EPA/600/R-92/080
September 1993
Methods for Aquatic
Toxicity Identification Evaluations
Phase II Toxicity Identification
Procedures for Samples Exhibiting
Acute and Chronic Toxicity
by
E. J. Durban1
T. J. Norberg-King1
L. P. Burkhard1
With Contributions from:
G. T. Ankley1
M. T. Lukasewycz2
M. K. Schubauer-Berigan2
J. A. Thompson2
1U.S. Environmental Protection Agency
2AScl Corporation - Contract No. 68-CO-0058
Previous Phase II Methods
by D. I. Mount and L. Anderson-Carnahan
EPA-600/3-88/035
National Effluent Toxicity Assessment Center
Technical Report 01-93
Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Duluth, MN 55804
/~x~y
{§&) Printed on Recycled Paper
-------
Disclaimer
This document has been reviewed in accordance with U.S. Environmental
Protection Agency policy and approved for publication. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
-------
Foreword
This document is one in a series of guidance documents intended to assist
dischargers and their consultants in conducting acute or chronic aquatic toxicity
identification evaluations (TIEs). TIEs might be required by state or federal agencies
resulting from an enforcement action or as a condition of a National Pollutant
Discharge Elimination System (NPDES) permit. The methods described in this
document will also help to determine the adequacy of effluent TIEs when they are
conducted as part of a toxicity reduction evaluation (THE).
This Phase II document is the second of a three phase series of documents
that provide methods to characterize and identify the cause of toxicity in effluents. The
first phase of the series, Phase I (EPA, 1991 A; EPA, 1992), characterizes the
physical/chemical nature of the acute and chronic toxicant(s), thereby simplifying the
analytical work needed to identify the toxicant(s). Phase II provides guidance to
identify the suspect toxicants, and the last phase, Phase III (EPA, 1993A) provides
methods to confirm that the suspect toxicants are indeed the cause of toxicity. The
recent TIE documents (EPA, 1991 A; EPA, 1992; EPA, 1993A; and this document)
have been produced or revised to include chronic toxicity recommendations and
additional information or experiences we have gained since the original methods were
printed (EPA, 1988A; EPA, 1989A; EPA, 1989B).
This Phase II document provides identification schemes for non-polar or-
ganic chemicals, ammonia, metals, chlorine, and surfactants that cause either acute
or chronic toxicity. The document is still incomplete in that it does not provide methods
to identify all toxicants, such as polar organic compounds. This Phase II manual also
incorporates chronic and acute toxicity identification techniques into one document.
While the TIE approach was originally developed for effluents, the methods
and techniques directly apply to other types of aqueous samples, such as ambient
waters, sediment pore waters, sediment elutriates, and hazardous waste leachates.
These methods are not mandatory protocols but should be used as general guidance
for conducting TIEs.
The sections of both Phase I documents (EPA, 1991 A; EPA, 1992) which
address health and safety, quality assurance/quality control (QA/QC), facilities and
equipment, dilution water, testing, sampling, and parts of the introduction are
applicable to Phase II. These sections, however, are not repeated in their entirety in
this document.
iii
-------
Abstract
This manual and its companion guidance documents describe a three phase
approach for dischargers to identify the causes of toxicity in municipal and industrial
effluents (Phase I, EPA, 1991 A; EPA, 1992; and Phase III, EPA, 1993A). In 1989, the
document titled Methods for Aquatic Toxicity Identification Evaluations: Phase II
Toxicity Identification Procedures was published as a guidance document for identi-
fying the cause of toxicity in acutely toxic effluents (EPA, 1989A). This new Phase II
document provides details for more types of samples, tests and test procedures that
can be used to identify the specific chemical(s) responsible for acute or chronic
effluent toxicity when the cause of toxicity is related to non-polar organic compounds,
ammonia, surfactants, chlorine, or metals. Phase I characterization and Phase III
confirmation manuals, the other guidance documents in the three phase TIE ap-
proach, have also been produced or updated to include both chronic toxicity
information and new developments made since the first set of documents were
printed. The TIE approach is applicable to effluents, ambient waters, sediment pore
waters or elutriates, and hazardous waste leachates.
IV
-------
Contents
Page
Foreword , iii
Abstract iv
Tables vii
Figures viii
Abbreviations : ix
Acknowledgments x
1.0 Introduction 1-1
1.1 General Overview 1-1
1.2 Biological Testing Considerations 1-2
2.0 Non-polar Organic Compounds 2-1
2.1 General Overview 2-1
2.2 Acute Toxicity: Fractionation and Toxicity Testing Procedures 2-2
2.2.1 Sample Volume 2-3
2.2.2 Filtration 2-3
2.2.3 Column Size 2-3
2.2.4 C18 SPE Column Conditioning 2-4
2.2.5 Elution Blanks 2-4
2.2.6 Column Loading with Effluent 2-4
2.2.7 C18 SPE Column Elution 2-6
2.2.8 Blank and Effluent Fraction Toxicity Tests 2-6
2.2.9 SPE Fractions: Concentration and Subsequent Toxicity Testing 2-7
2.2.10 HPLC Separation 2-9
2.2.11 HPLC Fraction Toxicity Tests 2-9
2.2.12 HPLC Fractions: Concentration and Subsequent Toxicity Testing ...2-10
2.3 Chronic Toxicity: Fractionation and Toxicity Testing Procedures 2-10
2.3.1 Sample Volume 2-11
2.3.2 Filtration 2-11
2.3.3 Column Size 2-12
2.3.4 C18 SPE Column Conditioning 2-13
2.3.5 Elution Blanks 2-13
2.3.6 Column Loading with Effluent 2-14
2.3.7 C18 SPE Column Elution 2-14
2.3.8 Blank and Effluent Fraction Toxicity Tests 2-14
2.3.9 SPE Fractions: Concentration and Subsequent Toxicity Testing 2-15
2.3.10 HPLC Separation 2-16
2.3.11 HPLC Fraction Toxicity Tests 2-16
2.3.12 HPLC Fractions: Concentration and Subsequent Toxicity Testing ...2-17
2.4 GC/MS Analyses 2-17
2.5 Identifying Suspect Toxicants 2-18
2.5.1 Identifying Organophosphate Pesticides 2-19
2.5.2 Identifying Surfactants 2-19
2.6 Alternate Fractionation Procedures 2-21
2.6.1 Modified Elution Method 2-21
2.6.2 Solvent Exchange, 2-21
2.6.3 Alternative SPE Sorbents and Techniques 2-22
-------
Contents (continued)
3.0 Ammonia 3-1
3.1 General Overview 3-1
3.2 Toxicity Testing Concerns 3-2
3.3 Measuring Ammonia Concentration 3-5
3.4 Graduated pH Test 3-5
3.4.1 pH Control: Acid/Base Adjustments 3-5
3.4.2 pH Control: CO Adjustments 3-6
3.4.3 pH Control: Buffer pH Adjustments 3-7
3.5 Zeolite Resin Method 3-8
3.6 Air-Stripping of Ammonia 3-9
4.0 Metals 4-1
4.1 General Overview 4-1
4.2 Analysis of Metals 4-2
4.2.1 Prioritizing Metals for Analysis 4-2
4.2.2 Metal Analysis Methods 4-2
4.2.3 Metal Speciation 4-4
4.2.4 Identification of Suspect Metal Toxicants 4-4
4.3 Additional Toxicity Testing Methods 4-5
4.3.1 EDTA Addition Test 4-5
4.3.2 Sodium Thiosulfate Addition Test 4-6
4.3.3 Metal Toxicity Changes with pH 4-6
4.3.4 Ion-Exchange Test 4-7
5.0 Chlorine , 5-1
5.1 General Overview 5-1
5.2Tracking Toxicity and TRC Levels 5-1
6.0 Identifying Toxicants Removed by Filtration 6-1
6.1 General Overview 6-1
6.2 Filter Extraction 6-1
7.0 References 7-1
8.0 Appendix A A-1
VI
-------
Tables
Number Page
2-1. Solid phase extraction (SPE) column fractionation information 2-4
2-2. Comparison of toxic units (TUs) in each toxic fraction to TUs of all
fractions combined and whole effluent 2-7
2-3. Information for concentrating SPE and HPLC fractions 2-8
2-4. Example HPLC elution gradients for four commonly toxic SPE fractions 2-9
2-5. Eluate volumes needed for chronic SPE fraction toxicity tests with
Ceriodaphnia dubia and Pimephales promelas 2-12
2-6. Approximate effluent volumes needed for the chronic non-polar organic
identification procedures 2-12
2-7. Example HPLC elution gradient for SPE fractions from chronically toxic
effluent samples '. 2-16
2-8. Composition of 11 recommended fractions in modified elution scheme 2-21
3-1. Percent un-ionized ammonia in aqueous solutions for selected
temperatures and pH values 3-2
3-2. Calculated un-ionized ammonia LCSOs (mg/l) based on 24-h and
48-h results of a Ceriodaphnia dubia toxicity test conducted
at pH 8.0 and 25°C 3-3
3-3. Calculated un-ionized ammonia LCSOs (mg/l) based on 24-h, 48-h,
72-h, and 96-h results of a fathead minnow (Pimephales promelas)
toxicity test conducted at pH 8.0 and 25°C 3-4
3-4. Un-ionized ammonia toxicity values for species frequently used in
effluent testing 3-4
3-5. Percent un-ionized ammonia in aqueous solutions at 25°C and various
IDS levels 3-7
4-1. Atomic absorption detection limits and concentration ranges 4-3
4-2. Estimated instrumental detection limits for ICP-MS and ICP-AES 4-3
4-3. Metal LCSOs with respect to test pH 4-6
A-1. Effluent volume calculation worksheets A-2
A-2, Effluent volume calculation worksheets (example) A-6
vii
-------
Figures
Number Page
2-1. Phase II schematic for the identification of non-polar organic toxicants 2-1
2-2. Procedures for eluting the SPE column with a gradient of methanoI/water
solutions .....: 2-5
2-3. Concentrating effluent on the C18 SPE column '. 2-5
2-4. Procedure to concentrate toxic SPE fractions 2-8
2-5. Procedure to fractionate acutely toxic SPE concentrates using HPLC 2-9
2-6. Procedure to concentrate toxic HPLC fractions 2-10
VIII
-------
Abbreviations
AA Atomic Absorption
C18 Octadecylcarbon Chain
C8 Octylcarbon Chain
CTAS Cobalt Thiocyanate Active Substances
DO Dissolved Oxygen
DOC Dissolved Organic Carbon
EDTA Ethylenediamine Tetraacetic Acid
ERL-D Environmental Research Laboratory-Duluth
GC Gas Chromatography
GC/MS Gas Chromatography/Mass Spectrometry
Heppso N-(2-Hydroxyethyl) Piperazine-N'-2-Hydroxypropane Sulfonic Acid
HPLC High Performance Liquid Chromatography
ICp Inhibition Concentration Percentage
ICP-AES Inductively Coupled Plasma-Atomic Emission Spectroscopy
ICP-MS Inductively Coupled Plasma-Mass Spectrometry
Kow Octanol-Water Partition Coefficient
LAS Linear Alkylbenzene Sulfonate
LC/MS Liquid Chromatography/Mass Spectrometry
LC Lethal Concentration
MBAS Methylene Blue Active Substances
Mes 2-(N-Morpholino) Ethane-Sulfonic Acid
Mops 3-(N-Morpholino) Propane-Sulfonic Acid
NETAC National Effluent Toxicity Assessment Center
NIST National Institute of Standards and Technology
NOEC No Observed Effect Concentration
NPDES National Pollutant Discharge Elimination System
PBO Piperonyl Butoxide
Popso Piperazine-N,N'-bis (2-Hydroxypropane) Sulfonic Acid
QA/QC Quality Assurance/Quality Control
SPE Solid Phase Extraction
SS Suspended Solids
Taps N-tris-(Hydroxymethyl) Methyl-3-Aminopropane Sulfonic Acid
TDS Total Dissolved Solids
TIE Toxicity Identification Evaluation
TOC Total Organic Carbon
TRE Toxicity Reduction Evaluation
TU Toxic Unit
YCT Yeast-CerophyP-Trout food
ix
-------
Acknowledgments
This document presents additional methods and improvements made to the
procedures of Methods for Aquatic Toxicity Identification Evaluations: Phase II
Toxicity Identification Procedures (EPA-600/3-88/035) by Donald Mount and Linda
Anderson-Carnahan. This manual reflects new information, techniques, and test
procedures developed by theNational Effluent Toxicity Assessment Center(NETAC)
since the previous Phase II document was printed in 1989. This Phase II document
is based on the efforts of both federal and contract staff of the NETAC group. We
gratefully acknowledge the following individuals' contributions to the research and
development of the methods for this document: Penny Juenemann and Sharieen
Schm'rtt (federal staff); Joe Amato, Lara Andersen, Steve Baker, Tim Dawson, Joe
Dierkes, Nola Englehorn, Doug Jensen, Correne Jenson, Jim Jenson, Liz Makynen,
Phil Monson, Don Mount, and Greg Peterson (contract staff).
Through the support of Rick Brandes and Jim Pendergast (EPA, Permits
Division) and Nelson Thomas (ERL-D), the NETAC staff members developed and
revised the TIE series of documents.
-------
Section 1
Introduction
1.1 General Overview
The major objective of Phase II is to identify the
suspected toxicant(s) in effluent samples using toxicity
identification evaluation (TIE) procedures. Some general
guidance to achieve this goal might be furnished by the
results of acute or chronic Phase I tests (EPA, 1991 A;
EPA, 1992), but for many effluents, such as those that
contain non-polar organic toxicants, both separation and
concentration steps will be needed to achieve the stated
objective. If metals are the suspect toxicants, atomic
absorption (AA) spectrometry should be sensitive enough
to measure toxic concentrations directly in the sample,
and the number of metals is small enough that toxicity
can be attributed without separating one from another.
The same principle applies to toxicants such as ammonia
and chlorine; measurements can be made without sepa-
rating or concentrating the effluent. However, if non-polar
organic chemicals are suspected, separation is usually
necessary for analytical and toxicological reasons.
Because there are often many constituents within
the classes of chemicals (e.g., non-polar organics) iden-
tified in Phase I, initial efforts are most productively
directed towards separating the toxic from the non-toxic
constituents. With the need to identify the toxicant(s)
quickly, comes the temptation to analyze too soon. Using
methods such as gas chromatography/mass spectrom-
etry (GC/MS) one can identify many non-polar organics
that are present in the whole effluent mixture, but the
association of toxicity with compound identification is
very difficult to make for several reasons:
• There can be hundreds of compounds present in
the mixture, and to investigate all of them would
be very time consuming.
• Toxicity data for many of the chemicals identified
are usually not available; chronic data are espe-
cially scarce.
• Separate constituents are often not commer-
cially available; therefore, their toxicities cannot
be measured and compared to effluent toxicity.
• Interactions (additivity, synergism, antagonism)
are not known for the given mixtures and one
must know interactions to apportion toxicity.
Therefore, it is suggested that the search for a
separation technique to simplify the mixture into toxic and
non-toxic subsamples be the first priority, rather than
spending time investigating non-toxic components. If there
is a single suspect toxicant such as ammonia, then sepa-
ration needs are limited largely by the analytical require-
ments. If the toxicity is caused by one constituent, the
number of other non-toxic constituents is irrelevant when
attributing toxicity. However, Phase I results do not usu-
ally lead to a single suspect toxicant and, therefore,
separation may be necessary.
