. Aquat. Toxicol. Hazard
Assessment, in press.
(18) Parris, C. E.; Blair, W. and Brinckman, F. E.
Chemical and physical considerations in the use of
atomic absorption detectors coupled with a gas
chromatograph for determination of trace
organometallic gases. Anal. Chem. 49: 378—386
(1977)
(19) Brinckman, F. E.; Blair, W. R.; Jewett, K. L. and
Iverson, W. P. Application of a liquid chrornatograph
coupled with a flameless atomic absorption detector
for speciation of trace organometallic compounds. J.
Chromatogr. Sci. 15: 493—503 (1977).
(20) Jewett, K. L. and Brinckman, F. E. The use of
element—specific detectors with high-performance
liquid chromatography. In, Detectors in Liquid
Chromatography, ed. T. M. Vicrey. Marcel Dekker,
New York, pp. 205—241 (1983)
(21) Jewett, K. L. and Brinckman, F. B. Speciation of di—
and triorganotins in water by ion—exchange HPLC—GFAA.
3. Chromatogr. Sci. 19: 583—593 (1981)
(22) weiss, C. S.; Parks, E. J. and Brinckman, F. E.
Speciation of arsenic in fossil fuels and their

-------
conversion process fluids. In, Arsenic: Industrial,
Biomedical and Environmental Perspectives, eds. W. H.
Lederer and R. J. Fensterheim. Van Nostrand
Reinhold Co., New York, pp. 309—326 (1983).
(23) Brinckman, F. E.; Weiss, C. S. and Fish, R. W.
Speciation of inorganic arsenic and organoarsenic
compounds in fossil fuel precursors and products.
In, Chemical and Geochemical Aspects Of Fossil Energy
Extraction, eds. T. F. Yen, F. K. Kawahara, and R.
Hertzberg. Ann Arbor Science, Ann Arbor, MI, pp.
197—214 (1983)
(24) Fish, R. H.; Brinckman, F. E. and Jewett, K. L.
Fingerprinting inorganic arsenic and organoarsenic
compounds in in situ oil shale process waters using a
liquid chromatograph coupled with an atomic
absorption spectrometer as a detector. Environ. Sci.
Technol. 16: 174—179 (1982)
(25) Fish, R. H.; Tannous, R. S.; Walker, W.; Weiss, C. S.
and Brinckman, F. E. Organometallic geochemistry.
Isolation and identification of organoarsenic
compounds from Green River formation oil shale. 3.
Chem. Soc. Chem. Commun. 1983: 490—492 (1983).
(26) Weiss, C. S.; Jewett, K. L.; Brinckrnan, F. E. and
Fish, R. H. Application of molecular substituent
parameters for the speciation of trace organometals
in energy—related process fluids by element—selective
HPLC. In, Proc. DOE/NBS Workshop on Environ.
Speciation and Monitoring Needs, NBS Spec. Publ. 618,
eds. F. E. Brinckman and R. H. Fish, pp. 197—216
(1981)
(27) Blair, W.; Iverson, W. P. and Brinckman, F. E.
Application of a gas chromatograph—atomic absorption
detection system to a survey of mercury
transformations by Chesapeake Bay microorganisms.
Chemosphere 4: 167—174 (1974).
(28) Schwedt, G. Gas Chromatography, Chapter 3 in
Chromatographic Methods In Inorganic Analysis.
Verlag, Heidelberg (1980).
(29) Olson, G. J.; Brinckman, F. E. and Jackson, J. A.
Purge and trap flame photometric gas chromatography
technique for the speciation of trace organotin and
organosulfur compounds in a human urine standard
reference material (SRM). Internat. J. Environ.
Anal. Chem. 15: 249—261 (1983)
(30) Jackson, J. A.; Blair, W. R.; Brinckman, F. E. and
Iverson, W. P. Gas chromatographic speciation of

-------
methyistannanes in the Chesapeake Bay using purge and
trap sampling with a tin—selective detector.
Environ. Sci. Technol. 16: 110—119 (1982).
(31) Matthias, C. L.; Olson, G. J.; Brinckman, F. E. and
Bellama, J. M. A comprehensive method for the
determination of aquatic butyltin and butylmethyltin
species at ultratrace levels using simultaneous
hydridization/extraction with GC—FPD. Environ. Sci.
Technol. 20: 609—615 (1986)
(32) Blair, W. R. Characterization of controlled release
dynamics and identification of species released from
OMP impregnated wood pilings. NBSIR 83—2733,
National Bureau of Standards, Gaithersburg, MD
(1983)
(33) Jewett, K. L.; Brinckman, F. E. and Bellama, J. M.
Influence of environmental parameters on
transmethylation between aquated metal ions. In,
Organometals and Organometalloids: Occurrence and
Fate in the Environment, eds. F. E. Brinckman and J.
M. Bellama, ACS Symp. 82, Am. Chem. Soc.,
Washington, D. C., pp. 158—187 (1978).
(34) Brinckman, F. E.; Jewett, K. L.; Blair, W. R. and
Iverson, W. P. Studies on the biomethylation of tin,
arsenic, and mercury. Abstracts of Papers, Amer.
Chem. Soc., Chem. Soc. Japan Chemical Congress,
Honolulu, HI, April 1—6, 1979, paper Inor 222.
(35) Ritcey, G. M. Hydrometallurgy — 10 years later and a
look to the future. Hydrometal. 15: 1—4 (1985)
(36) Lazaroff, N.; Melanson, L.; Lewis, E.; Santoro, N.
and Pueschel, C. Scanning electron microscopy and
infrared spectroscopy of iron sediments formed by
Thiobacillus ferrooxidans . Geomicrobiol. J. 4: 231—
268 (1985)
FIGURE CAPTIONS
1. Generalized scheme for gas and liquid sample work up
and analysis for chemical speciation of metals by
means of chromatographic separation coupled with
element—selective detection. Not shown is direct,
nondestructive molecular characterization by the
latest reflectance or imaging spectroscopies (FTIR,
EMI) , described in the last section of this paper.
2. Tap water leachate from a shipyard sandblasting grit,
used to remove weathered antifouling paints from ship
hulls, was directly speciated. After two days

-------
contact with mild agitation, comparison of leachate
samples (bottom) with calibration solutions
(illustrated at top for 100 ng each of di— and tn—
butyltin cations) indicated that Ca. 6.0 and 3.0 ppm,
respectively, of the organotins were released.
3. Scheme for the extraction of organometals from
sediment pore waters using N2 was used to force
sample through a demountable HPLC precolumn for
subsequent elution and HPLC—GFAA analysis.
4. Dual element-selective detection, for Fe and As, of
the size exclusion chromatograms of various shale
oils.
5. GC—FPD chromatograms of surface water (w) and surface
microlayer Cm) samples from sites in Chesapeake Bay.
Standards (approximately 1 ng each): 1— MeSnH , 2—
MeSnH 2 , 3— Me 2 S, 4— Me 3 SnH, 5— Me Sn, 6— BuSnH3, 7-
Me 2 S 2 (approx. 5 ng), 8—Bu 2 SnH 2 . Substantial amounts
of monobutyltin are found in Black Marsh and Mary’s
Delight microlayer samples, along with tn— and
tetramethyltin in the latter sample. The hydrophobic
microlayer may be a site of organometal accumulation.
6. Purge/trap gas chromatography with flame photometric
and atomic absorption detection system (P/T—GC—FPD-
AA) for tandem detection of inorganic and methylated
tin and mercury species. The magnesium perchlorate
trap between the detectors removes H 2 0 vapor
generated by the FPD.
7. Summary of demonstrated interactions between reported
biogenic methylelements suggesting transmethylation
and redox as prevalent pathways in bioactive, metal—
rich wastes.
8. P/T-GC-FPD-AA tandem chromatograms of pore waters
obtained from lake basin sludge contaminated with tin
and mercury from industrial and mining operations.
The negative peak in FPD at Ca. 6.5 mm may
correspond to a hydrocarbon which characteristically
quenches the flame.
9. Detection of tin and mercury species (ppb) in lake
and river waters and sediment pore waters. *ND
denotes “not detected.”
10. Air sampling for subsequent GC—FPD analysis.
Volatile metal species can be trapped on
chromatographic packing (Tenax-GC) in the field. The
tube is then returned to the laboratory and desorbed
into the GC—FPD-AA system.

-------
11. Vapor phase GC—AA analysis of gases above cultures of
T. ferrooxidans growing on a pyrite—cinnabar (HgS)
mixture. The gases were trapped on a tenax—GC column
(see Figure 10) and desorbed into the GC—AA system.
The bacteria produce HgO but no other volatile (i.e.,
methylated) Hg species.
12. Microscopic reflectance Fourier transform infrared
spectra of iron pyrite oxidizing bacterium T.
ferrooxidans . Biogenic jarosite precipitates are
indicated (33)

-------
MICROWAVE PROCEDURE FOR THE DETERMINATION OF
METALS IN OILY WASTE
THOMAS R. COPELAND, PH. D., ERCO DIVISION OF ENSECO, INC.,
CAMBRIDGE, MASSACHUSETTS
ABSTRACT
Microwave ovens have been used for the wet ashing of
biological and geological samples for over a decade. The
primary advantages of these digestions are the rapidity of
the digestion process and the freedom from atmospheric con-
tamination. This paper presents the development of a micro-
wave digestion procedure for six metals (As, Be, Cd, Cr, Ni,
and Pb) in oils, waste oils, and oily wastes. Digestion
parameters which were investigated include sample size,
microwave power, digestion time and digestion reagents. The
optimum procedure used 1 gram of sample, nitric acid, and
700 watts (full power) for 15 minutes. Recoveries of
organometallic standards spiked into new and used motor oils
ranged from 85% for As to 127% for Cr. This procedure was
then used to digest waste oils and oily wastes. The digest—
ates were analyzed by flame (5 metals) or furnace (As)
atomic absorption, and these results were compared to
results obtained for the same samples by ICP (EPA Method
3040/6010) and by AA using a modified EPA Method 3050 for
the digestion.
INTRODUCTION
Microwave ovens have been used for acid digestions of bio-
logical and geological samples for over a decade. The
stated advantages of using a microwave oven rather than a
conventional hotplate or heating mantle are that the micro-
wave digestion is more rapid and the digestion takes place
in an essentially closed system, thus minimizing airborne
contamination. The primary disadvantage of the microwave
digestion is that the oven must be protected from the acid
fumes. Several means of protection of the oven are avail-
able —— continuous flushing, partial evacuation of a
secondary containment vessel, or protective (Tef1on ) coat-
ing of the oven. Modification of an oven to afford this
protection is relatively simple. Alternatively, a system
specifically designed for acid digestions is available (CEM
Corp., Indian Trail, NC).
Methods currently available in SW 846 for preparation of
oils, waxes, and oily wastes suffer some limitations.
Method 3030 is an acid digestion that uses sulfuric acid in
the digestion mixture. Metals that form insoluble sulfates
(barium and lead) are not recovered by this method. Method
3040 is a dilution procedure in which an oil sample is
diluted with an organic solvent (xylerie, kerosene, etc.) and

-------
analyzed by Inductively Coupled Plasma Emission Spectrometry
(ICP). This method is insensitive to arsenic, in general,
and is unable to detect metals that are contained in large
particles in a waste oil (sludge particles, wear metal frag-
ments, etc.). It is useless for oily sludges or multiphase
samples.
Method 3050 is an acid digestion intended for use with soils
and sludges. It does not use sulfuric acid in the procedure
and, therefore, does allow analysis of lead and barium.
However, volatile elements, such as arsenic, may be lost.
In order to minimize losses of volatile elements, a modi-
fication to Method 3050 may be made. If the flask/condenser
digestion apparatus specified in Method 3030 is used with
the acid digestion procedure specified in Method 3050, lead
and barium are not precipitated and volatiles such as
arsenic are not lost. A complete digestion of an oil or
oily waste by this procedure may take several days.
The microwave procedure used in this study provided
recoveries of the metals of interest (As, Be, Cd, Cr, Ni,
and Pb) comparable to or better than those achieved by the
modified Method 3050, and required only 15 mm in the micro-
wave oven. The method has been applied to new and used
motor oils, oils from a waste oil reclaiming facility, oily
sludges from the waste oil reclaiming facility, a waste oil
from a metal grinding operation, and a contaminated jet
fuel. The microwave results for all of these samples are
comparable to or better than results obtained by Method
3040, Method 3050, or modifications of these methods.
MATERIALS AND METHODS
MATERIALS
o All acids and the hydrogen peroxide used for the oil
digestions were Baker Instra Analyzed (J.T. Baker).
o Organometallic standards used to prepare spiked samples
were Conostan (Dupont) standards.
o Kerosene used for dilution of the oils for analysis by
Method 3040 was from Fisher Scientific.
o The ethoxylated primary alcohol used for the Modified
Method 3040 dilutions was Neodol 91—6 (Shell Chemical
Co.).
o The microwave oven was a 700—watt Quasar oven
manufactured for home use.
o Teflon PFA (perfluoroalkoxy) 180 ml screw—cap lars
were obtained from Cole—Parmer.
o Vacuum desiccators and glassware were acid—cleaned
Pyrex.

-------
METHOD
“ Modified” 3050
This modification utilizes a condenser (recommended in
Method 3030) to prevent loss of volatile elements such as
arsenic. It uses the nitric acid and peroxide reagents of
Method 3050, thereby avoiding the use of sulfuric acid and
the concomitant loss of lead. One gram samples of oil were
digested using nitric acid in the round—bottom
flask/condenser apparatus of Method 3030. Complete diges-
tions required 72—96 hr.
Method 3040
Samples were diluted with kerosene (1:10 on a weight:
weight basis) and analyzed by ICP.
“ Modified” Method 3040
Several articles were found during a literature search (all
from Air Force—supported research at Wright Patterson AFB)
which recommended the addition of a small quantity of an
acid mixture to a waste oil sample to facilitate decomposi-
tion of particulate matter in the sample. These articles
also recommended the addition of an ethoxylated primary
alcohol prior to dilution. The alcohol is somewhat polar
and has slight surfactant characteristics which presumably
stabilize the acid/oil/kerosene emulsion. The procedure
used for this study was adapted from the Air Force research
and may be summarized as follows:
A 1.00 g oil sample in a polyethhylene vial was
mixed with 250 mg of an acid mixture of
HF/HC1/HNO 3 (1:8:1) by shaking on a vortex mixer
for approximately 2 mm. The sample was then
diluted to a total weight of 10 g with a 3:1 mix-
ture of kerosene:alcohol. All mixtures and dilu-
tions were made on a weight basis. (The alcohol
used is NeodolR 91—6 obtained from Shell Chemical
Co., Geismar, LA). These samples were then
analyzed by ICP. Calibration standards were made
with Conostan organometallic standards and
baseoil. Conostan base oil was added to the
standards so that the final weight of standard and
base oil was one gram prior to dilution (to 10
grams) with the kerosene:alcohol mixture. The
calibration tIblank i was one gram of base oil
diluted to 10 grams with kerosene:alcohol.
Standards were made at 1, 5, 10, and 50 ug/g con-
centrations.

-------
Microwave Digestion
The literature search identified several papers which
utilized a microwave oven as the heat source for acid
digestions of sample matrices, such as biological tissues
(NBS orchard leaves, tomato leaves, fish tissue, human
teeth) and rocks and sediments. All these papers made
statements as to the rapidity of the digestion using a
microwave. Since a total digestion of a waste oil by
Methods 3030 or 3050 may require several days, the microwave
digestion was considered an attractive alternative. The
microwave procedure finally employed (the development is
discussed below) is summarized as follows:
A 1.00 g oil sample is weighed into a Teflon
digestion jar (180 ml) and 25 ml of concentrated
nitric acid is added. The jar is sealed and
placed in a Pyrex R vacuum desiccator. The
desiccator is partically evacuated ( 25 in Hg)
and sealed. The disiccator was used to prevent
acid fumes from damaging the electronics of the
microwave oven. The sample was then placed in a
standard microwave oven and heated with maximum
power (700 watts for the unit used) for 15 mm.
For most oils, the desiccator had to be removed
and vented in a fume hood several times during the
heating cycle due to the large quantity of acid
fumes generated. That is, the 15 mm heating
actually occurred in several 3— to 4—mm periods.
After 15 mm in the oven, the sample was rinsed
into a 25—mi volumetric flask and diluted to
volume with ASTM Type II water.
RESULTS AND DISCUSSION
DEVELOPMENT OF THE MICROWAVE DIGESTION METHOD
Initial experiments were performed using a single waste oil
(10W40 motor oil) and several acid mixtures including pure
HNO 3 ; HNO 3 :HC1; HNO 3 :HC1O 4 ; and HNO 3 :H 2 0 2 . No differences
were observed in the appearance of the digestates, nor was
there any difference in the recoveries of lead (the only
metal found in measurable amounts in this oil) in the
various digestates. Four aliquots of oil were digested
using only nitric acid and two aliquots using each of the
other digestion mixtures. The range of results for the
nitric acid digests was 28—35 ug/g. all values for the
other acid mixtures fell within these limits. Since high
chloride concentrations are known to cause loss of arsenic,
and since use of perchloric acid requires special hoods,
pure nitric acid was chosen as the digestion acid. Aliguots
of the waste motor oil sample were then subjected to varying
digestion times and again analyzed for lead to ascertain the
minimum acceptable digestion period. The results are pre-
sented below:

-------
Lead Recovery vs. Digestion Time
Time (mm) % Pb Recovered
9 43
10 93
11 90
12 101
14 97
15 100
18 107
22 104
55 106
It was concluded that a 10—15 mm. digestion was adequate
and that 15 nun, probably provided a sufficient margin of
safety.
To confirm these preliminary findings and to determine
recoveries for the other five metals of interest, a series
of experiments was conducted using three oil samples. The
samples were a new 10W40 motor oil, the same oil after 3,000
mi of use in an automobile (unleaded gasoline) and a
National Bureau of Standards Reference Material (SRM1634a,
Trace Metals in Residual Fuel Oil). The oils were digested
unspiked and spiked with Conostan organometallic
standards at three concentration levels (5, 10, and 20
pg/g). Peroxide was added to some samples after digestion
to determine if it had any effect. The digestion times were
held at 15 minutes to confirm that 15 mm was an adequate
digestion interval except for one oil which was digested for
nine minutes to determine the effect of digestion time on
metals other than lead. Results of these experiments are
presented in Table 1.
All analytical results presented in Table 1 were obtained by
atomic absorption spectrometric (AAS) methods. Arsenic
results were obtained by graphite furnace AAS, and the
remaining five elements were analyzed by flame AAS. Lower
limits of detection could be obtained by analyzing all ele—
merits by graphite furnace procedures. However, the flame
detection limits are adequate to detect five of the six
metals at levels which are below the level of concern in
waste oils. Analysis by flame AAS is considerably faster
than graphite furnace analysis.
The laboratory spikes of digestates are aliquots of the
microwave digestates, spiked with aqueous laboratory stan-
dards immediately prior to analysis. These spikes are
intended to detect any influence the digestate matrix may
have on the analytical technique. Two laboratory spike
recoveries were outside the range of 95—105 percent and
warrant discussion. The laboratory spike recovery for chro-
mium averaged 105 percent. Strong acid solutions (such as

-------
Table 1 • Microwave oven digestions of oils — concentrations found end recoveries
As
As
— Be
Be
Cd
Cd
Cr
Cr
Ni
Ni
Pb
Pb
(pg/g)
found 0
S
recovery
(pg/g)
found
S
recovery
(ug/g)
found
S
recovery
(ug/g)
found
S
recovery
(ug/g)
found
S
recovery
( g/g)
found
S
recovery
New 10140
New 10140
(0.25
<0.25
(1.25
(1.25
0.58
0.48
5
<5
<2.5
(2.5
(5
<5
New 4 5 ig/g
New + 5 wg/g 11202
4.51
4.69
89
93
5.28
5.01
105
100
4.59
4.46
92
88
<5
6.58
-—
130
4.59
5.14
92
102
7.19
(5
143
- —
New + 10 g/g
New + 10 iig/g 11202
8.41
7.15
84
71
10.4
10.4
104
104
8.81
8.39
88
84
11.0
11.0
110
110
10.2
7.36
102
74
10.2
8.39
102
84
New + 20 pg/g
New + 20 jg/g
15.7
15.3
78
76
22.4
22.4
112
112
17.4
17.3
87
86
21.3
22.5
107
113
16.6
17.7
83
88
21.3
19.9
107
100
Used 10140
Used 10140
(0.25
<0.25
<1.25
<1.25
1.35
1.35
<5
<5
<2.5
<2.5
34.5
29.8
115
99
Used • 5 ig/g
Used + 5 ig/g” 11202
5.14
<0.25
103
——
5.78
3.24
120
65
6.12
3.05
96
34
7.53
<5
150
——
6.95
<2.5
139
——
34.5
15.9
90
0
Used + 10 g/g 11202
Used + 10 g/g
8.86
9.31
89
93
11.5
12.1
115
121
10.6
11.9
92
106
15.2
16.2
152
162
11.2
11.2
112
112
42.4
43.8
102
115
Used 4 20 ug/g
Used + 20 g/g 11202
16.7
13.5
83
68
23.2
17.0
116
85
19.5
14.9
91
68
28.2
19.1
141
96
17.7
15.6
88
78
51.3
49.8
96
88
NBS1634a
ND81634a + 20 ug/g
<0.25
17.9
——
89
<1.25
28.3
——
142
<0.5
19.1
-—
95
20.2
33.0
325
160
27.2
38.9
94
79
<5
21.3
——
107
Average recovery
(excluding the incom-
plete digestion)
85
111
89
127
94
104
Lab spike of
digestates
92
101
102
105
101
101
‘As analysis by furnace LAS;
“Incomplete digestion.
all others by flame IAS.

-------
the digestates) are known to cause a slight enhancement of
the chromium response in flame AAS and is the likely
explanation of the slightly high recoveries. The laboratory
spike recoveries for arsenic averaged 92 percent. Elemental
arsenic and many arsenic compounds (particularly halides)
are volatile and may be lost during drying and ashing in the
graphite furnace. Nickel nitrate was added to all solutions
to minimize this loss. It appears as if some loss of
arsenic due to premature volatilization still occurs.
Average recoveries for the spiked samples were aenerally
good. Low recoveries occurred for the sample digested for 9
mm., near complete recoveries occurred for samples dicested
14—15 mm. The addition of peroxide after the acid
digestion did not influence recoveries. For these samples,
peroxide was added after 15 mm. of acid digestion and re-
heated for 2 mm. in the microwave.
As with the laboratory spikes of digestates, the lowest re-
coveries for spikes carried through the digestion procedure
were for arsenic (average, 85 percent), the highest
recoveries were for chromium (average, 127 percent). The
highest chromium recovery is for the NBS reference material
(230 percent) which is based on a non—certified chromium
value.
All of the microwave digests produced a precipitate when
diluted to final volume. This precipitate (white)
redissolved after approximately 1 hr. for all oils except
the NBS reference material. The NBS oil contained a larqe
amount of a red—brown precipitate which did not redissolve.
These (NBS) samples were filtered (gravity filtration
through Whatman 42 paper) prior to analysis. All of the oil
samples left an organic residue in the Tef1on digestion
vessel. This residue varied in color from yellow to rust—
brown and was wax—like (almost asphaltic) in appearance.
Repeated attempts to digest this material in nitric or
nitric/perchioric acids were unsuccessful. Recoveries of
the spiked metals were, however, unaffected by the presence
of the residue.
COMPARISON OF THE SELECTED METHODS
In order to compare the methods discussed above, six waste
oil samples were selected for analysis. Three of these
samples had been obtained by ENSECO for an earlier EPA/OSW
program. These oils had been analyzed by Method 3040 for
that program. These oils had also been analyzed for vola-
tile organics, sulfur, and chlorine. Three sludges from a
refining operation obtained for a different Work Assignment
under this contract were also selected. These samples are
essentially oily wastes containing substantial amounts of
solid material. All six samples were digested in duplicate
by the microwave procedure. Three of the samples had been

-------
previously analyzed (over 1 yr. ago) by Method 3040. All
six samples were analyzed by the modified Method 3040 (acid
added prior to dilution) and the three samples previously
analyzed by 3040 were reanalyzed by 3040 (no acid) to assure
that no degradation of the sample had occurred. The three
oily waste samples were also digested by the modified Method
3050 using a condenser. Analytical results are presented in
Table 2.
Oil 254, from a waste oil reclaiming facility, was rather
clean (pale yellow, not black) and had less than .1 percent
bottom sediment and water. No volatile organics were
detected (including volatile organohalogens) when previously
analyzed. Results for this oil are in rather good agreement
by all analysis techniques. The ICP results for chromium
are near the method detection limit and may be subject to
interference.
Oil 63, from a waste oil collection facility, was dark brown
in color and contained no solid material. This oil con-
tained substantial amounts (over 1,000 j&g/g) of chlorinated
solvents, benzene xylene, and toluene. Again, results for
all analysis techniques are in good agreement, with the
exception of ICP values for Cr being slightly higher than
the AAS results.
Oil 4 was a waste oil from a grinding machine operation of
an auto manufacturer. The oil was extremely dark in color,
viscous, and contained substantial amounts of suspended par-
ticulate matter, including metallic fragments. Results for
this oil are not in agreement for Cd, Cr, Ni, or Pb. The
ICP results by Method 3040 (no acid) are the lowest values
for all four metals. Method 3040 (with acid added) yielded
higher concentrations for all metals and the microwave
digestate concentrations were even higher. These results
are consistent with a major portion of the metals content
being in the particulate matter. The ICP used in the
previous analysis (by 3040, no acid) utilized a conventional
nebulizer system and would discriminate against metals
contained in large particulates. The nebulizer used in the
ICP for this study is of a more efficient design and, in
fact, the results for this study are higher than for the
previous analysis. ICP results for the sample which con-
tained acid are higher than those for the same ICP and
nebulizer for the sample which did not contain acid. For
this oil, the acid addition appears to facilitate the dis-
solution of some of the particulates with a concomitant
increase in metals results. This oil (from a metal grinding
machine) is most like the jet engine lubricating oils
containing “wear metals” analyzed in the Wright Patterson
studies and these results support their conclusion that acid
additon is beneficial. However, the acid addition

-------
Table 2. Analytical Results
Concentration
Waste Oil/Method As Be Cd
( g/g)
Cr Ni Pb
Oil 254:
ICP (previous study) 6.0 <0.2 <0.5 1.3 1.2 52
ICP (this study), no acid <5 0.02 0.3 0.7 1.3 42
ICP (this study), with acid <5 <0.01 0.3 0.7 1.3 41
Microwave <1.25 <0.5 <0.5 <0.5 <2.5 49
Microwave, duplicate <1.25 <0.5 <0.5 <0.5 <2.5 46
Oil 63:
ICP (previous study) 7.6 <0.2 1.1 2.7 1.1 490
ICP (this study), no acid <5 <0.01 0.7 1.9 0.8 476
IC? (this study), with acid (5 <0.01 0.8 1.9 0.8 610
Microwave <1.25 <0.5 1.0 <0.5 <2.5 503
Microwave, duplicate <1.25 <0.5 1.0 <0.5 <2.5 455
Oil 4:
IC? (previous study) <5 <0.2 <0.5 560 1.8 6
IC? (this study), no acid 1.1 0.02 1.2 784 4.1 21
IC? (this study), with acid <1 0.04 2.0 1,030 6.8 35
Microwave (1.25 <0.5 2.5 1,500 7.2 54
Microwave, duplicate <1.25 <0.5 2.6 1,440 7.7 49
Oil 17501:
IC? (this study), with acid 3.2 <0.01 <0.1 5.4 <0.1 <0.5
Microwave 2.6 <0.5 2.5 160 41 275
Microwave, duplicate 2.4 <0.5 2.1 150 34 222
3050, Condenser 2.3 <0.5 2.0 150 40 170
Oil 17502:
ICP (this study), with acid <5 0.2 7.1 902 69 352
Microwave 4.0 <0.5 3.2 250 59 255
Microwave, duplicate 4.0 <0.5 3.3 258 59 258
3050, Condenser 2.8 <0.5 5.0 640 70 300
Oil 17503:
IC? (this study), with acid <5 0.18 3.3 178 59 206
Microwave 4.2 <0.5 3.6 330 67 310
Microwave, duplicate 4.6 <0.5 3.9 323 70 328
3050, Condenser 3.3 <0.5 3.8 360 81 330

-------
apparently does not completely dissolve the particulate
metals, since the results by microwave digestion are still
higher than the ICP with acid values.
Oils 17501 to 17503 were oily wastes from a reprocessing
facility. These samples were taken at various depths in a
storage tank which feeds a rotary filter press. The
contents of the tank were the residuals remaining after the
reprocessing. Oil sample 17501 was taken from the top to
middle of the tank and was an oil/water/sludge emulsion.
Oil 17502 was the sludge from the bottom of the storage
tank, and Oil 17503 was the sludge remaining after rotary
filtration. Sample 17501, when acidified and subsequently
diluted with alcohol/kerosene, formed a gelatinous layer
with a yellow liquid layer on the surface. Only the liquid
layer was analyzed by ICP. Apparently, only the arsenic was
solubilized and present in this layer. This gel formation
did not occur with the other two oily wastes, and the ICP
and microwave results are in reasonable agreement.
The ICP nebulizer did clog several times during the analysis
of these oily waste samples. These samples had a very high
particulate content and this problem was expected. For
sample 17502, the ICP (Modified Method 3040) results are
higher than the microwave results, while for oil sample
17503 the reverse is true. Results from the Method 3050
condenser digestion are generally intermediate between the
microwave and ICP results for these two samples. The
variability in the analytical results for these samples is
most likely due in part to the inhomogeneity of the samples
which contained particulates including metal shavings (these
samples ranged from 7,400 to 15,000 ug/g iron). However,
the agreement between duplicate determinations by the micro-
wave procedure does not reflect homogeneity problems. The
results for As by the Modified Method 3050 are consistently
lower than than the ICP (where available) or microwave
values, indicating that even with the condenser, some
arsenic is volatilized. In fact, the spike recovery from
another similar oily waste prepared (by Modified 3050) at
the same time as were these samples was only 70 percent for
As. Recoveries of the analytical spike for the other five
metals of interest ranged from 95 percent to 102 percent.

-------
INTERLABORATORY EVALUATION OF ICP METHOD 6010
THOMAS A. HINNERS, ENVIRONMENTAL MONITORING SYSTEMS
LABORATORY, U.S. ENVIRONMENTAL PROTECTION AGENCY, LAS VEGAS,
NEVADA; CLIFTON JONES, HOMIGOL BIESIADA, DONALD SCHOENGOLO,
THOMAS STARKS, JOSEPH CAMPANA, ENVIRONMENTAL RESEARCH
CENTER, UNIVERSITY OF NEVADA, LAS VEGAS, NEVADA
ABSTRACT
Inductively coupled plasma emission spectroscopy (Method
6010) is undergoing interlaboratory evaluation for sludge-
type wastes. The wastes included in the study are furnace
ash, treatment sludge, mine waste, electro-plating sludge as
well as contaiminated soil and sediments. Twenty-three
elements will be determined by conventional ICP-OES in
digests in the spiked and unspiked waste materials.
The raw materials as well as bulk digests well be analyzed
by all participating laboratories so that the variation con-
tributed by sample preparation can be distinguished from the
variation contributed by the measurement process. Both
sequential and simultaneous instruments will be used in the
study to allow an assessment of performance as a function of
instrument type. The study will include an analysis-of-
variance to characterize the homogeneity of each waste
material before distribution to the participating laborator-
ies. Heterogeneous mateirals are not suitable for inter-
laboratory studies.
Method 6010 includes quality control tests to assess the
data quality. If the results for diluted samples or for
spiked digest are not within the control limit of 10%, the
method of standard additions is specified to provide data of
acceptable quality. For this interlaboratory study, a 4-
point method of standard additions is specified to allow a
least-squares analysis to obtain the sample concentrations.
Use of alternate wavelengths is recommended to confirm data
quality when spectral overlap could be present.
While Method 6010 will accurately reflect the element quan-
tities solubilized by the digestion procedure used, it will
not reflect the waste content of an element when components
in the waste form insoluble compounds (such as barium and
lead with sulfate). However, the quality control spike
addition in Method 6010 will indicate a data quality problem
for affected elements in such a situation so that data of
known quality is obtained. Methods that are applied
directly to solids (such as neutron activation analysis and
x-ray fluorescence analysis) are called for to assess waste
content (but not availability) when solubility limitations
are present.