When a method for separating the toxicant(s) is
found, concentration might be an inherent part of the
procedure (as in solvent extraction) which will simplify the
problem of finding a method to concentrate the toxicity. At
each stage of the separation and concentration process,
measurement of toxicity is the best way to evaluate the
success or failure of the manipulations.
The interpretation of TIE results can be different
than in the classical research approach, where experi-
ments are designed to either accept or reject a hypoth-
esis. In TIE work, an experiment usually permits
acceptance but not rejection of the hypothesis. For ex-
ample, if ammonia is the suspect toxicant, it can be
removed using zeolite resin. If the post-zeolite effluent is
still toxic, you can conclude that there are additional
toxicants present. If the post-zeolite effluent is not toxic,
you cannot conclude that there are no additional toxicants
because the zeolite might have removed other toxicants
in addition to the ammonia.
The always present question of whether or not
there is more than one toxicant immensely complicates
data interpretation. Phase I results might not give an
indication of multiple toxicants unless the toxicant classes
change over time or from sample to sample. Phase II
results are often such that one cannot tell whether the
situation is one of partial removal of a single toxicant or
1-1
-------
toxlctty resulting from multiple toxicants. The issue might
be resolved when one toxicant is identified and measured
analytically. Experience shows that the best choice is to
try to focus on the toxicant that appears easiest to iden-
tify. Usually that will be a toxicant that can be separated
from the sample (e.g., extracted or recovered from a
sorbent that reduces the toxicity) and for which there is a
broad spectrum analytical identification method. Above
all, data should always be interpreted under all probable
scenarios, i.e., one toxicant, multiple toxicants, and even
different toxicants from sample to sample.
Experience gained since the first Phase II (EPA,
1989A) document was printed has shown that effluent
toxicants are not always strictly additive. When they are
not additive, the toxicant present in the largest number of
toxic units (TUs)1 will determine the toxic units of the
effluent. Non-additive toxicity will not be reduced by ma-
nipulations that remove toxicants present in fewer TUs
than the major toxicant. Two or more toxicants might be
present In approximately equal TUs, however, the ratio of
TUs might change over different sampling times so that
different chemicals determine the toxicity of the effluent.
These important problems can be dealt with in Phase III
(EPA, 1993A). In Phase II, the objective is to find which
toxicants are present in toxic concentrations. However,
failure of addttivity may confuse Phase II results. Minor
toxicants might not be noticed until the major one has
been removed. In addition, additivity cannot be deter-
mined until at least one toxicant has been identified.
Usually Phase II and Phase III merge and overlap, there-
fore such concerns regarding non-additivity must be in-
corporated in Phase II, at least in the latter stages.
As effluent constituents are identified, a sorting
process begins in which a decision must be made as to
whether or not each one identified contributes to the
toxicity of the effluent. Usually, this is based on the
estimated concentration and the constituents' toxicities.
Analytical error in quantttation might be large (10-fold or
more) because recoveries and instrument response fac-
tors probably will not yet have been determined on a
particular chemical. Uncertainty about toxicological data
Is caused by differences in species sensitivity and water
quality effects, when literature values are available. Con-
fidence in an acute toxicity value (LC50) will vary depend-
ing on the quality of the test, the number of times it was
repeated, and the completeness by which the results and
conditions were described. Data on chronic effect levels
are often scarce and rarely have tests been repeated.
Species sensitivity frequently varies from 100-fold to 1,000-
fold; an error will likely be introduced when the published
TU calculations are described in EPA, 1992. The TUs of whole
effluent equals 100% divided by the LC50, NOEC, or ICp (IC25.IC50)
of the effluent The TU of a specific chemical equals the concentration
of the compounded divided by the effect level of the compound.
toxicity data for species other than the test species are
used. When the uncertainty of the toxicity data is high, a
maximum of 100-fold difference between measured con-
centrations and literature effect would be acceptable to
classify a chemical as a suspect. If one has good data for
the test species being used, then this difference might be
reduced (e.g., to 10-fold). Since these decisions are
always subjective, they will sometimes be wrong no mat-
ter how carefully they are made. Perhaps most important
is use of an iterative process to make these decisions.
First evaluate candidates that have concentrations higher
than or closest to their chronic or acute effect levels and if
these prove to be negative, then examine those that have
concentrations below their effect levels. Remember that
the suspected toxicant concentrations at the dilution equal
to the effect level concentration are the important concen-
trations to compare. At some point, a decision must be
made whether the true toxicants have not yet been iden-
tified or measured and that different sample preparation
or analyses must be used.
For some effluents, Phase I results might not
have provided any guidance for selecting the appropriate
Phase II procedures to follow. Other characterization
steps that might be helpful are solvent extraction (acidic
or basic), sample evaporation, size exclusion chromatog-
raphy, lyophilization, and vacuum or steam distillation
(Jop et al., 1991; Walsh et al., 1983). We have little
experience upon which to recommend procedures in
these cases. It is most important to realize that the more
severe the effluent treatment, the more likely it is that
toxic artifacts will be created. These toxic artifacts could
then be confused with effluent toxicity; therefore, artifac-
tual toxicity must be monitored for each technique by
using blanks.
Phase II efforts should develop into Phase III
confirmation (EPA, 1989B; EPA, 1993A) as soon as good
evidence is obtained that one or more candidates are
probable toxicants. The primary product of Phase II is the
chemical identification of the suspected toxicants to fur-
nish the basis for Phase III testing. The techniques de-
scribed in this document are useful for TIE work with
effluents as well as ambient waters (Norberg-King et al.,
1991) and sediment pore water or elutriates (EPA, 1991B).
1.2 Biological Testing Considerations
The Phase I characterization documents (EPA,
1991 A; EPA, 1992) provide detailed discussions of vari-
ous issues that are important in decision making through-
out the TIE. The guidance covers use of various species,
test concentrations, effluent sample types, testing re-
quirements for quality assurance (QA), test endpoints,
frequency of changing the test solutions, and more. All of
these issues will not be discussed at length here and the
user is encouraged to refer to the acute Phase I or the
chronic Phase I as companion documents for the TIE
process. As the Phase II identification and Phase III
confirmation steps are initiated, QA requirements should
1-2
-------
be revisited and the types of tests modified as needed.
Several of these testing concerns are addressed below.
During Phase I, the analyst is searching for an
obvious alteration in effluent toxicity, which might be
obtained by using modified acute or chronic test methods.
Confirmation testing (Phase III) conducted according to
the standard methodologies will confirm whether the sus-
pect toxicant(s) detected in the characterization and iden-
tification steps (Phases I and II) is the true toxicant.
In characterizing the toxicity in Phase I, factors
such as time requirements, number of tests and the test
design had to be considered and weighed against the
type of questions that are posed. EPA has published
manuals that describe the acute or chronic test methods
to determine the toxicity of effluent or receiving waters to
freshwater and marine organisms (EPA, 1991C; EPA,
1993B; EPA, 1993C), and these tests are typically those
that indicated the presence of toxicity which the TIE
initiated. Deviations from these standard effluent testing
protocols were discussed in both the acute Phase I (EPA,
1991 A) and the chronic Phase I (EPA, 1992) manuals.
For either the acute or the chronic Phase I procedures,
the test volumes, number of test concentrations, and
number of replicates were ail reduced from the standard
test methods (EPA, 1991C; EPA, 1993B). Additional modi-
fications for the short-term chronic tests (EPA, 1993B)
including shorter test duration and a reduction in the
frequency of the test solution renewal are suggested.
Throughout this document the TIE procedures for
acutely toxic samples are based on the following species:
Ceriodaphnia dubia, Pimephales promelas, Daphnia magna,
Daphniapulex, Hyalella azteca, and Chironomus tentans. Al-
most all acute tests have been conducted using 10 ml of test
solution in a 1 oz plastic cup (or 30 ml glass beaker). TIE
procedures with chronically toxic effluents are based on
tests using either C. dubia or larval fathead minnows
(P. promelas). In our laboratory, the chronic tests with C.
dubia generally are conducted using 10 ml of test solution
in 1 oz cups and the chronic tests with fathead minnows
are conducted using 50 ml of test solution in a 4 oz plastic
cup (10 fish per cup). Use of other species is constrained
only by availability, size, age, and adaptability to test
conditions, and the threshold levels for additives and
reagents for the other organisms must be determined.
As soon as good evidence is obtained to impli-
cate a suspect toxicant(s), the procedures for performing
the toxicity tests can be changed. Therefore in Phase II,
the time to modify the -tests from the way they were
conducted in Phase I may depend on when the toxicant is
identified, and generally there is more flexibility for this in
Phase II than in Phase 111. The quality control (QC)
measures in Phase I were not very strict because the data
are primarily informative rather than definitive. The iden-
tity of the suspect toxicant(s) furnishes the basis upon
which Phase III testing will be conducted, which will
require stricter QC measures.
Initially, the use of modified protocols in Phase II
may continue; however, once specific toxicant(s) identifica-
tion has been made, Phase II (and Phase III) testing
conditions should be similar to the methods described in
the protocol that was used to trigger the TIE. Although a
shortened version of the 7-d C. dubia test (which is referred to
as the 4-d test) may have been used in Phase I, the use of this
test changes in Phase II (and Phase 111). In order to use the 4-d
test in Phases I and II of the TIE, the 4-d test must detect similar
trends of toxicity as the 7-d test does. However, in Phase 111 the
7-d test is required because the toxicity as measured in the 7-d
test (with additional replicates, more test concentrations, addi-
tional volume) was used to detect toxicity for the permit, and
should be used to confirm the cause of toxicity. In the early
Phase II chronic toxicity evaluation steps, the qualitative evalu-
ation of toxicity might be useful and there is no reason why a
toxicity test could not be terminated sooner than day seven, if
the answer to a particular question has been found.
Information obtained from all toxicity tests should
be maximized. For instance, in acute toxicity tests, moni-
toring time to mortality might be useful. In chronic toxicity
tests, time to young production of the cladocerans or the
lack of food in the stomach of the larval fish might be
useful parameters. Observations such as these made
during a test might be subtle indications and quite infor-
mative of small changes in toxicity. For example, if there
is complete mortality on day four of the baseline effluent
test, and in the EDTA addition test (Section 4) the animals
either do not reproduce or grow yet they are alive at day
seven of the exposure, the indication is that the toxicity
was reduced. These results suggest that either an addi-
tional toxicant is present or the EDTA concentration was
not sufficient to remove all cationic metal toxicity. These
types of observations in the short-term tests might be just
as useful as reductions in young production or growth.
For evaluating whether any manipulation changed toxic-
ity, the investigator should not rely only on statistical
evaluations of test endpoints (see below and Phase I;
EPA, 1992). Some treatments may have a significant
biological effect that was not detected by the statistical
analysis. Judgement and experience in toxicology should
guide the interpretation.
In addition, for acute or chronic toxicity tests,
randomization, careful exposure time readings, use of
animals of uniform narrow-age groups (i.e., C. dubia
neonates 0-6 h old rather than 0-12 h old) might assist in
detecting smaller differences in tests. For example, in the
chronic C. dubia tests, it is important to use organisms of
known parentage (EPA, 1993B) when the number of.
replicates is reduced from ten to five. For C. dubia, daily
renewals of the test media (as required in the chronic
manual; EPA, 1993B) might not be necessary in Phase I
or early Phase H testing as long as the toxicity of the
1-3
-------
effluent can be measured with one or two renewals.
However in Phase III, tests must be conducted with daily
renewal of test solutions similarto the routine biomonitoring
Although reference toxicant tests are not recom-
mended for each set of Phase I manipulations, when a
toxicant has been identified in Phase II and some Phase III
confirmation tests indicate it is the toxicant, that chemical
should become the reference toxicant with the species
used In the TIE. In Phase II, the reference toxicant data
are useful for identification interpretation and provide
Information on the quality of the test organisms and
general test procedures. Reference toxicant tests should
be conducted routinely and control charts should be
generated (EPA, 1991C; EPA, 1993B). If a toxicant has
not been identified, standard reference toxicants should
be used, but as soon as a toxicant is identified, that
compound should be used as the reference toxicant for
the TIE tests.
The Phase I procedures frequently rely only on
one test species, but in Phase III of the TIE the use of
more than one species is recommended. This will be
useful In determining whether or not the cause of toxicity
Is the same for other species of the aquatic community. In
Phase I, we recommended that the species that detected
the toxicity be the first choice as the TIE species. If an
alternative species is chosen one must prove in Phases II
and ill that the species that initially detected the toxicity is
being impacted by the same toxicant as the alternate
species. Both species need not have the same sensitivity
to the toxlcant(s), but each species' threshold must be at
or below the toxicant concentration(s) present in the
effluent. One method of proving that the two species are
being affected by the same compound(s) is to test several
samples of the effluent over time with both species. If the
effluent possesses sufficient variability, and the two spe-
cies LCSOs or IC25s change in proportion one to another
as would be expected, the analyst may assume that the
organisms are reacting to changing concentrations of the
same compound. Further proof that the two species are
responding to the same toxicant should surface during
Phases II and/or III. If the toxicant is the same for both
species, then characterization manipulations that alter
toxfcfty to one species should also alter toxicity to the
second species. The extent to which toxicity is altered for
each will depend upon the efficiency of the manipulation
to remove toxicity and the organism's sensitivity to the
toxicant. Approaches used in Phase III will confirm whether
the two species are indeed sensitive to the same toxicant
In the effluent. If in Phase III, the organism of choice is not
responding to the same toxicant as the species that
triggered the TIE, extensive time and resources might
have been wasted.
The type of sample to use in the Phase II identifi-
cation stage most often will be similar to those used in
Phase I. The use of multiple (daily samples for acute and
chronic tests or the minimum of three for chronic tests)
effluent samples for each chronic test should nor be used
in the early stages of Phase II (EPA, 1992). The use of
one grab or composite sample for the Phase II identifica-
tion procedures is needed until some suspects have been
identified. For instance, if several effluent samples are
used for renewals during the chronic Phase I and/or
Phase II TIE and the toxicants are different or change in
their ratios one to another, interpreting the results will be
difficult. Indeed, such variability must be identified but it
should be done after at least one, or preferably most of
the toxicants are known. The use of one sample is even
more important in Phase III, (EPA, 1993A) where toxicity
data are correlated to the measured concentrations in the
effluent. If multiple samples are used for one toxicity test,
this correlation cannot be readily done because the same
toxicant may not be present in each sample, it might be
present in varying concentrations, or other toxicants may
appear. For the acute TIE, one composite or grab sample
has been used for identification and confirmation steps.
However, since the permit test may require daily samples
for an acute static renewal or 7-d short-term tests, once
the toxicant is identified each daily sample may have to
be analyzed for that toxicant.
Sample degradation is a concern that should be
addressed. The toxicity of the whole effluent can be
monitored by conducting toxicity tests upon sample ar-
rival and at periodic intervals throughout the TIE. For
some types of toxicants degradation or loss might be
expected (i.e., chlorine, see Section 5) but for toxicants
such as non-polar organic compounds, this may not
readily be known. Since the toxicant identification stage
might be lengthy, it is important to know that the toxicity
remains in the effluent even at a lower concentration.
When a toxicant is identified, further analyses and toxicity
tests can be conducted on effluent samples and any
toxicant degradation or loss evaluated.
As discussed in the Phase I manuals (EPA, 1991 A;
EPA, 1992) if the level of toxicity for any given effluent has
been established with some degree of certainty from
previous tests, it might be adequate to use four effluent
dilutions and a control to follow toxicity changes of the
sample to reduce the cost of the tests. As the toxicant is
identified, test concentrations should be selected to de-
tect small changes in toxicity. We are assuming that if
effluents have inhibition concentration percentage (ICp)
(or no observed effect concentration (NOEC)) values
below 10%, the effluent is likely to show acute toxicity and
if so, an acute TIE approach can be used. If chronic TIE
work is to be done on a highly toxic effluent, the same
recommendations given in the chronic Phase I manual
should be followed; that is, use concentrations of 4x, 2x,
1x and 0.5x the IC25 or IC50 value. For example, if the
IC25 is 5% effluent, we would suggest using a test
concentration range such as 20%, 10%, 5% and 2.5%
effluent for the various tests. With chronic toxicity data
where the NOEC is 12% (or the IC25 is 10%), a concen-
tration series such as 6.3%, 12.5%, 25%, and 50% would
be logical. If closer concentration intervals are desired,
1-4
-------
using 20% effluent as the high concentration and a dilu-
tion factor of 0.7, the concentrations to test would be 7%,
10%, 14%, and 20%. If the NOEC (from historical data) is
40-50% (or above 50%), then the concentration series to
test might be either 25%, 50%, 75%, and 100% or 40%,
60%, 80%, and 100%. Choice of dilution factor and test
concentration range is a matter of judgement and de-
pends on precision required and practicality.
After conducting Phase I procedures on an efflu-
ent sample, the amount of effluent available for subse-
quent identification work can be sufficiently reduced so
that it may be impractical to try to conduct each step as
described in this manual. This is most likely to be a
concern for the non-polar organic identification techniques
and other methods that require large volumes of effluent
to identify the toxicant. Therefore, when the volume of an
effluent sample is limited, it might be possible to track
toxicity through the non-polar identification steps without
quantifying the amount of toxicity that is being tracked.
Essentially, this means that the toxicity tests are done
without dilutions and the results would indicate only that
toxicity was present or absent; the degree of toxicity
present would not be measured. Once a suspect toxicant
is identified, it is important that the amount of toxicity
removal is known (through the use of dilutions) because
this information can be used to correlate a suspect toxi-
cant to the effluent toxicity in the Phase III confirmation.
If the number of replicates per test concentration
is reduced, one must assume that precision is sufficient to
decipher changes in toxicity that must be measured. One
problem in using reduced replicates and low numbers of
test concentrations in chronic tests is that this smaller
data.set is not amenable to all statistical requirements as
recommended for the short-term tests (EPA, 1989C; see
Section 5.8). Use of more organisms and more replicates
than in the Phase I modified tests might be preferable if
Phase I and/or Phase II data are likely to be used in
Phase III confirmation (See Sections 2.2 and 2.3).
For acute toxicity tests, usually the LC50 or EC50
is reported for the toxicity data (calculated as recom-
mended in EPA, 1991C). Endpoints for the most com-
monly used freshwater short-term chronic tests are growth,
reproduction, and survival. The no effect level (the NOEC),
and the effect concentration (the lowest observed effect
concentration (LOEC)) are determined using the statisti-
cal approach of hypothesis testing to determine a statisti-
cally significant response difference between a control
group and a treatment group. The NOEC/LOEC are heavily
affected by choice of test concentrations and test design
(see Phase I; EPA, 1992). The linear interpolation method
(EPA, 1993B) provides a point estimate of the effluent
concentration that causes a given percent reduction based
on organism response. To calculate the inhibition concen-
tration percentage (ICp), a computer program (Norberg-
King, 1993; DeGraeve et al., 1988; EPA, 1989C) is
available and the assumptions for the method are not the
same as the test design requirements for hypothesis-
based analyses. This point estimation method is particu-
larly useful for analyzing the type of data obtained from
chronic TIE tests using dilutions (see Phase I; EPA,
1992). Confidence intervals are calculated using a boot-
strap technique and might be useful in determining if
significant toxicity alterations have been observed. A
significant reduction in toxicity and the precision of refer-
ence toxicant tests must be determined by each labora-
tory for each effluent. The use of the IC50 for Phase I
TIEs might be more useful in correlating the characteriza-
tion test results to the effluent toxicity than an IC25.
However, there are situations when an IC50 may not be
able to be estimated while the IC25 can. Above all, it is
most important to use a consistent effect level for TIE
toxicity testing (EPA, 1992A; EPA, 1992B). When sub-
stantial toxicity reductions occur in the toxicity tests, it
may not always appear to be a significant reduction when
the IC25s are compared. In order to further evaluate
whether toxicity reductions occurred, the sample size
(number of replicates, number of concentrations) should
be increased in subsequent testing in an effort to differen-
tiate toxicity responses from the sample size limitations.
The dose response curves should then be compared to
see if responses are similar. Once the toxicant is identi-
fied, the number of replicates should be increased, the
dilution sequence might be modified and more dilutions
used (see Phase III; EPA, 1993A). This should increase
the confidence in the IC25 (or any other ICp value cho-
sen) estimate.
1-5
-------
-------
Section 2
Non-Polar Organic Compounds
2.1 General Overview
The procedures described in this section pre-
sume that the results of Phase I tests have implicated
non-polar organic compounds as the cause of acute or
chronic toxicity. Results of Phase I tests that clearly
implicate a non-polar organic toxicant typically are (1) all
toxicity is removed by the C18 Solid Phase Extraction
(SPE) column and (2) toxicity was observed in the metha-
nol eluatetest (see Section 8.6, EPA, 1991 A; and Section
6.7, EPA, 1992). In some instances, toxicity might not be
removed completely by the SPE column, but sufficient
toxicity is recovered in the methanol eluate to suggest a
non-polar organic toxicant is present. While toxicants
other than non-polar organic compounds might be re-
moved by the column (e.g., metals), the elution for such
toxicants is unlikely to be similar to that of the non-polar
organic compounds. However, there is also the possibility
that non-polar toxicants such as surfactants will be re-
moved by the column and not recovered in the methanol
eluate. The goal in this section is to separate the non-
polar organic toxicants from the many non-toxic compo-
nents of the sample to simplify the analytical work needed
to identify the toxicant.
This section provides the general background
information on non-polar organic compounds along with
methods for concentrating and separating the toxicants
for samples with acute and/or chronic toxicity. While the
method provides a stepwise procedure, there are in-
stances where the investigator may have to modify the
approach to achieve the best results.
Also provided in this section are procedures that
might prove helpful in less common situations. Metabolic
blockers can be used to reduce or eliminate certain
organophosphate compounds from exhibiting their toxic-
ity (Section 2.5.1). In the instance when toxicity is not
recovered in the methanol eluate and toxicity is not evi-
dent in the post-column effluent, alternate SPE elution
procedures might be needed (Section 2.6).
A flow diagram of the general procedures fol-
lowed in identifying non-polar organic toxicants is shown
in Figure 2-1. In this procedure, the C18 SPE column is
used to extract non-polar organic compounds from efflu-
ent samples. These compounds are then selectively
Effluent Sample
Toxicity Test
SPE Fractionation
Toxicity Tests
r
Concentrate and Combine Toxic SPE Fractions
Toxicity Tests
GO/MS Analysis (optional)
HPLC Fractionation
Toxicity Tests
Concentrate and Combine Toxic HPLC Fractions
Toxicity Tests
GO/MS Analysis
GC/MS Identification
Compare Concentrations to Toxicity Values
Figure 2-1. Phase II schematic for the identification of non-polar
organic toxicants.
2-1
-------
stripped off the column by eluting the C18 sorbent with
solvent/water mixtures that are increasingly less polar. As
a series, the "fractions" resulting from column elution
contain analytes that are decreasingly polar and decreas-
ingly water soluble. Each fraction is then tested for toxic-
Hy. The fractions that exhibit toxicity are concentrated,
and chromatographed using reversed phase High Perfor-
mance Liquid Chromatography (HPLC). The resulting
HPLC fractions are collected and tested for toxicity. The
toxic HPLC fractions are concentrated into methanol by
using another CJ8 SPE column. The concentrates are
toxicity tested as before and analyzed using gas chroma-
tography/mass spectrometry (GC/MS). Those constitu-
ents that are Identified by GC/MS are roughly quantitated,
by assuming that the identified constituents and the inter-
nal standard have the same response factor, and the
estimated concentrations are compared to available tox-
fcity values for each chemical. If this process reveals
strong suspect toxicants, mass balance testing (Phase
III; EPA, 1993A) could be started to determine whether
additional toxfcants are present. If no suspect toxicants
are identified by GC/MS, a longer analysis time on the
HPLC might help the identification by increasing the
separation between toxic and non-toxic components, es-
pecially if there are many constituents present. Also,
additional constituents might be identified by increasing
the concentration factor by using larger effluent samples.
At some point, the probability that the toxicants are not
chromatographing on the gaschromatograph or the mass
spectrometer is not detecting the toxicants must be con-
sidered if no suspect toxicants are identified. Use of other
types of mass spectrometry, such as liquid chromatogra-
phy/mass spectrometry (LC/MS) or direct probe mass
spectrometry may be useful. Some effluents might re-
quire the SPE fractionation of several different samples
before good suspect toxicants are found (Burkhard et al.,
1991; Lukasewycz and Durhan, 1992).
The sorbents that we recommend for use in SPE
and HPLC columns are chemically identical. The column
packing is composed of silica gel which has been reacted
with octadecyl silane to produce a covalent bonded phase
one layer thick. The mechanism of extraction with C
sorbents is relatively simple. Extraction of effluent com-
pounds occurs because the C1S sorbent competes for the
non-polar compounds more strongly than the surrounding
water molecules of the effluent. Sorption of non-polar
organlcs is also influenced by ionic strength, pH, and total
organic carbon (TOC) levels. The same compounds will
partition on both SPE and HPLC columns and the order of
elution of chemicals will be approximately the same. The
major difference between the SPE and HPLC columns is
the amount of resolution achieved. The particle size em-
ployed In HPLC columns is smaller, providing a greater
surface area and better component resolution. Despite
less resolution, SPE columns have the advantage of
possessing a higher loading capacity in general than
HPLC columns. The SPE column could be considered as
a preparatory column for sample cleanup while the HPLC
column gives far more refined and controlled separation
of sample constituents.
To elute non-polar organic toxicants extracted by
the C18 SPE column, the sorbed compounds must have a
higher affinity for the eluting solvent than for the octadecyl
functional group (C18). Choosing a solvent for elution is
complicated because the toxicants' identities are not
known, in general, the solvent should be less polar than
water and more polar than the C18 functional group. The
degree of solvent strength required to elute the toxicants
is also unknown. Since methanol is less polar than water,
has a very low toxicity (EPA, 1991 A; EPA, 1992) and
elutes chemicals from CL sorbents, it has been a good
solvent choice for most TIE purposes to date.
During sequential column elutions with succes-
sively increasing methanol in water concentrations, the
relatively hydrophilic, polar compounds are eluted first,
and the more hydrophobic non-polar compounds are
eluted last. Given the strength of methanol as a solvent
for non-polar compounds, it is possible that very hydro-
phobic (octanol water partition coefficient (log K ) >4)
effluent compounds will not be eluted from the c"" sor-
bent. If toxicity caused by a very hydrophobic compound
is extracted by the SPE column but not eluted by metha-
nol, less polar solvents might be used to elute the SPE
column (Section 2.6.1).
Once toxicity is found in one or more C18 SPE
effluent fractions, the toxic fractions can be concentrated,
then fractionated using HPLC. HPLC separation is used
to reduce the number of non-polar organic chemicals
associated with the toxicant(s) and to simplify analytical
identification. The toxic HPLC fractions are concentrated
and then analyzed by GC/MS. The estimated concentra-
tions of constituents in the final concentrate (based on an
internal standard) are then compared to their toxicity
values to decide which may be sufficiently high in concen-
tration to cause toxicity. If none are found, higher concen-
tration factors, other analytical methods (e.g., LC/MS),
and better separation are recommended.
Fractionation and toxicity testing procedures for
non-polar organic toxicants causing either acute or chronic
toxicity are presented in different sections of this chapter.
The acute toxicity (Section 2.2) and chronic toxicity (Sec-
tion 2.3) sections, have similar outlines and were written
so either section could be used independently. As a
result, some details that apply to both acute and chronic
toxicity are repeated. After using Section 2.2 or Section
2.3 the investigator should then follow the identification
techniques described in Sections 2.4 and 2.5. Additional
identification techniques are included in Sections 2.5.1
and 2.5.2, and alternate fractionation methods are found
in Section 2.6.
2.2 Acute Toxicity: Fractionation and
Toxicity Testing Procedures
In the initial stages of Phase II, toxicity tests may
be conducted on C18 SPE effluent fractions and blank
fractions to detect the presence of toxicants and not to
quantify the magnitude of the toxicity in each. As in the
toxicity tests conducted during Phase I, careful measure-
2-2
-------
ment of test solution water chemistry parameters is not
required, and duplicate exposures are not needed during
initial stages of Phase II. The major purpose of this step is
to assess whether or not acute toxicity is present in the
effluent fractions and the blank fractions. However, as
suspect toxicants are identified, quantitative toxicity mea-
surements will be needed to compare with the analytical
measurements. If Phase II data will be used to correlate
effluent toxicity to toxicant concentrations (Phase III),
then more replicates per concentration, randomization of
test concentrations, careful observation of organism ex-
posure times, and organisms of approximately the same
age should be used (Section 1.2). Also, the amount of
eluate that is collected from the SPE fractionation, SPE
concentration, and the amount of eluate used for testing
and GC/MS analysis should be measured at all steps.
The volume of eluate must be measured to determine the
actual toxicity concentration in each step of the proce-
dure. If it is expected that the Phase II data will be needed
later, it is prudent to measure the degree of toxicity in the
eight SPE effluent fractions at the onset of testing. We
rarely see blank fraction toxicity; therefore, there is little
need to evaluate the toxicity of the blank fractions with
dilutions.
2.2.1 Sample Volume
The volume of effluent needed depends on its
toxicity, the toxicity of the chemicals causing effluent
toxicity, and the sensitivity of the analytical method. Since
only the first of these will usually be known when Phase II
begins, trial and error will dictate volume size. For acutely
toxic effluents with LC50 values in the range of 25-100%,
2,000 ml have usually been adequate to perform one
complete Phase II procedure, i.e., C18 SPE and HPLC
fractionations, and GC/MS identification. Examples of the
variables that should be considered when deciding what
volume of effluent to fractionate are provided in Appendix
A, Tables A-1 and A-2.
2.2.2 Filtration
For acute tests, glass fiber filter(s) (1 u.m nominal
pore size) should be prepared as described in Section 8
of Phase I (EPA, 1991 A). Both 45 mm and 90 mm
diameter filters have been used routinely, the 90 mm filter
allows about four times more effluent to be passed over
the filter than one 45 mm filter. All filters and glassware
should first be pre-rinsed with pH 3 high purity water (e.g.,
Milli-Q® Water System, Millipore Co., Bedford, MA) to
remove any metal residues followed by a high purity
water rinse which is discarded. The filter should then be
rinsed with 200 ml of dilution water (rather than with high
purity water) and a sample collected after most of the
volume has been filtered, to provide the filter toxicity
blank. In subsequent steps, a dilution water column blank
will be collected after passing the filtered dilution water
through the SPE column. The same type of dilution water
should be used for the filter blank as for the column blank
(Section 2.2.4). Usually, a reconstituted water is used for
these procedures (EPA, 1991C).