-------
INTRODUCTION
Method 6010, an inductively coupled plasma optical emission
spectroscopy (ICP-OES) method, is undergoing interlaboratory
evaluation for sludge-type wastes. The wastes included in
the study are a fly ash, a waste treatment sludge, mine
tailings, an electroplating sludge, a contaminated soil plus
a river and an estuarine sediment. The twenty-three ele-
ments to be determined by conventional ICP-OES in digests of
the spiked and unspiked waste materials are given in
Table 1.
TABLE 1.
23 Elements Included in this Method 6010 Study
Aluminum Cobalt Potassium
Antimony Copper Selenium
Arsenic Iron Silver
Barium Lead Sodium
Beryllium Magnesium Thallium
Cadmium Manganese Vanadium
Calcium Molybdenum Zinc
Chromium Nickel
Boron and silicon are excluded from this list because the
digestion procedure (Method 3050) specified in Method 6010
for sludges, soils and sediments is not suited to dissolving
refractory compounds of these elements. Method 3050
involves a digestion of samples with nitric acid and
hydrogen peroxide.
The raw wastes as well as spiked bulk digests will be
analyzed by all participating laboratories so that the
variation contributed by sample preparation can be
distinguished from the variation contributed by the measure-
ment process. Both sequential and simultaneous instruments
are desired in the study to allow an assessment of perform-
ance as a function of instrument type.
Because heterogeneous materials are not suitable for inter-
laboratory studies, this project included an analysis-of-
variance by the coordinating laboratory to characterize the
homogeneity of each waste material before distribution to
the participating laboratories.
The variance of the results to be obtained from a
homegeneity evaluation will be a combination of three varia-
bilities: sample homogeneity, subsampling, and analysis
(including digestion). The experimental design allows the
analytical variance to be distinguished from sample/sampling
variance.
Each of the seven waste material was remixed using a V-
blender to assure that each was in the same state of homo-
geneity that it was prior to shipment. Settlement and

-------
partitioning that could occur during shipment could affect
homogeneity among the several aliquots. Samples were
aliquoted by multiple splitting using a V-blender to assure
maintaining homegeneity during aliquoting.
In order to determine the homogeneity and subsampling variance
of each of the seven samples in the coordinating lab study,
two subsamples were analyzed from random aliquots. Fro the
homogeneity/subsampling evaluation, five aliquots were chosen
randomly from the 13 aliquots of each material. A four-gram
subsample was removed from each aliquot and then three of the
same five aliquots were subsampled randomly a second time.
This scheme of eight subsamples of each of the seven wastes
allows a satistical description of the subsampling
variability, the homogeneity, and the analytical variability.
To control the subsampling variance, the same quantity of
subsample or subaliquot must be taken from each aliquot by the
participating labs. The labs having sequential instruments
will require an upper limit of 200 mu of digestate, and the
project plan requires the maximum dissolved-solids levels of
Method 3050 (2 g/100 ml). Consequently, all labs were
instructed to digest 4-g subsamples using the Method 3050
procedure.
Before any field samples were analyzed, the instrument
detection limits (IDLs) were determined within 30 days of the
start of the analyses. The instrument detection limits (in
ugh) were determined by multiplying by three, the average of
the standard deviations obtained on three nonconsecutive days
from the analysis of a standard solution at a concentration
three to five times the expected IDL, with seven consecutive
measurements per day.
In addition to conducting the homogeneity evaluation, the
coordinating laboratory (Environmental Research Center,
University of Nevada) was directed to:
1. Conduct a large-scale digestion of each waste
material (proportional to 2 g of waste to 100 ml of final
digest) to produce a volume of at least 2.0 liters.
Participating laboratories needed to receive a sufficient
volume of each digest (after spiking by the coordinating
laboratory) to allow for a direct analysis, two dilutions, and
potentially for three standard additions (if the method of
standard additions is required). At least 1500 ml were
required in order to provide 100 ml to five laboratories with
simultaneous instruments and 200 ml to five laboratories with
sequential instruments.
2. Spike the bulk digests (one per waste) to provide
two liters that has all 23 elements (Table 1) present at no
less than 100 times the instrument detection limits (when
consistent with solubilities) while diluting each digest
minimally (so that the matrix concentrations were not altered
significantly). After the digests were spiked and mixed, the

-------
solutions were stored at room temperature for at least 48
hours. Solutions were clarified by filtration when
precipitation or turbidity was observed.
3. Prepare 1000 ml of three quality control (QC)
solutions for the 23 elements with concentrations near the
middle of the linear calibration ranges. These solutions were
used to evaluate the calibration uniformity and digestion
recoveries of the laboratories participating in the Method
6010 evaluation study. Incompatibility of the standards and
interference interests made more than one QC solution
necessary.
4. Analyze by ICP-OES for the 23 specified elements in
the 7 spiked digests and all QC solutions at 4 + 2 days, 2
weeks ‘. 2 days and 6 weeks ‘. 2 days following preparation to
monitor the stability of the element concentrations. Tests
for the need for the method of standard additions (MSA) and
apply MSA (where needed) following the procedure below.
Test for the need to use the method of standard additions (as
mandated in Section 7.6.5 of Method 6010) by diluting
solutions 1:4 (i.e., 1 part sample plus 3 parts solvent) with
a solvent consisting of 2.5% (v/v) HC1 and 5% (vlv) UNO. . If
the concentration of an element in these diluted soluti8ns is
at least 10 times the participating laboratory’s instrument
detection limit (IDL) and multiplication of the diluted
concentration by 4 does not produce a value that is within 10%
of the concentration obtained on the undiluted digest, apply
the method of standard additions for that element using (where
feasible) the undiluted digest. For element concentrations
that are measurable but too low for this dilution test, spike
a portion of the undiluted digest to produce a concentration
that is 20 times the laboratory’s IDLs (or below any expected
solubility limits) while keeping dilution of the digest to a
minimum. If the expected and observed concentrations for the
spiked digest differ by more than 10%, apply the method of
standard additions for determining that element in that waste.
The method of standard additions is not required when an
element concentration is undetected in an undiluted digest or
a spike added to a digest is precipitated. Flag latter data
with “pcp” on Form I.
If the method of standard additions (MSA) is used for deter-
mining the concentration of an element in a digest, the fol-
lowing conditions apply:
a. data must be obtained that is in the linear
calibration range,
b. the digest must be analyzed unspiked and at
three spike levels (with the unspiked portion diluted to the
same extent as the spiked portions but no more than 5%)
designed to increase the apparent digest concentration by (but

-------
not beyond the linear range) and no less than lx IDL, 2x IDL,
and 4x IDL,
c. the MSA data will be subjected to least squares
(linear regression) analysis to obtain the sample
concentration (i.e., the intercept), the slope, and the cor-
relation coefficient; and the results reported on CLP Form
VIII and the concentration flagged on CLP Form I with the
letter to signify determination by standard additions,
d. if the correlation coefficient (r) for an MSA
analysis is less than 0.995, flag the data on CLP Form I with
the symbol “+“.
5. Remove 4-g portions from two of the aliquot bottles
for each waste and spike the separate portions with mixed
standard solutions such that the resultant digests will be
expected to have concentrations for all 23 elements that are
at least 100 times the coordinating laboratory’s instrument
detection limits.
Spiking solutions will also be sent to the participating
laboratories to allow for spiking of 2 portions of each waste.
Conduct ICP-OES analyses on each spiked waste after following
the digestion procedure in Method 3050. Test these digests to
determine the need for MSA (as specified above) and, if
necessary, conduct MSA analyses.
6. Determine the solids content of each waste material
in duplicate using a 2-g portion from 2 aliquot bottles
following the procedure in Method 3050. Method 3050 specifies
that element concentrations will be reported on a dry-weight
basis.
7. After the second set of analytical measurements
specified for the spiked digests was obtained (2 weeks + 2
days after preparation), the following materials were sent to
the laboratories participating in the Method 6010 evaluation:
a. 200-ml portions of the spiked digests to
laboratories with sequential instruments,
b. l00-ml portions of the spiked digests to
laboratories with simultaneous instruments,
c. one aliquot of each waste containing at least
16 grams,
d. 100 ml of each of three sigle-blind QC solu-
tions for sequential-instrument labs and 50 ml for
simultaneous-instrument labs to be analyzed by participant
labs both with and without digestion. For sequential
instruments 50 ml will be digested to yield 200 ml of digest

-------
and for simultaneous instruments 25 ml will be digested to
yield 50 ml of digest,
e. spiking solutions and spiking instruments for
portions of each waste to be subsampled in duplicate by
participating labs (using 4-g portions) such that the
resultant digests will be expected to have concentrations for
all 23 elements that are at least 100 times the instrument
detection limits.
By having each participating laboratory add the spiking
solution after the portions are removed, the potential for
non-homogenous spiking of the solid materials in bulk is
avoided as is this source of variability for results from the
laboratories. If the coordinating laboratory spiked portions
for the participating laboratories, the evaluation would not
reflect the aliquoting and weighing by each lab that are parts
of Method 3050 in preparation for measurements by Method 6010.
Participating laboratories in the ICP Methods 6010 study were
directed to:
1. Analyze 10 solutions (provided by the coordinating
laboratory) requiring no digestion. These solutions are the
seven spiked digests and the three single-blind QC samples
prepared by the coordinating laboratory. The same criteria
was required for MSA tests and application as for the co-
ordinating laboratory.
2. Weigh four 4-gram portions of each waste (to the
nearest 0.01 g) into 4 acid-cleaned beakers for the prepara-
tion of 200-ml digests. Record the gross weight for one of
the beakers with the sample portion to be dried to determine
the solids content (per Method 3050). Total weighings for
this subsampling are 4 x 7 wastes = 28. Although Method 3050
allows 1-to 2-gram portions for each 100 ml of digest, use of
the higher proportion is specified here to provide an
evaluation of the method at the maximum dissolved-solids
level.
3. For each waste, add spiking solution (in the volumes
to be specified by the coordinating laboratory) to the sample
aliquots in two of these beakers. Total predigestion spiking
is 2 x 7 wastes = 14.
4. For each waste determine the percent solids on one
of the weighed (unspiked) portions following the procedure in
Method 3050. Total solids determinations are seven (one per
waste). Digest these dried portions along with the other
three portions per Step 5 below.
5. Digest the two spiked and the two unspiked portions
of each waste plus three single-blind QC solutions (provided)
and a blank according to the procedure for ICP analysis in

-------
4et.’iod 3050. Note that Step 7.6 in Method 3050 is to be
excluded because it is not specified for ICP analyses.
Because Method 3050 describes the digestion procedure when 2
grams of sample are used to produce 100 ml of digest, twice
the reagent volumes specified in sections 7.2-7.5 in Method
3050 need to be used to yield 200-mi digests from the 4-gram
portions specified for this study (Step 2 above). Digest
either 50 ml of each of the three blind QC solutions
(provided) to poduce three 200-ml digests for sequential
instruments or 25 ml of each of these QC solutions to produce
100-ml digests for simultaneous instruments, unless directed
otherwise by instructions received with the samples.
Store the resultant 200-ml digests in acid-cleaned plastic
bottles, and conduct ICP-OES analyses within seven days for
the 23 elements listed in Method 3050. If the concentration
of any required element in these digests exceeds the
calibrated range, repeatedly dilute (with a solution that has
2.5% (v/v) HC1 and 5% (v/v) HNO ) and analyze portions of
those digests until all 23 elemenf’s have been measured in the
calibrated ranges. Report the results on the CLP Cover Page
and on revised Forms 1, V and VI. Total digestions are 32
(4 x 7 wastes + 3QC + blank).
Method 6010 includes quality control tests to assess the data
quality. If the results for diluted samples or for spiked
digest are not within the control limit of 10%, the method of
standard additions is specified to provide data of acceptable
quality. For this interlaboratory study, a 4-point method of
standard additions is specified to allow a least-squares
analysis to obtain the sample concentrations. Use of
alternate wavelengths is recommended to confirm data quality
when spectral overlap could be present.
While Method 6010 will reflect accurately the element
quantities solubilized by the digestion procedure used, it
will not reElect the waste content of an element when com-
ponents in the waste form insoluble compounds (such as barium
and lead with sulfate). However, the quality control spike
addition in Method 6010 will indicate a data quality problem
for affected elements in such a situation so that data of
known quality is obtained. Methods that are applied directly
to solids (such as neutron activation analysis and x-ray
fluorescence analysis) are called for to assess waste content
(but not availability) when solubility limitations are
present.

-------
METHODS FOR EVALUATING SOLIDIFIED WASTE
P. HANNAK AND A. J. LIEM, ALBERTA ENVIRONMENTAL CENTRE,
VEGREVILLE, ALBERTA, CANADA; P. COTE, WASTEWATER TECHNOLOGY
CENTRE, ENVIRONMENTAL PROTECTION SERVICE, ENVIRONMENT
CANADA, BURLINGTON, ONTARIO, CANADA
ABSTRACT
A number of methods for characterizing solidified waste is
currently being assessed in the EC/EPA/AEC/Industry Co-
operative Study. These are short-term laboratory tests
which yield information on material properties (bulk
density, specific gravity and moisture content), ability to
maintain physical integrity (unconfined compression strength
and weathering resistance - freeze/thaw and wet/dry), poten-
tials for release of contaminants by convective transport
(permeability) and dissolution (acid neutralization
capacity, Toxic Characteristic Leaching Procedure- TCLP -and
equilibrium leaching), indication of long-term rate of con-
taminant release (dynamic leaching) and containment
mechanisms (sequential extraction). A brief description is
given of these tests, their rationale and utilization.
Some of the methods were developed elsewhere and have been
adopted or adapted with minor modifications. Others have
been developed specifically for the co-operative study.
Emphasis is thus given to the latter, specifically to the
method for quantifying resistance to wet/dry cycling. The
development of this method and preliminary results are
described, showing the reproducibility and ability to detect
differences in weathering resistance of various solidified
products.
INTRODUCTION
Landfilling or land disposal in general is widely used in
waste management. It is perhaps the only practical option
for dealing with wastes that contain inorganic contaminants,
such as heavy metals.
A major concern with landfilling is the potential release of
contaminants and the consequent contamination of ground and
surface waters. Past incidences have shown that unless pre-
cautionary measures are taken, long-term adverse environ-
mental impacts could result and remedial actions are very
costly, if not impossible.

-------
Various precautionary measures could be and have been taken.
These include disposal site selection (eg. climatic and
geologic considerations) and the use of protective liners.
A different but complementary measure is to treat the waste
prior to disposal. The treatment, commonly referred to as
solidification, stabilization and fixation, provides con-
tainment of contaminants within the treated waste. Briefly
described, the containment is effected by forming an
“impermeable” solid matrix, or lowering the soluhility of
the contaminants, or both. The development of test methods
that could indicate the effectiveness of such containment
should therefore be of interest to researchers, regulatory
agencies and industry.
A co-operative study( 1 ) is being undertaken to assess the
applicability of such test methods to a variety of wastes
treated by commercial processes. These are short-term
laboratory test methods for characterizing the intrinsic
properties of solidified waste that are related to contain-
ment effectiveness. Some of these methods have been adopted
or adapted with minor modifications from elsewhere, such as
those for soils, concrete and nuclear waste testing. Others
have been developed specifically for the co-operative study.
This paper presents an overview of all the test methods that
are being assessed, their rationale, utilization and limita-
tions. A more detailed description is given of a method
that has been developed for the co-operative study for
quantifying resistance to weathering (wet/dry cycling).
Results are presented to show how readily differences in
resistance of various solidified wastes could be detected,
and hence the utilization of these methods for process corn-
par i son.
RATIONALE
The effectiveness of containment would obviously depend on
the waste, the process used and the conditions at the dis-
posal site. As a generalization, and only in terms of the
properties of the treated waste, the following gross para-
meters could be expected to be important:
(1) contaminant solubility
(2) diffusive and convective transports through the
treated waste interstices, and
(3) ability to maintain physical integrity
The test methods have therefore been developed to yield in-
formation that are directly or indirectly related to these
parameters. Although the relationship may not he known
quantitatively, or still needs to be developed, such test
methods will have immediate and useful applications. The
test results could be used for comparison purposes - for
example, assessment of a new process, optimization of an

-------
existing process and selection of alternative processes.
Furthermore, these methods could be standardized and used
for regulatory purposes. In fact 1 , the Toxicity
Characteristic Leaching Procedure (TCLP)(’) is included in
the co-operative study.
TEST METHOD DESCRIPTION, SIGNIFICANCE AND LIMITATIONS
A detaile 1 description of the test methods is given
elsewhere,1 .3, 4 ) hence only a summary is presented, as shown
in Table I. The following discussion therefore deals only
with the significance and limitations of the test results.
(1) Bulk density, specific gravity and moisture content
( Tests #1, 2 and 3) : By itself, bulk density has no bearing
on containment efficiency. However, combined with specific
gravity and moisture content, it could be used to compute
porosity, which is related to permeability and hence poten-
tial release of contaminants by convective transport of
liquid through the treated waste. Moreover, bulk denisty
could be used to estimate landfill size requirement or life
time.
(2) Permeability (Test #4) : As previously indicated perme-
ability is directly related to convective transport of con-
taminants.
(3) Unconfined compression strength freeze/thaw and
wet/dry resistance (Tests #5, 6 and 75 : The treated and
landfilled waste could be subjected to conditions which
might lead to disintegration. This could result in forma-
tion of cracks and generation of particulates, which in turn
would increase permeability and liquid-solid contact area.
Unconfined compression strength and weathering resistance
tests are therefore used to indicate ability to maintain
physical integrity under pressure and adverse climatic
changes, respectively.
(4) Acid neutralization capacity, equilibrium leaching and
sequential extraction, and TCLP (Tests #8, 9, 10 and 12) : A
measure of containment efficiency in contaminant concentra-
tion corresponding to “complete” dissolution under
equilibrium conditions. This upper limit concentration will
depend on, amongst other factors, the pH and nature of the
leaching agent. The values in distilled water and in an
acidic or buffered leaching agent are obtained from equili-
brium leaching and TCLP, respectively. In sequential ex-
traction, progressively more aggressive leaching agents are
used to indicate also the mechanisms by which the con-
taminants are contained. Since heavy metals are typically
the contaminants of concern, and since their solubility is
highly pH-dependent, capacity to neutralize acidic leaching
agents is therefore also an important property.

-------
(5) Dynamic Leaching (Test #11) : The concentrations
measured in equilibrium leaching, TCLP and sequential
extraction are “equilibrium” or “static” values. No infor-
mation could be deduced on the rate of release or its
dependence on time. In dynamic leaching, “fresh” leaching
agents are used after certain periods of liquid-solid
contact have elapsed and the solid and leachate are
separated. Hence time-related behaviour (diffusive
transport) could be dpduced and used for long-term predic-
tion of leaching rate.1.S, 6 )
It should be immediately apparent that the results of these
tests could not and should not be directly used for preclic-
tion of contaminant release under actual field conditions.
Such prediction is beyond the scope of the purpose of
developing the test methods. Even for process comparison
purposes, the interpretation of the test results should he
cautiously made. On the basis of the results of a specific
test, one could compare or rank different processes.
However, the relative significance of different tests is
site-specific. For example, the importance of freeze-thaw
resistance would obviously depend on local climatic conch-
tions. In some areas, it is an irrelevant property since
freezing never occurs. Such an extreme case can be dealt
with somewhat straight forwardly. Consider, however, a case
where two processes are to be compared: one produces a
“high” permeability but “low” solubility product, and the
other the converse. Assessment of the “better” process
could not be universally made. It must be made on a site-
specific basis. Further work is thus still needed to assess
the feasibility of developing “performance indices” based on
all the test results and the site-specific conditions of the
landfill.
PRELIMINARY RESULTS
A sub-project was carried out to familiarize the partici-
pating laboratories in the co-operative study with the test
methods and to refine or modify these. Two wastes -an
“artificial” aqueous waste and rolling mill baghouse dust -
were solidified by “generic” processes and the treated waste
specimens were tested in three laboratories.
The details of the results are still to be published. A
partial summary is given in Table II only to show the ranges
of values obtained. This sub-project proved to be useful in
identifying areas of refinement and modifications. The pro-
cedures adopted in the co-operative study are described in
Ref. 3.

-------
METHOD FOR MEASURING WET/DRY RESISTANCE
Method Description
To our knowledge, there are two methods for quantifying
resistance to wet/dry cycling. These methods are used for
testing soil and cei ent samples, and are described in ASTM
D559-76 standard.(lO)
The following features of the existing methods have
been adopted:
(1) Soaking in water and oven drying to represent
wet/dry cycling.
(2) Specimen weight loss as a measure of wet/dry
resistance, and weight loss measurement after each
cycle.
(3) Twelve as the maximum number of cycles.
There are, however, major modifications as described
below:
(1) Use of control specimens : These undergo identical
treatment except the drying phase, during which they are
kept in a moist atmosphere (> 96% RH). Weight losses of the
test specimens could then be corrected with those of the
control specimens. This approach has been adopted to elimi-
nate the effect of other factors that might contribute to
weight loss, such as matrix dissolution.
(2) Measurement and expression of weight loss : A wire
brush is not used. “Loose particulates” are defined as the
residue that remains after a specimen is removed at the end
of soaking (wetting phase). Measurements of weight loss are
made directly on the dried residue, and not indirectly on
the wet test specimens. The weight loss is then expreseed
on a dry basis relative to the original dry weight of the
specimen. This approach has been adopted to improve re-
producibility and precision, and to eliminate the need for
accounting for saturation changes in the test specimens.
(3) Test specimen preparation : Small size specimens, 45 mm
dia. x 74 mm long, are used to reduce space and material re-
quirements. Thes are identical to those used in the
freeze/thaw test,( ) thus allowing the use of common control
specimens to save both time and effort. In addition, a
minimum of 28 days of curing time was used to “stabilize”
the specimens.
(4) Materials and wet/dry conditions : Distilled water is
used to standarize dissolution conditions and to eliminate
differences in solid contents of the soaking fluid. Soaking
and drying times are set at 24 + 1 h to fit to a “normal”

-------
laboratory schedule. The drying is carried out in a vacuum
oven at 60 ± 3°C to prevent irreversible loss of water of
hydration. The oven is evacuated to a minimum of 70 kPa and
purged with nitrogen at a flow rate corresponding to a
nominal residence time of 30 + 3 mm. These conditions are
set to minimize difference in oven size and drying capacity.
A summary of the procedure is described below:
(1) Pr9pate specimens according to ASTM C305-82( 1 . 1 ) and
Cl92 81U 2) standards, then mold and cure as previously out-
lined.
(2) Use one specimen for moisture content determination as
specified in Ref.3.
(3) Weigh “as is” the test and control specimens.
(4) Place both control and test specimens in pre-weighed
beakers, dry test specimens in a vacuum oven, hut keep the
control specimens in moist atmosphere as previously
described (This step constitutes the drying phase).
(5) Remove test specimens from the oven, and soak both the
test and control specimens by adding distilled water into
the beakers (This step constitutes the wetting phase).
(6a) After 24h, remove both control and test specimens and
repeat steps (3) to (6) until twelve cycles are reached or
tests specimens destroyed.
(6b) Dry and weigh the residues that remain in the beakers
after Step (6a) is completed.
(7) Define and express the results as follows:
(1) Relative weight loss of specimen i (control
or test) at cycle j, is defined.
= Rj,j/(Wj(l-a) (1)
where = dry weight of residue of specimen i at cycle j
measured in Step (6b)
Wj = initial weight (wet basis) of specimen i
measured in Step (3)
a = weight fraction of water measured in Step (2).
(2) Cummulative average weight loss of the
control at cycle n, CL , is computed by
CL =1 (2)

-------
where c = superscript denoting control specimen
NC = number of replicate control specimens
(3) “Corrected” cummulative average weight loss
attributed to wet/dry cycling at cycle n, CL , is
computed by
CL =j, , >2 >2 (3)
and CL = CL - CL (4)
where t = superscript denoting test specimen
NT = number of replicate test specimens
CL = cummulative average weight loss of the test
specimens. For convenience, CL will be
referred to simply as “corrected cummulative
weight loss.”
Ma t e r i a is
Three types of solidified product were prepared to
demonstrate the applicability of the method to common binder
systems. In addition, two types of coating were used to
determine their effect on wet/dry resistance. Table ITT is
a summary of the materials used.
Results and Discussion
Reproducibility
For each of the solidification products shown in Table III,
three replicates were used for test specimens, and two for
the control specimens. The ratio of the computed standard
deviation to mean of CLn (see Eqn (‘4)) was in the range of
0.2 to 1.5% for all the products tested. (These re-
producibility values were obtained in one laboratory by one
operator using specimens prepared from one batch.) Thus the
method shows promise in terms of reproducibility. However,
the effect of other sources of variation (eg. inter-labora-
tory and batch-to-batch) is still to be established.
Isolation of wet/dry cycling effect
As previously mentioned, weight loss could be caused by
factors other than wet/dry cycling. Figure 1 shows the “un-
corrected” cummulative weight loss of the control (CL , see
Eqn. (2)) and that of the test specimen (Clx, see Eqn.(3)).
The additional weight loss, CL 11 , could be readily discerned,
which could be attributed to wet/dry cycling.

-------
Applicability for process comparison
Figure 2 shows a comparison of the three binder systems
tested. Large differences in wet/dry resistance were
obtained. The corrected cummulative weight losses for pro-
ducts PA, PB and PC were 2.5%, 1.5% after twelve cycles and
28% after three cycles, respectively.
Figures 3 and 4 show the effect of coating on products PA
and PC. Improvements in wet/dry resistance were obtained in
both cases. However, larger benefits were obtained for the
lower resistance product PC. It suggests therefore that the
degree of improvement is dependent on the product itself, as
well as on the type and thickness of coating.
These results have been presented to show the potential of
the method for ranking or comparing processes on the basis
of wet/dry resistance.
Comparison with freeze/thaw resistance
A similar method ha been developed for quantifying
freeze/thaw resistance.’ ) Comparison of these two types of
weathering resistance is shown in Figure 5 for products PB
and PC. To both types of weathering, PB had the higher
resistance. However, the difference was more pronounced in
PC than in PB. Therefore, it is not presently clear whether
process comparison results based on one type of weathering
would automatically apply to the other.
CONCLUDING REMARKS
An overview has been given of the rationale, utilization and
limitations of the methods for evaluating solidified wastes
that are currently being assessed in a co-operative study
involving US and Canadian agencies and industry. A more
detailed description is given of a method that has been
developed for quantifying resistance to wet/dry cycling.
The preliminary results thus far obtained show that (1) the
method is reproducible and can be used for process com-
parison purposes, (2) coating is a promising means of im-
proving resistance to wet/dry cycling, and (3) wet/dry and
freeze/thaw tests may or may not produce similar results.
ACKNOWLEDGEMENTS
The authors gratefully acknowledge the support of Dr.
Malcolm Wilson (Alberta Environmental Centre) and Mr. Trevor
Bridle (Environment Canada, EPS).
The project managers of this project are Nancy Cathcart and
Julia Stegemann (Environment Canada). Julia Stegemann’s

-------
contribution to refine the testing methods is also ap-
preciated. The project would not be possible without the
support of the US EPA officials Messrs. David Friedman and
Canton Wiles.
REFERENCES
1. “Investigation of Test Methods for Solidified Waste
Characterization - a Co-operative Program,” (Draft),
Environment Canada, EPS Wastewater Technology Centre
(1985).
2. “Toxicity Characteristic Leaching Procedure,” US EPA
Method 13XX (revised) (1985).
3. “Test Methods for Solidified Waste Characterization”
(Draft), Environment Canada, EPS Wastewater Technology
Centre and Alberta Environmental Centre (1985).
4. Hannak, P. and Liem, A. J., “Development of a Method
for Measuring the Freeze-Thaw Resistance of Solidified
Wastes,” in “Proceedings of the International Con-
ference on New Frontiers f Hazardous Waste Manage-
ment,” EPA 600/9-85/025 (1985).
5. “Measurement of the Leaching of Solidified Low-level
Radioactive Wastes,” American Nuclear Society, ANS-]6-1
(1981).
6. Cote, P.L. and Hamilton, D.P., “L achability Comparison
of Four Hazardous Waste Solidification Processes,”
Paper Presentation at the 38th Annual Purdue Industrial
Waste Conference, May 10 - 12, 1983.
7. ASTM D854-83 “Test Methods for Specific Gravity of
Soils,” American Society of Testing and Materials,
Annual Book of ASTM Standards; vol. 04.08(1984).
8. “Absolute Volume and True Density of Solid Wastes,”
(Second Draft), ASTM D-34-02 Task Group(1985).
9. ASTM C109-80 “Test Methods for Compressive Strength of
Hydraulic Cement Mortars,” American Society of Testing
and Materials, Annual Book of ASTM Standards; vol.
04.01 (1984).
10. ASTM D559-82 “Methods for Wetting-and-Drying Tests of
Compacted Soil-Cement Mixtures,” American Society of
Testing and Materials, Annual Book of ASTM Standards;
vol. 04.08 (1984).

-------
11. ASTM C305-82 “Method for Mechanical Mixing of Hydraulic
Cement Pastes and Mortar of Plastic Consistency,”
American Society of Testing and Materials, Annual Book
of ASTM Standards; vol. 04.01.
12. ASTM C192-81 “Method for Making and Curing Concrete
Test Specimens in the Laboratory,” American Society of
Testing and Materials, Annual Book of ASTM Standards;
vol. 04.02.

-------
TABLE I SUMMARY OF TEST METHODS ”
IESI
BRIEF DESCRIPTION
SIGNIFICANCE
1. Bulk Density
2. Specific gravity’ 7 ’
2a. True density ’
3. Moisture content
4. Permeability
5. Unconfined
compression strength”’
6. Freeze/Thaw
resistance ‘
7. Wet/Dry resistance
8. Acid neutralization
capacity
9. Equilibrium leaching
10. Toxic Characteristic
leaching Procedure ’
11. Dynamic leaching
12. Sequential extraction
Length and weight measurements
Pycnometer — water displacement
Pycnometer — gas displacement
Drying and weight measurements
Flow rate measurements under tightly
specified conditions
Load required to “destroy” specimen
Weight loss after cycling; or number
of cycles to “destroy” specimen
Weight loss after cycling; or number
of cycles to “destroy” specimen
Acid addition — pH measurements
“Extraction” with distilled water to
equilibrium conditions
“Extraction” with acidic or buffer
solution to equilibrium conditions
“Extraction with distilled water
(monolithic specimen; leachate
replacement)
“Extraction” with progressively more
aggresive solutions
Tests #1 to 3 are used to compute porosity, which could be related to
permeability. (Bulk density could also be used to estimate landfill
size requirement or life—time.)
Conventive transport of contaminants
Tests #5 to 7 indicate ability to maintain physical integrity.
Disintegration could lead to formation of cracks and particulates.
hence increased permeability and liquid — solid contact area.
Tests # B to 10 indicate potential dissolution of contaminants. High
buffering capacity could reduce dissolution of heavy metals in acidic
leaching agents.
Mobility of contaminants could be related to rate of release.
Indication of bonds with which contaminants are contained.

-------
TABLE II Preliminary Test Results
1.
2.
2a.
3.
4.
5.
TEST
Bulk Density
Specific Gravity
True density
Moisture content
Cc)
Permeability
Unconfined compression
strength
6. Freeze/Thaw
resi stance
7. Wet/Dry resistance
8. Acid neutra1ization
capacity
9. Equilibrium Leaching
As ( e)
Cd
Cr
Pb
(e)
Phenol
10. Toxic Characteristic
Leaching Procedure
11. Dynamic Leaching
12. Sequential Extraction
RANGE OF VALUES
1.4 — 1.6 g/cm 3
2.7 — 3.2 g/cni 3
2.5 — 2.6 g/cm 3
39—47 %wt.
—6
4x10 cm/s
2000 — 4300 psi
(13 — 30) (1O kPa)
50% weight loss after
5 — 12 freeze/thaw cycles
(see Section 5)
4 meq/g dry waste results in one
pH—unit decrease from Initial pH of
13 and 12
0.08 — 0.10 ppm
0.001 — 0.06 ppm
0.01 — 20 ppm
1.5 — 2.7 ppm
690 — 870 ppm
(not performed, the
EP test was used)
(modifications to procedure
(f)
made)
(modifications to procedure
made)

-------
NOTES:
(a) Artificial waste contained 0.04 M of As, Cd, Cr, Pb and Phenol;
Industrial Waste contained Cd, Cr and Pb at 170, 670 and 9200 ppm wt.,
respectively; additive to raw waste ratios: artificial waste (1.0), cement
(0.35), bentonite (0.60); industrial waste (1.0), soluble silicate (1.2),
cement (1.3), water (1.8)
(b) Lower values for true density, measured by air pycnometer, show matrix
dissolution.
(c) Only measured for one specimen and only in one laboratory; gas evolution
observed during measurements; material for permeability cell changed to
stainless steel.
(d) Modifications made to the procedure, Including varying the quantity of
acid to be added to cover a wider range of pH change (previously quantity of
acid was preset).
(e) Presence only in artificial waste.
(f) Modifications included removal of thimble (suspected of contributing to
diffusive resistance).
(g) Modifications included increase in reagent—solid ratio to ensure complete
extraction.

-------
TABLE III SUMMARY OF MATERIALS USED
( 1) Composition of waste/binder systems, % wt .
Product
Code
Mater i a 1
PA
PB
PC
Baghouse dust*
32
—
—
Water
20
30
46
Cement
41
—
21
Soluble silicate
7
—
—
Lime
—
16
—
Flyash
—
54
—
Bentonite
—
—
33
*lndustrjal Waste
( 2) Coating
(1) Type: — Urethane concrete floor coating
— Marine paint
(Ii) Application rate, mg dry paint/cm 2 surface area
— 1.2 to 2.4
— 3.1 to 5.8

-------
4
3
2
1
o 1 2 3 4 5 6 7 8 9 10 11 12
WET/DRY CYCLE
FIGURE 1. Cumulative weight loss of test and control specimens for unpainted product PA,
showing the contribution of wet/dry cycling (“corrected” weight loss).
C,)
0
-J
I-
I
IL l
U I
>
I-
-J
C)
LU
I-
C)
LU
0
C)
z
0_._I
ri
“— ‘-I
“CORRECTED” WEIGHT LOSS
• • .0 • — — —
SPEC lME
0 . —o •
CONTROL
I I I
I
I I I I
I

-------
a
C l)
Cl)
0
30-
0
C,
/
25- /
W / LEGEND FOR PRODUCTS (TABLE III)
20- /
/ 0-PA
/ 0-PB
15-
o /
10-
0 .1
Ui •1
5—.
•1
o :, ___ _
0 — 0 0 0 — D 0 — — 0 .I
a 0 P _ P
- 01 234 5 67 89101112
C)
WET/DRY CYCLE
FIGURE 2. Wet/dry resistance of three uncoated products, showing marked differences
due to composition.

-------
WET/DRY CYCLE
FIGURE 3.
Wet/dry resistance of uncoated and coated product PA, showing marginal
improvement resulting from coating.
U
(1)
0
. .1
I-
I
0
LU
LU
>
I- .
4
-J
C)
a
LU
I-
U
LU
0
C)
S
C
. 1
U
001 2 34 5 6 7 891011
12

-------
01 2
WET/DRY CYCLE
FIGURE 4.
Wet/dry resistance of uncoated and coated product PC, showing marked
improvement resulting from coating and the effect of type and thickness
of coat.
v i )
(I)
0
-I
I-
I
0
LU
LU
F-
4
-j
D
C)
LU
I-
C.,
LU
0
C-)
a
C
-J
0
40
30
20
10
0
34 5 6 7 8 9101112

-------
S
40
30 ___
20
10
00
WET/DRY OR FREEZE/THAW CYCLE
FIGURE 5. Comparison between wet/dry and freeze/thaw resistance of products
PB and PC, showing dependence of correspondence on the product.
PRODUCT PC
U
C l)
0
1
I-
I
C,
LU
L i i
>
I-
-J
0
0
L i i
I-
C-)
LU
0
C-)
C
-J
0
0
S
S
S
S
S
S
S
I
WEATHERING
I
S
S
I
WET/DRY
FREEZE/THAW
I
I
I
I
I
•
PRODUCT PB
1 2 3 4 5 67 891011
12

-------
FOURTH SESSION
Thursday, July 17, 1986
8:00 a.m. - 12:00 p.m.
Chairperson:
Duane Geuder
Chemist
Office of Emergency Response
and Remediation
USE PA
401 “M” Street, S.W.
Washington, D.C. 20460

-------
DYNAMIC VALIDATION OF SUPERFUND/RCRA ANALYTICAL METHODS
J. G. PEARSON AND L. R. WILLIAMS, TOXICS AND HAZARDOUS WASTE
OPERATIONS BRANCH, QUALITY ASSURANCE DIVISION, ENVIRONMENTAL
MONITORING SYSTEMS LABORATORY, U.S. ENVIRONMENTAL PROTECTION
AGENCY, LAS VEGAS, NEVADA
ABSTRACT
The fundamental goals of the method selection and validation
process are (1) to select a method of measurement that is
capable of producing measurements of the type and quality
needed for a particular application, and (2) to verify that
the method selected will reliably produce adequate measure-
ments during its use. The major steps in this process will
be briefly described. These six steps are: (1) de’ermina-
tion of method requirements, (2) method selection/develop-
ment, (3) single-laboratory testing, (4) confirmatory test-
ing, (5) final method description, and (6) formal
collaborative study or alternative validation procedure.
This paper will focus on one alternative validation
procedure, namely the process of dynamic validation. The
strengths and weaknesses of this procedure will be discussed
and an example of how the process has been used in the
Superfund Contract Laboratory Program will be presented.
The example used will be the dynamic validation of the GC/MS
method for volatile organic compounds in water and soil.

-------
THE SUFFICIENCY, REDUNDANCY, AND APPROPRIATENESS
OF SURROGATE AND MATRIX SPIKE COMPOUNDS IN ORGANIC ANALYSIS
F. C. GARNER, M. T. HOMSHER, LOCKHEED ENGINEERING AND
MANAGEMENT SERVICES CO, LAS VEGAS, NEVADA; J. G. PEARSON,
TOXIC AND HAZARDOUS WASTE OPERATIONS BRANCH, QUALITY
ASSURANCE DIVISION, ENVIRONMENTAL MONITORING SYSTEMS
LABORATORY, U.S. ENVIRONMENTAL PROTECTION AGENCY, LAS
VEGAS, NEVADA
ABSTRACT
This paper investigates the mathematici relationship between
the surrogate analyses and the matrix spike analyses used in
the organic analytical method of the United States Environ-
mental Protection Agency Contract Laboratory Program.
Surrogate spike componds are added to each sample and
analyzed in order to observe method performance for each
sample. Matrix spike compounds are added to one sample out
of every twenty samples, and are analyzed to observe method
performance for each amtrix spike compound for each sample
batch. Surrogate spike compounds which do not correlate
with any matrix spike compounds are inappropriate.
Surrogate spike compounds which are highly correlated with
other surrogate spike compounds are redundant because they
contain the same information. Matrix spike compounds for
which there are no correlated surrogate spike compounds
reflect an insufficiency in the list of surrogate spike com-
pounds, and the need to use new 5urrogates. This paper
investigates the organic analytical method for the
sufficiency, redundancy, and appropriateness of surrogate
and matrix spike compounds in soil analysis, and suggests
which compounds should be removed from or added to the
surrogate and matrix spike compound lists.

-------
STATISTICAL APPROACH TO MIJLTI-ANALYTE DATA QUALITY
IVAN 1. SHOW AND LLEWELLYAN R. WILLIAMS, S-CUBED, LA JOLLE,
CALIFORNIA
ABSTRACT
An EPA-sponsored work group meeting will address the
statistical analysis of multianlayte chemical data. The
meeting is motivated by the need for clear recommendations
on the “best” and most useful statistical methods applicable
to environmental and human health data. To reach a
concensus that is both statistically valid and useful for
policy and management decisions, the meeting brings together
a panel of experts with a unique combination of experience
and perspective; the panel consists of policy makers and
managers, chemists, and statisticians. The panel will
suggest methods and define needs and applications prior to
the meeting. The statistical bases for suggestions are
that: (l)constituents of multianalyte chemical samples are
generally not independent, and (2) univariate statistical
methods are not appropraite due to the lack of independence.
From the management perspective, any method considered must
be comprehensive and must address significant issues.
Suggested methods will be evaluted during the meeting;
following the meeting, analyses will be performed using EPA
(RCRA, Superfund, TSCA) and other data sources. Finally,
the panel will evaluate the analyses for validity and
usefulness and the results will be consolidated into one or
more recommended procedures for application to the EPA
Testing and Monitoring Program.

-------
ANALYSES FOR CHLORINATED DIBENZO-P-DIOXINS AND DIBENZOFURANS
IN ENVIRONMENTAL SAMPLES AND EMISSIONS FROM COMBUSTION
AND INCINERATION PROCESSES
ROBERT HARLESS, ENVIRONMENTAL MONITORING SYSTEMS LABORATORY,
U.S. ENVIRONMENTAL PROTECTION AGENCY, RESEARCH TRIANGLE PARK,
NORTH CAROLINA; AUBRY DUPUY, JR., DANNY MC DANIEL,
ENVIRONMENTAL CHEMISTY LABORATORY, U.S. ENVIRONMENTAL
PROTECTION AGENCY, NASA/NSTL, BAY ST. LOUIS, MISSISSIPPI
ABSTRACT
Extraction, clean-up and high resolution gas chromatography-
high resolution mass spectrometry (HRGC-HRMS) methods of
analysis used in the U.S. Environmental Protection Agency
National Dioxin Study for determination of 2378-TCDD and other
tetra- through octa-chiorinated dibenzo- -dioxins (CDDs) and
dibenzo-furans (CDFs) are described. Many of these compounds
were found in matrices such as soil, sediment, fish, water,
fly ash, and stack gas emissions at various concentration
levels. Minimum limits of detection were in the range of 1 to
S ppt for 2378-TCDD in fly ash and environmental samples and
50-200 pg/sample train for stack gas emissions. Quality
assurance procedures, analytical criteria used for
confirmation of these compounds, QA requirements for
analytical data and some typical results are described.

-------
SAMPLE AND ANALYSIS FOR DELISTING DATA
VERIFICATION/DELISTING SPOT CHECKS
MYLES MORSE, AND JOHN WARREN, U.S. ENVIRONMENTAL PROTECTION
AGENCY; AND WILLIAM SPROAT, TECHNICAL RESOURCES, INC.
ABSTRACT
As part of the Delisting Spot Check Program, the Office of
Solid Waste has found that 74% of all facilities sampled
under this program have failed to adequately characterize
their waste through representative sampling. Currently, a
minimum of four representative samples are required to
characterize the hazardous or non—hazardous nature of a
waste. However, the use of only four samples precludes any
valid statistical assessment of the waste. This is
especially true when the waste is disposed in a surface
impoundment or lagoon and samples are collected only from
the perimeter of the impoundment. As an alternative, the
use of a nonparametric tolerance limit is suggested for
determining the number of samples necessary to adequately
characterize a hazardous waste. In particular, this
methodology is applied to the sampling of surface
impoundments containing hazardous wastes.

-------
QUALITY ASSURANCE IN THE GROUND WATER MONITORING
TASK FORCE FACILITY ASSESSMENT PROGRAM
MICHAEL 3. KANGAS, JO ANN DUCHENE, AND TIMOTHY E. TYBURSKI,
ICAIR LIFE SYSTEMS, INC., CLEVELAND, OHIO; AND PAUL H.
FRIEDMAN, OFFICE OF SOLID WASTE, U.S. ENVIRONMENTAL
PROTECTION AGENCY, WASHINGTON, DC
ABSTRACT
The U.S. Environmental Protection Agency (USEPA) Hazardous
Waste Ground Water Task Force (HWGWTF) is conducting ground-
water sampling at sites where 58 commercial land disposal
facilities are located under its Facility Assessment Program
to determine whether or not they are contaminating ground
water with hazardous wastes. Laboratory analyses of ground-
water samples from these sites are being performed for
organic compounds, metals and inorganic and indicator para-
meters. This paper provides an overview of initial quality
assurance (QA) activities applied to these analyses.
A Data Evaluation Committee (DEC) consisting of HWGWTF mem-
bers from USEPA headquarters, regional offices and Office of
Research and Development CORD) laboratories provides QA
overview for Facility Assessment Program activities. The
DEC developed a Quality Assurance Project Plan (QAPP) speci-
fically for the Facility Assessment Program. The QAPP esta-
blishes data quality objectives (DQO) which define the
quality of the data desired from the ground-water sampling
and analytical efforts. These DQO address data accuracy,
precision, representativeness, completeness and
comparability. The QAPP also defines roles and responsibil-
ities of HGWWTF members and contractors and procedures to be
followed in achieving the DQO.
This paper provides an overview of QA activities performed
by the HWGWTF DEC and technical support contractors during
evaluation of the first six facilities sampled under the
program (Phase I) and provides examples of improvements
planned for subsequent QA efforts based on experience gained
during Phase I. Evaluation of analytical data and labora-
tory performance through integration of information from
performance evalutation samples, analytical chemistry
audits, statistical summaries and control charts of labora-
tory quality control data generated by Superfund Contract
Laboratories is emphasized.
INTRODUCTION
In 1985, the U.S. Environmental Protection Agency (USEPA)
Administrator established the Hazardous Waste Ground Water
Force (HWGWTF) to evaluate the level of compliance with
applicable hazardous waste regulations at 58 existing
commercial hazardous waste land disposal facilities.

-------
A Facility Assessment Group (Figure 1-1) was formed within
the HWGWTF to evaluate potential ground-water contamination,
assess compliance status and take actions necessary to
resolve identified problems and areas of noncompliance at
the 58 facilities (USEPA 1985a).
This paper focuses on one element of the Group’s activities,
providing quality assurance (QA) for laboratory analyses of
ground-water samples collected under Phase I of the UWGWTF
Program.
LABORATORY QUALITY CONTROL FOR PHASE I MONITORING
The Phase I HWGWTF Program involved evaluation of six of the
58 facilities. The remaining 52 facilities will be
evaluated during subsequent phases of the Program. Phase I
was a “shakedown” effort to identify and correct initial
Program problems. Phase I involved analysis of ground-water
samples for approximately 150-200 compounds (organic, metal,
inorganic and indicator parameters and other tentatively
identified compounds). The number of compounds will be
expanded in subsequent phases to include more of the com-
pounds listed in Appendix VIII of 40 CFR Part 261.
Phase I laboratory analyses were performed using Invitation
For Bid (IFB) Regular Analytical Service protocols for
metals, cyanide and organic parameters from USEPA’s Contract
Laboratory Program (CLP). Protocols specified in supplemen-
tal Special Analytical Service (SAS) contracts were used for
inorganic and indicator parameter analyses. The list of
analytes will be expanded during subsequent phases through
use of SAS contracts.
The laboratory quality control (QC) requirments applied
during Phase I are summarized in Table 2-1 for metal,
inorganic and organic analyses.
PHASE I QUALITY ASSURANCE ACTIVITIES
The general and specific QA/QC requirements which CLP
laboratories are required to follow are defined in the Scope
of Work of their IFB (EPA 1985 b,c). This section describes
the supplemental QA/QC procedures developed and applied by
the I-IWGWTF Data Evaluation Committee (DEC) specifically for
evaluation of laboratory performance and quality of data
generated under the Facility Assessment Program.
QUALITY ASSURANCE PROJECT PLAN AND HWGWTF DATA EVALUATION
COMMITTEE
Current USEPA policy requires that every monitoring and
measurement project have a written and approved QA Project
Plan (USEPA 1983). A QA Project Plan addresses 16 required
elements (Table 3-1). The HWGWTF DEC is primarily respon-

-------
FIGURE 1—1 PHASE I HAZARDOUS WASTE GROUND WATER TASK FORCE FACILITY
ASSESSMENT GROUP ORGANIZATIONAL AND FUNCTIONAL RELATIONSHIPS
USEPA Administrator
Lee Thomas
I
Hazardous Waste Ground
Water
Task Force
Facility
Assessment Group
State Teams
• Team Leader
• Field Inspectors
• Permit Writers
• Enforcement Staff
• State Counsel
• Sampling and Analytical
Personnel
• Communications Expert
I I I I
National Enforcement
Investigations Center (NEIC)
• On-Site Coordinators
• Field lnspectorsfSamplers
• Laboratory Auditors
Headquarters Core Team
• Director
• Deputy Director
• Communications Advisor
• Technical Coordinator
• Document Control Officer
• Met hods/QA Support
• Field Assessment Team
Regional Teams
• Regional Team Leaders
• On-Site Coordinator
• Document Control Officer
• Field Inspectors
• Permit Writers
• Enforcement Staff
• Regional Counsel
• Sampling and Analytical
Personnel
• Communications Expert
On-Site
Facility Data
Site Inspection
Evaluations
Procurement
Samp g_
EMSL•LV
PRC
PRCNersar
I I I I I : I
SamplelData
Routing and
Administration
Sample Mgrnt
Office (VIAR)
Sample Analysis
• California Analytical,
Organics
• Rocky Mountain
Laboratory, Inorganucs
Performance
Data Evaluation Evaluation Studies Technical Advisory
Life Systems. Inc EMSL-CIN Panel
EMSL-LV

-------
TABLE 2-1 MAJOR CLP QUALITY CONTROL ELEMENTS FOR ANALYSIS OF
METAL, INORGANIC AND ORGANIC PARAMETERS
Parameter Type Quality Control Element
Metal and . Initial Calibration and Calibration Verification
Inorganics
• Continuing Calibration Verification
• Preparation Blank Analysis
• Interference Check Sample Analysis
• Matrix Spike Analysis
• Duplicate Sample Analysis
• Furnace Atomic Absorption (AA) Quality Control (QC)
Analysis for Metals (Method of Standard Additions required
under certain conditions)
• Laboratory QC Sample Analysis
Organic • Documentation of Gas Chromatography/Mass Spectrometry
(GC/MS) Mass Calibration and Abundance Pattern
• Documentation of GC/MS Response Factor Stability
• Internal Standard Response and Retention Time Monitoring
• Reagent Blank Analysis
• Surrogate Spike Response Monitoring
• Matrix Spike and Matrix Spike Duplicate Analysis
• Specific Quality Assurance (QA)/QC for Pesticide Analysis

-------
TABLE 3—I REQUIRED ELEMENTS OF A QA PROJECT PLAN(a)
Title Page with provisions for approval signatures
Table of Contents
Project Description
Project Organization and Responsibility
QA Objectives for measurement data in terms of precision, accuracy,
completeness, representativeness and comparability
Sampling Procedures
Sampling Custody
Calibration Procedures and Frequency
Analytical Procedures
Data Reduction, Validation and Reporting
Internal Quality Control Checks and Frequency
Performance and System Audits and Frequency
Preventive Maintenance Procedures and Schedules
Specific Routine Procedures to be used to assess data precision,
accuracy and completeness of specific measurement parameters
involved
Corrective Action
Quality Assurance Reports to Management
(a) Source: USEPA 1983.

-------
sible for developing and implementing the QA Project Plan
for the HWGWTF Facility Assessment Program. Members of the
DEC include representatives from USEPA Headquarters, the
Office of Research and Development (ORD) Environmental
Monitoring Systems Laboratory - Las Vegas (EMSL-LV), the ORD
Environmental Monitoring and Support Laboratory - Cincinnati
(EMSL-CIN) and a representative from each USEPA regional
office. Several contractors provide support and participate
in DEC activities. Major QA activities of the DEC are high-
lighted in the following sections.
ANALYSIS OF PERFORMANCE EVALUATION SAMPLES
Performance Evaluation (PB) samples produced and distributed
by EMSL-CIN were analyzed along with field samples and
blanks from each Phase I site. Results of PE sample
analyses from the CLP laboratories supporting the Phase I
effort were evaluated by EMSL-CIN against results from two
referee laboratories and EMSL-LV. A report summarizing CLP
laboratory performance was prepared and distributed to DEC
members and support contractors.
LABORATORY QUALITY CONTROL EVALUATION
The DEC received three separate analyses of evaluating
laboratory QC performance:
o A data audit performed by EMSL-LV.
o A Contract Compliance Screen performed by the
USEPA CLP Sample Management Office (SMO)
o A contractor-prepared analysis of laboratory QC
data on a case and program basis
The following sections summarize briefly each of these
analyses.
THE EMSL-LV DATA AUDIT
The Phase I EMSL-LV data audit focused on evaluation
of: (1) identification of organic compounds by GC/MS; (2)
inductively coupled plasma, flame AA and furnace AA analyses
of metal parameters; and (3) SAS protocols for inorganic and
indicator parameters. Results were reported to the DEC ver-
bally during regularly scheduled teleconferences.
THE SMO CONTRACT COMPLIANCE SCREEN
The SMO routinely assesses deliverable completeness and
technical compliance with contract requirements to
facilitate determination of payment recommendations and to
identify problems in laboratory compliance. Copies of
screening worksheets and summary forms are sent to each
laboratory screened to resolve identified problems.

-------
Copies of the worksheets, summary forms and additional
laboratory subinittals were also distributed to the EMSL-LV
and the HWGWTF support contractor responsible for laboratory
QC data analysis.
CASE AND PROGRAM LABORATORY QC DATA ANALYSIS
Following guidance provided in the QA Project Plan, the
HWGWTF support contractor (ICAIR, Life Systems, Inc.)
prepared an analysis of laboratory QC data generated for
each facility/site (case) evaluted and a summary of accuracy
and precision performance for all analyses performed under
the Phase I Program. Performance was analyzed relative to
Data Quality Objectives (DQO) for:
o Accuracy
o Precision
o Representativeness
o Completeness
o Comparability
The Program DQO were established and documented in the QA
Project Plan before Phase I sampling was initiated. Program
DQO are expressed as limits (average values) defining
acceptable levels of performance on a case and program
basis. Control limtis established from historical CLP data
generally served as DQO for this program.
Tables 3-2 through 3-6 summarize Phase I DQO for accuracy
and precision. Evaluation of case and Program accuracy,
precision and completeness illustrate use of DQO.
ACCURACY PERFORMANCE
Accuracy of analytic methods is expressed as percent
recovery of spiked compounds, both analytes (inorganic,
indicator and organic parameters) and surrogates (organic
analyses only) and known analyte concentrations in
laboratory control samples (LCS) (inorganic and indicator
parameters only). Percent recovery (%R) is determined as
follows:
Observed Concentration
(%R) Recovery = _________________________ x 100 (1)
True Concentration
Accuracy is evaluted in terms of program DQO (Tables 3-2 and
3-3) and actual performance. The actual accuracy achieved
by laboratory analyses is summarized quantitatively as an
average %R + standard deviation and displayed in tabular
form for eaTh case (site) and graphically for the Program
(Figure 3-1).

-------
TABLE 3—2 ACCURACY GOALS FOR ORGANIC SURROG T S AND
MATRIX SPIKE RECOVERY IN WATER SAMPLES’ ’
Fraction Compound Average Recovery, Z
VOA Toluene D8 88—110
VOA 4—Bromofluorobenzene 86—115
VOA 1 ,2—Dichioroethane D4 C 76—114
VOA 1,1—Dichioroethene 61—145
VCA Trichioroethene 71—120
VOA Chlorobenzene 75—130
VOA Toluene 76—125
VOtd\ Benzene 76—127
BN / Nitrobenzene D5’ ’ 35—114
BN 2_Fluorobiphe J C 43—116
BN Terphenyl D14’ 33—141
BN 1,2,4—Trichlorobenzene 39—98
BN Acenaphthene 46—118
BN 2,4—Dinitrotoluene 24—96
EN Di—n—butylphthalate 11—117
BN Pyrene 26—127
BN N—Nitroso—di—n— 41—116
propylamine
BN 1 ,4_Dichl?rlbenzene 36—97
Acid Phenol D5 C 10—94
Acid 2—Fluorophenol’ ’ ‘ 21—100
Acid 2,4,6—Tribromophenol’ / 10—123
Acid Pentachlorophenol 9—103
Acid Phenol 12—89
Acid 2—Chiorophenol 27—123
Acid 4—Chloro—3—inethylphenol 23—97
Acid . . 4—Nitrophenol 10—80
Pest. e, Dibutylchlorendate C 24—154
Pest. Lindane 56—123
Pest. Heptachlor 40—131
Pest. Aidrin 40—120
Pest. Dieldrin 52—126
Pest. Endrin 56—121
Pest. 4,4’—DDT 38—127
(a) Source: ICAIR, Life Systems 1985.
(b) Volatile organics.
(c) Surrogate compound.
(d) Base/neutrals.
(e) Pesticides.

-------
TABLE 3—3 ACCURACY COALS FOR INORGANIC AND INDICATOR ANALYSES(a)
Average
Method or Parameter Recovery, %
Atomic Absorption method 90—110
Inductively Coupled Plasma method 90—110
Mercury by Cold Vapor 90—110
Cyanide 90—110
Sulfate 80—120
Total Organic Halide (TOX) 80—120
Purgeable Organic Halide (POX) 80—120
Total Organic Carbon (TOC) 80—120
Purgeable Organic Carbon (POC) 80—120
Chloride 90—110
Nitrate 90—1 10
Ammonia Nitrogen 90—110
Total Phenol 80—120
(a) Source: ICAIR, Life Systems 1985.

-------
130 -
120 - -UCL
110 - -UWL
0
0
100- 0
a
0 00
-GM —
§90-
0
0
D
80-
- LWL
a
70 - - LCL
60 -
50 —
0 2 4 6
Case
FIGURE 3-1 ACCURACY CONTROL CHART FOR TOTAL COBALT

-------
PREC IS ION PERFORMANCE
Precision of analytic methods is evaluated using results of
duplicate analyses to calculate the relative percent
difference (RPD) for each duplicate pair:
D 1 -D 2
RPD = _______________ x 100 (2)
(Di + D 2 )/2
where:
RPD = Relative Percent Difference
D 1 = First Duplicate Concentration
D 2 = Second Duplicate Concentration
Percent recovery is used in place of the duplicate concen-
tration to calculate RPD for matrix spike and surrogate
spike compounds. The RPD provides a quantitative estimate
of precision which can be evaluated against CLP control
limits, where applicable, and Task Force DQO (Tables 3-4,
3-5 and 3-6). Performance is summarized in tabular form for
the case and graphically for the Program (Figure 3-2).
COMPLETENESS PERFORMANCE
Completeness is evaluated in terms of the total number of
samples taken (from sample traffic reports) and the number
of acceptable analyses performed (from laboratory QC
reports). The number of acceptable analyses completed
divided by the number of samples taken times 100 is an index
of completeness. The DQO for data completeness based on
percent analyses completed is 90% for this program.
For each facility sampled during Phase I, a laboratory QC
Data Evaluation Report summarizing accuracy, precision and
completeness performance was prepared and distributed to the
DEC. Each report also provided:
o An analysis of laboratory control sample perfor-
mance
o A summary of reported laboratory blank contamina-
tion
o An analysis of reported versus contract require
detection limits
o A tabular summary of compounds and concentrations
found in field, PE and blank samples
o Control charts summarizing program accuracy and
precision performance to date for all parameters
analyzed (Figures 3-1 and 3-2)

-------
TABLE 3—4 PROGRAN PRECISION GOALS FOR !bJATRIX
SPIKE/MATRIX SPIKE DUPLICATE ANALYSES ’
(b)
Average RPD
Fraction Compound Limit, %
V0A 1,1—Dichloroethene 14
VOA Trichloroethene 14
VOA Chlorobenzene 13
VOA Toluene 13
VOA(d. Benzene 11
B/N’ / 1,2,4—Trichlorobenzene 28
B/N Acenaphthene 31
B/N 2,4—Dinitrotoluene 38
BIN Di—n—Butylphthalate 40
BIN Pyrene 31
B/N N—Nitroso—di—n—propylamine 38
B/N 1,4—Dichlorobenzene 28
Acid Pentachiorophenol 50
Acid Phenol 42
Acid 2—Chlorophenol 40
Acid 4—Chloro—3—methylphenol 42
Acid 4—Nitrophenol 50
Pest.’ / Lindane 15
Pest. Heptachlor 20
Pest. Aidrin 22
Pest. Dieldrin 18
Pest. Endrin 21
Pest. 4—4’DDT 27
(a) Source: ICAIR, Life Systems 1985.
(b) Relative percent difference.
(c) Volatile organics.
(d) Base/neutrals.
(e) Pesticides.

-------
TABLE 3—5 PROGRAN GOALS FOR PREç 9 ION BY
SURROGATE COMPOUND FRACTION a
Average RPD 1 ’
Fraction Limit, Z
Volatile Organics 15
Base/Neutrals 50
Acids 40
Pesticides 30
(a) Source: ICAIR, Life Systems 1985.
(b) Relative percent difference.

-------
TABLE 3-6 PROGRAM PRECISION GOA 9 OR
INORGANIC AND INDICATOR ANALYSES a,
Average RPD
Method or Parameter Limit, %
Atomic Adsorption method 30
Inductively Coupled Plasma method 30
Mercury by Cold Vapor 30
Cyanide 20
Sulfate 20
Total Organic Halide (TOX) 20
Purgeable Organic Halide (POX) 20
Total Organic Carbon (TOC) 10
Purgeable Organic Carbon (POC) 10
Chloride 10
Nitrate 40
Anmionia Nitrogen 10
Total Phenol 20
(a) Source: ICAIR, Life Systems 1985.
(b) Relative percent difference.

-------
140 -
130 -
120- 0
110 - -UCL
100 -
90-
0
80 —UWL
•L5 70-
0
60-
50-
40 -
10 -
20 -
0 0
-GM
10- 0
0 Q OQ 0 0
0-
0 2 4 6
Case
FIGURE 3—2 PRECISION CONTROL CHART FOR ENDRIN

-------
Upon receipt of all reports for a case, the DEC conducted a
teleconference to discuss their findings.
THE DEC TELECONFERENCE AND REPORT TO DATA USER
The DEC conducted a teleconference for each case to:
o Discuss laboratory performance and data quality
o Develop consensus regarding data useability for
the case
o Develop recommendations for ways to improve
laboratory performance and data quality in subse-
quent cases
Upon completion of the DEC teleconference, the DEC prepared
and distributed a report to the appropriate data user (the
USEPA National Enforcement Investigation Center or Regional
Office) summarizing limitations of and appropriate use for
laboratory data in required technical site evaluations.
FEEDBACK TO CLP LABORATORIES
Feedback to laboratories regarding problems arid potential
improvements resulting from Phase I activities included:
o Discussions between the SMO and the laboratory
regarding reconciliation of problems identified
through contract compliance screening
o Communications between the DEC and SMO, and subse-
quently between the SMO and the laboratory,
regarding deficiencies identified during the DEC
teleconference
o Direct communication between the EMSL-LV and the
laboratory regarding technical problems and issues
identified during the teleconference
Additional changes in IFB protocols and requirements result-
ing from Phase I QA activities are discussed in sections
below.
EVALUATION OF THE PHASE I EFFORT
Following completion of Phase I sampling and analysis, the
DEC met with all Phase I participants to discuss results of
QA activities and to develop recommendations for implementa-
tion during subsequent phases of the program. This group
concluded that the Phase I QA effort was generally a
success, but reached specific conclusions regarding desired
areas of improvement. Examples of the conclusions reached
during the meeting included:

-------
1. Analysis of PE sample results indicated consistent
performance by laboratories among cases. Because
of this consistency and the high cost of prepar-
ing, distributing, analyzing and evaluating PE
samples, the Phase I PE analysis frequency (every
case) was considered excessive.
2. Phase I PE samples were not prepared to match
matrix constituents found at each site.
Therefore, results of analysis of PE samples by
participating laboratories were not considered
representative of performance achieved in analysis
of actual field samples.
3. Laboratory spike and duplication analyses were
assigned by the laboratory during Phase I with no
input from HWGWTF personnel familiar with the
samples of greatest interest (those most likely
contaminated) in each case.
4. Laboratory reported detection limits reported did
not represent method detection limits.
5. Laboratory calculations for spike recoveries and
duplicates were not routinely checked for arith-
metic errors.
6. Retention factors used by the laboratories were
not routinely monitored as part of laboratory
performance evaluation.
7. The Laboratory QC Data Report format used during
Phase I did not follow the order of discussion
during the DEC teleconference.
IMPROVED QA FOR SUBSEQUENT PHASES
Recommendations for improvements developed during the DEC
Phase I evaluation meeting generally paralleled specific
conclusions regarding weaknesses in initial QA procedures.
Examples of improvements include:
1. Reduction in the frequency of PE sample analysis
and evaluation. Laboratories will be required to
analyze PE samples at the beginning of their
contract only, instead of one with each set of
samples from a site.
2. Spike and duplicate analyses will be performed on
samples from wells suspected of being
contaminated. Laboratories will perform spike and
duplicate analyses of samples selected by the
HWGWTF personnel reviewing site background infor-
mation instead of selecting these analyses accord-
ing to in-house procedures.

-------
3. At least one DEC representative will be included
in routine on-site evaluations conducted at
laboratories supporting the HWGWTF Facility
Assessment Program.
4. Expand the QA evaluation process to include field
measurements of pH and conductivity.
REFERENCES
ICAIR, Life Systems, Inc. 1985. Hazardous Waste Ground
Water Task Force Facility Assessment Program Quality
Assurance Project Plan. ICAIR, Life Systems, Inc. Project
No. 211393 (PRC Engineering Project No. 277). Life Systems,
Inc. September, 1985, Cleveland, OH.
ICAIR, Life Systems, Inc. 1986. Laboratory Quality Control
Data Evaluation Report for GSX Services of South Carolina,
Inc. Facility, Pinewood, South Carolina. ICAIR, Life
Systems, Inc., Project No. 211393 (PRC Engineering Project
No. 277). Life Systems, Inc. January, 1986, Cleveland, OH.
USEPA. 1983. Interim Guidelines and Specifications for
Preparing Quality Assurance Project Plans. U.S. Environ-
mental Protection Agency. USEPA Office of Exploratory
Research. February, 1983. Washington, DC.
IJSEPA. 1985a. Hazardous Waste Ground Water Task Force
Facility Assessment Program Plan. U.S. Environmental Pro-
tection Agency. USEPA Hazardous Waste Ground Water Task
Force. April, 1985. Washington, DC.
USEPA. 1985b. Contract Laboratory Program Statement of
Work for Inorganics Analysis, Multi-Media Mutli-
Concentration (SOW No. 785). U.S. Environmental Protection
Agency. USEPA Contract Laboratory Program. July, 1985.
Washington, DC.
USEPA. 1985c. Contract Laboratory Program Statement of
Work for Organics Analysis, Muliti-Media-Concentration
(Attachment A). U.S. Environmental Protection Agency.
USEPA Contract Laboratory Program. July, 1985. Washington,
DC.

-------
SAMPLING AND ANALYSIS FOR DELISTING
PETITIONS/STATISTICAL BASIS FOR SAMPLING
Myles Morse: 1 William Sproot: 2 John Warren 3
Since the inception of the Hazardous Waste Delisting Program in 1980, 30
facilities that have petitioned the Environmental Protection Agency for exclusion
from the hazardous waste regulations under 40 CFR 260.22 have been visited for
sampling and data verification as part of the Delisting Spot Check Program.
These facilities have included electroplating operations, petroleum refinery
operations, steel finishing operations, and multiple waste treatment facilities.
Among the hazardous wastes generated at these facilities are wastewater
treatment sludges, petroleum refinery wastes, electric arc furnace dust, brine-muds
from chlorine production, and stabilized/solidified wastes from multiple waste
treatment facilities. Disposal methods have included drum storage with ultimate
landfill disposal, direct landfilling of sludge, storage and disposal of sludge in
surface impoundments or lagoons, drying beds and piles and in one case, using a
treated sludge/asphalt mixture to pave a parking lot.
While the Delisting Spot Check Program has produced much valuable
information, the Office of Solid Waste has found that a majority of the facilities
sampled have failed to adequately characterize their waste through representative
sampling and analysis. Characterization of the hazardous nature of the waste
involves collection and analysis of an adequate number of representative samples
from the disposal site. Currently, the regulations require that an adequate
number of representative samples (but never less than four samples) be collected
to characterize the hazardous or non-hazardous nature of a waste. The use of
only four samples, however, precludes any valid statistical assessment of the
waste. This is especially true when the waste is disposed in a surface
impoundment or a lagoon.
of Office of Solid Waste, U.S. Environmental Protection Agency.
2 of Technical Resources, Inc., Rockville, Maryland.
3 of Office of Policy, Planning, and Evaluation, U.S. Environmental
Protection Agency.
Disclaimer
*The views and perspectives put forth in this paper are those of the authors and
do not necessarily reflect official agency policy.

-------
Collection of representative samples at surface impoundments can be
particularly difficult for a variety of reasons. Many impoundments are quite
large, with surface areas covering 20 acres or more. In addition, the surface
impoundment usually consists of a sludge layer of varying depth with an overlying
liquid layer. Typically, the sludge layer can average 4 to 6 feet with a 6 to
10 foot Liquid layer on the surface.
Accordingly, in terms of efficiency, perimeter sampling is the method of
choice at most surface impoundments. Perimeter sampling offers a convenient
method for sample collection with a minimum of effort. It may not, however,
adequately characterize the hazardous wastes contained within the impoundment.
Hazardous wastes disposed in surface impoundments are rarely homogeneous.
These wastes demonstrate variability over both time and space. Variability over
time can occur for a variety of reasons including the following:
• Process changes
- Segregation of specialized operations
- Addition or removal of operations, (i.e., plating, etc.)
• Schedule changes
- Change from an 8-hour day to 24-hour operations
• Manufacturing changes (product change)
- Job shop operations (i.e., continuously variable product line)
- Addition of new product lines
• Raw material changes as a result of altering:
- Suppliers
- Expenditures due to variable cost
- Product line
• Management practices
- Offloading of other wastes into impoundment or disposal site

-------
• System aberrations
- Variations in treatment process (e.g., pH fluctuations)
- General system breakdown
- Failure of bacteria (in biological treatment system).
Spatial variability can occur in both a horizontal and vertical direction.
Variability in the horizontal direction can result from the use of a single
discharge point, the movement of a single discharge point, or from the use of
multiple discharge points into an impoundment. In all cases, an uneven horizontal
distribution of waste or sludge can occur, with the greatest concentration of
sludge occurring near the discharge point. This is a direct result of the heavier
sludge particles settling out of the liquid phase first. In addition, manmade
obstructions such as weirs or barriers within the impoundment can allow uneven
horizontal distribution to occur. Sloped sides can allow greater sludge
accumulation in the center of the impoundment. The use of natural structures,
such as quarries, also allows for uneven sludge accumulation on the bottom of the
impoundment, depending on the topography of the impoundment site. These
horizontal variations will not be adequately represented by perimeter sampling.
Variability in the vertical direction is affected by changes in the waste over
time as well as the accumulation of the heaviest particles of waste near the
discharge point. These two factors can produce layers of waste within the
impoundment of varying concentrations that perimeter sampling again, will not
adequately address. Accordingly, the collection of representative samples from a
surface impoundment or lagoon must be done in such a way as to ensure that
data are collected from all parts of the holding area, not just from the perimeter.
To adequately characterize the waste, a sufficient number of representative
samples must be collected to demonstrate the hazardous or non-hazardous nature
of the waste. As indicated earlier, the regulations presently specify that no less
than four samples are required for this purpose. A sample size of four, however,
is totally inadequate for any meaningful statistical analysis of the waste unless
the waste is completely homogeneous. Classically, statistical distributions are
based on sample populations of 30 or more observations. Delisting petitions
typically contain less than 20 data points and in many cases less than 10.