The volume of effluent that can be passed through
a single filter is sample specific. If more than one filter is
needed (as is often the case) a single filter blank can be
prepared by stacking three to eight pre-rinsed filters in
one filter holder, followed by a dilution water rinse. The
filters are then separated and used one at a time to filter
the effluent sample. If samples are high in suspended
solids additional pre-filtration may be needed. Centrifuga-
tion may also be useful for reducing solids in the sample.
The decision to use a vacuum or a pressure system for
filtering should have been made during the filtration tests
of Phase I. If a volatile chemical is indicated in Phase I,
pressure filtration should be used.
Filtration equipment should be thoroughly cleaned
before use to prevent any toxicity carry-over or particle
buildup from previous samples. We have found that glass
vacuum filtering apparatus with stainless steel filter sup-
ports (for samples without pH adjustments), or plastic
pressure filtering devices are the most useful. We have
also found that if removable glass frits are used, they can
be rigorously cleaned with aqua regia for 20-40 min
followed by rinsing with copious amounts of water to
remove residual effluent particles, since glass frits may
act as a filter. The removable stainless steel filter sup-
ports do not require as rigorous cleaning as fritted glass-
ware, and therefore are a good substitute.
A portion of the filtered sample must be reserved
for toxicity testing while the rest is used for C18 extraction.
If the filtration toxicity blank exhibits slight or complete
toxicity, but the post C18 SPE column effluent is not toxic
(and effluent toxicity was unchanged after filtration), the
blank toxicity can be ignored since the effluent toxicity
was removed (see Phase I). However, as the identifica-
tion process continues, the blank toxicity will have to be
eliminated, or it could lead to a misidentification of the
cause of toxicity.
When effluent samples are readily filtered
(-2,000 ml for one 90 mm 1 urn filter) it may be possible
to filter the effluent for the filtration test of Phase I but then
use unfiltered effluent with the C18 SPE column test and
the methanol eluate test (Phase I). Once it has been
demonstrated that filtration does not reduce toxicity, rou-
tine filtering of these effluents (before passing the effluent
through the SPE column) can be eliminated. This will
reduce the amount of toxicity testing required.
2.2.3 Column Size
Various sizes of C18 SPE columns are available
ranging from 100 mg to 10,000 mg packing material. We
routinely have used Baker® 1,000 mg columns for 1,000
ml of effluent (J.T. Baker Chemical Co, Phillipsburg, NJ).
Volumes for a 1,000 mg C18 SPE column are used in the
following description, since this is the size most often
used for acutely toxic effluents. Other available column
sizes and the appropriate volumes to be used in their
preparation are listed in Table 2-1. Positive pressure
pumps (EPA, 1991 A) are convenient for the large volume
2-3
-------
effluent samples because flow rate can be controlled.
Vacuum manifolds can be used for drawing the small
samples and solvents through the column. Whichever
system is used, it must be made of materials that dilute
acid and solvents do not destroy, or from which chemicals
are not leached that are toxic or that interfere with analyti-
cal measurements. Teflon, glass, and stainless steel are
all acceptable choices.
2.2.4 C18 SPE Column Conditioning
The 1,000 mg C18 SPE columns are conditioned
by pumping 10 ml of 100% methanol through the column
at a rate of 5 ml/min. The pumping rate can be increased
to 40-50 nWmin when using the larger C SPE columns
(e.g., 5 g or 10 g). The volumes of conditioning solvent
recommended for other size columns are shown in Table
2-1. We most commonly use methanol as the condition-
ing solvent but other water miscible solvents such as
acetonitrile, ethanol, or isopropanol may be substituted.
Before the packing goes dry, 10 ml of high purity water
must be added. As the last of the high purity water is
passing through the column, 25 ml of filtered dilution
water is added. The last 10 ml of dilution water is col-
lected for a dilution water column blank. After the dilution
water has been collected, pumping is continued until no
dilution water emerges from the column.
2.2.5 Elution Blanks
To generate elution blanks from a 1,000 mg
column, two successive 1.5 ml volumes of 25% methanol/
water (%v/v) are pumped sequentially through the condi-
tioned column and collected in one analytically clean,
labeled glass vial to produce a 3 ml sample. This proce-
dure is repeated with two successive 1.5 ml volumes of
50%, 75%, 80%, 85%, 90%, 95% and 100% methanol/
water. The column should be allowed to dry for a few
seconds between each elution with the different 3 ml
volumes of methanoI/water solutions. This will result in
eight 3 ml SPE fraction blanks (Figure 2-2). The volume of
methanol solutions used for elution will vary depending on
column size as shown in Table 2-1.
2.2.5 Column Loading with Effluent
The same column is then reconditioned with 10
ml of 100% methanol and 10 ml of high purity water, as
described in Section 2.2.4. Without allowing the column to
dry, 1,000 ml of filtered effluent is pumped through the
column at a rate of 5 ml/min (Figure 2-3). The pumping
rate can be increased to 40-50 ml/min when using the
larger C18SPE columns (e.g., 5 g or 10 g). Three samples
(-25 ml) of the post-C18 SPE column effluent are collected
after 25 ml, 500 ml and 950 ml of the sample has passed
through the column. Each post-column aliquot is toxicity
tested to determine the presence of acute toxicity in the
post-column effluent. This information can be used to
determine whether the toxicant is removed from the efflu-
ent by the column. As Phase II progresses, the recom-
mendation is to increase the volume of post-column effluent
collected to 50-60 ml so that dilutions can be made and
LC50 values obtained. Pumping is continued until no
effluent emerges from the column.
The efficiency of the C18 SPE column is deter-
mined by the extraction efficiency (i.e., how well the
column sorbent removes the effluent components), and
the elution efficiency (i.e., how well sorbed effluent com-
pounds are removed from the column by the solvent
elution). For purposes of the TIE, "efficiency" applies only
to recovery of those compounds causing or affecting
effluent toxicity. Since most acute effluent tests do not
require large volumes of post-column effluent, the ques-
tion of extraction efficiency can be determined by measur-
ing the toxicity of the post-C18 column effluent sample.
The toxicity of each aliquot collected after different vol-
Tablo2-1. Solid Phase Extraction (SPE) Column Fractionation Information1
C SPE
Sorbent
Amount8 (mg)
100
500
1,000
5,000
10.000
Volume
Conditioning
Solvent (ml)
2
5
10
50
100
High Purity
Water
Volume (ml)
2
5
10
50
100
Maximum
Volume
Effluent (ml)
100
500
1,000
5,000
10,000
Minimum (500x)
Elution
Volume3 (ml)
2x0.1
2x0.5
2x 1.0
2x5.0
2x 10
Suggested (333x)
Elution4
Volume3 (ml)
2x0.15
2 x 0.75
2x1.5
2x7.5
2x15
'The Information is based on manufacturer's guidance and experimental data from ERL-D.
*The smaller columns (100, 500. and 1.000 mg sorbent) are available pre-packed from J.T. Baker Chemical Co., the larger columns (5 000 and
10,000 mg sorbent) are available pre-packed from Analytichem International. -a i -
'Elution with two successive allquots of the volume listed.
*Tho 333x concentration factor is most often used for acute work. >
2-4
-------
C18 SPE Column
Conduct Toxicity Test on Each Fraction
Figure 2-2. Procedures for eluting the SPE column with a gradient of methanol/water solutions.
Filtered
Effluent
Concentration Factor
1X
C-| g SPE Column
Test Post C-| g
Effluent
1X
Figure 2-3. Concentrating effluent on the C]8 SPE column.
2-5
-------
umes of effluent have passed through the column can be
compared. If there is toxicity in these aliquots, but it is
independent of the volume of effluent previously passed
through the column, then the post-column effluent toxicity
is probably caused by toxicants that are not extracted by
the column. If toxicity increases as the volume of post-
column effluent passed through the column increases,
the capacity of the column to sorb the toxicants was
probably exceeded.
In some post-column effluent, a biological growth
may occur during toxicity testing which may result in
artifactual toxicity. Such growth can make it appear as if
the toxicant is not removed by the column. While this
growth does not occur in all effluents, when it does occur
with one post-column effluent sample, the growth often
occurs in each subsequent post-column effluent sample
from the same preparation. The growth may appear to be
filamentous and give a milky appearance in the test
vessel. This effect has been linked to methanol stimula-
tion of bacterial growth. Methanol is present in the post-
column samples because a small amount of methanol is
constantly released from the column during the sample
extraction. Effluents from biological treatment plants may
develop this characteristic more readily than industrial
effluents.
Additional filtering of the post-column effluent
sample through a 0.2 urn filter before testing to remove
bacteria and eliminate the growth has been helpful. To
avoid art'rfactual toxicity as much as possible in the post-
column effluent, initiate the tests with the post-column
samples on the same day the effluent is extracted even if
fractions are not tested simultaneously. For those few
effluents where we have not eliminated this type of arti-
factual toxicity, holding the post-column effluent is prob-
lematic in that more time is available for bacteria to cause
problems in the post-column sample matrix. When post-
column artifactual growth is not readily eliminated, a
different solvent (e.g., acetonitrile) to condition the col-
umn (but not for eluting) may be useful in reducing the
post-column artifactual bacterial growth. This artifactual
growth has not occurred in the toxicity tests with methanol
SPE fractions (Section 2.2.8).
2.2.7
C1g SPE Column Elutlon
Once the effluent sample has been loaded onto
the column, elution can begin. To elute a 1,000 mg
column, two successive 1.5 ml volumes of the 25%
methanol/water mixture are pumped through the column
and collected in one labelled, analytically clean vial to
make a 3 ml sample. Subsequently, two successive 1.5
ml volumes of each of the 50%, 75%, 80%, 85%, 90%,
95% and 100% methanol/water are pumped through the
column and collected in separate vials (Figure 2-2). The
next elution volume should be added when no more of the
preceding one is emerging from the column.
This entire procedure (conditioning through elu-
tion) Is repeated using a second 1,000 mg C18 SPE
column for the second 1,000 ml of filtered effluent. The
dilution water column blank samples should be kept sepa-
rate. The corresponding fractions from the blank and the
sample from each 1,000 ml fractionation can be com-
bined. For example, the 3 ml 100% methanol sample
fraction from the first column and the 3 ml 100% methanol
sample fraction from the second column are combined to
produce a total of 6 ml. There will be eight 6 ml blank
fractions and eight 6 ml effluent fractions.
The vials containing the methanol/water fractions
are tested immediately or sealed with perfluorocarbon or
foil-lined caps and stored under refrigeration. These frac-
tions represent a "first cut" separation of effluent compo-
nents. Elution volumes will vary if columns of different
sizes are used or if the particular effluent under study or
the research question being posed dictates .method modi-
fication.
2.2.8 Blank and Effluent Fraction Toxicity
Tests
The next step is to determine the toxicity of the
blank and effluent fractions. While the choice of test
concentration depends on the toxicity of the effluent in
most instances, we have used a high test concentration of
2x or 4x (the LC50 or 100% effluent) for acutely toxic
effluents. The methanol content in the fractions limits the
concentration that can be tested, and at this point the
amount of methanol is assumed to be 100% in all the
fractions for dilution calculations; however, this is not
assumed for add-back tests (described below). Usually
120 u.l of each blank and sample fraction (333x) is in-
jected into separate 10 ml aliquots of dilution water to test
at 4x the 100% effluent2. This will give a 1.2% methanol
concentration which is below the methanol LC50 for both
C. dubia and fathead minnows in the 100% methanol
fraction. The resulting methanol concentration must be
adjusted for the species tested (see Section 8 of EPA,
1991 A). During the initial stages, five animals in each 10
ml aliquot are used without duplicates. Using the above
volumes, the tested solution is more concentrated (i.e., 4x) than
100% effluent, assuming 100% extraction and 100% elution
in one fraction. These test solutions can be diluted to
provide an LC50 for each sample fraction. Blank fractions
need not be diluted, since hopefully they are nontoxic.
Individual chemicals in the fractions could be
toxic even when they are not toxic in the whole effluent,
since the concentration tested may be as high as 4x
whole effluent. Therefore, to be toxic at whole effluent
concentrations, an individual fraction must have an LC50
of 25% or less. Since there is no way to know whether the
toxicant(s) eluted over more than one fraction or what the
percent extraction and elution efficiency are, fraction tox-
2ln the Phase II document published in 1989, testing at 5x was
recommended, the methanol level was 1.5% at this concentration. In
order to lower the methanol level, this was changed to 4x in this
document.
2-6
-------
icily up to 100% (4x whole effluent) should not be disre-
garded.
If toxicity occurs in any of the fraction blank tests
and it is small relative to the toxicity in the corresponding
sample-fraction (e.g., 20% mortality versus 80%), the
sample fraction results should not be dismissed. If all
organisms die in the blanks and the effluent fractions,
dilutions should be tested to make sure the sample frac-
tion is substantially more toxic than the blank. In general,
blanks should not have measurable toxicity.
If the SPE fractions are toxic at effluent concen-
trations of 1x or 2x and toxicity is reduced in two of the
three post-column effluent samples, the toxicant could
still be a non-polar organic compound. If the effluent
fractions are not toxic individually and the post-column
samples are non-toxic, it is possible that the toxicity has
been spread across several fractions or has not been
recovered from the column. Combining and concentrating
fractions may be useful or other elution procedures may
be necessary. If toxicity is observed in the fractions at 1x,
2x, or 4x and in the post-column effluent samples, it is
possible that not all the toxicity is caused by non-polar
compounds, that break-through of the toxicant has oc-
curred, or that the toxicity is artifactual.
In addition to concentrating column artifacts to
toxic levels, effluent constituents present at nonlethal
levels may be concentrated to toxic levels in this test if
they have a relatively high recovery value. Actual effluent
toxicants with poor recovery may not be present in these
test solutions at toxic levels. Spurious results of this
nature will be identified in the later stages of Phase II and/
or in Phase III.
Elution efficiency may be approximated by sum-
ming the amount of toxicity (i.e., TUs) in the toxic fractions
(provided dilutions are tested) and comparing this value
to whole effluent toxicity expressed as TUs. When sum-
ming acute toxicity, it is important that all values are for
comparable endpoints (i.e., LCSOs). Adding of TUs may
be somewhat imprecise for several reasons. A single
toxicant may occur in more than one adjacent fraction, in
which case a small amount of the toxicant in one fraction
may not be detectable because it is present below the
effect concentration. For acute toxicity, this problem may
be solved by combining a portion of each effluent fraction
(and separately testing the corresponding blank fractions)
and measuring total toxicity at 1x. If more than one
toxicant is present, the effluent fraction toxicity may not be
strictly additive in their toxicities, and when separated into
different fractions the sum of the fraction toxicities may be
low even if extraction and elution efficiencies were 100%.
Table 2-2 illustrates a hypothetical example. The toxicity
test results from the test with a portion of all fractions or a
few of the fractions may show somewhat greater toxicities
than those of the whole effluent. This may be caused by
enhanced toxicity due to matrix effects. When this occurs,
it may be possible to compensate for toxicity enhance-
ment by methanol, by adding methanol to the whole
effluent and evaluating the toxicity. This methanol addi-
tion may in turn stimulate
Table 2-2. Comparison of Toxic Units (TUs) in Each Toxic Fraction to
i TUs of All Fractions Combined and Whole Effluent
Toxic Fraction (% Methanol)
75
80
85
SUM
Combined Fractions
Whole Effluent
TUs
0.5
1.2
0.6
,2.3
2.7
2.5
biological growth, and if this happens, the test is negated.