-------
The current policy of the Office of Solid Waste regarding the number of
samples that are required to characterize a hazardous waste site is based on the
size of the disposal area. According to Petitions to Delis: Hazardous Wastes: A
Guidance Manual (EPA/530-SW-85-003, April 1985), the disposal area should be
divided into four equal quadrants, each quadrant not exceeding 10,000 square feet
in size. For disposal areas larger than 40,000 square feet, the total area should
be divided into equal segments, with each segment not exceeding 10,000 square
feet. For disposal areas less than 10,000 square feet, the total area should be
divided into four equal quadrants.
Within each quadrant or segment of equal size, five random complete-depth
core samples are collected and combined into one composite sample. For disposal
areas up to 40,000 square feet, a minimum of four composite samples would be
collected. For areas larger than 40,000 square feet, the number of composite
samples would be equal to the number of segments within the disposal area.
(These areas are based on the average size of typical surface impoundments or
landfills.)
Evaluation of the data for delisting purposes is based on the use of the
maximum reported EP leachate value for each of the inorganic constituents of
concern as an input parameter for the Vertical and Horizontal Spread (VHS)
model. 4 This model is used to predict the concentration of a hazardous
constituent at a compliance point (in this case, a drinking water well) 500 feet
from a waste disposal site (see 50 f . 7882, February 26, 1985 and 50 EE 48896,
November 27, 1985). If’ the maximum reported EP leachate value generates a
compliance point concentration that exceeds the health-based standard for the
appropriate constituent of concern, then delisting is not recommended.
The maximum reported EP leachate value is used for several reasons. First,
because of the small sample size (n = 4), rarely can meaningful statistical analyses
‘ Depending on the characterization of the waste and the constituents
being evaluated, other test results may be used as input parameters to
the VHS model, i.e., (the use of the EP Toxicity Test for Oily Wastes
where total oil and grease content exceeds one percent or the use of
total constituent analysis results for organics with subsequent input into
the General Linear Model (see November 27, 1985, 50 . 48886)).

-------
be performed. Further, since the waste is usually not homogeneous, the use of
the mean value instead of the maximum value would tend to obscure any large
variations in constituent concentration in the waste. Without a larger sample
size, 5 the distribution of the population cannot be determined with any
confidence. Second, use of the maximum EP leachate value represents a
reasonable worst-case situation and allows for some margin of error in the
assessment. Third, the use of composite samples formed from individual core
samples also tends to average the constituent concentrations present in the waste
and obscure any variation in concentration where small data sets are being
evaluated.
As an alternative to the maximum reported EP leachate value, we suggest, on
a case-by-case basis, the use of the sample mean (or median) as an input
parameter to the VHS model. However, using the sample mean involves making a
determination as to the number of samples necessary to adequately characterize
the waste, and whether the data elements are normally distributed. Therefore, in
order to determine if some value other than the maximum should be used to
assess the data elements, we need to determine how many samples are needed,
and how this population is distributed.
Since the wastes that are being sampled are usually heterogeneous mixtures,
samples collected from the waste disposal area may not be normally distributed.
To assess data from non-normal populations, statistical analysis must be conducted
using tests that do not rely on normality assumptions. These tests are called
distribution-free or nonparametric since they can be applied to populations with
almost any distribution and are not confined to the parameters of a population.
Accordingly, nonparanletric procedures can be used to determine the number
of samples that need to be collected from a waste site. Through the use of One-
Sided Nonparametric Tolerance Limits, the number of samples necessary to
establish a degree of certainty for a percentage of all values in the sample size
can be established (Table I). For example, in the case of a 40.000-square-foot
surface impoundment, to be 90% certain that 85% of the total population is less
For our purposes, the sample size refers to the total number of data
elements (composite or single core samples) collected from an
impoundment.

-------
than the largest sample value obtained, a total of 15 random complete-depth core
samples would have to be collected.
Table 1. Sample Sizes for One—Sided Nonparametric Tolerance Limits
Percent of All Values
- n q .500 .700 .750 .800 .850 .900 .950 975 .980 990
500 1 2 3 4 5 7 14 28 35 69
.700 2 4 5 6 8 12 24 48 60 120
.750 2 4 5 7 9 14 28 55 69 138
.800 3 5 6 8 10 16 32 64 80 161
.850 3 6 7 9 12 19 37 75 94 189
.900 4 7 9 11 15 22 45 91 144 230
0 .950 5 9 11 14 19 29 59 119 149 299
.975 6 11 13 17 23 36 72 146 183 368
.980 6 11 14 18 25 38 77 155 194 390
.990 7 13 17 21 29 44 90 182 228 459
.995 8 15 19 24 33 51 104 210 263 528
.999 10 20 25 31 43 66 135 273 342 688
If the sample mean is to be used as an input parameter to the VHS model,
we feel that the sample size should consist of 45 random complete-depth core
samples collected from the waste site. This value, obtained from Table 1,
provides a 90% certainty that at least 95% of the total population is less than the
largest value obtained. We feel that the 95% value is sufficient to adequately
characterize the waste population. In addition, the sample size is sufficient to
show that significant variations in constituent concentration should not be
obscured by either incorrect sampling procedures or poor analytical technique.
The use of 45 samples to characterize the waste population is based on a
surface impoundment that is 40,000 square feet or less in surface area. When
larger impoundments are involved, an additional 12 samples would be required for
each 10,000-square-foot segment. Very large impoundments should be dealt with
on a case-by-case basis because of the large number of samples that could be
required.
After determining the number of samples necessary to adequately
characterize the nature of the waste, the data can be evaluated to determine if it
is normally distributed (Figure 1). Standard statistical procedures such as the

-------
Goodness of Fit tests can be performed, although these tests require very large
sample sizes (i.e., where a — 100). As an alternative, Lillefor’s test may be used
on smaller sample sizes to assess probable normality.
If the data are normally distributed, 6 then an appropriate test can be used
to determine if any reported data values are true outliers. One test procedure
available for the determination of outliers is Dixon’s Test, also called the Extreme
Value Test. 7 If true outliers are identified by this procedure, their values will
not be used in the determination of the sample maximum or mean.
If a sample of sufficient size is determined to be normally distributed, then
the sample mean could be used as an input parameter in the VHS model to
determine the compliance point concentration of the constituents of concern. 8 It
should be noted that the sample mean or the sample median could also be used
for certain non-normally distributed populations. For example, a bi-normal
distribution (Figure 2) or a platykurtic (flattened) distribution (Figure 3) could
justify the use of a population mean. Sample populations that are skewed to the
6 Distributions such as a binormal distribution could also justify the use of
the mean and will be determined on a case-by-case basis.
Other mechanisms for determining true outliers could be used.
8 Other factors may preclude the use of the mean (e.g., historical
information, sampling location, management practices, etc.)
Figure 1. Normal distribution

-------
left or right (Figure 4) could justify the use of the sample median. The
determination of the appropriate statistical parameter for non-normally distributed
samples will be made on an individual basis.
Given the increased validity of these procedures over the use of only four
composite samples, a quantitative assessment of the hazardous or non-hazardous
nature of the waste, its variability, and the likelihood of using a representative
value such as the mean as an input value to the VHS model, can be made. 9
Using the same area requirements as described in the Petitions so Delis:
Hazardous Wastes: A Guidance Manual, characterization of a 40,000-square-foot
disposal area would require the collection of 45 random complete-depth core
samples to achieve a 90% certainty that at least 95% of all values at the site are
less than the sample maximum. Since the sampling procedure described in the
guidance manual only required 20 samples to be composited into 4, the use of this
suggested procedure will more than double the sampling effort required. In
addition, if this procedure is followed, analytical costs will probably increase
significantly. However, adequate characterization of hazardous wastes is the
largest single problem facing the regulated community as well as the Delisting
Program staff. The use of the statistical procedures described in this paper is
seen as an alternative method to increase the reliability of the collected data.
Figure 2. Mixture of two unequal normal distributions
9 See Appendix 1 for examples.

-------
Figure 3. Skewed dlsfrlbutions
Skewed to the left
(negatively skewed)
Skewed to the right
(positively skewed)
Figure 4. Platykurtic distribution

-------
APPENDIX 1
The following examples each show two potential Agency conclusions
regarding the petitioner’s delisting effort. In the first example, the petitioner
initially submitted analytical results for only four samples, claimed to be
representative of the waste contained in a 40,000-square-foot impoundment with a
sludge layer 10 feet in depth. The EP toxicity data for the original four samples
were .0630, 0.0820, 0.195, and 0.378 mg/i. For the volume of waste contained
within this impoundment (14,815 cubic yards), the maximum allowable leachate
concentration for lead, predicted by the VHS model, would be 0.3 15 mg/i. The
Agency’s review involved the use of the maximum extract value for lead to be
used in the VHS evaluation, resulting in a concentration of lead at the compliance
point in excess of the National Interim Primary Drinking Water Standard
(NIPDWS).
Forty-one further samples were requested, and from the total of 45 samples
the following data characteristics were obtained:
Sample mean .170 mg/i
Sample median .195 mg/i
Sample standard deviation .015 mg/I
Sample minimum .032 mg/i
Sample maximum 0.378 mg/i
Sample second-highest maximum 0.284 mg/I
Tests for Normality gave no indications of marked departures from the
assumption of Normality. The maximum value of 0.378 mg/l when used as an
input parameter for the VHS model generates a compliance point concentration for
lead of 0.060 mg/I, exceeding the NIPDWS for lead.
Dixon’s Test gives:
Maximum - 2nd highest maximum
D= _________________
Maximum - Minimum
.378 - .284 0.094
__________ _____ = 0.272
.378 - .032 = .346

-------
and since this exceeds the tabulated value for 5% significance, it can be regarded
as an outlier, and is excluded from further consideration in the Delisting review
process. The remaining data indicates that the sample mean does not exceed the
NIPDWS for lead. A recommendation to grant the petition could be made based
on this data, providing that all other delisting criteria are satisfied. If the
original data set had been used to evaluate the petition, a false positive error
would have occurred and the petition would have been incorrectly denied.
In the second example, the EP test result for four samples collected from a
40,000-square-foot impoundment with a depth of ten feet produced extract values
for lead of 0.195, 0.221, 0.296, and 0.284 mg/I. Due to questions regarding
historical variation in the manufacturing process, the Agency requested additional
samples. The petitioner submitted an additional 41 samples. From the entire data
set, the following characteristics were obtained:
Sample mean 0.296 mg/I
Sample median 0.328 mg/I
Sample standard deviation 0.239 mg/I
Sample minimum 0.145 mg/i
Sample maximum 0.428 mg/i
Sample second-highest maximum 0.391 mg/I
Analysis of data indicated that the sample was non-normally distributed and
this, to a certain extent, negates the use of Dixon’s test. However, if Dixon’s
test is used, then:
0.428 - 0.391 = .038
D= _________ ___ =0.133
0.428 - 0.145 = 0.284
which is not significant and therefore the maximum value cannot be regarded as
an outlier. Accordingly, the median is the more appropriate characteristic to be
used as an input parameter for the VHS model and, as it generates a compliance
point concentration for lead of 0.052 mg/I, exceeding the NIPDWS, the petition
would be recommended for denial.

-------
If only the original data had been used to evaluate the petition, a false
negative error would have occurred and the petition would have been incorrectly
recommended for exclusion.

-------
FIFTH SESSION
Thursday, July 17, 1986
1:30 p.m. - 5:00 p.m.
Chairperson:
Denise Zabinski
Chemist
Offices of Solid Wastes
USEPA
401 “M” Street, S.W.
Washington, D.C. 20460

-------
DEVELOPMENT OF PERFORMANCE EVALUATION SAMPLE
SERIES FOR THE GROUNDWATER MONITORING SURVEY AT DUMP SITES
EDWARD L. BERG, QUALITY ASSURANCE BRANC}4, ENVIRONMENTAL
MONITORING AND SUPPORT LABORATORY, U.S. ENVIRONMENTAL
PROTECTION AGENCY, CINCINNATI, OHIO
ABSTRACT
Development of performance evaluation (PE) sample series for
the Groundwater Monitoring Survey of Dump Sites - PE samples
are developed and provided to U.S. Environmental Protection
Agency (USEPA) contract laboratories analyzing ground water
monitoring samples collected from hazardous waste sites,
using SW-846 methods. The PE sample series for the first
six sites consisted of three groups: (1) full-volume
organic samples, volatiles and base/neutrals analyzed by the
gas chromatograph/mass spectrometer (GC/MS) capillary column
technique and direct injection, (2) full-volume trace metal
samples analyzed by inductively-coupled plazma (ICP) and
atomic absorption (AA) methods and (3) general chemical
analyses for minerals, total organic carbon (TOC), total
organic halides (TOX), purgeable organic hildes (POX),
purgeable organic carbon (POC), phenols, cyanide, etc.
Beginning in FY86, dioxins and herbicide PE samples were
added to the above three groups. The PB samples are sent to
the USEPA contractor field sampling team who incorporates
the organic and trace metal full-volume samples into the
stream of samples collected at each site, thereby providing
double blind PB samples. The general chemical analytes are
provided in ampuls and the contract laboratories are
instructed to remove an aliquot and dilue to one liter. The
results of the PE Program for the first 15 sites will be
discussed.

-------
OSWER LABORATORY EVALUATION PROGRAM - A PROGRESS REPORT
FLORENCE M. RICHARDSON, OFFICE OF SOLID WASTE, U.S.
ENVIRONMENTAL PROTECTION AGENCY, WASHINGTON, D.C.
ABSTRACT
The Office of Solid Waste and Emergency Response has an
ongoing laboratory evaluation program for EPA Regional,
State and Contractor Laboratories. The program is voluntary
and entails the periodic analysis of performance samples
using specified methods to allow laboratories to evaluate
their capability to analyze RCRA/CERCLA samples using SW-846
methods.
The program is structured so that it is a self-auditing
operation. Samples are periodically sent to the designated
laboratory contact along with specific instructions and
analytical standards necessary for the analyses. The
samples range from very simple aqueous solutions to more
complex matrices characteristic of wastes. Participating
laboratories receive four sets of samples per year. Results
are submitted to EPA for evaluation against referee values.

-------
IMPACTS AND INTERFACE OF CERCLA MONITORING REQUIREMENTS
WITH OTHER STAFF AND FEDERAL PROGRAMS
KATHE L. STAUBER, DENNIS M. STAIKEN, PH.D., ROBERT L.
FISCHER, PH.D., AND ROBERT R. E-IIRST, OFFICE OF QUALITY
ASSURANCE, NEW JERSEY DEPARTMENT OF ENVIRONMENTAL PROTECTION
TRENTON, NEW JERSEY
ABSTRACT
The New Jersey Department of Environmental Protection is
responsible for implementing environmental policy and
administering a regulatory structure in accordance with
Federal and State statutes. The Department’s quality
assurance program provides a mechanism which enables the
Department to administer programs based on reliable monitor-
ing and anlaytical data. The QA program uses several ele-
ments to control QA activities in adminstering the
Federal/State NPDES, Safe Drinking Water Act (SDWA), and
RCRA programs. These elements include the use of a State
Laboratory Certification Program, specialized QA permit re-
quirements, use of QA project plans, and use of laboratory
and field audits and perforamnce evaluation (PE) samples.
Implementation of the CERCLA program has required an inte-
gration of its unique QA requirements with the other State
programs. A conrerstone of the U.S. EPA CERCLA program is
use of the Contractor Laboratory Program (CL?). The CLP
program requires participating laboratories to use specified
CLP procedures which are periodically revised. In addition,
CLP labs analyze CLP performance evaluation samples and
undergo EPA audits. The analytical and QA deliverables from
CL? labs are also unique and require extensive analytical/QA
documentation. In addition, CERCLA monitoring requirements
require specified procedures/methods for groundwater
monitoring, sampling, and analysis, and for soil and air
analysis.
The comprehensive nature of the CERCLA program spans multi-
media environmental matrices (i.e. groundwater, soil, air,
etc.). Consequently, administration of the program often
impacts or directly transgresses other programs. As an
example, a site may overly an aquifer which requires CERCLA
groundwater monitoring. The water table may also supply or
impact potable water under the SDWA program. The site may
also be or impact a RCRA or ECRA (NJ Environmental Cleanup
Responsibility Act) site, and if continuous surface
discharge wells are installed, they may need a NPDES permit.
These situations commonly occur. Unfortunately, the
monitoringf and analytical requirements of each program are
different.

-------
The Department QA program had to integrate aspects of the
CERCLA program within the context of a NJDEP QA program to
avoid undue program fractionation. To accommodate the needs
of the CERCLA program, a State analytical services contract
was established. This State Contract Laboratory Program
interfaces with much of the EPA CLP program with rigid QA
requirements (known as Tier I deliverables) when the labs
are analyzing CERCLA samples. To avoid statutory contradic-
tions within programs, labs which analyze drinking water
samples in connection with a CERCLA case must be a NJDEP-
certified drinking water laboratory. Other portions of the
State CLP span non-CERCLA analytical tasks for aqueous and
nonaqueous matrices, and require a modified QA deliverables
package (knows as Tier II). This Tier II data deliverables
package was adopted as a NJDEP QA policy which is upgrading
the QA procedures in other programs. The information
required with Tier II packages includes results of blanks,
duplicates, tune performance checks (GC/MS), surrogates,
etc. State groundwater discharge permits now require that
self-compliance monitoring data be submitted to the Depart-
ment in the Tier II format. On occasion, laboratories
outside the State CLP program may be used by site
contractors. In these cases, laboratories are audited and
must adhere to the same QA requirements as CLP labs.
Several key issues remain unresolved. There are differences
in groundwater sampling requirements between CERCLA and
NPDES/NJPDES groundwater monitoring. There are also
differences in analytical approaches and methods in evaluat-
ing groundwater contamination as it affects CERCLA, RCRA,
ECRA, NJPDES, and SDWA programs. The Department QA program
will continue to integrate these issues as they affect State
programs.
When Congress enacted legislation in 1981 which created the
Superfund program for cleanup of hazardous contaminated
sites, a new realm of monitoring requirements and analytical
methods was also created. The U.S. Environmental Protection
Agency was delegated the responsibility for administering
the funds for the cleanups, and consequently established the
Contract Laboratory Program (CLP) to manage the analytical
and quality assurance aspects of the program and to maintain
sufficient qualified laboratories on a term contract to
handle the analytical workload. The samples to be collected
at Superfund sites were determined to contain innumerable
contaminants at unknown concentration levels in matrices
that were complex and variable in nature. Although metho-
dologies for analysis of inorganic and organic contaminants
existed at the time, the methods were either in a proposed
form or were not validated completely. The USEPA chose to
establish a set of methodologies which were modifications of
existing methodologies. The methods were published in the

-------
CLP’s Information for Bid (IFB) document as required methods
to be used by contract laboratories when providing
analytical services to the Agency under the Superfund
program. The advent of these methodologies and the ensuing
confusion over their application in some states has led to a
number of conflicts with other Federal and State regulatory
programs. These conflicts often manifest themselves at
sites where program jurisdictions overlap. The following
discussion includes general comments on conflicts which have
arisen in New Jersey.
FUNDAMENTAL DIFFERENCES IN MONITORING REQUIREMENTS
In New Jersey, as in many other states, the state has been
delegated primacy for administering regulations governing
drinking water, surface and subsurface disposal of municipal
and industrial wastes, and maintenance of quality of air and
water resources. The Department of Environmental Protection
is the agency which regulates these areas in New Jersey.
These programs are generally fee—based permitting programs
which incorporate routine compliance monitoring requirements
designed to eliminate sources of contamination of the
State’s resources. The monitoring segments of these pro-
grams also provide a substantial database for evaluating
long—term trends in environmental quality.
Monitoring requirements at Superfund sites differ substan-
tially from the situations described above. The require-
ments are site—specific, intensive, and necessarily cumber-
some to achieve the goals of the program. Analytical data
generated at the sites must be beyond reproach in a court of
law in order to recover the cost of the cleanup from respon-
sible parties. Since the objective of the program is to
clean up sites expeditiously, there is no long—term trend
analysis of the site, nor is there any form of compliance
monitoring. The data available at the site is generally
short—term at best. Therefore, it is crucial that the regu-
latory agency gather as much data and supporting documenta-
tion as possible.
DIFFERENCES IN MIALYTICAL REQUIREMENTS
Most environmental management programs are mandated by
Federal regulations and legislative acts, including the Safe
Drinking Water Act (40 CFR Part 141), the Clean Water Act
(40 CFR Part 136), and the Resource Conservation and
Recovery Act (40 CFR Parts 260-265). Consequently, the
Federal regulations also mandate specific analytical metho-
dologies and monitoring requirements. Analytical methodo—
logies used to report data to the NJDEP are regulated by the
Laboratory Certification Program, which monitors the use of
the required analytical procedures. Enforcement of environ-
mental management regulations is often dependent on reliable

-------
data derived from sample analysis using the required
methods.
The methods employed by CLP laboratories for the Superfund
program are integral parts of a contract between the U.S.
Environmental Protection Agency and its contractors.
Defensibility of the analytical methods is dependent upon
compliance with the terms and conditions of the CLP’s IFB
documents rather than with a Federal or State regulation.
Although the actual differences in analytical procedures are
often minor, the IFB methods are intended to be used for
samples with moderate or high levels of contaminants. An
example of this can be found in the level of contamination
allowable in TFB method blanks. The IFB methods allow for
up to 50 ppb of certain contaminants in the method blanks,
while the methods required by 40 CFR Parts 141 and 136
require that NO contamination be present above the method
detection limit (MDL). MDL’s themselves differ substantial-
ly as the IFB’s “Contract Required Detection Levels” are
suited to moderate or high levels of contamination. The
MDL’S for drinking water or water/wastewater analysis must
be capable of detecting trace quantities of contaminants.
In the case of drinking water, the IFB criteria for method
blanks and MDL’S are significantly above the existing or
proposed maximum contaminant levels (MCL’s) or health-based
criteria. Spiking levels for determining analytical
accuracy and precision are set at a level which is too high
to be representative of trace contaminant levels (MCL’s) or
health-based criteria. Spiking levels for determining
analytical accuracy and precision are set at a level which
is too high to be representative of trace contaminant
analysis.
Sample holding times are another notable difference between
the IFB methods and regulatory methods. For example, the
maximum allowable holding time for purgeable organics as
required by NJDEP’s Division of Hazardous Site Mitigation is
7 days. This contradicts the holding time for the same
samples covered by 40 CFR Parts 136, 141., and 260-265, which
is established at 14 days. Analytical results from samples
held for more than 7 days are rejected for use under
Superfund, but are acceptable for other regulatory programs.
The IFB document specifies the required surrogate and inter-
nal standard compounds to be used for each analysis, when
applicable. These compounds, in most cases, are different
than those required and/or recommended in 40 CFR Parts 136,
141, and 260-265 for similar methodologies. This difference
separates the QAIQC measures for the analyses.
Nowhere are the differences in analytical, data handling,
and deliverables requirements more evident than in a
laboratory that is certified by an agency, such as NJDEP, to

-------
report compliance data under the regulations and also
provides analytical support for Superfund projects. In a
commercial laboratory where economics play a major role in
determining the laboratory’s analytical specializations, the
Superfund program is a source of motivation. The laboratory
concentrates its efforts on becoming a “CLP Laboratory,”
thus dedicating its instruments to analyses of waste samples
collected from Superfund sites. Although the effort pays
off in several ways, economically as well as technically,
the change in methodologies, regardless of how subtle,
jeopardizes their legal status as a “certified” laboratory.
In situations such as this, clients requesting analytical
services for projects not related to Superfund or the CLP
often have no choice but to accept and pay for deliverables
not required for the data’s intended use. In other labora-
tories, where both sets of requirements are adhered to, the
cost of doing so can be prohibitive. This often includes
dedication of separate instrumentation, personnel, and other
costly resources.
There are conflicts between CERCLA monitoring and other
monitoring programs when dealing with the data usage. If a
facility applies for a RCRA permit and is regulated as such
for several years, a data bank is collected on that site.
The level of data deliverables and validation for the site
is not intensive due to the historical data collected on the
site. If the RCRA site is later regulated as a Superfund
site, then the previous data is invalidated for use in site
character izat ion.
ADMINISTRATION AND ENFORCEMENT
Some similarities exist between the CLP and the New Jersey
Laboratory Certification Program in the area of administra-
tion. Both programs require formal application and submis-
sion of a fee (the fee is collected annually by the
certification program). The laboratories must then perform
acceptably on a set of proficiency evaluation samples.
Assuming satisfactory performance on the PE samples, the
laboratories then undergo a thorough on—site audit and
evaluation of instrumentation, personnel gualifications,
quality control, etc.
After contract award (CLP) or issuance of certification,
enforcement of the provisions of either program differ.
Because of the regulatory nature of the laboratory
certification program, laboratories which do not comply with
the applicable regulations are subject to fines and suspen-
sion or revocation of their certification. Suspension or
revocation of certification disqualifies a laboratory from
reporting data in connection with SDWA, NPDES/NJPDES, or
RCRA until such time as corrective action measures have been
implemented. In addition, the Department may audit any of

-------
the laboratories’ records at any time if it determines that
such action is warranted. Any data which is determined to
be unacceptable to the Department may be rejected. In
contrast, the recourses available to the CLP contract user
are generally limited to non—payment for deliverables for a
given sample or set of samples. If the USEPA cites a CLP
Laboratory for consistent violations of the terms of the
IFB, the laboratory may be placed in a “cure program” until
performance improves or other terms are achieved.
INTERFACES OR CERCLA WITH OTHER PROGRAMS
Over the past two years, the New Jersey Department of
Environmental Protection has attempted to reconcile some of
the differences in monitoring and analytical requirements
between CERCLA and the major Federally—mandated regulations.
The Department initially embarked on the development of a
“Uniform Laboratory Standards” document in 1984. The
document’s intent was to standardize as much of the
analytical and quality control processes as possible without
leading laboratories out of compliance with the applicable
methodologies. The document also standardized sampling
equipment cleaning procedures for the Department. By mid—
1985, the CLP program had incorporated substantial revisions
in the IFB document, making the Uniform Laboratory Standards
document somewhat outdated. At the same time, the
Department initiated development of a new analytical
services contract to replace the contract which was about to
expire. The Department seized the opportunity for a second
attempt at reconciling some of the differences.
The scope of work in the new analytical services contract
(known as X—085) reflected much of this effort while still
maintaining differentiation in analytical methods. The
contract is divided into four “tasks:”
Task I Air Analysis
Task II Aqueous Sample Analysis
(SDWA, NJPDES)
Task III Nonagueous Sample Analysis
(RCRA, NJPDES)
Task IV USEPA CLP Analysis
(CERCLA)
The contract scope of work concentrated on applying quality
control and data handling requirements from the IFB methods
which were beyond those addressed in the regulatory methods
and did not conflict with those methods. Examples of this
are use of system performance check compounds (SPCC’s) and
calibration check compounds (Ccc’s), handling of surrogate
and matrix spike data outside of control limits, and
definition of deliverables.

-------
The deliverables requirements are probably the most
significant extract from the CLP. All laboratories are re-
quired to supply deliverables in two formats, called Tier I
and Tier II. Tier I deliverables are an adaptation of the
CLP deliverables specifications, including methodology sum-
maries, laboratory chronicles, chain of custody documenta-
tion, chromatogranis, raw data, and tabulation of all results
and quality control data. Tier II is an abbreviated version
of the CLP deliverables package, which includes only enough
information and data to perform a cursory evaluation of the
results. This format was developed in order to decrease
turnaround times and to accommodate the lesser documentation
needs of the SDWA, NJPDES, and RCRA programs.
CONCLUS ION
The IFB methodologies and Federally—mandated methodologies
for SDWA, NPDES, and RCRA all have merits for their intended
uses. The IFB document provides for full documentation of
sample results and more stringent quality control for all
analyses connected with CERCLA short—term site—specific
projects. Their application to moderately and highly
contaminated multi—media samples has been executed and rea-
sonably well validated to the USEPA’s credit. However,
their use for analysis of trace contaminants and for com-
pliance analysis for SDWA, NPDES, and RCRA awaits final
resolution.
The Federally-mandated methodologies from 40 CFR Parts 136,
141, and 260—265 have undergone extensive validation studies
and are continually supported by additional long—term data
generation. The Department’s interpretation of the regula-
tions has resulted in promulgation of these methodologies
and a certification program to monitor and enforce their
use. The Department will continue to seek and implement
common procedures, but will do so in cooperation with the
U.S. Environmental Protection Agency. The Department sup-
ports further method research in hopes of eventually elimi-
nating some of the existing fractionations between the
various program analytical requirements.

-------
Ane]ysis of Non—Homogeneous Mixtures
Robert L. Fischer, Ph.D
Dennis Stainken, Ph.D
Robert Hirst
Katlie Stauber
Susan Dengler
Ne Jersey Department of En ’ironmnentaa
Pr o t e c t I on
Office of Quality Assurance
CN—A 02
Trenton, NJ 08625
July, 1986

-------
The New Jersey Department of Environmental Protection)
Office of Quality Assurance has administered a laboratory c.erti ficati on
prog?aTr fot the Safe Drinking Water Act and the Clean Water Act
(New Jersey’s Pollutant Discharge Elimination System) for many
yearb. As ret. progr’ns were enect ed by the various state
agerci es’ they i ec 1 uii cc 1 that the n’onitorir.g data reported for
ccirp] 5 ancc n UF t be produced by laboratoi i es c c i t 5 fi e by the
Office of Qt.Elity Assurance. These programs have required the
analysis of n’rre c or.pi e san p] es and the development of new
methods ii ’ cases tJ.eie eti;ods were not available. To accomplish
this t,isk, the Cffice of Quality Assurance has increased the
nunhet of catecoties for which laboratories can he certified, end
the an 3)tacal pioceciutes that are required in the analysis of
the san pIes, by pron u3 ga t inc these procedures in the Regu] at ions
(‘,overni Hg I chri nUn y N rti fi votiot and Standards of Performance.
Some piogicurs requited the analysis of substances that did not
have val Idet ed rcetl’ods while others required methods specific N ’
a cert aft matrix. The Department has resolved these issues In
several weys.
Cur first exen.j-le is New Jersey’s RCRA program. Tie req iirewents
for the analysis of these samples wet e quite simply solved by the
departrent h> pron:ulgat ing the USE]’A ‘s Test methods for

-------
Evaluation of Sc,l:id Waste (SW866) and by providing for certifica-
tion of lahoi t on es for RCRA analysis in the New Jersey Labo—
rat ory Certi firat ici’ 1 egulat ions.
The second example is a program created by an amendment to
the Ne Jersey Safe Drinking Water Act; known as Assembly Bill
A—280. This program requires the owner or operator of each
potable water supply to undertake the periodic testing of the
water provided to it customers for various hazardous
contaminai’ts whici inc]uc 4 e purgeable organics, certain aromatic
at’d aliphatic hydrocarbons, a pesticide, PCB’s an aldehycle, a
1
-------
The second groups will tlier use these risk factors to set the
maximun’ c.oi:taminant levels. Unfortunately the riethod detection
levels for the GOC series irethods are genera] ly above the pro—
je-cred niaxiuri contaminant levels. A third work group has been
asserbled to modify methods 502.1 and 503.1 to obtain lower
detection levels. These ]evels should lower the Practical
Quantit itiori ]evel to a point that will satisfy the intent of the
act. If th modifications proves successful, the A—280 appro .ed
]ahoratorie. will gather data to ‘a] id te this procedure.
Since this act also gives the Drinking Water Quality
]nstitute the power to include other substances in the ]ist of
analytes, ar additiona] research project has been started for the
analysis of asbestos 5n water. Initial results of this project
indicate that this method will require less ana]ysis time arid
thereforE he I ess costly than the presert electron microscopy
method. Ne Jersey does not have a sufficient number of
laboratories, to validate this method so it will he presented to
EPA whE-ii it is ready for validation.
Our next example is the Sludge Nanagement Program, part of
which is regulated through the Sludge Quality Assurance Regu—
lations (SQAF). These regu]ations require sewage treatment
plants to analyze sludge periodically for selected chemical
paran eters and certain physicai properties. The reporting
frequency is dependent upon the total flow entering the sewage
treatment plant. At the time of SQAR promulgation there were no
-------
standardized methods for the analysis of sludge and a lack of
control over the laboratory performing the sludge analysis. As a
result, the data submitted to the Department was difficult to
iriterpet and as difficult to use in the broad sense as was
originally intended.
Hence the Sludge Nethods Task Force was formed to eva]uate
and establish the methods arid QC procedures needed for the
ani]y .cis cf s]udge as required by SQAR. The task force is
composed of individuals chosen from industry, sewage treatm r’t
plants, acedeiric arc.! state cud federal governments. This Task
Force 1 as developed genet a] procedure for the adoption of
stardarized n:etiodologies. This procedure consists of five
steps. The first step i a ]iterature review and a survey Cuí
]abotpto ies tc identify candidate procedures. Step two of the
proceduies is task force evalu tion of each candidate method
mc] uding re ’t i tirg ivtc a St ardardized format. After the method
is given tentative approve] by the Task Force, step three of tie
procedure is circulation of the methods to a panel of
]abo atori es for their review and comnerit . During step four each
]aboratory conlrert is reviewed by the Task Force and appropriate
charges are nacle to the method. In step five the methods ate
subject to inter and intra laboratory validation. Once the Task
Force is satisfied with the validation results, the methods wi]]
he recommended to tie department for use in sludge analysis.
Presently, there are site methods ready for va]idatlon. These
methods are oi the analysis of p11, total residue, volatile and
ash of total residue, oi] and grease, pheno]s, and metals. The
validation study will consist of the
-------
co]]ectioi’ of liquid anc solid sludges at two concentrations,
homogenizing ecch batch, and delivering samples to participating
laboratories fo replicate ar.e]ysis. A local laboratory wi]] be
enip]ciyed fc.r the homogenization process and for referee analysis.
Prior to ir;itietion of the study, the laboratories will partici-
pate in a symposiun where the methc ds and va].idation procedure
wi]] be discussed tc assure that a]] the methods are clearly
understood ard uriforirly applied.
(“ur ne ct example is the EnvirorLmental C]eanup Responsibility
Act (E(’} A) which w s signed into law in September of 1983. This
act pro ’ides tie Departmeiit with th statutory ability to ensure
that. industria] e&-tahlishmeits involved with hazardous substance
aud ‘aEtc re rot sold, transfered oi closed without proper
cleanup. The Departsient corducts a review of the facility
iricltiding ar cu—site inspection. If DFP finds that the site is
riot er vironirer ta] !y acceptub] e th company must develop and
implement a DIP approved cleanup plan. The company must al c
provide financial ass starice for the full estimated cost of the
cleanup. Thus this program requires the analysis for every
conceivable compound in any matrix. At present we have used
SW846 as the major source of methodologies and any other EPA
Method avaiiab].e.
-------
To Summarize
The IcJrEP requirer that eli analytical data submitted to DEF
within the KFDIS end the EDWA programs be from state certified
labs. The State Lab Certification Regulations were recently
rEv ised to upgrade and extend the coverage of the Lab
Certifiatioi Frogram to include PCRA and various State
Program. Provision was made within the new Certification Program
to mi nate inp] ementat ion of the recommended s]udge methods and
to further validate the procedures. Laboratories performing
sludge ana lyses wili have to he certified within the
Certification Program. The Program certifies by category of
ana] ysis
Category 1—Drinking Water (SDWA)
Category 2— Water/Wcstet..aterlS]udge (NJPDES)
Category 3—Waste Analysis (PCRA,NJPDFS)
Within each category, a lab is certified by methods (appropriate
for subcategories of inorganics, organics, limited chemistry,
etc.)
Laboratories conducting these analyses will be required to
use recomnended procedures and submit results to the State with
specific dEta de]i ’erabies requirements. This specific QAIQC
-------
date wi]] b ircorpr’rated into a data base to establish yearly
statistical acceptance parameters for the nethods. During this
process, we will also initiate establishment of sludge solid
waste anc 7 A2SC analytical performance evaluation samples to
survey the performance or participating laboratories and enforce
Department standards.
-------
ANALYSIS OF NON-HOMOGENEOUS MIXTURES
ROBERT L. FISCHER, PH.D., DENNIS M. STAINKEN, PH.D., KATHE
STAIJBER, AND SUSAN DENGLER, OFFICE OF QUALITY ASSURANCE, NEW
JERSEY DEPARTMENT OF ENVIRONMENTAL PROTECTION, TRENTON, NEW
JERSEY
ABSTRACT
The Office of Quality Assurance (OQA) within the New Jersey
Department of Environmental Protection administers a labora-
tory certification program for the Safe Drinking Water Act
and the Clean Water Act (New Jersey Pollutant Discharge
Elimination System). All laboratories which report environ-
mental measurement data to the Department in compliance with
the Acts must be certified by OQA. In the NJPDES Program,
permits are issued to each facility which discharges
industrial or municipal waste to sewage collection systems,
surface waters, or groundwater. The NJDEP Division of Water
Resources writes specific monitoring requirments into each
permit and reviews periodic discharge monitoring report sub-
mitted by the dischargers. In recent years, many facilities
in New Jersey, some of which are also NJPDES facilities,
were designated as RCRA facilities. Many additional
monitoring requirements were included into the facilities’
monitoring requirements, including analyses for contaminants
in complex, nonaqueous matrices (i.e. soils, sludges, non-
aqueous liquids). The NJPDES program also began to place
much emphasis on disposal problems associated with sludge,
which also added to the need for validated analytical
methods.
Many analytical procedures exist for analysis of con-
taminants in complex matrices. Some of these methods, such
as those published in the USEPA’s Test Methods for Evalua-
tion of Solid Waste (SW-846), have been promulgated by the
Agency and are required methods for programs like RCRA.
Many analytes remain without validated methods while others
require methods specific to the matrix under examination
(such as sludge). The Department of Environmental Protec-
tion seeks to regulate as many nonaqueous analytical pro-
cedures as can be validated. It therefore became crucial
that a protocol for validating new or modified methods for
nonaqueous sample analysis be established to assure reliable
monitoring of RCRA and NJPDES facilities.
To accomplish this task, the Office of Quality Assurance, in
conjunction with the Division of Water Resources and the
Division of Waste Management, has developed a procedure to
validate new or modified methods. The process includes
research of existing literature, single laboratory assess
merit of the methods, and interlaboratory studies of method
-------
performance. The goal of these protocols is to provide a
mechanism for NJDEP to “approve” methods for use with
nonaqueous samples and to enforce the methods to assure the
maximum possible comparability of results between
laboratories.
INTRODUCTION
The New Jersey Department of Environmental Protection,
Office of Quality Assurance has administered a laboratory
certification program for the Safe Drinking Water Act and
the Clean Water Act (New Jersey Pollutant Discharge Elimina-
tion System) for many years. All monitoring data reported
to the State for compliance with the various programs
administered by the Department must be produced by labora-
tories certified by the Office of Quality Assurance. As the
number of programs in the Department has increased, many
additional monitoring requirements and the analysis of more
complex samples has been required of these certified labora-
tories. To accomplish this task the Office of Quality As-
surance has increased the number of programs for which
laboratories can be certified and the analytical procedures
that are required in the analysis of the samples by promul-
gating such methods as those in the !JSEPA’s “Test Methods
for Evaluation of Solid Waste” (SW 846) and the certifica-
tion of laboratories for RCRA analysis. However, since many
programs require the analysis of substances that do not have
validated methods or require methods specific to a certain
matrix, the Department has developed a procedure to review
and validate new or modified methods.
One program manages the disposal of sludge under the New
Jersey Sludge Quality Assurance Regulations (SQAR). These
regulations require sewage treatment plants to analyze
sludge for selected chemical parameters and certain physical
properties. After the regulations were promulgated, the
Department found that there were no standardized methods for
sludge analyses and that the quality of the data submitted
could not be interpreted for the ultimate disposal of the
sludge. The Department therefore established a Sludge
Methods Task Force to evaluate and establish the standard
validated methods needed for the analysis of sludge as
required by SQAR. This Task Force is composed of indi-
viduals chosen from industry, sewage treatment plants,
academia, EPA and volunteers from several state agencies.
This Task Force developed a general procedure for the
adoption of the standardized methodologies. This procedure
consists of a review of the literature and a survey of
laboratories to find candidate methodologies. The methods
are then reviewed by the other members of the Task Force and
rewritten into a standard format. After the method is given
-------
temporary approval by the committee, the method is circu-
lated to a panel of laboratories for their review and corn-
rnents. Each comment that is returned to the Task Force is
then evaluated; the method is again reviewed and given a
final approval for inter and intra-laboratory validation.
The analysis of metals in sludge was the first procedure
that was subjected to the validation procedure. Intra-
laboratory testing evaluated four methods for sample pre-
paration; dry ashing, matrix acid digestion, high pressure
decomposition (PARR Bomb) and matrix acid/hydrogen peroxide
digestion. Atomic absorption spectroscopy was used for
quantitative analysis.
Five replicates of each of two sludge samples were digested
using each of the digestion methods. The digestate was then
analyzed using a model 360 Perkin-Elmer atomic absorption
spectrometer. Matrix acid/hydrogen peroxide digestion and
the high pressure decomposition methods yielded higher
recoveries and were consistent with each other. Economic
and availability considerations made the matrix
acid/hydrogen peroxide digestion procedure the method of
choice.
The revised method was then subjected to an inter-laboratory
study in which fourteen laboratories were given a dried
municipal sludge obtained from the EPA. Each laboratory was
asked to analyze five replicates and report the results to
the Task Force. The results of this study indicated that
the data generated was within the EPA established ranges.
Therefore, it was concluded that the matrix acid/hydrogen
peroxide digestion would be used for the determination of
metals in sewage sludge and was recommended for final
approval by the Task Force for inclusion into the Regula-
tions Governing Laboratory Certification.
Presently, there are five more methods ready for validation.
These are methods for the analysis of pH, total residue,
volatile and ash of total residue, oil and grease, and
phenols. The Sludge Methods Task Force has designed an in-
terlaboratory study for the validation of these methods.
The study will consist of the collection of liquid and solid
sludges at two concentrations, (i.e. high and low contamina-
tion) homogenizing each batch and delivering to participat-
ing laboratories for replicate analysis. A local laboratory
will be employed for the homogenization process and for
referee analyses. Prior to initiation of the study the
laboratories will participate in a symposium where the
method and validated procedure will be discussed to assure
that all the methods are clearly understood and uniformly
applied.
-------
The Sludge Methods Task Force is in the process of preparing
a report presenting the progress of the development of these
methods. The report discusses the background of the task
force and the necessity of uniform methods for sludge
analysis. The analytical parameters needed by the NJDEP for
sludge regulation are addressed as well as the procedure to
develop methods for accurate assessment of these parameters.
The methods outlined in the report are “proposed and recom-
mended.” The proposed methods are those methods designed by
the task force which have not been through interlaboratory
validation. The “recommended” methods are those methods
which have been fully validated and are recommended to the
NJDEP as the preferred methods for these analyses. The
Sludge Methods Task Force report will be issued yearly to
reflect the progress of the task force and to update pro-
posed methodologies.
The NJDEP requires that all analytical data submitted to DEP
within NPDES and SDWA programs be from certified labs. The
State Lab Certification Regulations were recently revised to
upgrade and extend the coverage of the Lab Certification
Program to include RCRA and various State Programs. Pro-
vision was made within the new Certification Program to
initiate implementation of the recommended sludge methods
and to further validate the procedures. Laboratories per-
forming sludge analyses will have to be certified within the
Certification Program. The Program certifies by category of
analysis:
Category 1 - Drinking Water (SDWA)
Category 2 - Water/Wastewater/Sludge (NJPDES)
Category 3 - Waste Analysis (RCRA, NJPDES, ECRA)
Within each category, a lab is certified by methods (appro-
priate for subcategories of inorganics, organics, limited
chemistry, etc.)
Laboratories conducting sludge analyses will be required to
use recommended procedures and submit results to the State
with specific data deliverables requirements. This specific
QA/QC data will be incorporated into a Quality Assurance
data base to establish yearly statistical acceptance para-
meters for the methods. During this process, we will also
initiate establishment of sludge analytical performance
evaluation samples to survey the performance of partici-
pating laboratories and enforce Department standards. It is
envisioned that the scope of analyses will expand as the
Sludge Methods Task Force and the Office of Quality Assur-
ance recommend additional methods for incorporation into the
regulations.
-------
HAZARDOUS WASTE ANALYSIS USING GAS
CHROMATOGRAPHY/MASS SPECTROMETRY (GC /MS) AND
LIQUID CHROMATOGRAPHY/MASS SPECTROMETRY (LC/MS)
JIM POPPITI, FINNEGAN CORPORATION, ROCKVILLE, MARYLAND
ABSTRACT
The Hazardous Waste Program has evolved over the past 10
years. During this time several additions and changes were
made to the regulations which require testing water,
leachates, and groundwater for organic toxic constituents.
Mass spectrometric methods provided in SW-846 are generally
adequate for the determination of gas chromatographable
Appendix VIII compounds, however, newer MS techniques can be
used to extend the number of analytes and eliminate chroma-
tographic interferences.
Newer MS techniques, such as LC/MS, can be used for
determining polar toxic organics in wastes and waste
leachates. Specifically, Thermospray/LC/MS offers the
ability to determine polar organics with high analysis speed
and good sensitivity. Therniospray/LC/MS is best suited for
the determination of water soluble, non-volatile compounds
which cannot be determined by conventional extraction and gas
chromatographic methods. This technique can be used to
rapidly screen samples since it produces molecular ions (or
adducts) and several fragment ions useful for identification
of organic compounds. Sample introduction can be achieved
through a sample loop (for rapid screening) or normal reverse-
phase liquid chromatography.
Both GC/MS and LC/MS are evaluated for the analysis of real
and simulated complex lecahates and waste extracts. A com-
parison of these methods is presented with respect to the
types of compounds determined, sample preparation, estimated
detection limits, analysis costs, and applicability of these
techniques for routine waste leachate analysis.
INTRODUCTION
The Hazardous Waste Program has evolved over the past 10
years. During this time several additions and changes were
made to the regulations which require testing wastes,
leachates, and groundwater for organic toxic constituents.
Mass spectrometric methods provided in SW-846 are generally
adequate for the determination of gas chrotnatographable
Appendix VIII compounds, however, newer MS techniques can be
used to extend the number of analytes and eliminate chroma-
tographic interferences.
Recently EPA solicited comment on reducing the number of
analytes for groundwater analysis to those amenable to GC/MS.
-------
While several polar Appendix vi ii compounds would rarely be
encountered in the environment (i.e., aflotoxins, mytomycin,
etc.) many environmentally significant compounds are quite
polar and not readily accessable via GC/MS. Phenols are one
example of large class or compounds which fall into this
category.
Newer MS techniques, such as LC/MS, can be used for
determining polar toxic organics in wastes and waste
leachates. Specifically, Thermospray/LC/MS offers the
ability to determine polar organics with high analysis speed
and good sensitivity. This technique can be used to rapidl
screen samples since it produces molecular ions (or adducts
useful for identification of organic compounds. Sample
introduction is achieved through a sample loop (for rapid
screening) or normal reverse-phase liquid chromatogrpahy.
EXPER IMENTAL
Chemicals and samples
Authentic standards of phenols used in this study were
obtained from Supelco Inc., and used without additional
purification. The toxicity leachate extract was provided by
ENSECO Laboratory, Cambridge, MA. Information provided by
ENSECO with the sample indicated that the sample contained
phenols and other low molecular weight organic acids.
Instrumentation
The instrument used for GC/MS was a Finnigan MAT 5100 EF
equipped with a 30 m, 0.25 mm Id, 0.25um film capillary
column. The column liquid phase was DB-1. Samples were
injected, using the splitless technique, at 50 deg C. The
column was programed to 270 deg at 10 deg/min.
The instrument used for LC/MS was a Finnigan MAT 4600 equipped
with a Finnigan Thermospray inlet. All data were obtained in
the negative ion mode using a conversion dynode multiplier
operated at 5Kv on the conversion dynode. The solvent system
used was 30% Methanol in 0.1 N ammonium acetate. The LC
column used was an RP 18.
RESULTS AND DISCUSSION
The reconstructed ion chromatogram for the GC/MS results is
presented in Figure 1. The major peaks in the chromatogram
correspond to cresols and low molecular weight organic acids.
Identification was made by comparison of spectra to the 42,000
compound EPA/NIH library. The sample was the acid extract
from a TCLP leachate from a hazardous waste sample. The
leachate (pH 5 acetate buffer) was extracted with methylene
chloride after pH adjustment to 2. The resulting extract
-------
4 TO 1711
Fi 9 urr 1: RlcDnztructcd 11111 chr11m to 9 r m
1119 ,rid 1in r rcali.
P lC TA TQ.P 1513
4 d ’eb IS:3 ; £II*ICLO ( 34 IT)
S FIis TQJ L1 TE
liii,.., I1I$
l4. Pbt Li N S. 4. 5 144. a 5. 1.5 J I U 41.
Sic
‘IC
144?
4 18
6:41
13 a
16:41
1288
1411
23:28
S
1
-------
contains a high level of acetic acid. The extract was reduced
to 1 ml using a Kuderna Danish concentrator.
The sample contained high levels of phenols (mostly cresols)
and low molecular weight organic acids. The high acid con-
centration did not appreciably effect the gas chromatography
on the system used. No other compounds besides cresols and
organic acids were identified in this extract. The estimated
detection limit for this analysis was 20 ppb based on the in-
jection of 10 ng of DFTPP.
Phenol was not observed in this chromatogram since the solvent
and acetic acid interfered with the mass spectra of early
eluting peaks (up to about 6 mm..) The analysis time was
about 35 minutes including data processing time.
The sample was also analyzed via LC/MS (Figure 2). Initially
the sample was introduced through the sample loop to determine
whether phenols were present. The spectrum obtained indicated
that phenols could be present. The spectrum shows that
Nitrophenol, Dinitrophenol, and Methyldinitrophenol may be
present since ions at the proper molecular weights are
present. The sample was then injected onto the LC column and
chromatographed. The resulting chroniatogram is presented in
Figure 3. This figure shows that molecular ions consistent
with the three phenols mentioned above elute in
chromatographic peaks within about 15 minutes.
The sample was then spiked with a phenol mix which contained
5 ng each of 4-Nitrophenol, 2,4-Dinitrophenol, and
2-methyl-4, 6-Dinitrophenol. The intensity of the single ion
chromatograms increased by 163,416, and 466 height units/ng
injected respectively for the three compounds (Figure 4). The
unspiked sample was therefore estimated to contain 16, 3, and
3 ng of the phenols respectively. (This type of calculation
is for estimation purposes only since it amounts to a one
point standard addition.) To relate these concentrations to
the original sample (500 ml leachate) the actual leachate
sample would contain about 6 ppb of 4-Nitrophenol, 1 ppb of
2,4-Dinitrophenol, and 1 ppb of the 2-Methyl-4,
6-Dini trophenol.
The system was also checked to determine whether there was any
carry-over in the sample loop from one sample injection to the
next by injecting a blank. The blank (Figure 5) shows some
small peaks corresponding to the molecular ions at the correct
retention times. The intensity of these ions are quite low.
The signal-to-noise ratio for each of these can be estimated
from the figure at 4:1, 2:1, and 4:1 respectively with peak
heights corresponding to 232, 602, and 201. The estimated
amount of material carried over therefore is about 10%.
The results of this preliminary study demonstrate the utility
and versatility of Thermospray/LC/MS. While no attempt was
-------
MASS SPECTRUM
05/86/86
SAMPLEi 21268 ACID FR
CUNDS. i TSP BUF/WJ C)I441EG
‘e.eeex
135
DATA: POP3 *189
CCLI: CAL42SII *1
161
BASE M’Z: 95
RIC: 435712
195
144
158528.
11?
MI’Z 188
II
183
18?
I I i. .1 I I .11
il .
II
II
148 168 228 248
Figure 2: Mass spectrum of leachate
obtained by Thermospray LC MS.
260
-------
RIC+PIASS Cè OMATOGWi 6
‘8?’86 14*81:90
S I.Ei PUX 21268
COPCS.s TSP BLF’3 EOH4€G
4CLa C 1 N
DATA: POP?Z *1
CALLs Ca429M 01
0 , 4.0 QLIAN 9 0,
5C 6 ITO 308
I.OJ 8 sU 3
2632.
138.041
± 0.508
1322.
183.055
± 0.580
1598.
197.059
* e.soe
300 SCAN
15 :80 TIME
138
50.2 71
164
185
213
183
68.
197
66 78
I I
50 100 150 208 250
2:30 5:08 7:30 10:90 12:38
Figure 3: Single ion chromatograms for the
phenols in
239
the leachate 5ample .
-------
5C6 ITO 308
RIC+MASS HROP TO
85’e?’86 14s38s08
DATAs POP9 *1
C .Is CAL4 II 01
COHOS. s TSP BLI/30R1€OH’NEG
RANGE:G 1, 74 L1 EL:H 8 4.8 QUAH:A
0 1.BJ 0 SEi U20 3
7g 30
ion chromatogram5 + or the
after addition of 5 ng of
3448.
138.041
± 8.580
3404.
183.855
± 0.508
3928.
197.859
± 0.508
: : : SCAN
15 ,00 TINE
8?.
138
86.
183
108.
19?
R _ — I—
50
2:30 5s80
Figure ‘i: Single
phenol5
150 200 250
12z30
10300
each.
-------
POP8 *1 SCANS I TO
RIC+I1 SS Cb O,1ATOGRf 15
05/07/86 14:20:00
S 1PLEs SOLI.EHT 8LA 4(
cOHOS. $ TSP F’3e 1EOI4’*.G
RN1I : C. 1 7 LABEL: H
50
2:38
DATA:
CAL4 M 01
8 4.0 OJN4 : A 8, 1.0 -J B SEi U 3
A J V
U I U
100
5:00
232.
138.041
± 0.508
602.
183.055
± 0.588
281.
197.059
± 8.588
300 SCAN
15:08 TI
I - U I I
- I I I I I 1 I I •
150
7:30
280
18:00
250
12:38
Blank
injection.
Some carry—over
1 8
183
33.4
197
I
A
Figure
is observed.
-------
made to optimize either chromatographic conditions or the
solvent system it was possible to analyze a complex leachate
sample for phenols in about 15 minutes. GC/MS of this sample
revealed high concentrations of organic acids and cresols,
however, the other phenols were not detected and the analysis
time was about twice as long as the LC/MS run.
The data indicate that the two techniques are generally
complimentary and that LC/MS is superior for determining
phenolic compounds, especially those that produce a strong
negative ion (i.e., nitrophenols, etc.) Furthermore, the
LC/MS system is capabale of delivering the entire sample to
the ion source through the sample loop thus allowing rapid
screening of samples for the possible presence of phenols
before proceeding with the factual analysis.
The sensitivity of the Thermospray LC/MS system is very high
for the compounds determined. The phenols containing
heteroatoms, other than oxygen, produce very strong molecular
ions. This has also been observed for halogenated phenols.
Petachlorophenol, for example, produces very strong molecular
ions at 263 and 265. This behavior is exactly as predicted
using a classical Hammet approach. The phenolic negative ion
is stabilized by groups which have electron withdrawing
characteristics. Moreover, nitro groups are able to resonance
stabilize the resulting negative ion as shown in Figure 6.
Similarly, chloro, bromo, cyano, etc. substituted phenols
would also be expected to exhibit electron withdrawing
characteristics and therefore stabilize the negative phenol
ion.
In the negative ion mode organic acids are not normally
observed. This is most likely due to low ionization effi-
ciency (i.e., R-COO- is not formed unless the resulting charge
can be stabilized) or, if the ion is formed it rapidly
fragments and no ions indicative of actual molecular species
are observed. The first case is probably the most likely.
The per analysis cost is determined from the sample prepara-
tion and analysis procedures used. Analysis time therefore
affects one part of the overall cost. Generally, the less
time an analysis takes the lower the cost, up to a point. In
the case of GC/MS of acid fractions the analysis portion of
the cost is based on an analysis time of about 35 minutes.
(This is about average for most labs we have spoken with.)
LC/MS can be done in about half that time and thus one would
expect the overall cost using LC/MS to be lower. The cost
reduction will not, however, be half that of GC/MS since the
sample preparation procedure for each method is the same. As
analysis time is reduced the overall cost of the test will
approach the cost of sample preparation. In any case LC/MS
should offer a less expensive alternative to GCIMS for many
phenolic compounds. Furthermore, if a screening procedure is
-------
Figure 6:
IC
Through resonance of
nitro group
-------
used the analysis time will be reduced by an order of
magnitude and hence the analysis cost will be even lower.
SUMMARY
The data presented here generally indicate that GC/MS and
LC/MS are complimentary techniques for hazardous waste and
leachate analsyis. While the data are limited, LC/MS was
demonstrated to be a viable technique for analysis of several
phenolic compounds that were not detected using GC/MS.
-------
UPDATE ON A COOPERATIVE INVESTIGATION OF TEST METHODS FOR
SOLIDIFIED WASTE CHARACTERIZATION
J. STEGEMANN, ENVIRONMENT CANADA WASTEWATER TECHNOLOGY
CENTRE; N. CATHCART, ENVIRONMENT CANADA INDUSTRIAL PROGRAMS
BRANCH; D. FRIEDMAN, OFFICE OF SOLID WASTE, U.S.
ENVIRONMENTAL PROTECTION AGENCY, AND A. LIEM, ALBERTA
ENVIRONMENTAL CENTRE
ABSTRACT
A cooperative program is being conducted to investigate the
suitability of a number of laboratory test methods for
determining the chemical and physical properties of a large
variety of solidified wastes. The participants in this
study are Environment Canada, the U.S. Environmental Protec-
tion Agency, Alberta Environment, and fifteen companies
involved in developing or marketing solidification
technology. The details of the study and its present status
will be reviewed.
The objectives of this study are: (1) to develop a uniform
testing protocol for solidified waste to be used to deter-
mine the degree of hazard reduction achieved by the treat-
ment, and (2) to create a data base of the properties of
solidified hazardous waste achieveable with present
technology which will assist in setting standards and
provide a basis for futher developmental work.
Each of the participating companies has applied the
solidification system of its choice to as many as five
wastes. Three laboratories have applied a protocol of
twelve tests to the products of these solidification treat-
ments to determine their intrinsic physical and chemical
properties. In addition, Louisiana State University has
performed microstructural characterization of selected
solidified products. A report of the final results is
expected to be available in the spring of 1987.
INTRODUCT ION
Landfilling is the chosen disposal method for many hazardous
wastes which are nonrecyclable, or nondestructable, or for
which the disposal options of recycling or destruction are
too costly. Solidification processes* are designed to
improve wastes for landfilling by ameliorating their physi-
cal properties, or immobilizing the contaminants to prevent
groundwater contamination, or both.
* In the context of this paper, the term “solidification”
refers also to processing sometimes termed “stabiliza-
tion” and “fixation”.
-------
The numerous vendors of solidification processes make
various claims as to the different degrees to which their
processes affect the properties of concern. At present, no
set of standard tests exists for measuring the efficacy of
solidification processes.
A study has been initiated by Environment Canada, in
cooperation with the U.S. Environmental Protection Agency,
Alberta Environment, and fifteen Canadian, American and
European companies involved in developing or marketing
solidification technology, to develop a uniform assessment
protocol for evaluating the properties of a variety of
solidified wastes (1)
OBJECTIVES
The objectives of the cooperative program are:
1) To assess the suitability of a protocol of 12 short—
term test methods for characterizing physical and
leaching properties of a wide variety of solidified
wastes,
2) To develop a data base of properties of solidified
wastes which will assist in setting standards and allow
comparative evaluation of solidified wastes and raw
wastes,
3) To provide a basis for further work towards the
development of accelerated test methods and mathemati-
cal models to estimate long—term stability, and
4) To promote the use of the test methods as standards
which would allow uniform evaluation of solidified
wastes.
BACKGROUND
A solidified waste matrix is typically formed by pozzolanic
reactions such as those which occur between lime and fly ash
and in portland cement. Other materials such as clays,
polymers and proprietary specialty sorberits may also be used
to effect contaminant containment. Immobilization of
contaminants in solidified wastes may be a result of
chemical reaction, physical entrapment, or adsorption to the
solidified matrix.
A common approach to evaluating the performance of a solidi-
fied waste is the attempted simulation of disposal site con-
ditions (related to climate, geology, etc.) with laboratory
testing. This type of approach has a number of disadvan-
tages:
-------
1) It is difficult to simulate field conditions with lab—
scale tests.
2) One test (or a small number of tests) cannot be
applicable to many disposal site conditions,
3) It is impossible to simulate the effects of long
periods of time, and
4) Specific reasons for the qualitative performance of the
waste form are not determined.
Another approach is to attempt to characterize the intrinsic
properties of the solidified waste in terms of the chemical
and physical factors which affect the leachability of the
waste form:
1) The degree of chemical immobilization of the
contaminants, i.e., the contaminants may be chemically
bound in a variety of compounds of varying solubilities
and reactivities, or they may be physically adsorbed or
trapped in the solidified waste matrix. The mechanism
of immobilization affects the response of the waste to
different groundwater characteristics, particularly
with regard to pH.
2) The potential for contact of the groundwater with the
waste, i.e., the hydraulic conductivity of a waste, as
well as that of its surroundings, determines whther
mobile contaminants are transported throuah the solidi-
fied waste matrix by advection or diffusion.
Resistance to weathering, and compressive strength may
also affect the surface area of waste in contact with
the leaching groundwater.
The approach of intrinsic property deteminatiori has ad-
vantages corresponding to the disadvantages of the first ap-
proach:
1) Wastes with particular intrinsic characteristics can be
matched to the disposal scenarios to which they are
most suited,
2) Intrinsic characteristics may be used as source terms
in mathematical models for estimating the effects of
long periods of time, and
3) The specific nature of a waste form’s qualitative
characteristics will be known, and subject to improve-
ment.
The protocol of twelve short—term laboratory tests proposed
in this study is designed to characterize the intrinsic
physical and chemical properties of solidified wastes.
-------
The results from the assessment protocol are not intended to
be interpreted directly in terms of site—specific environ-
mental impact. This would require additional information
regarding disposal site conditions and long—term stability
of the waste (from accelerated testing and mathematical
modelling).
TESTING
The twelve tests in the assessment protocol may be classi-
fied as physical or leaching tests as listed in Table 1.
Table 1
THE TEST METHODS(2)
Physical Tests Leaching Tests
Bulk Density Sequential Chemical Extraction
Water Content Equilibrium Leach
Solids Specific Gravity US EPA Toxicity Characteristi.c
Permeability Leaching Procedure (TCLP)
Unconfined Compressive Acid Neutralization Capacity
Strength Dynamic Leach
Freeze/Thaw Weathering
Wet/Dry Weathering
The physical tests and the TCLP are discussed elsewhere (3).
A short description of the remaining leach tests follows:
SEQUENTIAL CHEMICAL EXTRACTION
This procedure is used to examine the speciation of heavy
metal contaminants in a waste. Knowledge of the bonding
characteristics of the metals leads to an improved
understanding of the environmental leachant characteristics
that could affect their mobility.
The test is conducted by extracting a sample of ground waste
with media of increasing aggressiveness which separate the
contaminants into five fractions:
A) ion-exchangeable metal ions
B) hydroxides, surface oxides and carbonate bound metal
ions
C) metal ions bound to hydroxides and iron and manganese
oxides
D) metal ions bound to organic matter and suiphides
E) residual metal ions
-------
EQUILIBRIUM LEACH TEST
This test is designed to measure the effective aqueous solu—
bility of the contaminants.
A sample of waste is ground to —100 mesh and continuously
mixed with distilled water at a liquid to solid ratio of
4:1. To ensure that equilibrium is attained, the test is
run for seven days, after which the extraction liquid is
separated from the solids and analyzed for the contaminants
of interest.
ACID NEUTRALIZATION CAPACITY
The ability of a waste to neutralize acid is important
because metals tend to become more soluble in low pH
environments.
To measure a waste’s capacity for acid neutralization, a
sample of solidified waste is ground to —100 mesh and
divided into subsamples which are placed in extraction
bottles containing increasing amounts of acid. The bottles
are tumbled until equilibrium is attained and the pH of each
of the solutions is measured.
DYNAMIC LEACH TEST
A small cylindrical specimen of solidified waste is immersed
in distilled water. The leachant is replaced at intervals
calculated according to a simple diffusion model, such that
the mass of contaminant leached in each interval is the
same. The results are used to calculate an apparent diffu-
sion coefficient and a leachability index which are in-
dicators of the contaminant mobility through the matrix
under diffusion control.
OTHER TESTS
In addition to the above tests, micromorphological and
microchemical characterization of the solidified products
for selected wastes will be carried out at Louisiana State
University (LSU), using chemical extractions, X—ray powder
diffraction, energy dispersive X—ray analysis (EDX), and
scanning electron microscope analysis (SEM).
STUDY OUTLINE
The study involves numerous participants with responsibili-
ties as listed in Table 2.
Five raw wastes were chosen for use in the study based on
suggestions from all participants.
An attempt was made to select wastes which would:
-------
1) Provide a comparison to results from previous work
undertaken at Environment Canada’s Wastewater
Technology Centre (WTC) (4) and U.S. Army Corps of
Engineers’ Waterways Experiment Station (WES) (5),
2) Include a range of wastes routinely solidified by a
number of vendors of solidification technology,
3) Include wastes containing problem contaminants, and
4) Maintain the international nature of the cooperative
program.
Table 2
PARTICIPANTS AND THEIR RESPONSIBILITIES
Environment Canada, — Project management
Conservation and Protection,
Industrial Programs Branch
USEPA Office of Solid Waste — Engage Dynamac Corporation
— Engage Radian Corporation
— Peer review of report
USEPA Water Engineering — Sponsor this and complimen—
Research Laboratory tary project at Waterways
Experiment Station
Dynamac Corporation — “third party” acting as
intermediary between testina
labs and industrial partici-
pants to protect proprietary
information
Radian Corporation — Assist with raw waste
collection and characteri—
zat ion
Ontario Ministry of the — Assist with raw waste
Environment collection
Laboratory Services Branch
-------
Table 2
PARTICIPANTS AND THEIR RESPONSIBILITIES (Cont’cJ)
Environment Canada,
Conservation and Protection,
Wastewater Technology
Centre (WTC)
Alberta Environmental
Centre (AEC)
U.S. Army Corps of
Engineers
Waterways Experiment
Station (WES)
Louisiana State University
(LSU)
Dept. of Mechanical
Engineering
— Assist in project management
— Provide background data
— Participate in preliminary
study of interlaboratory
rep rod uc lb i 1 i ty
— Assist with raw waste
collection and chatacteri—
zation
— Perform testing on
solidified wastes
- Major contribution
report preparation
— Provide background data
— Organize and participate in
preliminary study of inter—
laboratory reproducibility
— Assist with raw waste
characterization
— Perform testing on raw and
solidified wastes
— Major contribution to final
report preparation
— Participate in preliminary
study of interlaboratory
reproducibility
— Peer review of report
— Carry our microstructural
and microchemical
characterization
Industrial Participants
- Perform
wastes
solidification
of
The fully characterized wastes
Radian Corporation and WES.
contaminants of interest for
described in Table 3.
raw and
to final
were
provided by
the
WTC,
The
five wastes
and
the
the
leaching
tests
are
-------
Table 3
WASTES AND CONTAMINANTS OF INTEREST
Contaminants of
Approximate
Waste Interest
Concentration
(ppm, wet
weight basis)
WTC Synthetic Solution Arsenic 2600
Cadmium 4400
Chromium 1600
Lead 8300
Phenol 3600
WES Synthetic Sludge Cadmium 4700
Chromium 22000
Mercury 400
Nickel 22000
Aluminum Coil Aluminum 24000
Plating Waste Arsenic 90
Chromium 1000
Lead 100
Thallium 50
Cyanide 2000
Dredge Spoil Chromium 100
Copper 70
Lead 140
Mercury 0.3
Zinc 700
Polychior mated
Biphenyls (PCB) 1
Wood Preservation Soil Pentachiorophenol 5500
Polyaromat ic
Hydrocarbons (PAH) 300
Arsenic 100
Lead 40
-------
A five gallon (U.S.) pail of each waste and information
regarding its chemical and physical characteristics were
sent by the waste collector to each of the fifteen
industrial participants, at a rate of one waste per month.
The industrial participant was given one month to experiment
with the waste, and was then sent a solidification kit (con-
taining molds and directions for specimen preparation) from
the third party, Dynamac Corporation.
It is the task of Dynamac to protect proprietary informa-
tion. The companies are participating under the agreement
that their identities will remain unconnected with any par-
ticular solidified products unless they choose otherwise.
After solidifying the waste by their desired process,
according to instructions provided in the solidification
kit, the industrial participant allowed the solidified
material to cure for 28 days on site before shipping it to
Dynamac.
Upon receipt of the solidified products, Dynamac relabelled
the product of each industrial participant with a uniaue
secret code, and distributed the samples to either WTC, AEC,
WES, or LSU for testing. In general, testing was initiated
after 56 days of curing, but some departures from the
intended schedule occurred. Figure 1 shows the schedule of
activities for the study.
WTC, AEC, and WES are applying the protocol of 12 test
methods to the solidified products, such that each test is
run in duplicate for each solidified product by two labora-
tories. Quality control samples consisting of blanks,
duplicate standards and split samples for each waste are
also being analysed in each laboraboty.
LSU is carrying out microstructural and microchemical char-
acterization for the solidified products from the Wood
Preservation Soil and the WES Synthetic Sludge.
After testing, the experimental results will be collected
and analysed by WTC. Dynamac will send each industrial par-
ticipant the results for their particular solidified
products to give them the opportunity to choose to have
their products identified with their corporate name in the
final report.
CURRENT STATUS
Testing of the solidified products is expected to be
completed by the end of August, 1986, with chemical analysis
of leachates to be finished by the end of October (Figure
1). The final report for the study will be made available
in May, 1987 at the Fourth International Hazardous Waste
Symposium on Environmental Aspects of Stabilization!
-------
Figure I: PROGRAM SCHEDULE - INVESTIGATION OF TEST METHODS FOR SOLIDIFIED WASTE CHARACTERIZATION
DATE 1985 1986 1981
ACT’v”v I £ I I A I I ft I I I C I N I I I I I I I I I I I A I N I A I I I C I I $ A I I S
‘‘ ‘ ‘U I S I SW, u sYI ’ I”I”I •V V5I 5 IVIIIIVIVIU 1 fl , 1H 5
Shipent of raw waste
and raw waste characteristIcs
to industrial participants
I
3 4
2
5
Shipsentof
solidification kits
to industrial partIcipants
I 3 1 5
2
SolidifIcation
I
I
I
—I— 3
4 5
Shipsent of solidified
saaples to testing
facIlities
1
:
2
3
4
5
I
(2)
(3)
Testing begins
1
1
:
2 3
4
5
i
(2)
(3)
TestIng ends
1
I
I
1315
Results sent to Industrial
participants for acknowledgesent
1
I
54231
• Response
FINAL REPORT
I
Where ni ers I to S refer to the following wastes: I. WTC Synthetic Solution
2. Dredge Spoils
3. AlulInu, Coil Plating Waste
4. Wood Preservation Soil
5. WES Synthetic Sludge
( ) 8rackets refer to sespies received late
-------
Solidification of Hazardous and Radioactive Wastes to be
held in Atlanta, Georgia.
In the course of the study, the assessment protocol will
have been applied to almost 300 samples, representing 69
different solidified products. This will provide a thorough
validation of the test methods under investigation and a
solid data base of solidified waste properties.
ACKNOWLEDGEMENTS
The authors would like to thank Mr. Trevor Bridle and Dr.
Pierre Cote from Environment Canada’s Wastewater Technology
Centre, Mr. Peter Hannak from Alberta Environmental Centre,
the Environmental Engineering Group from US Army Corps of
Engineers’ Waterways Experiment Station, Dr. Harvill Eaton
from Louisiana State University, and Dr. Carleton Wiles fom
US EPA for their contributions to the program. Our thanks
also to the industrial participants for the solidified waste
products.
REFERENCES
(1) Cathcart, N., “Investigation of Test Methods for Soli-
dified Waste Characterization: A Cooperative Program”,
Presented at the 6th National Conference on Waste
Management in Canada , Vancouver, B.C., November 5—7,
1984.
(2) EPS Wastewater Technology Centre/Alberta Environmental
Centre, “Test Methods for Solidified Waste
Characterization”, Draft Report, January, 1986.
(3) Hannak, P., Liem, A.J. and Cote, P., “Methods for
Evaluating Solidified Waste”, Presented at the USEPA
Second Annual Symposium on Solid Waste Testing and
Quality Assurance , Washington, D.C., July 15—18, 1986.
(4) Cote, P.L. and Hamilton, D.P., “Leachability Comparison
of Four Hazardous Waste Solidification Processes”,
Proceedings of the 38th Annual Purdue Industrial Waste
Conference , May 10—12, 1983, pp. 221—231.
(5) Jones, J.N., Bricka, R.M., Myers, T.E. and Thompson,
D.W., “Factors Affecting Stabilization/Solidification
of Hazardous Waste”, Proceedings: International
Conference on New Frontiers for Hazardous Waste
Management , Pittsburgh, PA, September ]5—18, 1985,
USEPA 600/9—85/025.
-------
SIXTH SESSION
Thursday, July 17, 1986
7:30 p.m. - 9:30 p.m.
Chairperson:
Kenneth Jennings
Environmental Scientist
Office of Waste Program
Enforcement
USEPA
401 “M” Street, S.W.
Washington, D.C. 20460
-------
CORRECTIVE ACTION UNDER RCRA, INTERIM MEASURES
JACQUELINE MOYA AND KENNETH JENNINGS, OFFICE OF WASTE
PROGRAMS ENFORCEMENT, U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C.
ABSTRACT
Section 3008(h) of RCRA gives EPA authority to issue
Administrative Orders or seek judicial relief requiring cor-
rective action or send other response measures deemed neces-
sary to protect human health or the environment. This
authority is based on information that a release has or is
occurring into the environment from a facility, and provides
for the broad applicability of corrective actions and
response measures.
Extensive and time consuming investigations may be required
to develop a comprehensive corrective measures study (CERCLA
Remedial Investigation equivalent) for a facility. During
this period, the release could continue unabated which could
allow the spread of contamination or the continuance of con-
ditions that may endanger human health or the environment.
Interim measures are actions that should be taken in advance
of longterm remedial measures to prevent releases or addi-
tional contamination, and to reduce, abate or remove the
exposure threat presented by releases.
INTRODUCTION
One of the administrative authorities granted the United