We have rarely used this approach since the fractions
have seldom caused more toxicity than the effluent itself.
At this point in Phase II, the effluent fractions should also
be tested in water with TOG and suspended solids which
mimics the effluerit'to lessen matrix effects on toxicity. As
the identification step moves into Phase III, it is better to
use dilution water that mimics effluent or receiving water
characteristics.
2.2.9 SPE Fractions: Concentration and
Subsequent Toxicity Testing
The SPE fractionation provides a general separa-
tion of non-polar organics and except in relatively
uncomplicated effluents, GC/MS analysis of the concen-
trates of toxic C18 SPE fractions will result in very compli-
cated chromatograms from which the toxicant(s) cannot
be distinguished from other effluent components. A sec-
ondary fractionation using HPLC is often needed to fur-
ther simplify toxic effluent fractions prior to component
identification by GC/MS analysis.
In order to maximize the chromatographic sepa-
ration capability of the HPLC, the volume of the sample
injected onto an analytical size HPLC column should be
as small as possible (i.e., <0.5 ml); therefore, the toxic
SPE fractions (usually >1 ml) must be concentrated prior
to injection onto the HPLC column. This concentration
step will provide the added benefit of an increase in
concentrations of constituents in the HPLC fractions as
well as rid the SPE fractions of water. The latter issue is
important if GC/MS analysis will be performed on the
concentrated SPE fraction prior to injection on the HPLC
(Durhan et al., 1990).
, The volume of the fractions from the initial SPE
fractionation procedure and the number of fractions to be
combined will determine the size of the SPE column to
use for the concentration procedure. Table 2-3 contains
information on the column sizes we have found to be
most useful. In the procedure outlined below (Figure 2-4),
we have used a 100 mg column which is the most
commonly used size for concentrating SPE fractions of
acutely toxic effluents. Most often the toxic effluent SPE
2-7
-------
Table 2-3. Information for Concentrating SPE and HPLC Fractions'
C,,SPE
Sorbont
Amount (mg)
100
200
500
1,000
Volume
Conditioning
Solvent (ml)
1
2
5
10
High Purity
Water
Volume (ml)
1
2
5
10
Maximum
Toxic Fraction
Volume (ml)
20
40
100
200
Maximum
Diluted Fraction
Volume (ml)
100
200
500
1,000
Minimum
Elution
Volume2 (ml)
3x0.1
3x0.2
3x0.5
3x1.0
Approximate
Eluate
Volume (ml)
0.22
0-44
1.10
2.20
"Concentration information is based on manufacturers guidance and experimental data from ERL-D.
*EIut!on with three successive aliquots of the volume listed.
Diluted Toxic Fraction(s)
1
100 mg Ctl SPE Column Sorption
I (discard post-C|S
T effluent)
Dry Ot, Column with Nitrogen
(optional)
Elute Column with three 0.10 ml volumes
of 100% Methanol
I
Collect Eluate (Concentrate)
Conduct Toxicity Test
Rguro 2-4. Procedure to concentrate toxic SPE fractions.
fractions are combined and diluted with high purity water
and the corresponding blank fractions are treated simi-
larly. In cases where there are multiple non-polar toxi-
cants, and when toxicity occurs in several fractions, it may
be more useful to concentrate each fraction separately for
subsequent HPLC separation. The percent methanol in
the diluted fraction sample should be <20% and the
volume to which the fractions can be diluted is dependent
on the amount of column packing. For example, the total
volume of the diluted fractions should not exceed 100 ml
for the 100 mg C18 SPE column (Table 2-3). No more than
three toxic fractions of 6 ml each can be combined and
concentrated on the 100mg column. When the total
volume of combined fractions or the individual fraction
volume is above 20 ml, larger columns should be used;
consult Table 2-3 for column size and elution volume
Information. The effluent and blank concentrates and the
column blank are tested for toxicity to ensure that the
toxicant Is still in the concentrate and that artifactual
toxicity was not introduced by the procedure. If there is
not measurable toxicity in the concentrate, it is possible
that the percentage of methanol in the diluted f ra'ction was
too high. The concentration procedure should then be
repeated with a new set of toxicity tested fractions diluted
to a lower methanol concentration, e.g., 10%.
Below is an example of how to prepare effluent
and blank fraction concentrates. First, a 100 mg C]8 SPE
column is conditioned with 1 ml of methanol and 1 ml of
high purity water similar to the procedures described in
the SPE Column Conditioning Section (2.2.4). Column
blanks for toxicity testing are obtained by rinsing the
column with at least 20 ml of dilution water. After collect-
ing the column blank, recondition the column with 1 ml of
methanol and rinse with 1 ml of high purity water. The
diluted blank fractions (for dilution guidance see Table 2-
3) are then drawn through the 100 rng C18 SPE column
under a pressure of 380 mm Hg using a vacuum manifold.
Unlike the first fractionation step (Section 2.2.6) the post-
column sample cannot be tested for toxicity because of its
high methanol concentration (i.e., 10-20%). The column
is then dried for 10 min using a gentle flow of nitrogen (10-
20 ml/sec). Drying the column usually increases the re-
covery of toxicity, but sometimes toxicity is not recovered
from the column, possibly as a result of volatilization. If
this occurs the concentration procedure can be repeated
without the nitrogen drying step.
After drying the sorbent, the luertip of the column
is fitted with a luer-Iock needle and 100uJ of 100%
methanol is placed into the column using a microliter
syringe. Nitrogen is then applied to the column at a rate of
~4 ml/sec to force the methanol through the sorbent. The
luer-Iock needle is needed to ensure the collection of
small volumes; when using larger column sizes (e.g.,
>500 mg) this is not necessary. The first 100 uJ aliquot of
methanol applied to the column will yield approximately
25 uJ of eluate. Two more 100 uJ aliquots of 100%
methanol are also forced through the column. The final
volume of eluate collected will be approximately 220 u.l. If
desired, measure the exact volume collected (using a uJ
syringe) to calculate concentration factors (Table 2-3). As
in most chromatographic separations and extractions,
three separate smaller elutions of methanol are more
efficient than one large one.
2-8
-------
The 100 mg C18 SPE column is reconditioned
with 1 ml of methanol and rinsed with 1 ml of high purity
water. It is then used to concentrate the diluted toxic SPE
column fractions, using the same procedure used for the
blank fractions (Figure 2-4). In lieu of reconditioning the
same column, two columns can be conditioned, one used
for the diluted blank fractions and the other used for
concentrating the diluted toxic SPE fractions. The result-
ing column blanks should be toxicity tested separately.
The original effluent volume of 2,000 ml is now
concentrated into a 220 u.l sample or a nominal concen-
tration of 9,091x (ignoring the amount used for testing).
As work progresses and more quantitative results are
needed, the eluate volume must be measured to provide
the correct concentration factor. If 9 u.l of concentrate is
diluted to 10 ml in dilution water, the resulting test con-
centration will be 8x whole effluent. Additional test con-
centrations (e.g., 4x, 2x, 1x) can be prepared to determine
an LC50 of the concentrate, and toxicity recovery can be
calculated by comparing this LC50 to the LC50 of the
effluent. The concentrate toxicity might be higher than the
sum of the individual toxic fractions because some of the
toxicant may have been in adjacent fractions that were
concentrated in the first step (Section 2.2.7) but not
detectable by the toxicity test of the single fraction. The
concentrate toxicity may also be lower .than expected
because of low extraction and elution efficiencies. Where
greater concentration factors are desirable, SPE fraction-
ation should be repeated with additional volumes of efflu-
ent, followed by combining the toxic fractions before
concentration. The size of the column used for concen-
trating may have to be increased, along with the appropri-
ate changes in .dilution and elution volumes (Table 2-3).
The important concern here is not 100% recovery
of toxicants but enough recovery for GC/MS analyses and
to obtain measurable toxicity in the HPLC fractions. If
recovery is too low, changing or eliminating the column
drying time may help. Sometimes recovery appears to
increase with drying time while other compounds are
volatilized from the column during the drying process. For
concentrates analyzed using GC/MS, column drying to
remove water is critical to GC column performance.
2.2.10 HPLC Separation
The same column packing functionality should be
used in the HPLC column as the SPE column. At later
stages, when more is known about the toxicants, other
sorbent types may be used.
The HPLC conditions presented in this section
are general. As more information on the effluent is gath-
ered, HPLC conditions should be modified to achieve
better separation and higher concentration factors. We
use a flow rate of 1 ml/min on an instrument equipped with
a 5 urn C reverse phase column (250 mm x 4.6 mm i.d.).
The HPLC elution conditions will change depending on
which SPE fractions have been concentrated. The HPLC
conditions for the four most commonly toxic SPE fractions
are listed in Table 2-4. Depending on the size of the
Table 2-4. Example HPLC Elution Gradients fpr Four Commonly Toxic
SPE Fractions
75% or 85% SPE Fractions 85% or 90% SPE Fractions
Time (min) % Methanol/Water Time (min) % Methanol/Water
0
1
13
20
25
50
60
90
100
100
0
1
13
'20
25
60
70
90
100
100
HPLC injector and column, more than one HPLC fraction-
ation run may be required to fractionate the entire blank
concentrate. When multiple HPLC fractionations are con-
ducted, collect all the corresponding HPLC fractions in
the same set of vials. For example, if two HPLC fraction-
ations were performed for the blank concentrate, 25-2 ml
HPLC fractions would be obtained.
Using the HPLC equipment described above, all
of the blank concentrate remaining after toxicity testing is
injected (<500 u.I) and 25-1 ml fractions are collected in
analytically clean glass vials (Figure 2-5). The same
procedure is followed using the effluent sample concen-
trate. The vials should be sealed (e.g., with foil lined caps)
and stored at 4°C after use. As soon as toxicant identifica-
tion is obtained by GC/MS, then HPLC conditions (gradi-
ent, fraction size, and number of fractions) can be optimized
for further fractionations.
2.2.11 HPLC Fraction Toxicity Tests
Before specific toxicants are identified, toxicity
tests on each HPLC blank fraction and sample fraction
are conducted using non-replicated exposures of five
animals each. The amount of methanol in the HPLC
fractions limits the concentration that can be tested. As-
sume that each fraction is 100% methanol to calculate the
necessary dilution. A methanol concentration of 1.2%
should not be exceeded for C. dubia and fathead minnow
acute toxicity tests.
Fractionate SPE Concentrate
Using HPLC
I
Collect 25-1 ml HPLC
Fractions
1
Conduct Toxicity Test on Each
Fraction
Figure 2-5. Procedure to fractionate acutely toxic SPE concentrates
using HPLC.
2-9
-------
For acute studies, when all of the SPE fraction
concentrate remaining after toxicity testing is injected
(one injection) on the HPLC (Figure 2-5), each resulting 1
ml HPLC fraction equals 2,000 ml of effluent (assuming
no loss and toxicant elution in only one fraction) or a
2,000-fold concentration. If each HPLC fraction is then
diluted for testing (80 uJ to 10 ml) the resultant concentra-
tion is 16x the original effluent concentration. In prelimi-
nary Phase II testing the HPLC fractions are tested without
dilutions. Only the toxic HPLC fractions are tested again
with dilutions to generate an LC50. Some loss of toxicant
tends to occur in each concentration step and the result-
ing toxic'rty may be decreased relative to the original
effluent.
The blank fractions should not be toxic. If they
are, then additional tests with dilutions must be conducted
on both blanks and toxic fractions to find out whether
there is enough additional toxicity in the sample fractions
to warrant analysis.
The toxicity of the HPLC fractions should be
tested at twice (at least) the concentration at which the
original SPE column fractions were tested because re-
covery of toxicity and analytical measurements indicate
that up to 50% of the initial concentration of toxic com-
pounds may be lost in this step (Durhan et al., 1990). The
amount of methanol should not exceed the amount used
in the SPE fraction tests described above (Section 2.2.8).
2.2.12 HPLC Fractions: Concentration and
Subsequent Toxicity Testing
The HPLC fractions that exhibit toxicity and their
corresponding blank fractions must be concentrated in a
solvent suitable for GC/MS or other analytical techniques.
The procedure is identical to that described in Section
2.2.9 and is depicted in Figure 2-6. Judgement must be
used to decide whether to concentrate each toxic fraction
separately or to combine various toxic and adjacent frac-
tions prior to concentration. If, for example, three succes-
sive fractions exhibit toxicity, there is a good chance that
the same toxicant is in all three. If there are other,fractions
that show toxicity but they are separated from the first set
by several non-toxic fractions, there is high probability
that the second set contains a toxicant different from the
first three. There is also a good chance that at least one
non-toxic fraction on either side of the toxic fractions
contains some of the tpxicant(s). The advantage of com-
bining fractions is to reduce the work load andjncrease
concentration In the final concentrate. The disadvantage
is that more constituents that are not the toxicant(s) will
also be concentrated. This decision is not always straight-
forward and must be based on trial and error, and experi-
ence. Blank fractions corresponding to the toxic fractions
are concentrated the same way.
The HPLC fraction and blank concentrates should
be finally checked for toxicity before GC/MS analysis.
This concentrate is now nominally 9,091x more concen-
trated than the effluent. If 18 jil is diluted to 10 ml, the
resultant test concentration will be 16x the original sample
Diluted HPLC Fraction^)
100 mg C18 SPE Column Sorption
| (discard post -C18
effluent)
Dry Column with Nitrogen (optional)
1
Elute Column with three 0.10 ml
Volumes of Methanol
i
Collect Eluate (Concentrate)
Tox
1
Conduct Toxicity Test
1
Analyze Concentrate on GC/MS
Figure 2-6. Procedure to concentrate toxic HPLC factions.
concentration. To quantitate toxicity and use the Phase II
data later, additional lower concentrations should be tested
(e.g., 8x, 4x, 2x); TUs of this concentrate can then be
compared to previous toxicity test results. The HPLC
concentrate should be tested at one to two times the test
concentration of the HPLC fraction tests (i.e., 16x or 32x).
This is the last opportunity to assure that the toxicant is
still present in the concentrate before it is subjected to
GC/MS analysis. Whether the toxicant is detected by the
analytical detector (mass spectrometry in our laboratory)
is always a question. Since GC/MS detects only about
20% of organic chemicals (EPA, 1989B), even such a
broad spectrum method is not certain to identify the
toxicant. As work progresses with more samples of the
effluent and quantitative results are needed, the amount
of eluate collected should be carefully measured and
recorded to accurately calculate the concentration fac-
tors. In addition, the volume of concentrate removed for
toxicity testing and analytical analyses should also be
recorded.
2.3 Chronic Toxicity: Fractionation and
Toxicity Testing Procedures
The chronic Phase II non-polar organic toxicity
identification follows the same general approach and
employs manipulations similar to those described for the
acutely toxic non-polar organic compounds (Section 2.2).