States Environmental Protection Agency by the Hazardous and
Solid Waste Amendments of 1984 is the 3008(h) corrective
action authority. This authority is extremely important in
that it empowers EPA and eventually the States to order the
owner-operator of a facility which has released contamina-
tion to the environment to undertake action(s) to arrest and
reverse the effects of that release in any environmental
medium. The nature of releases is often extremely complex
and potentially far reaching, especially where several media
are involved. It is, therefore, prudent often times to take
immediate action to remove the source of contamination or to
take immediate action to remove the source of contamination
or impede its progress until a final, comprehensive solution
is designed. Immediate actons of this kind are called
interim measures. The guidance developed by OWPE draws
heavily from the experience of the Office of Emergency and
Remedial Response in implementing the immediate/planned
removal portions of the National Contingency Plan. What
follows is a summary of that guidance.
-------
TYPES OF INTERIM MEASURES
Corrective action orders should incorporate actions (interim
measures) necessary to protect human health or the environ-
ment. Interim measures are actions that should be taken in
advance of long term corrective measures to prevent releases
or additional contamination, prevent or reduce the further
spread of contamination, and reduce, abate or remove the
exposure threat presented by releases. During the selection
of an interim measure, the Agency should consider the magni-
tude of the potential threat to human health or the environ-
ment. The Agency’s authority to seek relief by requiring an
owner/operator to perform specified activities is directly
correlated to the protection of human health or the environ-
ment. Therefore, if the threat is minimal or the risk has
yet to be determined, simple monitoring of ground water,
surface water, soil or air may be the types of action
required. For example, if a release to ground water is
minimal and the aquifer is not used by the nearby popula-
tion, a program to pump and treat may not be appropriate.
If the threat is greater or as more information becomes
available through initial or additional sampling and
analysis, more serious actions should be contemplated either
by incorporating actions into a single “phased” order or by
issuing separate orders.
Attached is a list of some possible interim measures. It
was compiled from several actions and past CERCLA remedial
guidance.
INTERIM MEASURES
Containers
a) Overpack/re-drum
b) Construct storage area/move to storage area
c) Segregation
d) Sample/analyze and dispose
e) Excavation/disposal
f) Temporary cap
Surface Impoundments
a) Reduce head
b) Remove free liquids and/or highly mobile wastes
c) Stabilize/repair side walls/increase
freeboard/install geotextile
d) In-Situ solidification
e) Cover (control air release or overflow due to
rain)
f) Interim ground water measures
g) Run-off/run-on control (diversion or collection
devices)
-------
h) Document the concentration of constituents left in
place when a surface impoundment handling charac-
teristic wastes is clean closed*
* When this type of surface impoundment is clean closed,
there may be constituents left in place. Some of these
health or the environment (e.g., corrosive waste may
contain heavy metals).
Landfill
1) Run-off/run-on control (diversion or collection
devices)
b) Reduce head on liner and/or in leachate collection
sys tern
c) Repair leachate collection/removal system or
french drain
d) Install new leachate collection/removal system or
french drain
e) Temporary cap/cover (asphalt, synthetic or clay)
f) Interim ground water measures
g) Excavation/disposal
Waste Pile
a) Run-off/run-on control (diversion or collection
devices)
b) Cover (polymeric membrane, geotextile or clay)
c) Solidification
ci) Interim ground water measures
e) Removal of the waste pile for more secure storage
f) Excavation/disposal
Ground Water
a) Sampling and analysis
b) Delineation of plume
c) Interceptor trench/sump/french drain
d) Pump and treat/in-situ treatment
e) Cut-off walls (slurry or bentonite)
Surface Water Release (point vs non-point)
a) Overfiow/underfiow dams
b) Filter fences
c) Run-off/run-on control (diversion or collection
devices)
ci) Regrading/revegetation
e) Cover with geotextile
f) Sample and analyze surface waters and sediments or
point source discharges
-------
Tanks
a) Leak or cracks detection/repair
b) Relining
c) Partial or complete removal
d) Pipeline removal or replacement
e) Secondary containment
Soils
a) Sampling and analysis
b) Run-off/run-on control (diversion or collection
devices)
c) Temporary cap/cover
d) Excavation/disposal
Gas Migration Control
a) Pipe vents
b) Trench vents
c) Gas barriers
d) Gas collection system
e) Gas treatment system
f) Gas recovery
g) Air monitoring sytem
Particulate Emissions
a) Truck wash (decontamination unit)
b) Re-vegetation
c) Application of dust supressant
Other Type5 of Action
a) Fencing to prevent direct contact
b) Alternate water supply to replace contaminated
drinking water
c) Temporary relocation of exposed population
d) Extend contamination studies to off-site areas
e) Other actions necessary to protect human health or
the environment
f) Temporary or permanent injunction
g) Suspend or revoke authorization to operate under
interim status
-------
SEVENTH SESSION
Friday, July 18, 1985
8:00 a.m. - 12:00 p.m.
Chairperson:
J. Howard Beard
Chief
Physical Science Section
Office of Waste Program
Enforcement
USE PA
401 “M” Street, S.W.
Washington, D.C. 20460
-------
ISSUES AND PROBLFI4S RELATED TO CERTIFICATION
AND PERFORMANCE EVALUATION OF hAZARDOUS WASTE
TESTING LABORATORI ES
FRED SETO, PH.D., KUSUM PERERA, PH.D., CINDY DINGMAN, BARTON
SIMMONS, AND ROBERT STEPHENS, PH.D., CALIFORNIA PUBLIC
HEALTH FOUNDATION, BERKELEY, CALIFORNIA
ABSTRACT
In the course of performing its duties as a lead agency to
administer the California Hazardous Waste Control Program,
the California Department of Health Services has received
several hazardous wastes analysis reports which showed
dubious analytical results. Consequently, a program for
certification of harzardous waste testing laboratories was
mandated in 1982 with the primary objective to improve the
quality of laboratory data. The program was implemented in
April, 1985. The response from the hazardous waste testing
laboratories has been enthusiastic. As of June, 1986, the
Department has received 115 certification applications
including 8 laboratories from outside California. Some
issues and problems are as follows:
(1) Number of Parameters and difference matricies —The
large number of analytical parameters and various
matricies require specific analytical methods or
modified methods in terms of sample preparation and
instrumentation conditions.
(2) Test Categories — The available categories for
certification do not include all tests required of
laboratories for waste classification and monitoring.
Many of the required methodologies have not been
validated.
(3) Proficiency Test Samples — Analyses of proficiency
test (PT) samples by an applicant laboratory provide
some indication of the laboratory’s performance
capability. However, the task of preparation and
validation of a PT sample library is enormous
considering the large number of parameters, matrices,
and concentrations.
(4) On—Site Visit — On site inspection of an applicant
laboratory’s operation and activities is a useful
measure of its capability.
-------
Generally, the program has been feasible and successful in
terms of improving the quality of data generated by
hazardous waste testing laboratories. Experience in the
establishment and conduct of this program will be presented.
-------
METHODOLOGY FOR THE D ERMINATION OF PRENALYTICAL
HOLDING TIMES
M.P. MASKARINEC, ANLYTICAL CHEMISTRY DIVISION, OAK
RIDGE NATIONAL LABORATORY
ABSTRACT
Because of the importance of analytical methodology to
public decision making in the environmental area, it
has become necessary to apply strict quality control
and quality assurance precedures to all aspects of the
aspects of the process, background data or valid
statistical procedures are available to assist in the
development of appropriate quality control procedures.
However, the period of time from which the sample is
collected until it is analyzed has largely been left to
empiricism, particularly with respect to organic
analysis. This paper is a discussion of the
experiments ongoing in an attempt to clarify and
document the preanalytical holding time associated with
the analysis of organic compounds in water and soil.
The experimental design consists of the analysis (by
USEPA methods) of volatile organic compounds and
semivolatile nitroorganic compounds in three water and
three soil matrices under three storage conditions
until all analytes fall outside the 90% confidence
limit of the analysis of the original sample.
In order to perform this experiment, methods have to be
developed for the preparation of a large, homogeneous
volume of original sample which can be aliquotted into
individual storage containers with a precision of less
than 5%. For volatile organic compounds in water, a
method has been developed which consists of filling a
Tedlar gas sampling bag with water, adding a methanolic
solution of volatiles, mixing, and aliquotting into
standard VOA vials. This procedure produces aliquots
which are about ± 3% by analysis. The method for soil
sample preparation involves precise weighing of a dried
soil sample, and addition of a known volume of water
prepared in the manner described above. For the
semivolatiles, ethanolic solutions of the compounds are
added to volumetric flasks containing the water and
aliquotted directly. Soil samples are prepared using a
solution of the semivolatiles in diethyl ether, with
the ether being allowed to evaporate.
Preliminary data on the holding time of various
analytes will be presented, as will the complete
experimental design. Precision and accuracy data for
the various preparation methods will be presented.
-------
HAZARDOUS WASTE REGULATIONS IN CANADA
N.L. PORTER-CATHCART, WASTE MANAGEMENT DIVISION, INDUSTRIAL
PROGRAMS BRANCH, CONSERVATION AND PROTECTION ENVIRONMENT,
CANADA
ABSTRACT
The Constitutions of Canada and the United States contain
similar provisions which divide legislative powers between
federal and state or provincial governments. Canadian pro-
vinces however, have a great deal more autonomy and au-
thority over many affairs than do their American counter-
parts. In Canada, the provinces exercise considerable con-
trol over land and natural resources and thus, their role in
environmental matters is extensive. They have significant
environmental responsibility and use several legislative and
regulatory instruments to carry out that responsibility.
The Canadian Council of Resource and Environment Ministers
(CCREM), is composed of federal and provincial Ministers of
the environment and natural resources. CCREM meets
regularly to discuss the state of the environment and to
reach consensus on those areas affecting both levels of
government. It reflects the division of powers governing
environmental legislation in Canada and the need for co-
operative action.
In 1978, in response to growing concern over the potential
environmental damage posed by uncontrolled disposal of
hazardous waste, CCREM recommended that the federal govern-
ment, in consultation with the provinces, initiate a program
to develop a transboundary control system to track hazardous
waste movements. Transborder environmental matters are
clearly federal jurisdiction and hence, the only appropriate
source of regulation was the federal government.
Accordingly, the Canadian government has developed legisla-
tion to control the international and interprovincial move-
ments of wastes. This has been done under the authority of
the Transportation of Dangerous Goods Act, 1980. After ex-
tensive federal/provincial consultation, the regulations
implementing the Act finally came into force July 1, 1985.
Two key areas covered are hazardous waste listing and
“cradle to grave” tracking (manifesting) for hazardous waste
movements. The hazardous waste listing system is based on a
combination of the existing hazard identification procedures
-------
and lists of hazardous and/or toxic substances, currently
widely used to describe hazardous materials in the trans-
portation field. The addition of other criteria/lists that
ensures appropriate and complete identification of all
environmentally hazardous properties of wastes completes the
system. Once listed or caught by a characteristic described
by the Transportation of Dangerous Goods Regulations, the
waste is defined as hazardous and becomes subject to the
prenotification (international only) and manifest controls
accordingly. Provincial legislation regulates the manage-
ment options at the waste’s ultimate destination.
At the present time there are four provincial control
systems for chemical wastes in Canada, which are for the
most part complementary to the national system. Some pro-
vinces are maintaining their own systems in the interim,
others have already or will soon be adopting the national
approach. All provinces agree that the ultimate goal is to
refine the existing and proposed system, and attain one
nation—wide system which will effectively address all
hazardous waste movements.
-------
OVERVIEW OF RCRA ENFOR MENT
KENNETH JENNINGS, OFFICE OF SOLID WASTE, U.S. ENVIRONMENTAL
PROTECTION AGENCY, WASHINGTON, D.C.
ABSTRACT
The Office of Waste Programs Enforcement (OWPE) is part of
the Office of Solid Waste and Emergency Response (OSWER).
The charge of OWPE is the enforcement of RCRA and CERCLA
which is accomplished through its two divisions. This
discussion is focused on RCRA enforcement. The various
administrative orders and procedures are discussed in some
detail.
INTRODUCTION
The Office of Waste Programs Enforcement (OWPE) is one part
of the larger Office of Solid Waste and Emergency Response
within the United States Environmental Protection Agency.
It is the charge of OWPE to enforce the regulations
pertinent to the Resource Conservation and Recover Act of
1976 (RCRA), includina the Hazardous and Solid Waste
Amendments of 1984, and the Comprehensive Environmental
Response, Compensation and Liability Act of 1980 (CERCLA).
On a more practical level, OWPE provides leadership in
policy and technical matters of enforcement to the EPA
Regions and States and strives to promote the coordination
of enforcement activities with those of the other offices
within OSWER. Of particular relevance to this symposium is
the effort of OWPE and the Office of Solid Waste (OSW) to
coordinate the needs and objectives of enforcement with the
promulgation of regulations and the issuance and maintenance
of permits under RCRA.
OWPE is divided into two divisions for RCRA and CERCLA. The
RCRA division is comprised of two branches, Compliance and
Implementation, and Guidance and Evaluation. The Compliance
and Implementation Branch provides technical support to the
Regions and States on specific enforcement cases and
facilitates the implementation of enforcement quidance in
the field. This Branch also manages the coordination of
Regional enforcement activities with the objectives of the
RCRA Implementation Plan. The Guidance and Evaluation
Branch (GEB) has two major duties. it is largely
responsible for working with OSW to ensure that regulations
promulgated are enforceable. Secondly, this Branch is
responsible for the development of technical and policy
guidance for use by EPA Regions and States. Two notable
examples are the RCRA Ground—Water Monitoring Technical
Enforcement Guidance Document (draft) and Compliance Order
-------
Guide (final). The former provides technical guidance to
compliance/enforcement personnel and permit writers on what
constitutes a compliant ground—water monitoring system. The
RCRA Ground—Water Monitoring Compliance Order Guide is a
description of how regulatory compliance should be sought
for facilities with inadequate ground—water monitoring
systems.
OWPE has several administrative order authorities at its
disposal with which to compel owner or operators to comply
with the regulations. The selection of which authority or
combination of authorities to use should be based on case
specific considerations (see the RCRA Ground—water
Monitoring Compliance Order Guide for more details). The
3008a, 3008h, and 3013 authorities are summarized below.
§3008(A) ORDERS
A 3008(a) order may be issued only for violation of one or
more Subtitle C requirements. Therefore, when enforcement
personnel and the permit writer determine a facility’s
ground—water monitoring program to be technically
inadequate, enforcement personnel should determine whether
any of the technical inadequacies constitute violations of
Part 265 Subpart F, Part 270, or Part 264.
In some cases the regulations are specific as to what
findings of fact would indicate violations. For example, if
an owner/operator has installed only two downgradient wells,
the facility is clearly out of compliance with 265.91(a)(2)
of the regulations, the section that requires installation
of at least three downgradient wells. Likewise, if a
facility does not have some of the records specified in the
regulations (e.g., an assessment outline), or has not
performed some of the required analyses, then the owner is
clearly in violation. The decision concerning the existence
of a violation becomes more involved when it is based upon
evaluating the adequacy of a facility’s ground—water
monitoring system beyond the minimum requirements.
In great part, the heightened level of analysis required to
evaluate the overall adequacy of a system evolves from the
regulations’ reliance on broad performance standards. Given
the great variability between sites in terms of wastes
handled, hydrogeology, and climate, it is impossible to
design a regulatory system that defines for all cases
exactly what constitues an adequate ground—water monitoring
program. As a result, the Agency relies on performance
standards to define “adequate.”
The performance—oriented provisions of Subpart F set high
standards for interim status ground—water monitoring
systems, and enforcement personnel should not underestimate
the power and applicability of this language. For example,
-------
even though the regulations establish a minimum of one
background monitoring well, a single well is seldom
sufficient because owner/operators must design their systems
to meet the background—well performance standard listed in
265.9l(a)(l). Section 265.9l(a)(1) requires appropriate
locations and depths to yield samples representative of
background water quality not affected by the facility. If a
facility’s well array does not meet this standard, the
owner/operator is out of compliance with the regulations.
5 3013 ORDERS
Section 3013 orders may be issued to a facility only when
the Administrator determines that the presence or release of
hazardous waste at the facility may present a substantial
hazaard to human health or the environment. The facility
need not be violating RCRA regulations to qualify for action
under 3013.
Prior sampling of contamination is not necessary to support
a 3013 order. In the case of a facility that has not
conducted any ground—water monitoring activities, the
potential for release of hazardous waste, the nature of the
site’s underlying hydrogeology and the proximity of an
aquifer or populated area will usually be sufficient, with
expert opinion, to support a 3013 order. In some cases,
the Region may wish to use 3007 authority to sample one or
more wells at a facility in order to provide direct evidence
of a release. Given that direct evidence is often
unnecessary to establish the applicability of 3013, the
Region should probably avoid direct sampling unless it is
confident that existing wells will intersect the suspected
plume. Guidance issued Septembr 26, 1984 provides further
discussion of the grounds for issuance of 3013 orders.
(See memo from Courtney Price and Lee Thomas entitled,
“Issuance of Administrative Orders Under Section 3013 of the
Resource Conservation and Recovery Act”).
S3008(H) ORDERS
Section 3008(h) of RCRA provides that the Administrator may
issue an order or file a civil suit requiring corrective
action or other appropriate response measures whenever (s)he
determines that there is or has been a release of hazardous
waste into the environment.
As described in the September 1985 draft guidance on the
scope and use of 3008(h), the Agency is interpreting the
term “release” to include any spilling, leaking, pumping,
pouring, emitting, erupting, discharging, injecting,
escaping, leaching, dumping, or disposing into the
environment. To show that a release has occurred, the
Administrator does not necessarily need sampling data. Such
evidence as a broken dike at a surface impoundment should
-------
also support a determination that a release has occurred.
In some cases, information on the contents of a land
disposal unit, along with information on the contents of a
land disposal unit, along with information on the site
hydrogeology and the design and operating characteristics of
the facility may be enough for an expert to conclude that a
release has occurred.
Section 3008(h) orders (and civil suits) may be used to
address releases not only to the ground water, hut to other
media as well. The draft 3008(h) guidance states that the
authority covers releases of hazardous wastes into surface
water, air, the land surface, and the sub—surface strata.
The term “hazardous waste” is not limited to those wastes
listed for identified in 40 CFR Park 261. For 3008(h)
purposes, the term hazardous waste also includes the
hazardous constituents identified in Appendix VIII of Part
261. Section 7003 orders may be used in the event of an
imminent and substantial hazard. CERCLA section 106 and 104
orders may also be used on a limited basis at RCRA—managed
facilities to take long term or immediate corrective actions
respectively.
The RCRA Division of OWPE is dedicated to the strict, timely
and constructive enforcement of the RCRA regulations. OWPE
has been and will continue to be a resource to the Regions
and States in both technical and policy areas. Through a
close working relationship with OSW, OWPE envisions
continued progress in safe guarding human health and the
environment from RCRA—managed hazardous waste.
-------
SN-846 UPDATE
PAUL FRIEDMAN AND DENISE ZABINSKI, OFFICE OF SOLID WASTE,
U.S. ENVIRONMENTAL PROTECTION AGENCY, WASHINGTON, D.C.
ABSTRACT
This will be an overview of SW-846, the Office of Solid
Waste’s sampling and analysis manual, giving special
consideration to format and content. It will also include a
discussion of the relationship between the guidance and
regulations from various OSW programs and the methods. This
will present OSW’s concept for the further development of
the manual and discuss issues concerning the methods.
-------
TESTING ISSUES ATTENDANT TO IMPLEMENTATION
OF THE RCRA REGULATORY PROGRAM
DAVID FRIEDMAN, OFFICE OF SOLID WASTE, U.S. ENVIRONMENTAL
PROTECTION AGENCY, WASHINGTON, DC
ABSTRACT
As part of its efforts to implement the 1984 amendments to
the Resource Conservation and Recovery Act (RCRA) and to
improve the implementability of the RCRA regulatory program,
EPA is developing increasingly more quantitative property
based regulatory definitions. Such regulations include the
expanded Toxicity Characteristic which was recently proposed
and OWS’s efforts to develop concentration based listings.
Implementation of such regulations brings with it a number
of technical issues and questions. This paper will address
several of the testing issues including:
1. when and how often should a facility test its
residuals to insure compliance with the
regulations;
2. how should the Agency establish regulatory
thresholds when the sensitivity of the analytical
methods are the controlling factor;
3. what type of quality assurance/quality control
program should be incorporated into the RCRA
regulatory program.
INTRODUCTION
Measurement plays a key role in all aspects of the hazardous
waste management program. As the program becomes
increasingly more quantitative with performance criteria for
facility operation and for identifying hazardous and banned
wastes the critical role of the procedures employed to
determine a given property will only increase. Questions
such as whether or not a waste is a hazardous waste; is the
waste banned from land management; is the permeability of
the facility liner sufficiently low to prevent leakage; is
the facility leaking; is the remedial engineering activity
accomplishing the job of cleanup as it was designed to do.
These are all questions which require that accurate measure-
ments be taken in order for the proper decisions to be made.
WASTE ANALYSIS
The first question facing generators and operators of waste
management facilities is whether the waste they are dealing
with is a hazardous waste within the meaning of RCRA.
Answering this question raises several issues. These
include:
-------
— When is testing required and under what circumstan-
ces can other knowledge of the waste be used to
classify the waste,
— What methods to use in sampling and testing the
waste,
— How much testing is enough (i.e., how many samples
need be taken to insure that the data is representa-
tive of the waste and, how often does the testing
have to be repeated due to process variability),
When is Testing Required When Characterizing a Waste ?
EPA does not require testing, instead it requires that a
good faith evaluation of the waste’s properties be made
before the waste is stored, treated, or disposed of.
However, what constitutes a good faith effort, or when
should one test is a constantly occurring question.
The RCRA regulations (40 CFR 262.11) require that persons,
other than households, who generate wastes must evaluate
their wastes to determine if they meet the definition of a
hazardous waste. In addition to testing to determine if a
waste is a hazardous waste, testing is also required of per-
sons treating, storing, or disposing of hazardous waste to
insure that the operations are not resulting in harm to the
environment.
EPA’s RCRA regulations have generally not required test
data. Rather, threshold levels of each property have been
established and it has been left up to the regulated commu-
nity to determine whether testing is needed to insure com-
pliance with the applicable standard.
This approach, offers the regulated community a great deal
of flexibility. However, with the flexibility comes the
responsibility to perform an adequate evaluation. Whether
or not to test then depends on how much information one has
on the waste or waste generation process. Specifically,
what raw materials are used, what contaminants might reason-
ably find their way into the waste during product purifica-
tion (e.g., distillation). If one can, through chemistry
principles and process engineering considerations, be
certain that none of the toxic constituents could reasonably
be expected to be present in the waste at a level which
would exceed the regulatory thresholds, then one has demon-
strated that the waste is not hazardous or banned from land
disposal. If, on the other hand, one does not have suffi-
cient data on the material to defend one’s judgment, then
testing should be done. The amount of testing then becomes
a function of one’s knowledge of the process and any
preliminary testing results.
-------
Under the framework proposed for the Toxicity
Characteristic, the determination of whether a waste is a
hazardous waste continues to depend on whether the concen-
trations of constituents in the TCLP extract exceed the
applicable regulatory levels. Given the importance that
hazardous waste identification plays in the regulatory
program, EPA is re—evaluating the above mentioned policy and
deciding whether or not to require periodic waste testing.
EPA is considering three general approaches to such a test-
ing requirement. First, EPA could continue just to require
generators to evaluate whether or not their wastes exceed
applicable regulatory levels, but not specifically require
testing to make this determination. Second, EPA could
require testing of wastes at a frequency specified by regu-
lation. Third, EPA could require the generator to test,
while documenting the determination of the appropriate test-
ing frequency based on guidance provided by the Agency.
Although not requiring testing places the least burden on
the regulated community, EPA is concerned that this approach
may not promote voluntary compliance and that it could
hamper enforcement efforts against those members of the
regulated community who do not comply voluntarily with the
regulations.
Another possible approach is to require periodic testing,
specifying in the regulations both the method and the fre-
quency of testing. Thus, testing might be required on a
serninannual or annual basis. This approach would make en-
forcement of the regulations easier and would likely induce
a higher level of voluntary compliance since the regulations
would be highly specific with regard to what constitutes an
acceptable testing program and what actions and inactions
would constitute violations.
There are, however, several problems with such an approach.
First, there are problems inherent in specifying an
appropriate testing frequency. Eased on data from our own
industry studies efforts and data from the Office of Water’s
Effluent Guidelines Program, it is clear that many waste
streams are extremely variable in concentrations of chemical
constituents from one plant to another, even when the same
general process is employed. Variability exists not only
from one generator to another, but also spatially and
temporally within a single plant.
A third possible approach is to require generators to
perform testing on their wastes, but not to specify a test-
ing frequency in the regulations. Rather, generators would
be required to determine an appropriate testing frequency
based on guidance developed by EPA and to document, in their
records, this frequency determination. The advantage of
this approach is that process-specific factors could be
-------
taken into account in determining the appropriate testing
interval. Thus, although there would be some additional
burden on generators to determine, based on the guidance,
the appropriate frequency for testing tailored to specific
factors relating to his process, there would be less of a
chance of requiring unnecessarily frequent testing. This
approach does, however, present greater enforcement diffi-
culties than does the approach of specifying generic
periodic testing intervals.
How Much Testing is Needed ?
Irrespective of which approach is eventually adopted, the
problem of developing a testing program appropriate to the
waste at hand needs to be addressed. The problem remains as
to how to assure that the waste sample subjected to testing
is representative of both the batch and the process from
which they are derived. This problem arises not only in
regard to hazardous waste identification, but also in
connection with other waste sampling requirements.
In this paper, I will attempt to address one aspect of this
question; how does one determine how many tests to perform
to ensure compliance with the regulation? I will limit my
discussion today to testing for purposes of complying with
the Toxicity Characteristic and the Land Disposal Ban. I
will not attempt to delve into the detailed statistics of
testing. Instead, I will discuss the philosophy of testing
as it applies to the RCRA program. For additional informa-
tion on RCRA testing requirements, I refer you to the EPA
manual “Test Methods for Evaluating Solid Waste,” SW—846.
This manual contains a more thorough discussion of the RCRA
testing regulations.
Hazardous waste evaluation (e.g., is a waste hazardous, is
the waste banned from land disposal) is a problem not of
determining what the actual value of a property is but
rather whether it is above some defined regulatory
threshold. From the standpoint of testing, this is a
critical distinction since it is much less expensive to
answer this question than to detemine actual values.
Although this discussion will not emphasize statistics, the
amount of testing that should be conducted is greatly
influenced by the fact that the absolute value of the
property is not critical. Examining the question from the
point of view of the generator evaluating a waste, the ques-
tion becomes one of determining: How certain do I have to
be that the property does not exceed the threshold value?
The question for the analyst then becomes: How precisely
does one have to determine the value to answer the question
with a high enough degree of confidence?
The correct amount of testing is the minimum amount which
will avoid the legal and environmental consequences of an
-------
incorrect determination; taking into account the expense of
testing. As SW—846 describes, the Agency has taken the
position that evaluation to the 98% confidence level is suf-
ficient when determining whether or not a waste is a
hazardous waste. What this means is that persons who
conduct enough testing to demonstrate that they have attain-
ed this level of confidence will be considered to have made
a good faith effort to comply with the regulations.
The next aspect of the how often to test question revolves
around the problem of process changes. Wastes should be
reevaluated whenever a significant change is made in the
process generating the waste. However, when does a process
change become sufficient to be considered to be significant
and thus necessitate re—testing? This is another area where
no clear answers can be given. Whether or not the change is
significant hinges on whether the change could affect the
properties of the waste that is being evaluated. For
example, toxic elements are generally more leachable from
acidic wastes than from alkaline wastes. Thus, if one
changed from using an alkaline boiler cleaner to an acidic
one, then it would be prudent to re-evaluate, or even re-
test, the boiler cleaning sludges that are generated.
Assuming the question is one of whether or not the waste
exhibits th Toxicity Characteristic.
For the owner/operator of a facility that handles wastes
from many generators the problem is much more difficult.
Here the facility is not in control of the generation
process and has much less knowledge of the wastets proper-
ties. Before an owner/operator treats, stores or disposes
of any hazardous waste he must obtain a detailed chemical
and physical analysis of the waste. This does not mean that
the facility has to personally test each waste. An owner!
operator may elect to rely on data supplied by the customer
for the initial waste characterization. However, the waste
management facility remains liable for any incorrect charac-
terizations.
In addition, off—site facilities must inspect, and if neces-
sary, analyze each hazardous waste movement received at the
facility to determine whether it matches the identity of the
waste specified on the manifest. This inspection can be
anything from visual inspection coupled with physical
analysis to a fingerprint type analysis to in—depth analysis
testing.
Given the cost and delays attendant to waste testing and the
liability improper decisions can present to the facility,
the Agency has presented a method for facilities to use in
determining recharacterization frequency. It is intended
primarily for facilities to use in determining recharacteri—
zation frequency. It is intended primarily for facilities
that receive wastes from off—site generators, however, it
-------
may also be modified for use by on—site facilities. The
method allows the following criteria to be evaluated when
determining how often to recharacterize the waste:
— The potential for restricted wastes to be combined
in a waste shipment that is normally permitted.
— The design limitations of the hazardous waste
management process.
— The likelihood of the waste undergoing changes that
will affect its manageability.
— The prior history of the waste generator performance
and reliability.
weighing factors, ranging from one to five, are assigned to
each of these criteria to assess its relative importance.
After assigning weights, probabilities ranging from zero to
four should be chosen for each criterion indicating the
likelihood of a given generator and waste meeting that
criterion. For example, what is the likelihood of a
contracted waste having a restricted waste mixed in its
shipment. The criterion weight and probability are then
used to calculate the percent of a generator’s shipments
that should be recharacterized each year. Further details
on conducting these calculations can be found in “Waste
Analysis Plans — A Guidance Manual” EPA Publication No.
EPA/530—SW—84—012, October 1984 (GPO No. 055—000—00244—4,
Telephone: 202—783—3238)
In addition to the material in SW—846 and the Waste Analysis
Plan Guidance Manual, OSW will be developing a guidance
manual on representative sampling that will address these
concerns and anticipates publication in mid 1987.
Detection Limits vs. Quantitation Limits for use as Regula-
tory Thresholds
The leachate test levels that the Agency has developed for
use in the Land Disposal Ban and the Toxicity
Characteristic based on toxicological considerations range
from the ppm to the sub—ppb level. This presents a problem
for the Agency since some of these concentrations are below
the measurement range of the currently available analytical
methods.
EPA believes that the appropriate way to deal with this
problem is to establish analytical method based regulatory
levels. Such levels could be set at the analytical
detection limit or, as an alternative, they could be set at
the limits of accurate quantitation.
-------
Use of either detection limits or quantitation limits would
allow for regulatory levels that fall below the analytically
measurable level to be periodically updated as advances are
made in analytical methodology. In evaluating how to
resolve this issue, the Agency has to weigh the fact that
the limit of detection for a given analyte and method would
tend to vary more from laboratory to laboratory than would
the quantitation limit which is significantly higher and
thus freer from interferences and less analyst/instrument
sensitive.
FACILITY OPERATION MONITORING
One of the major issues facing the hazardous waste manage-
ment industry is how to monitor a facility’s operation to
insure that it both meets the terms of its permit and is not
causing environmental damage. Our ability to monitor many
of the factors affecting facility operation is in its
infancy. Techniques are expensive and, in some cases, are
of unproven accuracy. Issues facing us here include:
— For each type of opreation, what parameters need to
be monitored in order to insure adequate performance
of the process,
- What methods are available for such monitoring and
can less expensive methods be developed to
accomplish the same task.
When an applicant designs or applies for a permit for a
hazardous waste management facility, questions relating to
waste properties, facility applicability, process effective-
ness and operational quality control need to be addressed.
All these require measurement of one or more properties.
Attendent to these measurements are a number of additional
issues. These include:
— How do facilities insure that client wastes arriving
at the plant for treatment or disposal have the same
properties as those initially evaluated,
- What testing should be performed to prevent wastes
banned from land disposal from being disposed of in
such facilities,
— What testing should be required of storage facili-
ties, especially short—term shortage units or
transit accumulation units,
— What testing should be required of a treatment
facility (e.g., incinerator, stabilization unit,
chemical destructor) prior to issuance of a permit
to determine the boundary conditions of the process
(e.g., how to test to determine the range of
-------
constituents and concentrations that can be fed to
the incinerator without causing the destruction!
removal efficiency to drop below that which is
acceptable),
— How to determine the background environmental levels
(e.g., ground water quality) prior to construction
of the facility (e.g., thoroughness of testing) and
what to do when the quality (e.g., sensitivity) of
the monitoring methods improves after the facility
is in operation (e.g., should one have to reevaluate
the site or should a facility be allowed to continue
to use the old methods).
ENFORC ENT
How to insure that the data being gathered by the regulated
community is accurate enough for correct decision making.
The RCRA program is a self implementing program where the
EPA or authorized State establishes action levels and the
regulated community takes action based on the properties of
the waste or facility.
Given the precision of the various aspects of the testing
that is performed and the problem of obtaining truly repre-
sentative samples with limited sampling, how high an excur-
sion above the regulatory threshold should be permitted
before enforcement action is initiated.
Given the large number of facilities and factors that need
to be evaluated when making compliance inspections, what
types of tests should be used for screening facilities in
order to reduce the cost of the inspections.
Finally, given the complexity of waste testing, what type of
quality assurance/quality control program should be incorp-
orated into the RCRA program to insure that persons and
companies conducting such testing both know how to do the
tests, and properly perform such testing.
-------
POSTER SESSION
-------
LIQUID RELEASE TEST (LRT) FOR LIQUID
LOADED SORBENTS
BEN H. CARPENTER RESEARCH TRIANGLE INSTITUTE RESEARCH
TRIANGLE PARK, NORTH CAROLINA
ABSTRACT
The Liquid Release Test (LRT) for liquid—loaded sorbents is
needed to detect the potential release of held liquids when
sorbents are subjected to pressures from landfill
overburdens.
This presentation describes the relation of the test to the
different mechanisms by which liquids may be released from
sorbents, and the development of a commercially available
compressed—air driven consolidometer test unit. Two com-
mercially available Zero Fleadspace Extractors were shown to
be suitable for the test: the model 3740—ZHB ZHE made by
the Associated Design and Manufacturing Company, and the
model SD 1P58 1C5 ZHE made by the Millipore Corporation.
Initial liquid release ranges were defined for five sorbents
and five organic liquids. Sorbents tested were Floor Dry,
Florco, S-N—D, Fuller’s Earth and bituminous fly ash.
Liquids tested were aqueous calcium sulfate solution,
aqueous acetone solution, diesel fuel, trichioroethylene,
and xylene.
The test protocol was subjected to a single laboratory
evaluation, including a ruggedness test. The ruggedness
test showed that changes in sorbent—sorbate equilibration
time prior to testing; test pressure, pressure application
mode, testor drive direction, and sample size did not affect
the test results significantly, when the changes were kept
reasonably small within the range of laboratory differences
to be anticipated. Small changes in the liquid loading
within the range of initial liquid detection did not affect
the results significantly either. A 5—minute change in test
duration produced a difference in results (increase in pro-
portion of liquid releases observed) that was very nearly
statistically significant at the 0.05 probability level.
Based on the ruggedness test, the standard deviation of a
Liquid Release Test (set of 3 tests) is estimated at 0.29
proportion releases.
-------
1NTKUUU T1UL
To carry out its responsibilities under RCRA, the Environ-
mental Protection Agency needs a test method to indicate
whether liquids can be released from sorbent materials under
landfill pressures. Although the Paint Filter Test is a
quick arid easy field test that gives reproducible results,
it does not address the release of liquids when sorbents are
compressed. 1 Since amendments to RCRA require regulations
that prohibit the landfilling of liquids absorbed in
materials that biodegrade or release the liquids when com-
pressed, an adequate Liquid Release Test (LRT) must be
developed. 2
Many sorbent materials of organic and mineral origin are in
commerical use for the containment and removal of liquids
from spill sites and for shipment of hazardous liquids.
Although the sorbtive capacity of these materials is well—
known, their ability to retain liquids under landfill condi-
tions has not yet been fully evaluated. For this reason,
the LRT test development required consideration of the
mechanisms of liquid release from sorbents in landfills and
of the characteristics of the relationship between liquid
concentration and its potential for release.
MECHANISMS OF LIQUID RELEASE
Soil mechanics theory points out four mechanisms by which
liquids may be released from sorbents: 3
1. The liquids may be squeezed out by consolidation of the
sorbents under over—burden pressures;
2. The liquids may drain from the sorbents under gravita-
tional suction;
3. The liquids may be leached or washed from the sorbents
by water percolation; and
4. The sorbents may biodegrade, thereby losing their
ability to retain liquids.
Test methodologies with principles of application related to
the first two mechanisms were considered for evaluation. A
consolidation test using simulated over—burden pressures was
investigated because it relates to the squeezing out of
liquids by consolidation. A centrifuge test using weights
on top of the samples was investigated because it relates to
the squeezing out and the gravitational suction mechanisms.
A methodology related to the leaching mechanisms was not
studied since leaching is covered by another test, the
Toxicity Characteristic Procedure (TCLP). Sorbent bio-
degradation was not used to define a test methodology be-
cause sorbents that biodegrade are specifically banned from
landfill use.
-------
TEST DEVELOPMENT
Because of the large number of variables involved in cal-
culating pressures to be expected in a given landfill, the
test was designed such that it could be applied over a range
of pressure. Studies by Peirce and Shah indicate that the
pressures to 45 lb/in , (310 kPa) are to be expected,
depending upon the density and depth.4,5
CONSOLIDATION TEST
Two types of test units were utilized in the development of
the consolidation test: the conventional consolidometer and
the Zero Headspace Extractors employed for the Toxicity
Characteristic Leaching Procedure (Figures 1, 2, and 3).
Consolidometers designed and used at Duke University to
measure hydraulic conductivity were modified for use in the
consolidation test (Figure 1). The top and bottom of the
unit were made of polyvinyl chloride, while the cylinder and
piston were made of transparent cast acrylic. Teflon disks
0.8 mm thick were placed above and below the sample.
Approximately 30 holes spaced from 0.5 to 1 cm apart were
drilled into the disks using a 1 mm diameter drill bit. A
filter paper was positioned next to each Teflon disk to
collect any liquid which moved from the test sample under
the vertical loading. The perforated disks were used to
facilitate movement of this liquid to the paper while pre-
venting the filter paper from collecting liquid from the
sample by capillary suction.
A 100—gram sample of the liquid loaded sorbent was placed in
the test unit between the perforated Teflon disks, with
absorptive filter papers placed against the opposite sides
of the disks. A compressive force was applied through the
piston stem for a specified time. Release of liquid was
indicated when a visible wet spot was observed on either
filter paper, or distinct droplets of liquid were present on
the inner surface of either Teflon disk.
Initial tests focused on two typical sorbent materials:
Fuller’s Earth and Floor Dry, arid two liquids: 0.0) N
aqueous calcium sulfate and 5—volume percent acetone in
water. These sorbent materials were selected to provide a
range of sorbent characteristics used commercially. 6 The
calcium sulfate solution was selected because it is widely
relied on by researchers and practitioners as a standard
water for investigation of landfill liner permeability. The
acetone solution was selected as a representative solvent
found at hazardous waste landfills. 7
Over—burden pressures cause the bulk densities of sorbents
to increase under compaction, reducing the effective pore
-------
VERTICAL LOAD PISTON
THREADCD ROD B
WING NUT
TUBE
CERAMIC
FILTER
STONE
PERF RATED
TEFLON
DISK
FILTER
PAPER
FILTER