One major difference is that the concentration of the
eluting solvent (e.g., methanol) must be lower in the
chronic toxicity tests than in acute tests. In the initial
stages of Phase II, toxicity tests may be conducted on C18
SPE effluent fractions and blank fractions to detect the
presence of toxicants, and not to quantify the magnitude
2-10
-------
of the toxicity in each. However, as suspect toxicants are
identified, quantitative toxicity measurements will be
needed to compare with the analytical measurements. If
Phase II data will be used to correlate effluent toxicity to
toxicant concentrations (Phase III), then more replicates
per concentration, randomization of test concentrations,
careful observation of organism exposure times, and
organisms of approximately the same age should be used
(Section 1.2). Also the amount of eluate that is collected
from the SPE fractionation, SPE concentration, and the
amount of eluate used for testing and GC/MS analysis
should be measured at all steps. If it is expected that the
Phase II data will be needed later, it is prudent to measure
the degree of toxicity in the SPE effluent fractions (Sec-
tion 2.3.5) at the onset of testing. We rarely see blank
fraction toxicity, therefore, there is little need to evaluate
the blank fraction toxicity with dilutions. Also, the volume
of eluate must be measured to determine the actual
toxicity concentration in each step of the procedure.
The following discussion is based on our experi-
ences with C. dubia and fathead minnows (see Section
1.2). The use of other species will require reconsideration
of the appropriate test volumes and methanol concentra-
tion for each step. Chronic testing is more labor intensive
and generally requires more effluent sample volume than
acute testing. For the most part, in the descriptions below,
for C. dubia there are five replicates containing 10 ml of
test solution and one animal per cup. For the fathead
minnow tests, two replicates of 10 animals per 50 ml and
the control are usually used (Section 1.2). Typically we
use four concentrations and a control.
As soon as the cause of toxicity has been deter-
mined to be a non-polar organic compound (e.g., metha-
nol eluate test; EPA, 1992) it is prudent to concentrate
large volumes of effluent for the subsequent analyses. By
concentrating large amounts of the effluent it is possible
to plan the optimal usage of the amount of column eluate
available for toxicity testing.
2.3.1 Sample Volume
The volume of effluent needed depends on its
toxicity, the toxicity of the chemicals causing effluent
toxicity, and the sensitivity of the analytical method. Since
only the first of these will usually be known when Phase II
begins, the volume of effluent to process should be
considered at the beginning of the identification process
to minimize the amount of re-fractionating and re-testing
of effluent and fractions. Ideally, fractionation should pro-
vide enough volume of post-column effluent (Section
2.3.6), C18 SPE fractions (Section 2.3.8), SPE fraction
concentrates (Section 2.3.9), HPLC fractions (Section
2.3.11), and HPLC fraction concentrates (Section 2.3.12)
to conduct all chronic toxicity testing and chemical analy-
ses. Because of the many factors affecting the amount of
effluent needed, a significant amount of thought should
be put into the volume of effluent to obtain and process at
one 'time. It is prudent for the investigator to anticipate
how many identification procedures will be done, and
then calculate the volume of effluent needed using the
particular test parameters desired, before extracting any
effluent to ensure that sufficient volume of fractions, con-
centrates, and post-column effluent is available for the
planned procedures. It may be best to perform these
calculations with .several different effluent volumes and
test conditions to ascertain the optimal volume of effluent
to fractionate. A worksheet to assist with these calcula-
tions and an example are provided in Appendix A.
The volumes of eluate needed for chronic toxicity
testing at 2x, 1x, and 0.5x are provided in Table 2-5 for
the C. dubia and fathead minnow short-term tests based
on the methanol concentration that can safely be used for
the chronic tests. The amount of SPE fractionation eluate
needed for toxicity testing is presented for the range of
tests that are commonly performed with C. dubia or
fathead minnows, these volumes can be used in the
calculation worksheets found in Appendix A (Table A-1).
The approximate volume of effluent that will be needed
for testing with C. dubia and fathead minnows is listed in
Table 2-6 for various fractionation schemes and toxicity
testing parameters. When only a portion of the TIE proce-
dures will be used, obviously less effluent volume will be
needed. In Table A-2, the example calculations are based
upon the use of minimum elution volume for the SPE
columns (Table 2-1), concentrating only one SPE fraction
(Section 2:3.9), and taking into account the toxicity testing
(Sections 2.3.8, 2.3.9, 2.3.11, and 2.3.12) and GC/MS
analysis volumes (Section 2.5). These parameters are
discussed in detail below. If additional eluate is needed,
the chronic tests must be repeated for each fractionation.
In Phase II and Phase III more confidence in the toxicity
estimates is needed than in Phase I, therefore tests may
require more replicates. The volumes needed for those
tests are also presented in Table 2-5. When only limited
amounts of effluent are available, one must be creative
and plan its usage very carefully to obtain meaningful
results.
2.3.2 Filtration
For filtration of chronically toxic effluents, the use
of glass fiber filters (1 u.m nominal pore size) is recom-
mended. Both 45 mm and 90 mm diameter filters have
been used routinely, but the 90 mm diameter filter allows
about four times more effluent to be passed over one filter
than the 45 mm filter. All filters and glassware should first
be pre-rinsed with pH 3 high purity water to remove any
residual metals followed by a high purity water (e.g., Milli-
Q® Water System) rinse which is discarded. Low levels of
metals (e.g., ug/l) from the filters may cause toxicity
interferences and pre-rinsing the filters may provide cleaner
blanks and less contamination in effluent samples. To
collect the dilution water filter blank, first pass a volume
(-200 ml) of dilution water over the filter and discard it.
Next, collect the volume of dilution water needed to
conduct the filtration blank test. It is a good idea to prepare
excess volume, at least 500 ml for the C. dubia 7-d test and 800
ml for the fathead minnow 7-d test. A portion of the filtered
dilution water is collected for testing and a portion is re-
served for the solid phase extraction test blank (Section
6.6; EPA, 1992).
2-11
-------
Table 2-5. Eluata Volumes Needed for Chronic SPE Fraction Toxicity Tests with Ceriodaphnia dubia and Pimephales promelas
Test
Species
C. dubia
C. dubia
C.dvbia
C. dubia
C. dubia
C. dubia
C. dubia
C, dubia
P. promelas
P. promotes
P. promotes
P. promotes
Test
Duration
4-d
4-d
7-d
7-d
4-d
4-d
7-d
7-d
7-d
7-d
7-d
7-d
Original
Sample &
No. Renewals
2
4
3
7
2
4
3
7
7
7
7
7
High Test
Cone, of
SPE Fraction
2x
2x
2x
2x
2x
2x
2x
2x
2x
2x
4x
4x
No.
Rep
5
5
5
5
10
10
10
10
2
4
2
4
Volume (ml) of 500x
Eluate Needed
for Testing1
0.70
1.40
1.05
2.45
1.40
2.80
2.10
4.90
4.90
9.80
9.80
19.60
Test
Concentrations
2x, 1x, 0.5x
2x, 1x, O.Sx
2x, 1x, O.Sx
2x, 1x, O.Sx
2x, 1x, O.Sx
2x, 1x, O.Sx
2x, 1x, O.Sx
2x, 1x, O.Sx
2x, 1x, O.Sx
2x, 1x, O.Sx
4x, 2x, 1x
4x, 2x, 1x
'Test volumes per replicate are 10 ml/cup for C. dubia and 50 ml/cup for P. promelas. The fraction test solutions are prepared as one solution and
divided into aliquots for the replicates. For the SOOx eluate concentration, this volume is based on the assumption that the C. dubia test solutions
are prepared as 200 uj of SOOx into 50 ml for 2x, 100 uJ into 50 ml for 1x, and 50 ul into 50 ml for 0.5x. More volume will be needed if serial
dilutions are prepared (400 uJ vs 350 uJ). For the fathead minnow tests this assumes test solutions are prepared as 400 ul into 100 ml for 2x, 200
ml into 100 ml for 1x, and 100 uJ into 100 ml for O.Sx. More volume will be needed if serial dilutions are prepared (800 ul vs 700 ul). For the 4x
fathead minnow test, 800 uJ per 100 ml can be prepared in a similar manner.
Table 2-6. Approximate Effluent Volumes Needed for the Chronic Non-Polar Organic Identification Procedures'
Test
Species
C. dubia
C. dubia
C. dubia
P. promotes
P. promelas
Test
Duration
4-d
7-d
7-d
7-d
7-d
Original
Sample &
Number of
Renewals
4
3
7
7
7
High Test
Cone, in
SPE Fraction
Test
2x
2x
2x
2x
2x
No.
Rep.
5
5
5
2
4
Are
Dilutions
Used?
Yes
Yes
Yes
Yes
No
Volume Effluent (ml)
Needed to Conduct
SPE & GC/MS2
Analyses
3,000
2,000
5,000
10,000
5,000
Volume Effluent (ml)
Needed to Conduct
SPE & HPLC & GC/MS3
Analyses
15,000
15,000
20,000
50,000
40,000
'Calculation of toxlcity testing volumes assumes that: 4x high concentration for SPE concentrate test (Section 2.3.9), 8x high concentration for
HPLC fraction test (Section 2.3.11), 16x high concentration for HPLC concentrate test (Section 2.3.12), concentration of only one toxic fraction
(SPE and HPLC), the maximum amount of sample is concentrated on the SPE columns and all SPE columns are eluted with the minimum elution
volume.
*TIE procedures used: SPE fractionation and GC/MS of SPE concentrate.
JTIE procedures used: SPE fractionation, GC/MS of SPE concentrate, HPLC fractionation, and GC/MS of HPLC concentrates.
2-12
-------
After the filtration blank has been obtained, the
effluent sample is filtered using the same filter, a portion
of the filtrate is collected for toxicity testing, and a portion
is set aside for concentrating on the C18 SPE column. For
some effluents, one filter will often not suffice. A tech-
nique we use to prepare several filters at once is stacking
three to eight filters together in one filter holder, followed
by sequential rinses with pH 3 high purity water, high
purity water and dilution water (using the same rinse
volumes as above). Finally, the filters are separated and
set aside, using one at a time for the effluent sample. If
the samples have high suspended solids concentration,
pre-filtering using a larger pore size filter may help, and
the appropriate blanks should be used. If the sample
cannot be effectively filtered due to the presence of many
fine particles, centrifugation may be used (of course,
blanks must be prepared).
The filter housing should be thoroughly cleaned
before use to prevent any particle build-up or toxicity
carry-over from previous samples. We have found large
filtration apparatus (1,000 ml), removable glass frits, or
plastic filtering apparatus (e.g., Millipore®) to be useful.
The glassware cleaning procedure that is described in the
acute Phase I TIE manual (EPA, 1991 A) is sufficient for
chronic TIE work.,The glass frits may require rigorous
cleaning (i.e., soak in aqua regia for 20-40 min) to remove
residuals that may remain after filtering, since the glass
frit may itself act as a filter. Also available are removable
stainless steel filter supports in a glass vacuum filter
apparatus (available from Millipore®). These filter sup-
ports do not require as rigorous cleaning as fritted glass-
ware, and therefore are a good substitute.
When effluent samples are readily filtered
(~2,000 ml for one 90 mm 1 u,m filter) it may be possible
to filter the effluent for the filtration test of Phase I but then
use unfiltered effluent with the C18 SPE column test and
the methanol eluate test (Phase I). Once it has been
demonstrated that filtration does not reduce toxicity, rou-
tine filtering;of these effluents (before passing the effluent
through the SPE column) can be eliminated, which will
reduce the amount of toxicity testing required.
2.3.3 Column Size
Available C18 SPE column sizes and the appropri-
ate water and solvent volumes used in their preparation
are listed in Table 2-1. Positive pressure pumps are the
most convenient to use for the large volume effluent
samples because the flow rate can be controlled. Pumps
and vacuum manifolds can both be used for eluting C18
SPE columns. Whichever system is used, it should be
made of materials that dilute acid and solvents do not
destroy, or from which chemicals are not leached that are
toxic or interfere with analytical measurements. Teflon,
glass, and stainless steel are all acceptable.
When SPE is used for isolating non-polar or-
ganic toxicants, use the maximum volume of effluent and
the minimum elution volume for the column size selected
to optimize the concentration of toxicants in the methanol
eluates. For example, if 6,000 ml is processed, it is best to
use one 5,000 mg column with 5,000 ml and one 1,000 mg
column with 1000 ml of effluent and elute both columns
with the minimum elution volumes (Table 2-1) and com-
bine eluates. It is always best to process the maximum
volume of effluent on each column to achieve the highest
concentration of toxicants in the eluate.
2.3.4 Cia SPE Column Conditioning
The 10,000 mg C SPE columns (Analytichem
International, Harbor City, CA) are conditioned by pump-
ing 100 ml of methanol through the column at a rate of 40-
50 ml/min. This size column can process 10,000 ml of
effluent and is the largest commercially pre-packed SPE
column available at this time. The example presented in
this section will be for 10,000 ml effluent using a 10,000
mg SPE column. The volumes of conditioning solvent will
change when other size columns are used, as shown in
Table 2-1. We most commonly use methanol as the
conditioning solvent but other water miscible solvents
such as acetonitrile, ethanol, or isopropanol may also be
used to condition columns. Before the packing goes dry,
100 ml of high purity distilled water must be added. As the
last of that water is passing through, filtered dilution water
is added. The volume of dilution water needed may vary
from 250 ml to 1,200 ml depending on the species tested.
The first 100 ml is discarded and the remainder is col-
lected for the dilution water column blank. After the dilu-
tion water has been collected, pumping is continued until
no water emerges from the column.
Low dissolved oxygen (DO) in the post-column
dilution water blanks (even in reconstituted waters) has
occurred during some chronic tests; therefore, we discard
the first 100-200 ml and collect the remainder of the post-
column dilution water. Low DO has been a problem,
particularly in the fathead minnow growth test, and is
attributed to the small amount of methanol that bleeds
into the post-column sample. This may be alleviated by
discarding the first post-column aliquots.
2.3.5 Elution Blanks
For chronic work, we have been using seven
methanol/water fractions (50%, 75%, 80%, 85%, 90%,
95%, and 100%) rather than the eight used in acute TIEs.
By eliminating 25% methanol/water fraction used in acute
work the toxicity testing workload is reduced, in turn a
reduction in separation of toxic and non-toxic components
can occur. _
To collect the fraction blanks from the 10,000 mg
column, two successive 10 ml volumes of 50% methanol
in water are pumped through the conditioned column and
collected in one analytically clean labeled vial, to make a
20 ml sample. This procedure is repeated six more times
with two successive 10 ml volumes of 75%, 80%, 85%,
90%, 95% and 100% methanol/water solutions. The col-
umn should be allowed to dry for a few seconds between
each elution with the different 20 ml volumes of methanol/
water mixtures. This will result in seven 20 ml blank SPE
2-13
-------
fractions. The volume of methanol solutions used for
elution will vary depending on column size as shown in
Table 2-1.
2.3.6 Column Loading with Effluent
The same 10,000 mg column is reconditioned
with 100 ml of 100% methanol and 100 ml of high purity
water, as described in Section 2.3.4. The sorbent must be
reconditioned when the maximum volume of dilution wa-
ter has been passed over the column, otherwise the
sorbents' capacity will be exceeded. After the high purity
water rinse and without allowing the column to dry,
10,000 ml of filtered effluent sample is pumped through
the column at a rate of about 40 - 50 ml/min.
Discard the first 100-200 ml of post-column efflu-
ent, to reduce the possibility of higher concentrations of
methanol in post-column samples, which may contribute
to art'rfaclual toxicity. To evaluate the post-C18 SPE col-
umn effluent for toxicity, collect at least two aliquots (e.g.,
beginning and end) separately. If only small quantities
(<500 ml) of post-column effluent are needed for toxicity
testing (e.g., C. dubiatesl), several separate post-column
effluent samples may be more helpful in determining if the
toxicants are retained by the column. About 800 ml of
post-column effluent is needed for the fathead minnow
test if only one concentration (100%) of post-column
effluent is tested for toxicity. If two concentrations (100%
and 50%) are used, then the required volume for that
species increases to 1,200 ml for each post-column ali-
quot. As Phase II progresses, the recommendation is to
collect enough post-column effluent to conduct toxicity
tests with dilutions.
2.3.7 C18 SPE Column Elution
To elute the C18 SPE column, two successive 10
ml volumes of the 50% methanol/water mixture are pumped
through the column and collected in one labelled, analyti-
cally clean vial. Subsequently, two successive 10 ml total
volumes of each of the 75%, 80%, 85%, 90%, 95% and
100% methanol/water solutions are pumped through the
column and collected in separate vials. The next elution
volume should not be added until no more of the preced-
ing one is emerging from the column. This results in
seven 20 ml SPE fractions. If one 5 g column and two 1 g
columns are used to concentrate 7,000 ml of effluent, the
corresponding fractions can be combined. For example,
the 10 ml eluate of the 80% fraction from the 5 g column
can be combined with the two 2 ml 80% fractions from the
two 1 g columns. This applies to both sample and blank
fractions for a total of 14 ml.
This entire procedure (conditioning through elu-
tion) is repeated using a second 10,000 mg C18 SPE
column for a second 10,000 ml of filtered effluent. The
dilution water column blank samples should be kept sepa-
rate. The corresponding fractions from both the blanks
and the sample from each 10,000 ml fractionation can be
combined as described above. There will be seven 40 ml
blank fractions and seven 40 ml effluent fractions, repre-
senting 20,000 ml effluent.
The vials containing the methanol/water fractions
are sealed with perfluorocarbon or foil-lined caps, and
stored at 4°C if not tested immediately. These fractions
represent a "first cut" separation of effluent components.
Volumes will vary if columns of different sizes are used or
if the particular effluent under study or the research
question posed dictates method modification.
2.3.8 Blank and Effluent Fraction Toxicity
Tests
While the choice of test concentration depends
on the toxicity of the effluent (Section 1.2), in most in-
stances we have used a concentration of 4x or 2x as the
high test concentration for testing SPE fractions. The high
test concentration of the SPE fraction is in part controlled
by the tolerance of the organisms to methanol. For chronic
testing the concentration of methanol should be less than
0.6% for C. dubia, and less than or equal to1 % for fathead
minnows (see Phase I; EPA, 1992).
If the minimum elution volumes are used, typi-
cally SPE eluates are 500x effluent concentration. For
fathead minnow testing, eluates can be toxicity tested at
4x effluent concentration by diluting 80 u.l to 10 ml, which
results in a 0.8% methanol concentration. For C. dubia,
eluates can be toxicity tested at 2x the 100% effluent
concentration by diluting 40 ul to 10 ml which results in a
methanol concentration of 0.4%. If there is the need to
toxicity test the 500x eluate with C. dubia at 4x then the
SPE eluates can be concentrated by gently airing the
eluate down (using nitrogen) to half its original volume.
However, by using this procedure you risk losing the
toxicant because of evaporation or insolubility. Also real-
ize that when a water and methanol mixture is aired
down, the percent methanol composition changes, be-
cause methanol will evaporate faster than water.
If toxicity occurs in any of the fraction blank tests
and it is small relative to the toxicity in the corresponding
sample fraction, the sample fraction results should not be
dismissed. If all organisms die in the blanks and effluent
fractions, dilutions of each should be tested to make sure
the sample fraction is substantially more toxic than the
blank. In general, blanks should not have measurable
toxicity.
When the post-column effluent sample is toxic
and the fractions are toxic at effluent concentrations of 1x
or 2x, the toxicant could still be a non-polar organic
compound. If the fractions are not toxic individually and
the post-column sample is non-toxic, it is possible that the
toxicity is spread out among the fractions. Combining and
concentrating these fractions may be useful or other
elution procedures may be necessary (Section 2.6). If the
fractions are toxic and the post-column effluent is toxic, it
is possible that the toxicant(s) is in the fractions, and that
either an additional toxicant(s) is present in the post-
2-14
-------
column effluent, that break-through of the toxicant(s) oc-
curred, or that the toxicity is artifactual. If toxicity is
recovered at 1x, 2x, or 4x and in one of the post-column
effluent samples, it is possible that not all the toxicity is
caused by non-polar organic compounds or the possibility
exists of break-through in the post-column sample.
For the chronic TIE, the question of extraction
efficiency cannot be as readily addressed as it is for the
acute TIE (Section 2.2.8). Measuring the chronic toxicity
of the post-column effluent will be limited by the species
tested, the test volumes required for the test and the
frequency of sample replacement. Without a measure of
the toxicity in the post-column effluent, conclusions re-
garding extraction efficiency are difficult to make. The
limitations created by this concern are addressed in Phase
III (EPA, 1993A). Artifactual toxicity in the post-column
effluent has been a problem in chronically toxic effluents
as it was in some acutely toxic effluents. For a detailed
discussion of this artifactual toxicity that appears as a
biological growth and suggestions to avoid it please refer
to Sections 2.2.6 and 2.3.4., and EPA, 1992.
At this point in Phase II, the effluent fractions
should also be tested in water with TOC and suspended
solids that mimic the effluent to lessen matrix effects on
toxicity. As the identification step moves into Phase III, it
is better to use dilution water that mimics effluent or
receiving water characteristics.
2.3.9 SPE Fractions: Concentration and
Subsequent Toxicity Testing
The SPE f ractionation provides a general separa-
tion of non-polar organics and except in relatively
uncomplicated effluents, GC/MS analysis of the concen-
trates of toxic C18 SPE fractions will result in very compli-
cated chromatograms from which the toxicant(s) cannot
be distinguished from other effluent components. A sec-
ondary fractionation using HPLC is often needed to fur-
ther simplify toxic effluent fractions prior to component
identification by GC/MS analysis.
In order to maximize the chromatographic sepa-
ration capability of the HPLC, the volume of the sample
injected onto an analytical size HPLC column should be
as small as possible (i.e., <0.5 ml); therefore the toxic
SPE fractions (usually >1 ml) must be concentrated prior
to injection onto the HPLC column. This concentration
step will provide the added benefit of an increase in
concentrations of constituents in the HPLC fractions as
well as rid the SPE fractions of water. The latter issue is
important if GC/MS analysis will be performed on the
concentrated SPE fraction prior to injection on the HPLC.
The volume of the SPE fraction and the number
of toxic fractions to be combined will determine which size
SPE column will be used for the concentration procedure.
Table 2-3 contains information on column sizes and the
appropriate volume of conditioning and eluting solvents
we have found to be most useful. In the procedures
outlined below we have used a 200 mg SPE column to
concentrate one 40 ml toxic fraction from two 10,000 mg
SPE columns. Often the toxic effluent SPE fractions are
combined and diluted with high purity water. If enough
toxicity occurs in each fraction it may be more useful to
concentrate each fraction separately for subsequent HPLC
separation. The corresponding blank fractions are simi-
larly treated. The percent methanol in the diluted fraction
sample should be <20% and the volume to which the
fractions can be diluted is dependent on the amount of
column packing. For example, the total volume of the
diluted fraction(s) should not exceed 200 ml for the 200
mg C18 column (Table 2-3).
A 200 mg C18 column is conditioned with 2 ml of
methanol and rinsed with 2 ml of water similar to the
procedures described in the SPE Column Conditioning
Section (2.3.4). The diluted blank fractions are then drawn
through the 200 mg C18 SPE column under a pressure of
380 mm Hg using a vacuum manifold. When processing
larger volumes, or using larger columns, positive pres-
sure can be used. The solution passing through the
column cannot be tested for toxicity because of its high
methanol concentration (e.g., 10-20% methanol). The
column is then dried for 10 min using a gentle flow of
nitrogen (10-20 ml/sec). Drying the column usually in-
creases the recovery of toxicity, but sometimes toxicity is
not recovered from the column possibly because of vola-
tilization. If this occurs, the concentration procedure can
be repeated without the nitrogen drying step.
After drying the sorbent, the luer tip of the column
is fitted with a luer-lock needle (to ease collection of small
volumes) and 200 u.l of 100% methanol is placed into the
column using a microliter syringe. Nitrogen is then applied
to the column at a rate of ~4 ml/sec to force the methanol
through the sorbent, which is then collected in a small
glass vial. The first 200 u.l aliquot of methanol applied to
the column will yield approximately 125 uJ of eluate. Two
more 200 uJ aliquots (applied separately) of 100% metha-
nol are also forced through the column. The final volume
of eluate collected will be approximately 440 uL. Measure
the exact volume collected using a u.I syringe or pipet. As
in most chromatographic separations and extractions,
three separate smaller elutions of methanol are more
efficient than one large one.
The 200 mg C18 SPE column is reconditioned
following the directions given above in Section 2.3.4. It is
then used to concentrate the diluted toxic SPE fractions,
using the same sequence used for the blank fractions
(Figure 2-4). The concentrated blank fractions will serve
as the dilution water column blank because it cannot be
obtained for chronic toxicity testing as it can for acute
testing.
When the total volume of fractions is above 40
ml, larger columns should be used; consult Table 2-3 for
column size and elution volume information. The size of
the column used for concentrating should be chosen to
maximize concentration in the eluate. Therefore, choose
the smallest column appropriate for the diluted fraction
volume.
2-15
-------
If there is a large toxicity loss after the concentra-
tion step, it is possible that the percentage of methanol in
the diluted fraction was too high. The concentration pro-
cedure should then be repeated with a new set of toxicity
tested SPE fractions diluted to a lower methanol concen-
tration (e.g., 10%). Both the effluent and blank concen-
trates are toxicity tested at each step to track toxicity.
Generally we suggest that this toxicity test be at least at
two times higher than the concentration used in the first
SPE fraction test. The tests are conducted exactly as the
SPE fraction tests.
if an original effluent volume of 20,000 ml (using
two 10,000 mg SPE columns) is now represented by a
440 uJ concentrate, then the sample is 42,670x more
concentrated than the effluent (accounting for volume removed
for toxic'rty testing, see Table A-2 example). If 1 |*l of
concentrate is diluted to 10 ml in dilution water, the resulting
test concentration will be about 4x whole effluent. How-
ever, the 4x test solution should be prepared as one
sample before solutions are spirt among replicates. For
example, 5 uj is diluted to 50 ml for five replicates with the
C. dubiaiest described above (Table A-2). Additional test
concentrations (e.g., 2x, 1x, 0.5x) can then be prepared to
determine an IC25 or IC50 of the concentrate, and toxicity
recovery can be calculated by comparing this value to the
toxfcity of the effluent. The concentrate toxicity might be
higher than the sum of the individual toxic fractions be-
cause some of the toxicant may have been in adjacent
fractions that were concentrated in the first step (Section
2.3.7) but not detectable by the toxicity test of the single
fraction. The concentrate toxicity may also be lower than
expected because of low extraction and elution efficien-
cies.
The Important concern here is not 100% recovery
of toxicants but enough recovery for GC/MS analyses to
be successful and to obtain measurable toxicity in the
HPLC fractions. If recovery is too low, changing or elimi-
nating the column drying time may help. Sometimes
recovery appears to increase with drying time while other
compounds are volatilized from the column during the
drying process. For concentrates analyzed using GC/MS,
column drying to remove water is critical to GC column
performance.
2.3.10 HPLC Separation
The same column packing functionality should be
used In the HPLC column as is used in the SPE column,
such as C1§. At later stages, when more is known about
the toxicants, other sorbents might be more appropriate.
The HPLC conditions presented in this section
are general. An important consideration of HPLC fraction-
ation Is the number of HPLC fractions to collect. Since
chronic toxicity testing is very time consuming, deciding
the appropriate number of fractions to collect is an impor-
tant step. However, when choosing which collection
scheme to use, keep in mind the trade-off between sepa-
ration and toxicity testing load. When the fraction volume
is increased (toxicity testing load decreases) the separa-
tion of the toxicants from the non-toxic components de-
creases. We have used a 20 min separation gradient with
the collection of 20-1 ml fractions. There are many other
collection options that could be used, such as 10-2 ml
fractions or 4-5 ml fractions using the same separation
gradient. As information on the effluent is gained, HPLC
conditions should be modified from the general conditions
described b°elow, to achieve better separation and higher
concentration factors.
We use a flow rate of 1 ml/min on an instrument
equipped with a 5 |om C1B reverse phase column (250 mm x 4.6
mm i.d.). The HPLC elution conditions will change depend-
ing on which SPE fractions have been concentrated. An
example of HPLC conditions for commonly toxic SPE
fractions is listed in Table 2-7. First, the blank concentrate
is injected (<500 \i\) and 20-1 ml fractions are collected in
analytically clean glass vials. Depending on the size of
the HPLC injector and column, more than one HPLC
fractionation run may be required to fractionate the entire
blank concentrate. When multiple HPLC fractionations
are conducted, collect and combine all the corresponding
HPLC fractions in the same set of vials. For example, if
two HPLC fractionations were performed for the blank
concentrate, 20-2 ml HPLC blank fractions would be ob-
tained. The same procedure is followed using the effluent
sample concentrate. The vials should be sealed (e.g.,
with foil lined caps) and stored at 4°C if not tested
immediately. As soon as toxicant identification is obtained
by GC/MS (Section 2.5), then HPLC conditions (gradient,
fraction size, and number of fractions) can be optimized.
Table 2-7. Example HPLC Elution Gradient for
SPE Fractions from Chronically Toxic
Effluent Samples
80 oV 85% SPE Fractions
Time (min)
% Methanol/Water
0
10
12
20
80
90
100
100
2.3.11 HPLC Fraction Toxicity Tests
In the HPLC fraction toxicity tests for chronically
toxic effluents, the methanol in the HPLC fractions is one
of the limiting factors of the concentration of the fractions
that can be tested. Each fraction is assumed to be 100%
methanol to calculate the necessary dilution. A 0.6%
methanol concentration or less can be tested with C.
dubia, while a 1% or less methanol concentration can be
tested with fathead minnows.
In a chronic TIE with C. dubia, when all of the
SPE concentrate remaining after toxicity testing from
20,000 ml effluent is injected on the HPLC (one injection),
2-16
-------
each resulting 1 ml HPLC fraction equals 15,575 ml of
effluent (assuming the toxicant elutes in only one fraction,
see Table A-2). If 11 uJ of each HPLC fraction is then
diluted to 10 ml, the test concentration is 16x the original
effluent concentration. However, the 16x solution should
be prepared as one sample before aliquots are split to
provide replicates. For instance, in the example used
above, 55 u.l should be diluted to 50 ml, which is then
equally distributed into five test cups. Additional concen-
trations are prepared in a similar fashion to estimate the
IC25 or IC50 and to compare toxicity recovery to the
toxicity of the sample. Of course, some loss of toxicant
will occur in each step and the toxicity may be less.
The blank fractions should not be toxic. If they
are, then additional tests with dilutions must be conducted
on both blanks and toxic fractions to find out whether
there is enough additional toxicity in the sample fractions
to warrant analysis.