F .
SAMPL.E
C IOO ms)
TEFLON
D S K
CERAMIC
FILTER
STONE
PVC
Figure 1. Pressure test consolidometer.
-------
Figure 2. ZHEV, Associated Design and Manufacturing Co. Model.
5.
-------
I ,
_ _ 4 __
____ - r w
_______________ - , - -___________ j;
J S•
______ S u •
________ : ;
______________ - S ___
S - ___ ___
___________ - - S S
- S I-
- - . 5 S
- - S - -
___ :J
r - 5 .S . 5 i.7
____ .. .- -
_____ - --
rit . 4 c — Li • ‘ ‘ ‘ - ; ;

-5 - _-- ‘ ;
.. -: - a -. . . 1L
Figure 3.
ZHEV, Associated Design and Manufacturing Co. Model, disassembled.
-------
volume of the material and allowing the liquid to be
released. The compaction was measured as the fall in height
of the protruding stem of the piston. The compaction of the
sorbents was measured at 15-second intervals for the first
minute of the test. After that, measurements were taken at
one minute intervals. The tests were continued until
equilibrium consolidation conditions were achieved. For
these sorbents, a consolidation rate of less that 0.001 cm
per minute was selected as the definition of a state of
equilibrium. This equilibrium was generally achieved 10
minutes after the load was applied. Accordingly, initial
test data were collected using a 10—minute testing time.
The testing time was later extended to 30—minutes to allow
more time for drainage of any released liquid.
Two commercially—available compressed-air driven Zero Head—
space Extraction units were used for tests of additional
sorbents and liquids. These units were the Model 3740—ZHB
Vessel made by the Associated Design and Manufacturing (ADM)
Company (Figures 2 and 3) and the model SD1P581C5 Vessel
made by the Millapore Corporation. With the test time
extended to 30 minutes, these units gave results equivalent
to those obtained with the consolidometer. The ADM unit was
easier to use, and most of the results reported herein were
obtained with it. Except for the longer testing time, the
test procedure was the same for these units as for the
consolidometer. These units have no stem on the pistons, so
that compaction rates cannot be measured with the accuracy
available with the consolidometer. On the other hand, the
units can be used for two tests of sorbents: the Liquid
Release Test and the TCLP.
A range of liquid loadings was investigated to establish the
liquid release characteristics of the sorbents under pres-
sure. Initially pressures of 103 and 310 kPa (15 and 45
lb/in 2 ) were both used. Later tests used only the higher
pressure.
Three tests cells were used to provide three replicate tests
for each liquid load. Each test result is a pass/fail
rating indicating whether the liquid was not released (test
passed) or was released (test failed). Replication of the
test is necessary in order to provide an estimate of the
percent of the tests that show release of liquid. As shown
in Figure 4, this percent is zero for low liquid loadings,
and 100 for excessive liquid loadings. Between these
extremes, there is a transition zone within which the per-
cent of tests showing release varies from 33 to 66 (for
greater numbers of replicates, the percentages would show
greater numbers of intermediate values). This transition
zone has been defined as the initial released liquid detec-
tion range of the test. These ranges are shown in Table 1.
Because the sorbent materials are natural materials, they
may be expected to vary somewhat in composition and physical
-------
4 —--.----.—__ _________ ---—-- -- -—— ——____
3-——-----—--—----- —
70
V)
L i i
V )
2 - _____________—- -________________
L ii
0
0 ::
L i i 1 - ---•- -- -- ------ — -----— Ii — -_____
z
0--- ---—- a’- —— -__________
60 65
—1 — — - 514 58 62 ‘ 70 74 78
(WEIGHT liquid/WEIGHT sorbent) X 100
Figure . Diesel Fuel/Florco Liquid Release Characteristics,
Sorbent Pressure Test
-------
TABLE 1. CONSOLIDATION TESTS OF SORBENTS AND SORBATES
INITIAL RELEASED LIQUID DETECTION RANGES. W/W PERCENTa
—-—____
Diesel
Fuel
Trichioroethylene
Xy lene
Acetone
soiutlonb
Calcium Sulfate
SolutloliC
Fly Ash,
BItuminous
<10
42 d
<10
—-
—-
S-N-D
40-45
78-83
45-50
--
--
Florco
65-68
1 15 e
69-68
--
--
Floor Dry
80 e
140-155
80-85
90-120
140-190
Fuller’s
Earth
--
—-
-
60-75
5565
‘W/W PERCENT = (wt llquId/wt sorbent) x 100
b 5 vol acetone In water
c 001 N CRSO 4 in water
d 011 liquid release in three tests
etwo liquid releases in three tests
-------
structure from lot to lot, and these ranges should be
expected to vary from lot to lot also.
CENTRIFUGE TEST
In evaluation the centrifuge methodology for sorbent test-
ing, the effects of a range of centripetal force from 95 to
968 g’s were explored. Since the higher portion of this
range was obtained using lead weights placed on top of the
sample, the range represents primarily the difference depths
of overburden that might be placed upon a loaded sorbent.
It also provides a range of hydraulic gradients (ratios of
head loss to length of flow path) both including and exceed-
ing the gradients expected over a depth of landfilled
material. The consolidation test does not provide this
gradient.
The tests were conducted using relatively small samples in
filtration tubes fabricated for use in a table model labora-
tory centrifuge. The tubes were fabricated using standard
24 mm i. d. Pyrex glass tubing, a coarse grade fritted glass
plate (40 to 60 micron pore size), and a standard taper
ground glass joint. The plate was sealed into the tubing to
provide a filtration base and support the sample. Figure 5
shows the construction details. A glass cup was fitted onto
the tapered joint beneath the plate to collect released
liquid. Lead disks, 24.0 g each were used to provide weight
loadings on top of the sample and were sized to provide 0.5
mm tolerance and thus fit easily into the sampler, on top of
the sample. A lid, or plastic seal, was used to prevent
evaporation of liquid from the sample during the test. This
filtration tube fitted into a standard 50—mi centrifuge
shield.
A piece of coarse—grade Fisher filter paper was placed in
the filtration tube to cover the fritted glass plate. The
sample, sorbent with known liquid content, was weighed into
the tube, and lead disks were placed on top. Pairs of
samples were placed opposite each other in the centrifuge.
The pairs were adjusted to essentially equal gross weight by
varying the sample size slightly. During the tests, the
samples were subjected to the lowest chosen g force by set-
ting the equivalent rpm. After maintaining this force for
15 minutes, the sample container was reweighed to determine
the weight loss. The reweighed sample was then subjected to
the next higher selected g force for 15 minutes, and the
weight loss again determined. This repeated application of
successively higher g forces was continued over the desired
range, with weight losses determined at each step. In
selected tests, a final centrifugation was carried out for
an additional hour in order to obtain data for analysis as
to ultimate weight loss. Centrifuge tests were conducted
both with the glass cup in place beneath the filtration
tube, and without the cup. In either case, the glass pieces
-------
Sample
Filter Paper
Fritted Glass
Plate
t
2.9cm
‘I
Lid
Lead Discs, 24.0 gm each
Sample Container
Liquid Collection Cup
- 23mm —
——24mm——
10cm
Figure 5. Sample cell for centrifuge test.
-------
Data Points
95 . 0.4525
574- 1.075
638- 1.1289
968- 1.2704
968- 1.4147
l.5
0
.21.O
-J
0
U,
In
0
-J
CD
E
0
12.7 lb/in 2
11.5 lb/in 2
1.9 lb/in 2
.5
Test run 1 additional hour
0
0
19.4 lb/in 2
Initial Sample
9.556 grams, total
5.444 grams liquid
4.112 grams dry sorbent
g’s
100 200 300 400 500 600 700 800 900
1000
Figure 6. Release of 132.4 w/w % calcium sulfate solution from floor dry under applied centrifugal forces.
-------
were supported within the centrifuge shield by a rubber
cushion.
Tests were carried out using Floor Dry samples loaded with
calcium sulfate solution or with acetone solution. Tests
were also run using the Fuller’s Earth sorbent, but only the
acetone solution was investigated. Liquid loadings tested
included those within the release range for the consolida—
tion test (Table 1) and those below this range. 8 Figure 6
shows the results for Floor Dry loaded with 132.4 w/w%
Calcium Sulfate solution. This sample was tested without
the use of the lead weights. During the first 15 minutes,
0.4525 gram of liquid was released at a g force equivalent
to 1.9 lbs/in 2 . Successive tests, at 11.5, 12.7 and 19.4
lb/in2 resulted in incremental additional liquid losses of
0.6225, 0.539, and 0.1415 grams respectively over a total
testing time of one hour. The sample was then tested for
another hour, giving an additional liquid release of 0.1443
grams. Floor Dry was also tested within the range of
acetone solution loadings for which consolidation test
showed liquid release. These tests confirmed the release.
Tests using Fuller’s Earth were carried out using only the
acetone solution. Loadings within the range of liquid
release shown by the consolidation test were confirmed. A
51.2 w/w% loading of liquid showed a slight release (0.0069
grams out of 2.3372 total contained in the sample) at 12.1
lb/in 2 in 15 minutes. This liquid loading is below the
range of 60—75 found with the consolidometer. Tests of
liquid loadings within this range showed releases at 1.9
lb/in 2 for 15 minutes, confirming the previous findings.
RUGGEDNESS TESTING
The consolidation test was chosen for ruggedness evaluation,
rather than the centrifuge test because the latter would
require a large model centrifuge to accommodate the desired
100 grams of sample, and the smaller sample sizes, 7 to 10
grams, that could be used in table model centrifuges were
considered too small to be representative of the sorbent
materials to be tested.
RUGGEDNESS TEST DESIGN
The ruggedness testing was done principally to determine
the procedure’s sensitivity to minor reasonable variations
in the different test conditions. These tests are necessary
so that the procedure may be specified and controlled as
closely as needed to avoid excessive variation among
different laboratories. In addition, this ruggedness test
also tested whether the storage of the sample would alter
its liquid release characteristics. Some of the samples
were stored for 15 days prior to testing, to provide for
further equilibration of the mixture so that any need for
-------
such time extension between mixing of the materials and
testing could be determined.
LIQUID RELEASE TEST PROTOCOL FOR RUGGEDNESS TEST
The ruggedness tests were run as described in this section,
with changes in the procedure deliberately introduced at
points indicated by an (*). Numbered subheadings 1.0
through 8.9 shown are those defining the test protocol.
1.0 SCOPE AND APPLICATION
1.1 The Liquid Release Test (LRT) is designed to indicate
whether liquids can be released from sorbents when exposed
to landfill pressures.
1.2 Any liquid—loaded sorbent which fails the EPA Paint
Filter Free Liquid’s Test (SW—846 Method 9095), shall be
assumed to release liquids in this test. Analysts should
make sure that material in question will pass the Paint
Filter Free Liquid’s Test.
2.0 SUMMARY OF METHOD
(*) 2.1 A 100 ± 0.1 gram representative sample of the
liquid—loaded sorbent is placed between twin perforated
Teflon disks in a device capable of simulating landfill
pressures. Absorptive filter papers are placed against the
opposite sides of these Teflon disks, and a compressive
force of 310 kPa (45 lb/in 2 ) is applied. Release of liquid
is indicated when a visible wet spot is observed on either
filter paper, or distinct droplets of liquid are present on
the inner surface of either Teflon disk.
3.0 INTERFERENCES
3.1 When testing sorbents loaded with liquids that are
capable of rapidly evaporating (e.g., solvents), any liquid
migrating to the filter paper may eventually evaporate. For
this reason, all filter papers shall be examined immediately
after the conduct of the test.
4.0 APPARATUS AND MATERIALS
4.1 Pressure Tester: For the purposes of this test, an
acceptable pressure tester is one that is capable of accom-
modating a pressure of up to 414 kPa (60 lb/in 2 ), and which
is capable of being quickly and easily dismantled for in-
spection of the filter papers. This pressure tester shall
have an internal volume of 500 to 600 ml, and be equipped to
accommodate fitted filter papers and Teflon disks (see
Figures 1 and 2). Suitable apparatus known to EPA are
identified in Table 2 The devices identified in Table 2 are
known as Zero Headspace Extraction (ZHE) Vessels. The same
-------
device (with several minor modifications) is used in the
Toxicity Characteristic Leaching Procedure (TCLP), for
evaluating the potential leachability of volatile compounds.
To avoid confusion, this method also refers to the device as
the ZHE or ZHE vessel.
4.2 Pressure Source: The ZHE is pressurized using
compressed gas (e.g., air, nitrogen.) This may be supplied
from a compressed gas cylinder equipped with a suitable
regulator, or may be supplied using a simple air pump, pro-
viding that such devices are capable of delivering the
required pressures to the ZHE. The ZHE or the pressure
source shall be equipped with a pressure gage accurate to
within ± 7 kPa (1 lb/in 2 ), to indicate when the desired
pressure has been attained.
(*) 4.3 Balance: A balance accurate to within 0.1 grams.
4.4 Teflon disks: Two 0.8 mm thick Teflon disks,
perforated with 1 mm diameter holes, spaced approximately 5
mm apart (can be ordered with ZHE).
4.5 Filter papers: Two 90 mm absorptive filter papers
(Whatman No. F24l0—9, or Millipore No. 8P4004705, H5D75931A,
or equivalent). One is cut to fit into the piston of the
ZHE; the other fits without adjustment over the Teflon disk
next to the ZHE end—piece.
5.0 REAGENTS
5.1 None required.
6.0 SAMPLE COLLECTION, PRESERVATION AND HANDLING
6.1 All samples shall be collected using a sampling plan
that addresses the considerations discussed in “Test Methods
for Evaluating Solid Wastes (SW—846).”
6.2 Preservatives shall not be added to samples, and
samples shall not be kept at freezing temperatures.
(*) 6.3 Samples shall be tested as soon as possible after
collection, but in no case after more than two days after
collection. If samples must be stored, they shall be stored
in sealed containers and maintained under dark, cool condi-
tions.
7.0 PROCEDURE
7.1 Place the piston within the body of the ZHE such that
it is approximately in the middle of the device. Secure the
gas inlet/outlet flange to the device in accordance with the
manufacturer’s instructions.
-------
TABLE 2 SUITABLE ZERO-HEADSPACE EXTRACTOR VESSELS
Company
Location
Model No.
Associated Design and
Alexandria,
3740-ZHB
Manufacturing Company
Virginia
(703) 549—5999
Millipore Corporation
Bedford
Massachusetts
(800) 225-3384
SD1P381C5
-------
7.2 Position the ZHE in an upright position, and place one
of the filter papers against the piston and one of the
Teflon disks on top of it.
(*) 7.3 weigh out a representative sample of 100 ± 0.1
grams, and transfer it into the test unit on top of the
Teflon disk.
7.4 Using a stirring rod, gently smooth the sample so that
it is distributed evenly across the diameter of the test
unit.
7.5 Place the other teflon disk on top of the sample, and
the other filter paper on top of the Teflon disk. Assemble
the liquid inlet/outlet flange to the device (with the
stainless steel screen provided with the device set in place
over the filter paper), in accordance with the
manufacturer’ s instructions.
(*) 7.6 Tighten all fittings, make sure that the ZHE is in
a vertical position, and open the liquid inlet/outlet valve.
NOTE : Some samples, upon testing, may produce enough
liquid in a strongly positive test such that it is
capable of flowing out of the valve. To prevent any
liquid flowing from this valve from contaminating the
surrounding area, a liquid collection device (e.g.,
expandable sample bag, syringe) should be attached to
the valve before proceeding to the next step.
(*) 7.7 Connect the pressure supply line to the gas
inlet/outlet valve, and begin applying pressure, increasing
the pressure to 69 kPa (10 lb/in 2 ) in 20 seconds. Slowly
increase the pressure in 69 kPa (10 lb/in 2 ) increments, to a
maximum of 310 kPa (45 lb/in 2 ). Allow 20 seconds to attain
each 69 kPa increment of pressure.
NOTE : Instantaneous application of high pressure can
cause the filter paper to burst. If the filter bursts,
the test must be redone using a fresh sample.
(*) 7.8 Let the test unit stand at 310 kPa (45 lb/in 2 ) for
30 minutes. Check the pressure at 5—minute intervals, and
adjust as necessary. Very little adjustment should be
necessary, unless the unit is leaking. If leaking is in-
dicated, check and replace the ZHE 0-rings, or other
fittings, as necessary, and redo the test with a fresh
sample.
(*) 7.9 After 30 minutes, turn off the compressed air
supply, release the pressure from the unit, and immediately
and carefully disassemble the device so that the filter
papers may be inspected.
-------
7.10 Examine the filters for the presence of any wet spots.
The presence of a wet spot(s) indicates a positive test
(i.e., liquid release). The presence of a distinct droplet
of liquid on the inner surface of either Teflon disk also
indicates a positive test.
8.0 QUALITY CONTROL REQUIREMENTS
8.1 Repeat any test twice, using fresh samples, whenever no
liquid release is indicated in the first or second test
(total of three tests), to insure that the test is negative.
A release of liquid in any one or more of the three tests
indicates a positive test result.
8.2 All data should be maintained for easy reference and
inspection.
RUGGEDNESS TEST DESIGN
The ruggedness test requires an experimental design
designating the total number of tests to be made, and in-
dicating how the minor variations in the test procedure are
to be applied throughout the tests to provide a set of data
adequate for analysis. The analysis should provide measures
of any excessive variations, and measures of within
laboratory precision.
The test design is based on the use of a Plackett-Burman
fractional factorial experimental design as prescribed by
Youden and Steiner. 9 The test matrix is pictured in Table
3. Each row prescribes values of a test condition. Thus
each column prescribes a set of conditions for each rugged-
ness test. Capital letters (A,B,...G) denote nominal
values; small letters (a,b,...g) denote slightly altered
values that might affect the test. Conditions altered in-
cluded sample equilibration time, test pressure, sample
size, test duration, pressure application mode, piston drive
direction, and liquid loading.
During development of the test, samples were prepared and
allowed to equilibrate for 24 hours before testing. The
samples were kept sealed to prevent loss of liquid. In
addition, the sealed containers were shaken periodically
while equilibrating to keep the particles mixed and to
insure that the liquid was distributed as evenly as possible
among them. Nevertheless, it was felt that a further test
of the adequacy of equilibration time should be made. For
this reason, the samples tested for ruggedness were divided
into two sets. One set was tested after the usual 24 h
waiting time, the other set, after 360 h (15 days).
Pressure settings of 310 kPa (45 lb/in 2 ) are the usual level
applied. To determine the sensitivity of the test to small
-------
TABLE 3. RUGGEDNESS TEST DESIGN
C: 100 g C: 105 g
e: Pressure to 69 kPa in 20 sec; 69 to 138 kPa in 20 sec; 138 to 207 kPa in 20
sec: 207 to 276 kPa in 20 sec;
F: Piston presses down to test
G: 96.5 w/w (5 vol acetone
g: 95.5 w/w % (5 vol % acetone
Test Condition
Value of the
1 2 3
of
4 5
Determination
6
7
Number
8
1
Sample equilibrium
time
A
A
A
A
a
a
a
a
2Pressure
B
B
b
b
B
B
b
b
3Samp leSize
C
C
C
c
C
c
C
C
4Duration
D
D
d
d
d
d
D
D
5
Pressure Application
Mode
E
e
E
e
e
E
e
E
6
Piston Drive
Direction
F
f
f
F
F
f
I
F
7
Liquid Loading
G
g
g
G
g
6
6
g
A:
24
h
a:
360
h
B:
310
kPa
b:
345
kPa
D:
30
mm
d:
35
mm
E:
Pressure
raised
to
set
in 10 sec
276 to set point in 10 sec.
f Piston presses up to test
solution)
solution).
-------
changes in the applied pressure, some of the tests were con-
ducted at 345 kPa (50 lb/in 2 ).
Sample size was set at the usual 100 g and also at 105 g so
that the sensitivity of the test to small changes in
weighings of the sample might be determined.
The normal testing time, 30 mm, was extended to 35 mm for
some of the tests in order to determine how critical exact
timing of the test would be.
In other applications of the test vessels, other researchers
had observed that the filters had become torn or ruptured by
quick application of the pressure. Therefore two modes of
raising the sample to test pressure were employed. The
usual procedure of raising the pressure to set point in 10
seconds was used in half the tests; the gradual increase of
applied pressure, reaching each additional 69 kPa (10
lb/in 2 ) in 20 seconds was used in the other half. All tests
were observed for damage to the filters as well as for
wetness.
The ADM tester can easily be used with the piston driven
downward or upward during the application of test pressure.
The Millipore vessel, however, cannot conveniently be used
except in the upward driven position. Since all previous
test data had been collected with the piston driven down-
ward, both directions of drive were used in the ruggedness
test to see if the direction changed the results signi-
ficantly.
During the preparation of samples for the ruggedness test, a
new lot of Floor Dry was used. It was found not to hold the
expected amount of acetone solution that had been observed
on previous tests of other lots of the material. Prelimi-
nary tests made with the new lot showed consistent liquid
release at 105 w/w % liquid loading. Therefore a new search
was made for the initial liquid release range of the new
lot, and it was found to be between 95 w/w % and 98 w/w %.
For this reason, the liquid loadings of 95.5 and 96.5 were
adopted for the ruggedness test. These loadings were found
to yield data within the range for which 1 to 2 releases in
three tests were obtained. This range was necessary to make
the data analyzable and the results interpretable.
Each column of Table 3 defines a set of conditions under
which 9 tests were conducted. The total number of tests
required for ruggedness determination was estimated 72. The
estimate is based on the needs to have a high probability,
P’ = 0.9, of finding a difference as great as 0.33 in a one—
tail statistical evaluation test at the 5 percent signi-
ficance level of the proportion, p, of tests showing liquid
release. The ruggedness test was conducted at a liquid
loading for which the expected proportion of liquid releases
-------
was approximately 0.5. Accordingly, a difference of 0.33
would be found with a high degree of certainty. This dif-
ference would be critical, since it would lead to acceptance
of loaded sorbent for landfill if the altered test condition
showed no liquid releases for a material loaded to
correspond with a 0.33 proportion of releases.
Applying these criteria, the number of tests, n, were
estimated using the formula:
= (tc< + tfi )2 (plql + p2q2)
(P1 — P2) 2
where:
t c = the deviate of the “t” distribution correspondina
to a one—tailed test at the 5 percent level;
t,g = the deviate of the “t” distribution corresponding
to a one—tailed 0.9 probability level;
P1 = fraction of tests showing liquid release;
q 1 = fraction of tests not showing liquid release;
n = tests required per condition i;
1 denotes standard condition;
2 denotes altered conditon.
Substituting the required values and solving for n gives:
(1.69 + 1.306)2 (0.33 x 0.66 + 0.66 x 0.33)/0.332 = 36
The total number of tests is twice the number per condition
(2 x 36 = 72)
This formula is based upon methodology described by C. W.
Snedecor and William G. Cochran.’° The original formula was
based on the use of a test with known standard deviation.
Since we are estimating the standard deviation from the test
data, we have substituted values of deviates of the “t” dis-
tribution for those of the normal distribution.
With 72 tests required, the Plackett and Burman experimental
design (Table 3) was repeated nine times.
RUGGEDNESS TEST RESULTS AND DATA ANALYSIS
The results of the ruggedness test are shown in Table 4.
For each of the 8 sets of test conditions employed in the
-------
TABLE 4. RUGGEDNESS TEST RESULTS
.
Test
Conditions
F
•
Test Results
A
B
C
D
E
F
G
Liquid
Releases
Total
Tests
Fraction
Release.
‘p”
Test
Set
Equilibrium
time, h
Pressure
kPaa
Sample
Size,
g
Test
Duration
mm.
Pressurl-
zation
Rateb
Piston
Driven
Liquid
Loading
w/w96
1
24
310
100
30
fast
down
96.5
4
9
0.4144
2
24
310
105
30
slow
up
95.5
6
9
0 6667
3
24
345
100
35
fast
up
95.5
5
9
0.5556
4
24
345
105
35
slow
down
96.5
6
9
0.6667
5
360
310
100
35
slow
down
95.5
7
9
0.7777
6
360
310
105
35
fast
up
96.5
4
9
0.4444
7
360
345
100
30
slow
up
96.5
2
9
0.2222
8
360
345
105
30
fast
down
95.5
3
9
0.3333
TOTAL
37
72
a 310 kPa = 45 lb/in 2 : 345 kPa = 50 lb/In 2
bfast prescribed pressure reached In 10 sec. slow = prescribed pressure reached in 90 sec.
-------
TAI3LE 5. ANALYSIS OF RUGGEDNESS TEST RESULTS
Test Condition
Difference
Tested
Difference in
Proportion of
Liquid Releases
Corresponding
ta
Statistical
Significance
Test duration, mm
35-30
0.6111—0.4167=0.1944
1.68
almost 005 a
Equilibration time. h
24-360
0.5833-0.4444=0.1389
1.19
not signif. (0.10)
Liquid Loading. w/w%
95.5-96.5
0.5833-0.4444=0.1389
1.19
not slgnif. (0.10)
Pressure. kPa
310-345
0.5833-0.4444=0.1389
1.19
not sIgnlt’. (0.10)
Pressure Application Mode
slow-fast
0.5833-0.4444=0.1389
1.19
not signlf. (0.10)
Piston Drive Direction
down-up
0.5555-0.4722=0.0833
0.71
not signif. (0.25)
Sample Size, g
105-100
0.5278_O.5000r0.0278
0.236
not slgnif. (0.25)
alhe ta value for a significance level of 0.05 is 1.69.
-------
ruggedness test, the table lists the number of liquid re-
leases observed, the total number of tests conducted, and
the fraction of them that showed liquid releases.
The effects of the changes in test conditions were cal-
culated from the data of Table 4 by first computing the
proportion of releases for each set of 36 tests obtained by
summing the results over all sets that employed a particular
level of each condition. These proportions were then
compared by taking their differences. For example, test
sets 1, 2, 3, and 4 were all made after 24 h equilibration
time. These when combined showed a 0.5833 proportion of
releases. Sets 5, 6, 7, and 8 showed a 0.4444 proportion of
releases. The difference, 0.5833 — 0.4444 = 0.1389 is a
measure of the effect of the change in equilibration time.
In this comparison, all the other changes average out, since
their levels both appear an equal number of times. The
differences observed for each of the conditions tested are
listed in descending order of magnitude in Table 5.
The statistical significance of these differences was tested
using the Student “t” Test. The value of ta was computed
using the following formula:
6 (P1 — P2)
tQC —1.306
(pig + p2q 2 )0.5
The values of “t” computed were compared with those for
significance levels of p = 0.05, 0.10, and 0.25. This com-
parison is summarized in the Statistical Significance column
of Table 5. In accordance with the planned decision to use
the 0.05 level as the criterion for significance of an
effect, none of the conditions tested showed a significant
effect. The test method may be considered rugged with
respect to these conditions. As noted in the table, how-
ever, the effect of test duration, 0.1944, was uncomfortably
close to being significant. This differnce showed a “t”
value of 1.68 versus a criterion value of 1.69. In view of
this, it appears that the test time, 30 mm, may not be suf-
ficiently long to give proper indication of the release
potential of loaded sorbents exposed to landfill pressures
continuously. Further investigation of the effect of test-
ing would be needed to determine whether, indeed, a 30 mm
test adequately distinguishes those sorbents which will
release liquids in landfills.
The standard deviation of a single test series (3 tests) was
determined by two methods. 1) Based on the average over
the entire set of data for the ruggedness test, 37 liquid
releases were obtained in 72 tests. The corresponding
standard deviation of a test series is 0.2885 proportion
releases, calculated by:
s.d. = (pq/n) 0 • 5 = (0.5138 x 0.4861)13 0.5 = 0.2885
-------
REFERENCES
1. Test Protocol for Free Liquids, FR 47,38, February 25,
1982, p. 8311.
2. Congressional Record, October 3, 1984.
3. Artiola, J.F.; K.W. Brown and D. C. Anderson,
“Theoretical Evaluation of Gravitational Forces and
Consolidation on the Behavior of Sorbents in Landfills.
4. Peirce, J. J. Soil Preparation and Permeability Testing
with Consolidation Cell Apparatus, final report, EPA
Contract 68—03—3149, 24—2, September 28, 1984.
5. Shah, S. I., letter of February 1, 1985, to David
Friedman, Testing Criteria for Free Liquid Test, Task
Group under ASTM Section 034.02.07, Physical Analytical
Methods.
6. Artiola, J. C., letter of March 21, 1985, to Steve
Piper, GCA Technology, concerning selection of sorbents
and liquids for study.
7. Brown, K. W., and D. G. Anderson, “Effects of Organic
Solvents on the Permeability of Clay Soils,” EPA—600/
2—83—016, U.S. EPA, 1983.
8. Carpenter, B. H. J. J. Peirce; T. A. Peele, and K. A.
Witter, “Liquid Release Test Development,” Final
Report, U.S. EPA Contracts 68—03—3149 and 68—01—7075,
Task 38, March 7, 1986.
9. Youden, W. J. and E. H. Steiner, Statistical Manual of
the Association of Official Analytical Chemists,
A.O.A.C., Box 540, Benjamin Franklin Station,
Washington, DC 40044, 1975.
10. Snedecor, G. W. and W. G. Cochran, Statistical Methods,
6th Ed., Iowa State University Press, Ames, Iowa, 1967.
-------
EVALUATION OF METHODS FOR DETERMINING CHLORINE IN WASTE OILS
ALVIA GASKILL, JR., EVA D. ESTES, DAVID L. HARDISON,
RESEARCH TRIANGLE INSTITUTE, RESEARCH TRIANGLE PARK, NORTH
CAROLINA AND PAUL H. FRIEDMAN, OFFICE OF SOLID WASTE, U.S.
ENVIRONMENTAL PROTECTION AGENCY, WASHINGTON, D.C.
ABSTRACT
The Environmental Protection Agency has issued a final rule
prohibiting the sale for burning in non—industrial boilers
of used oils contaminated above specified levels with
certain metals and total chlorine. When burned as fuel in a
small boiler, the contaminants may be emitted to the ambient
air at hazardous levels. This regulation establishes a
rebuttable presumption that used oil containing more than
1,000 ppm total chlorine has been mixed with halogenated
solvents and is a hazardous waste. Rebutting the
presumption requires the seller of the oil to prove that
this chlorine is not due to halogenated solvents or other
hazardous halogenated organics. If the rebuttal is
successful, the oil can be sold as fuel up to a level of
4,000 ppm total chlorine.
To provide enforcement authorities and the regulated
community with appropriate methods to meet the chlorine
testing requirements of this regulation, an interlaboratory
evaluation of test methods and instrumentation was
conducted. The objectives were to assess the precision,
accuracy, detection limit, matrix effects, interferences,
field portabilities, and cost of this testing.
Methods and instrumentation evaluated included classical
ASTM bomb oxidation followed by graviinetric, titrimetric or
ion chromatographic analyses; instrumental microcoulometric
titration using a chlorine analyzer; energy and wavelength
dispersive x—ray fluorescence; a field kit based on a
chemical co].orimetric reaction; and a test device based on a
flame photometric response.
The evaluation was carried out by nearly 20 cooperating
laboratories who performed more than 120 analyses on around
40 samples of spiked virgin and waste oils. Spike
constituents included water; volatile, semivolatile, and
inorganic chlorine compounds. In addition, oil fuels and
blends with waste oils were evaluated.
-------
EVALUATION OF SELECTED DIGESTION METHODS FOR
DETERMINING HEXAVALENT CHROMIUM IN
SOLID WASTE MATRICES
JERRY D. MESSMAN, MARK B. CHURCHWELL, AND JOAN LATHOUSE,
BATTELLE COLUMBUS DIVISION, ANALYTICAL AND STRUCTURAL
CHEMISTRY CENTER, COLUMBUS OHIO; AND THEODORE D. MARTIN,
ENVIRONMENTAL MONITORING AND SUPPORT LABORATORY, U.S.
ENVIRONMENTAL PROTECTION AGENCY, CINCINNATI, OHIO
ABSTRACT
The analyses of solid waste materials for hexavalent
chromium present formidable challenges to the analytical
scientist. A metal speciation scheme such as the
differentiation between trivalent and hexavalent chromium
species, Cr(III) and Cr(VI), must address the capability to
maintain the integrity of the individual species during all
the sample manipulation phases of the overall analytical
method. Whereas much reseach has focused on the separation
and detection of dissolved chromium species in synthetic
aqueous mixtures or relatively clean liquid environmental
samples, the chemical solubilization and determination of
insoluble chromates in solid waste materials have not been
adquately addressed.
The present study has focused on an investigation of
selected digestion methods for the chemical solubilization
of insoluble Cr(VI) in barium chromate test compounds and in
real environmental samples. An alkaline digestion medium,
consisting of an aqueous solution of sodium carbonate and
sodium hydroxide, and acid digestion media, consisting of
nitric acid alone and in the presence of potassium
persulfate, were studied. The digestion methods were
evaluated in regard to their capabilities to solubilize
solid test samples without reducing Cr(VI) or oxidizing
Cr(III) species. A spectrophotometric method, specific for
Cr(VI) using the diphenycarbazide (DPC) color reagent, was
employed to measure concentration changes in hexavalent
chromium for each test sample solution resulting from
chromium redox phenomena occuring during the digestions.
The relative merits of the digestion methods based on the
analytical results and redox considerations of these
experiments will be discussed in the present paper.
-------
DEVELOPMENT OF A FIELD TEST FOR MONITORING
ORGANIC HALIDES
A. RAY TARRER, 3. GREGORY PERRY, WILLIAM M. HOLLOWAY, HUGH
L. FELLOWS, AUBURN WASTE HYDROCARBON REPROCESSING
LABORATORY, AUBURN UNIVERSITY, AL
ABSTRACT
In the past, used oil has provided a valuable energy source
in its use as a fuel for industrial and commercial boilers.
There has been recent concern, however, as to the health
risks involved in burning recycled oil containing
halogenated compounds. As reported in the November 29, 1985
edition of the Federal Register, these health risks have
prompted the EPA to impose new regulations involving the use
and management of used oil. Specifically, levels for total
halogen content have been set between 1000 ppm and 4000 ppm.
Above this concentration range, the oil is considered to be
a hazardous waste. Since about 500 million gallons of used
oil are burned each year, an inexpensive and reliable test
method is needed for determining the halogen content of a
used oil at the site of its generator.
In response to this need, Auburn University, under
sponsorship of the EPA, has been working on the development
of a field test to determine halogen concentrations in waste
oils. This test is an extension of the Beilstein flame
emission test. The test is very inexpensive, easily
performed and agrees with the standard ASTM oxygen bomb
technique within ± 10 to 15 percent.
The test is a very simple procedure requiring only hydrogen
and copper wool as replenishable materials. A copper probe
is first burned to remove any oxidation which may interfere
with the test. The clean probe is then dipped into the oil
sample and placed in the flame. Copper halides radiate
light at about 436 nm; therefore, the presence of halides is
indicated by a blue—green emission. Since the intensity of
the blue—green emission is a function of the halide
concentration, this concentration can be determined by
measuring the flame intensity using a photocell. The halide
content of very non—volatile samples can be determined as
low as 50 ppm. More volatile samples require dilution with
a low volatile oil. For most waste oils, a dilution ratio
of 10:1 diluent oil to waste oil is sufficient; this results
in a lower threshold detection limit of 500 ppm halide.
-------
The testing procedure works well for most waste oils. The
reproducibility is poor for oils which contain significant
amounts of water or for those that are highly volatile. With
experience the test can be used to screen even these types
of oil, but the results are not very quantitative.
This test has shown great promise for on—site determinations
of halide concentrations in waste oils. Inconclusive tests
are obtained for only a small percentage of oils, but even
these samples can usually be screened qualitatively for the
1000 ppm limit. This test can be performed by waste oil
users, dealers, and transporters as well as EPA enforcement
personnel thus making the new regulations more easily met.
This test represents a simple, inexpensive, and reliable
method of screening waste oils for halide contamination in
the field.
-------
RCRA LABORATORY CERTIFICATION
ROBERT R. HIRST, DENNIS M. STAINKEN, PH.D., ROBERT L.
FISCHER, PH.D., AND KATHE STAUBER, NEW JERSEY DEPARTMENT OF
ENVIRONMENTAL PROTECTION, OFFICE OF QUALITY ASSURANCE,
TRENTON, NEW JERSEY
ABSTRACT
The Office of Quality Assurance (OQA) within the New Jersey
Department of Environmental Protection develops laboratory
standards for the Department’s Safe Drinking Water and NPDES
(NJPDES) programs, and administers a laboratory
certification program. The existing certification program
was first promulgated for drinking water in 1977.
Certification for water and wastewater was offered beginning
in 1981. The USEPA Region II RCRA staff incorporated a work
output in the FY 1986 USEPA/NJDEP RCRA Interagency Agreement
which requires the Department to begin evaluating compliance
and performance of analytical laboratories which report
measurement data to the Agency and the Department with
respect to the RCRA program. OQA made a decision to develop
a laboratory certification program for waste analyses (RCRA)
and to incorporate such a program into its current
regulations.
The RCRA laboratory certification program, as it is
currently proposed, would provide laboratories with
certification in four (4) major categories: waste
characterization, inorganics, organics and miscellaneous
analyses. Laboratories will be certified by individual
parameter or analysis under the categories of waste
characterization, inorganics (AAS), and miscellaneous
analyses. For organics analyses, laboratories will be
certified for entire analytical methods rather than by
individual analytes (for example, a laboratory will be
certified for all parameters covered under SW—846 Method
8240 rather than by each individual analyte listed in the
method). Laboratories will be requested at the time they
apply for certification to designate which matrices among
(aqueous and/or nonaqueous) will be analyzed under their
RCRA certification. The program will reguire, along with
NJDEP’s RCRA Regulations, that all laboratories which submit
RCRA compliance data to the Department must be certified for
the applicable parameters or methods.
The new regulations will cover administrative procedures for
the program, such as annual fees for certification,
laboratory personnel qualifications, proficiency evaluation,
on—site inspections, and enforcement. Many new
administrative proposals are incorporated into the proposed
regulations, including a more efficient enforcement
-------
procedure which will apply to drinking water,
water/wastewater, and waste analysis certifications.
Development of the waste analysis certification program has
also raised many other issues which are under review at OQA,
such as development of performancce evaluation samples,
analytical method development and validation, method
equivalency procedures, and, in the long—term, establishment
of laboratory performancce standards and uniform quality
control procedures.
INTRODUCTION
Since its establishment in November, 1983, the Office of
Quality Assurance has been the focal point within the New
Jersey Department of Environmental Protection (NJDEP) for
development of laboratory standards, quality assurance, and
laboratory certification. The Office currently reviews
quality assurance project plant and standard operating
procedures for the Department’s varied environmental
monitoring programs, audits and advises NJDEP’s divisional
QA programs, administers the Department’s contract for
analytical services, and develops and administers the
Department’s regulations for certification of environmental
testing laboratories.
The existing laboratory certification regulations were first
promulgated by NJDEP for drinking water only in 1977. At
that time, the program was based primarily upon the U.S.
Environmental Protection Agency’s certification program for
the proposed National Interim Primary Drinking Water
Regulations, 40 CFR Part 141. The State Department of
Health, the agency delegated the authority for the
certification program at that time, offered drinking water
laboratory certification in five categories, microbiology,
limited chemistry, atomic absorption, gas chromatography,
and radiology. In 1978, the program was transferred to the
Department of Environmental Protection, where it was
expanded to include many parameters added by the passage of
New Jersey’s Safe Drinking Water Act by the State
legislature. The additional parameters are referred to as
the “secondary drinking water parameters.”
Certification of laboratories reporting data to the
Department for compliance with the National Pollutant
Discharge Elimination System (NPDES, now called NJPDES after
delegation to New Jersey by the USEPA) was offered beginning
in 1981. This program for water and wastewater testing was
divided into four categories: microbiology, limited
chemistry, atomic absorption, and gas chromatography.
Soon after promulgation of the water/wastewater
certification regulations, the Department’s, as well as the
USEPA’s, emphasis on safe drinking water and controlled
-------
elimination of discharges was matched by the advent of the
Federal CERCLA (Superfund) and RCRA programs. In addition,
New Jersey enacted its own legislation which requires
sellers of commercial and industrial properties to
demonstrate that their sites are free from hazardous
contamination. This bill, known as the Environmental
Cleanup Responsibility Act (ECRA), combined with the Federal
Acts, brought much attention to the analysis of multi—media
environmental samples for organic and inorganic
contaminants. CERCLA analytical work provided to the
Department by contractors is generally performed in
accordance with the USEPA Contract Laboratory Program’s
Information for Bid (IFB) documents. Use of the analytical
methods in these documents, which are revised frequently, is
made possible by the absence of legislation regarding
analytical requirements such as those imposed by two 40 CFR
Part 141 (drinking water), 40 CFR Part 136
(water/wastewater), and 40 CFR Part 261 (RCRA waste
analysis). The Federal regulations compensate for their
decreased flexibility by providing for standardization of
analytical procedures over an extended period of time and,
as a result, continous generation of additional data in
support of the methods’ validations from a growing
nationwide pool of government, industrial, and commercial
laboratories.
Because of the substantial need for analysis of multi—media
samples for hazardous contaminants, the Office of Quality
Assurance sought to improve and maintain the quality of data
reported for compliance with the NJPDES, ECRA, and RCRA
regulations. In late 1985, OQA began to study the
feasibility of certifying laboratories for multi—media waste
analysis. The release of 40 CFR Part 261 as final rule
provided the impetus for the design of a new laboratory
certification program for laboratories which report
compliance or investigation data to the Department of RCRA.
Also, because of the similarities and overlaps of some
NJPDES and ECRA moriitorinq requirements with those of RCRA,
OQA proposed to expand the applicability of the
certification regulations to those programs. In January,
1986, the first draft of the new regulations was distributed
for peer review within the Department. Many modifications
were incorporated and a final proposal of the Regulations
Governing Laboratory Certification and Standards of
Performance (N.J.A.C. 7:18—1.1 et seq.) will be available
for public comment through the New Jersey Register during
the summer of 1986.
ORGANIZATION OF THE RCRA LABORATORY CERTIFICATION PROGRAM
The Laboratory Certification Program, as it is proposed, is
made of up three major “categories”. Each category is
further divided into several “subcategories”. The following
-------
outline of the program illustrates these categories and
subcategories:
TABLE 1
ORGANIZATION OF THE LABORATORY CERTIFICATION PROGRAM
DRINKING WATER (including SDWA)
Microbiology
Limited Chemistry
Inorganics (AAS and “wet chemistry’ t methods)
Organics (THM’S, pesticides, herbicides, and other
organics)
Radiology
WATER/WASTEWATER (including NJPDES)
Microbiology (including recreational bathing water
testing)
Ames Testing
Limited Chemistry
Inorganics (AAS, ICP, and “wet chemistry” methods)
Organics (priority pollutant organics)
Radiology (including Radon)
Bioassays
WASTE ANALYSIS (including RCRA, NJPDES, and ECRA)
Waste Characterization
Inorganics
Organics
Miscellaneous
Each subcategory contains numerous individual parameters
and/or analytical methodologies for which certification is
offered. Certification by parameter or by method is
dependent upon practicality factors and instrumentation.
Generally, inorganics certification for laboratories
utilizing atomic absorption spectroscopy is issued by
individual analyte because each analytical measurement is
considered to be an individual analysis. On the other hand,
analysis of inorganics by inductively coupled plasma (ICP)
and organics by gas chromatography (GC or GC/MS) are handled
as measurements of groups of analytes using one method at a
time; thus, certification is issued by analytical method
rather than by analyte.
The Waste Analysis category, which includes analyses for
RCRA compliance and investigations, is illustrated in detail
in Table 2 below. The category is organized to correspond
directly with “Methods for Evaluation of Solid Waste,”
USEPA, SW—846. A system of offering combined sample
preparation/extraction/analysis procedures was initially
-------
considered (e.g. combinations of 3000—series procedures with
7000— and 8000—series procedures). However, due to the
overwhelming number of combinations possible, OOA decided to
offer certification for the analytical procedures along with
general certification for “aqueous” and “nonaqueous”
matrices. At the time laboratories apply for certificaton
in the Waste Analysis category, they will be required to
specify one or both of the two general matrices for which
they are requesting certification. For example, if a
laboratory requests certification for halogenated volatile
organics and specifies nonaqueous matricies on the
application, and assuming that the laboratory has met all
other conditions for certification, the laboratory will be
certified for “Nonaqueous Halogenated Volatile Organics.”
TABLE 2
WASTE ANALYSIS CERTIFICATION
WASTE CHARACTERI ZATION
Ignitability
Corrosivity
Reactivity
EP Toxicity
INORGANICS
Each 7000—series analyte by AAS
ORGANICS
Halogenated volatile organics
Nonhalogenated volatile organics
Aromatic volatile organics
Acrolein, acrylonitrile, and acetonitrile
Phenols
Phthalate esters
Organochiorine pesticides/PCB’ s
Nitroaromatics and cyclic ketones
Polynuclear aromatic hydrocarbons
Chlorinated hydrocarbons
Organophosphorus pesticides
Chlorinated herbicides
Volatile organics (GC/MS)
Semivolatile organics (GC/MS)
MISCELLAENOUS
Total and amenable cyanide
Total organic halides (TOX)
Sulf ides
pH
-------
APPLICABILITY OF THE RCA LABORATORY CERTIFICATION REGULATIONS
The proposed RCRA laboratory certification regulations were
developed with specificity targeted toward compliance
monitoring of New Jersey RCRA facilities. However, the
current need for regulation of waste analysis points to a
much broader scope of applicability. For example, the
Department’s NJPDES program administers permits and
regulates the State’s subsurface disposal facilities (such
as leaking, underground storage tanks, groundwater
injection, landfill leachates, etc.). The monitoring
requirements for such facilities generally mandate
collection and analysis of multi—media samples to
demonstrate compliance with their permit requirements. The
methodologies currently referenced by the Department’s
NJPDES regulations are the wastewater methods promulgated by
40 CFR Part 136 (October 26, 1984). These methods are not
applicable to more complex matricies which may be
encountered with monitoring of wastes. The NJPDES program
will therefore benefit from being enabled to require that a
laboratory reporting compliance data for these facilities be
certified in the Waste Analysis category.
The New Jersey Environmental Cleanup Responsibility Act
(ECRA) will also be able to mandate the use of certified
“waste analysis” laboratories because of similar monitoring
requirements.
ADMINISTRATION OF THE RCRA LABORATORY CERTIFICATION PROGRAM
The administration of the laboratory certification
regulations in the Office of Quality Assurance is funded by
annual fees which are assessed for each category. Since
there are four subcategories under the Waste Analysis
category, the proposed regulations provide for four distinct
fees.
The laboratory certification program regulates education and
experience requirements for laboratory managers,
supervisors, and technical personnel. For example, those
personnel directly involved with analysis of environmental
analysis of environmental samples associated with RCRA must
meet the following requirements:
1. GC operators must have at least nine (9) months
experience in the operation of the CC on
environmental samples. A formal training course in
the operation of the GC may be substituted for three
(3) months of experience.
2. CC/MS operators must have completed a formal
training course in GC/MS operation and have at least
nine (9) months experience in the operation of the
GC/MS on environmental samples.
-------
3. Extraction—concentration specialists must have at
least one (1) year experience in the preparation of
extracts from environmental samples.
4. Purge and trap specialists must have at least six
(6) months experience using the purge and trap
technique for volatile organics.
5. Pesticide and herbicide residue specialists must
have at least two (2) years experience in organo—
chiorine/organophosphorus pesticide, herbicide, and
PCB analyses, including method—specified clean—up
procedures on environmental samples.
6. Mass spectral interpretation specialists must have
at least two (2) years experience in the
interpretation of mass spectra generated from GC/MS
analysis of environmental samples.
7. Atomic absorption spectrometer operators must have
at least six months experience in the operation of
atomic absorption equipment or have completed a
formal training course in the operation of atomic
absorption equipment.
Until such time that proficiency evaluation samples for RCRA
are validated, the certification program will base a
laboratory’s eligibility for certification in the Waste
Analysis category on the information submitted with the
application and on its findings from on—site evaluations of
the laboratory. These on—site inspections of the laboratory
cover all areas regulated by the Department and generally
include audits of actual data reports submitted to the
Department by RCRA facilities. The laboratories are
afforded thirty (30) days from notification of deficiencies
in which to correct those deficiencies and notify OQA of
such. OQA plans to conduct annual announced on—site
evaluations of “waste analysis” laboratories. The Office
has also begun to conduct unannounced inspections of
laboratories, especially if a problem is exhibited by the
data reported to the Department.
The regulations provide for broader, more defined
enforcement of the regulations than in the past. The Office
of Quality Assurance may suspend a laboratory’s “waste
analysis” certificaton indefinitely for such infractions as
unsatisfactory performance (when PE samples become readily
available), data reporting deficiencies, failure to respond
to deficiencies noted by OQA, and noncompliance with any
requirements of the regulations. OQA may also decertify a
laboratory for serious violations of the regulations
including, but not limited to, direct or indirect
misrepresentations to the Department and falsification of
-------
records. In addition to the above, OQA may assess deficient
laboratories for fines of fifty (50) dollars to five
thousand (5,000) dollars for each deficiency or violation.
The enforcement process has been streamlined to permit rapid
turnaround for handling of deficiencies and violations, and
to expedite the corrective action process.
FUTURE DIRECTION OF THE RCRA LABORATORY CERTIFICATION
PROGRAN
As mentioned earlier, the Office of Quality Assurance does
not currently evaluate the performance of RCRA laboratories.
NJDEP’s needs for validated multi—media performance
evaluation samples cannot be over—emphasized. The Office is
eager to participate in the development of these samples and
has documented proposals for such.
The agencies within the NJDEP which administer the RCRA pro-
gram have established certain deliverables requirements
which, in addition to being used for validation of sample
results, are being stored in a database. OQA plans to use
the database to develop uniform laboratory performance
standards for such things as surrogates, matrix spikes
(accuracy), and matrix spike replicates (precision).
The Office of Quality Assurance advocates the development of
performancce standards, along with standardized test
procedures, as vital steps in achieving data comparability,
which is of great importance to any long—term monitoring
program. Although the Office receives numerous requests for
alternative test procedures, each proposal is assessed
against the existing methodologies. It is far more
important to a program such as RCRA to maintain consistency
in test procedures than to continually update “state of the
art” procedures, which damages data comparability. The RCPA
Laboratory Certification Program is designed to assure data
comparability and to control the approval of alternate test
procedures.
-------
U.S. EPA REFERENCE STANDARDS AND QUALITY ASSURANCE
MATERIALS FOR THE ANALYSIS OF ENVIRONMENTAL CONTAMINANTS
R.E. THOMPSON, P.A. WYLIE, NORTHROP SERVICES, INC.,
ENVIRONMENTAL SCIENCES RESEARCH TRIANGLE PARK, NORTH
CAROL INA
ABSTRACT
To support the need for a certified, quality—contolled,
common database in the analysis of contaminating chemical
agents in environmental substrates, U.S. EPA’s Environmental
Monitoring Systems Laboratory (Las Vegas) and Environmental
Monitoring and Support Laboratory (Cincinnati) jointly main-
tain several repositories of analytical grade refernece
materials under the Agency’s quality Assurance Reference
Materials (QARM) Project. Operated under contract by
Northrop Services, Inc. in Research Triangle Park, NC, the
project currently offers standards of over 2000 compounds of
environmental concern including pesticides and their
metabolites and degredation products, PCBs and other
halogenated organics, plasticizers, nitrosamines,
polynuclear aromatics, and heavy metals. Included amongst
these are those Agency—identified compounds commonly known
as the “priority pollutants,” materials regulated under
Appendix VIII (RCRA) and CERCLA (Superfund) legislation, and
groundwater monitoring compounds.
The program is currently being supplied by over 180 chemical
manufacturing companies and 23 chemical supply houses.
Additional compounds, especially select environmental reac-
tion and degradation products, are synthesized and purified
in—house. The program currently offers standards of over
2000 compounds of environmental concern. During 1985, more
than 100,000 standards were distributed in response to some
6000 requests.
Maintenance of such a repository for both neat (essentially
pure) materials and certified solutions necessitates a com-
prehensive analytical quality assurance (QA) program. The
analytical QA support includes component identification,
purity assays, concentration verifications, and stability
studies using a variety of instrumental methods, for most
reference materials, QA protocol comprises identification by
mass spectrometry or IR, with subsequent purity assay by GC,
HPLC, DSC, or elemental analysis. Additionally, isotoeic
purity of compounds enriched with stable isotopes ( 1 C,
‘ 7 C1, 21.1) is established by mass spectrometry. Typical
purity of materials distributed exceeds 99%. Prepared solu-
tions of single and multiple components (in sealed ampuls)
are analyzed (by GC, HPLC, UV—Vis) to verify concentration
immediately after preparation and again at established time
intervals to assess stability.
-------
The program supports laboratory quality assurance efforts
with standardized column packings and spiked media reference
materials. Technical manuals on analytical methods,
analytical quality control, standards preparation, and
related topics are also available. In addition, the project
staff operates a technical information and assistance ser-
vice, providing chemical/physical/toxicological information,
literature searches, methodology, instrumental troubleshoot-
ing, and a wide variety of other types of technical informa—
tion. More than 4000 monitoring, enforcement, and research
laboratories in 92 countries are currently utilizing these
services, all of which are provided by the Agency without
charge.
-------
THE QUANTITATION OF PCB’S AND PCDD’S BY ELECTRON IMPACT
GC/MS USING RESPONSE FACTOR ESTIMATION
A. D. SAUTER AND 3. J. DOWNS, A. D. SAIJTER CONSULTING,
HENDERSON, NV; AND J. D. BtJCMNER AND V. A. PIZZITOLA, EXTREL
CORPORATION, PITTSBURGH, PA
ABSTRACT
U.S. EPA methods for the determination of polychlorinated
dibenzofurans (PCDF’s) and polychlorinted dibenzodioxins
(PCDD’s) in hazardous wastes (1) and for the determination
of polychiorinated biphenyls (PCB’s) in samples originating
at Superfund hazardous sites (2) require the determination
of the concentration of all isomers of a given compound
class. For technical, practical, and economic reasons, it
is effectively impossible to standardize a GC/MS system for
all analytes of interest in such applications of GC/MS.
Therefore response factor estimation procedures are required
for the quantitation of these important environmental
pollutants when various isomers are identified in samples,
but standards are not available. The approach currently
proposed for the guantitation of PCB’s in Superfund analyses
utilizes the concept of “isomer group” quantitation. This
procedure employs a mean response factor for congeners of
the various degrees of chlorination to quantitate “unknown”
species. Likewise, the RCRA method for the quantitation of
PCDD’s and PCDF’s also employs mean response factors for
quantitation of various tetra, penta, hexa, hepta and octa
species. Obviously, the accuracy of “statistical”
approaches to RF value estimation depends on how well the RF
values utilized to estabish the mean values. Clearly, a
formalism which could provide RF value estimates for the
quarititation of various “unknowns” would be preferred from a
technical accuracy and a QA/QC standpoint, as “The object of
all science, whether natural science or psychology, is to
co-ordinate our experiences and to bring them into a logical
system” (3)
Recently, we have published a model for the estimation of
electron impact GC/MS RF values (4). In this work, we have
applied our model to estimating RF values for various PCB’s,
PCDF’s and PCDD’s with good results. For the thirty
compounds listed in Table 1, the average estimated/observed
response factor was found at 1.07 ±0.114. These data
included 25 RF values determined using full scan data at
nanogram levels and 5 RF values determined at picogram
levels in the selected ion monitoring mode. While the
application of the model is limited to situations where mass
and/or chromatographic discrimination are not severe,
properly applied, our model provides a formalism which can
yield estimated RF values with an average accuracy of
-------
approximately ± 10 percent at one standard deviation.
Within a proper QA/QC context, we assert that similar
accuracy can be expected for quantitation of various PCB’s,
PCDD’s and PCDF’s for which standards are not available. We
show how our model can be employed to this effect, and we
demonstrate how our model could unify the environmental
Quality Control and Quality Assurancce of the GC/MS
determination of PCB’s, PCDD’s and PCDF’s.
REFERENCES
(1) RCRA Method 8280, Draft, May 12, 1986.
(2) Proposed Analytical Method for Pesticides and PCB’s,
Exhibit D, presented at CLP Method Review Conference,
Atlanta, Ga, March 1985.
(3) Einstein, A., The Meaning of Relativity , Princeton
University Press, Princeton, N.J.
(4) Sauter, A. D., Downs, J. J., Buchner, J. D., Ringo,
N. T., Shaw, D. L., and Dulak, J. G., Model for the
Estimation of Electron Impact Gas Chromatography/Mass
Spectrometry Response Factors for Quadrupole Mass
Spectrometers, Analytical Chemistry, in press.
-------
Table 1
Estimated/Observed Response Factors for
Selected PCB’s, PCDD’s and PCDF’s
PCB’s Estimated/Observed
2 1.13
3,3’ 1.02
2,4,5 1.09
2,2 ’4,4’ 0.974
2,3’4,5’,6 0.961
a pentachiorobiphenyl 0.866
2,2’3,4,5,5’ ,6 1.08
2,2’3,3’ ,4,4’ ,5,5’ 1.10
2,2’ ,3,3’ ,4,4’ ,5,5’ ,6 1.10
2,2’ ,3,3’ ,4,4’ ,5,5’ ,6,6’ 1.13
PCDD’ sa
2,3,7,8 0.950
1,1,3,7,8 0.925
1,2,3,4,7,8 1.00
1,2,3,4,6,7,8 1.19
1,2,3,4,5,6,7,8 1.08
PCDD’ s
2,3,7,8 1.10
1,2,3,7,8 1.10
1,2,3,6,7,8 1.04
1,2,3,4,6,7,8 0.975
1,2,3,4,5,6,7,8 0.925
TCDD’ s
2,3,7,8 1.25
1,2,3,8 0.983
1,2,3,4 0.953
1,2,7,8 1.16
1,2,8,9 1.29
PCDF’s
2,3,7,8 1.14
1,2,3,7,8 1.09
1,2,3,4,7,8 1.01
1,2,3,4,6,7,8 1.32
1,2,3,4,5,6,7,8 1.29
x
ff0 = 1.07 + 0.114
a SIM RF values provided by D. Catalano,
TN.
IT Corp., Knoxville,
-------