The toxicity of the HPLC fractions should be
tested at twice (at least) the concentration at which the
SPE fraction concentrates were tested because recovery
of toxicity and analytical measurements indicates that up
to 50% of the initial concentration of toxic compounds
may be lost in this step (Durhan et al., 1990). The
concentration of methanol should not exceed the amount
used in the SPE fraction tests described above (Section
2.3.8).
2.3.12 HPLC Fractions: Concentration and
Subsequent Toxicity Testing
The toxic HPLC fractions and their corresponding
blanks must be concentrated in a solvent suitable for GC/
MS or other analytical techniques. Use the procedure
described in Section 2.3.9, Concentration of Fractions
(Figure 2-6). Judgement must be used to decide whether
to concentrate each toxic fraction separately or to com-
bine various toxic fractions prior to concentration. If, for
example, two successive fractions are toxic, there is a
good probability that the same toxicant is present in both.
If one toxic fraction is separated from the other by several
nontoxic fractions, there is a high probability that they
contain different toxicants. There is also a good probabil-
ity that at least one nontoxic fraction on either side of the
toxic fractions contains some of the toxicant. The advan-
tage of combining fractions is to reduce the workload and
to increase the amount of toxicant in the concentrate. The
disadvantage is that more constituents that are not the
toxicant will be included. The decision has to be based on
trial and error and experience. Blanks corresponding to
the toxic fractions are concentrated the same way.
The HPLC fraction and blank concentrates should
also be checked for toxicity before analysis on the GC/
MS. Generally, we suggest that these toxicity tests be
done at concentrations at least 2x higher than the con-
centration used in the previous HPLC fraction tests. Hope-
fully, the amount of concentrate available will be enough
to conduct the toxicity test and perform a GC/MS analy-
sis. Dilutions of the concentrate may be useful to compare
toxicity of this concentrate to each previous toxicity test
result. The HPLC concentrate (of 20,000 ml effluent) is
now 48,495x more concentrated than the effluent (see
Table A-2). If 3 uJ is diluted to 10 ml the resultant test
concentration will be about 16x the original sample con-
centration. This 16x solution should be prepared as one
solution before aliquots are removed for the replicates.
For instance, 15u.l is diluted to 50 ml for use in the
example given above, then split into five 10 ml test cups.
It is prudent to verify toxicity in the HPLC concentrate
before it is subjected to GC/MS analysis. Whether the
toxicant is detected by the analytical detector is always a
question. Since GC/MS detects only about 20% of or-
ganic chemicals (EPA, 1989B), even such a broad spec-
trum method is no guarantee that the toxicant will be
identified.
2.4 GC/MS Analyses
Procedures and methods provided in this section
are based upon our experience in performing GC/MS
analyses on fractions from numerous effluents and are
applicable to both acute and chronic toxicity identification.
In general, these procedures should be used.
A GC/MS system equipped to perform standard
chemical residue analyses is suggested; i.e., a 30 m
capillary column, electron impact ionization, scan range
of 50-500 amu, scan rate of 1 or 2 scans/sec, a GC
temperature program of 50 to 300°C at 5°C/min, and a
data system with library searching capability.
Prior to GC/MS analysis, the prepared blank and
toxic fraction concentrates should be tested for toxicity
(Figure 2-6). After verification of the toxicity in the metha-
nol concentrate, inject 1 or 2 \i\ of the concentrate (to
which an internal standard has been added) and collect
the mass spectral data. Note, methanol is not a typical
solvent for GC analysis and the injection of methanol on a
capillary column will shorten the column's life. Therefore,
routine GC/MS QA/QC procedures should be followed
closely to monitor the performance of the column.
The mass spectral data should be collected, the
chromatogram integrated, and all detected peaks library
searched. Reverse search is preferred. Concentration
estimates for all chromatographic peaks can be obtained
by using the response factor of the internal standard.
Usually the internal standard is added to a small aliquot
(10-20 |j.l) of the concentrate prior to GC/MS analysis. The
selection of internal standard to use is an individual
choice, and many different standards are available. An
external standard method could also be used for deriving
concentration estimates.
The MIST (National Institute of Standards and
Technology, Gaithersburg, MD) mass spectral library has
been used in ERL-D for performing library searches.
Other mass spectral libraries are available, but some of
the larger libraries contain multiple spectra for some of
the compounds in the database. Library searching results
that contain multiple identifications of the same com-
2-17
-------
pound are not as useful as those obtained using the NIST
library.
Once the library search results are available, the
search report for each peak must be examined to decide
whether the identification by the search is valid and
reasonable. The help of a trained GC/MS chemist is
required to do this evaluation. Questions we consider in
our laboratory when performing this process include: A)
are all major ions present in the correct proportions?, B) is
this identification consistent with other information about
the fraction?, C) do forward and reverse searching pro-
vide similar fits? and D) are the library searching fits
greater than 70%? Factor A must be met! Consistency,
factor B, considers circumstances such as "has the iden-
tified chemical been found in vastly different fractions," or
"has the same identification been given to numerous
peaks in the same chromatogram?" Both factors C and D
are somewhat relative and depend a great deal on the
sample and "rts matrix. In addition, the toxicants are often
very minor components in the GC/MS total ion chromato-
gram and thus, the quality of the mass spectral data even
after background subtraction can lead to poor results for
factors C and D.
After examination of the library search results, a
list of identified chemicals is assembled and evaluated
using the methods in the following section. For the confir-
mation analyses we suggest EPA method 625 (EPA,
1982).
2.5 Identifying Suspect Toxicants
If one toxicant is identified, then the goal of the
rest of Phase II is to determine if there are any other
toxicants contributing to effluent toxicity. Two parallel
lines of investigation should be pursued to achieve that
goal. The first is to determine whether or not the concen-
tration of the suspect toxicant is sufficient to cause toxicity
(EPA, 1993A). The second is to estimate the proportion of
the effluent toxicity that is caused by the suspected
toxicants, so that a decision can be made as to whether
other toxicants are present in the effluent.
The first line of investigation should begin by
comparing the estimated concentrations of identified
chemicals in the SPE or HPLC concentrate to their known
toxicity values. Recovery of 100% of each effluent toxi-
cant In the C18 SPE fractions may not be crucial, because
at this stage, only the estimated concentration of com-
pounds in the fraction and the toxicity of the fraction are
compared. Assumptions about the concentration of
toxicant(s) in the whole effluent are not made at this point,
nor is any statement made regarding recovery of whole
effluent toxic'rty in C18SPE column fractions. In later stages
of Phase II, inferences regarding the relationship between
the concentration of the suspected toxicant(s) in whole
effluent and the observed toxicity in the SPE fractions are
made. At this step, the compound quantification will have
been performed using an internal or external standard
response and since the compound's recovery is unknown,
considerable error may be involved in the concentration
estimate. Secondly, the toxicity data, if available, may be
for a different species than that used in the TIE. Species
differences are usually as large as 100-fold and often
1,000-fold. Given these two sources of uncertainty and
the chance that they may reinforce one another, certainly
if the estimated concentration of a chemical accounts for
the toxicity within a factor of 100, the chemical should
remain a suspect. To the extent that data for either
quantitation or toxicity values of the compound are known
to be better, concentration differences of smaller magni-
tude may be used to eliminate suspects.
Once a list of suspects is available, the measure-
ments for both concentration and toxicity should be re-
fined. This will usually require obtaining pure compound
to make better analytical measurements and to establish
acute or chronic toxicity estimates for the species of
concern. This step requires as much separation as practi-
cal before analysis so that the list of suspects is small.
At this stage, only the concentration of the sus-
pected toxicant(s) in the concentrate is known; until re-
covery through all the fractionation and concentration
steps is complete, suspect compound concentrations in
whole effluent are not known. Since the concentrate is
virtually devoid of suspended solids and much of the
effluent TOC, both of which may dramatically affect toxic-
ity of non-polar organics, the toxicity of non-polar chemi-
cals may be quite different in the fraction tests than in the
effluent test. Therefore, the toxicity of suspects in the
fraction test should be compared to the suspect's toxicity
in a relatively pure water, such as reconstituted water.
During this same stage, the steps leading to the
final concentrate should be checked for toxicity recovery.
The objective is to place a good estimate on how much of
the whole effluent toxicity is contained in the final concen-
trate. This is best done by testing the toxicity of the
concentrate at concentrations near those of whole efflu-
ent, correcting for volume losses due to toxicity testing
SPE column fractions (which was previously ignored). If
the toxicity of the final concentrate is similar to that of
whole effluent, allowing for losses, and if the concentra-
tion of the suspect(s) is sufficient to account for the
concentrate's toxicity, it is time to begin Phase III (EPA,
1993A). If multiple toxicants occur, the toxic units of each
are compared to the whole effluent toxic units.
If the concentrations from quantitation and toxic-
ity measurements are close to one another, Phase III
procedure should be started, recognizing that other toxi-
cants may yet be identified. If no suspects are found,
more concentration, more separation, and possibly differ-
ent or more sophisticated analytical methods must be
used. In some of the effluents we have tested, finding
other candidates has taken months and concentration
factors of >100,000 have been required. Since few labo-
ratories will have all the needed analytical equipment,
instrumentation from other sources should be considered.
2-18
-------
Because artifactual toxicity that equals toxicity
due to lost or unidentified toxicants can be created, as
one progresses to Phase III the suspect toxicant should
be identified. One purpose of Phase III is to identify such
errors. Should this error occur, one must start again at the
beginning of Phase II, or even return to Phase I. If several
different effluent samples were evaluated during Phase II,
redoing Phase I on additional samples may be time well
spent since the effluent may have changed in the interim.
In practice therejs no sharp boundary between
Phases II and III. In general, as soon as a probable
suspect is identified, confirmation procedures of Phase III
should begin. If a toxicant has been assumed to have
been identified when it has not, the identification of other
suspected toxicants can be hampered.
A final suggestion is to investigate the additivity of
toxicity for several constituents, if all toxicity is not ac-
counted for. Enhancement of toxicity by methanol should
also be checked.
2.5.1 Identifying Organophosphate Pesticides
Certain compounds must be metabolically acti-
vated by the test organism before they become toxic.
These activation reactions consist of oxidative metabo-
lism by a family of enzymes collectively known as cyto-
chrome P-450. Compounds such as piperonyl butoxide
(PBO) can block the toxicity of metabolically activated
toxicants making it a useful tool in the TIE. PBO is a
synthetic methylenedioxyphenyl compound that effectively
binds to and blocks the catalytic activity of cytochrome P-450.
Thus, when a nontoxic amount of PBO is coadministered with
the effluent or the effluent fractions that exhibited toxicity,
the toxicity of the compound requiring metabolic activa-
tion is greatly reduced or completely blocked (Ankley et
al., 1991).
Phosphorothioates are organophosphates known
to require cytochrome P-450 activation before expressing
toxicity and include common insecticides such as diazinon,
malathion, parathion, methyl parathion andfenthion. There
also are a number of organophosphates that are toxic in
the absence of metabolic activation; these include insec-
ticides such as dichlorvos, mevinphos and chlorfenvinphos.
We have found organophosphate insecticides
present in effluents and ambient waters at acute and
chronic toxicity levels (Amato et al., 1992; Norberg-King
et al., 1991). The toxicity of most organophosphates will
be removed from the sample by the C SPE column, and
they are typically recovered in the methanol eluates (see
EPA, 1991 A; EPA, 1992). The addition of PBO to the
effluent before addition of the test organisms was used as
a subsequent test in Phase I (EPA, 1991A; EPA, 1992). In
addition to the C18 SPE column removing the toxicity, a
reduction in toxicity with the addition of PBO would sug-
gest the presence of metabolically activated compounds
such as organophosphates. PBO has similar utility in
Phase II of the TIE in that either SPE fractions (Sections
2.2.8 and 2.3.8) or HPLC fractions (Sections 2.2.11 and
2.3.11) can be tested for toxicity both in the presence and
absence of PBO. A reduction in toxicity of the test fraction
would suggest the presence of a metabolically activated
chemical, and together with chemical analyses, can pro-
vide powerful evidence along with GC/MS data, for spe-
cific organophosphates as the toxicant(s). While PBO
should be useful for both acute and chronic TIE work,
most of our experience has been in the area of acute
toxicity. Thus, guidance presented below is based mainly
on acute tests.
Toxicity values for PBO are presented in Phase I
(EPA, 1991 A; EPA, 1992). In acute toxicity tests, concen-
trations of PBO ranging from 250-500 u.g/1 have effec-
tively blocked the acute toxicity of relatively large
concentrations of metabolically activated organophos-
phates to cladocerans (Ankley et al., 1991). In chronic
toxicity tests with C. dubia, PBO concentrations of 50 u.g/1
have been effective in blocking toxicity in the SPE fractions.
Detailed information on stock solution preparation is pre-
sented in the Phase I documents and is not repeated
here.
When toxicity tests are conducted on SPE frac-
tions or HPLC fractions, aliquots of the PBO solution are
added to the test solutions and mixed well before the test
organisms are added. As for any TIE manipulation, the
successful use of PBO is dependent upon the use of
appropriate controls and blanks. Effluent fractions and
blank fractions with and without the addition of PBO must
be tested simultaneously. A reduction in toxicity of the
effluent fraction occurring with the PBO added, and no
toxicity exhibited in either of the blanks, indicates that the
toxicant requires metabolic activation to exhibit toxicity. If
toxicity associated with the PBO addition is observed in
the blank fraction, either PBO was present at toxic con-
centrations or the methanoi concentration (from fraction
and/or PBO stock addition) in the test was too high. If
toxicity is observed in the effluent fraction with PBO
added, but not in the effluent fraction without the PBO or
in either of the blank fractions, this result is essentially
meaningless, in the latter situation it is possible that the
PBO has interacted in a synergistic fashion with another
compound present in the test effluent that normally would
not be toxic.
2.5.2 Identifying Surfactants
The goal in this section of Phase II is to identify
the toxicants when surfactants are implicated by Phase I
and Phase II results. The Phase I procedures of filtration,
aeration, and C18 SPE all affect surfactant toxicity, and
effluent samples that exhibit several or all of these behav-
iors may contain toxic concentrations of surfactants (EPA
1991 A).
Surfactants are surface active agents that have a
molecular structure that includes a polar, hydrophilic seg-
ment (either ionic or nonionic) and a relatively large non-
polar, hydrophobic, hydrocarbon segment. Surfactants
are used for a variety of household and industrial pur-
poses and therefore are ubiquitous in effluents, particu-
2-19
-------
(arly in untreated wastewater, and potentially could be
present at toxic concentrations in effluents (Ankley and
Burkhard, 1992). Some examples of surfactants are soaps,
detergents, charged stabilization polymers, and coagula-
tion polymers used in chemical manufacturing processes.
The molecular structure of surfactants causes them to
congregate at interfaces between water and other phases
such as air, oily liquids and particulate matter. This
congregative characteristic is responsible for the cleans-
ing and dispersive properties of surfactants.
There are many different kinds of surfactants and
they are classified by the nature of their polar segment.
When in aqueous solution, the polar segment of a surfac-
tant molecule can be either nonionic (not charged) or
ionic (charged). The ionic polar segment can be either
negatively charged (anionic), positively charged (cationic),
or both negatively and positively charged (amphoteric).
Based on this, surfactants are classified into the following
major classes: nonionic, anionic, cationic, and ampho-
teric.
Surfactants physical/chemical properties set them
apart from both strictly polar or non-polar organic com-
pounds and these properties uniquely influence the re-
sults of Phases I and li procedures for surfactants.
Experiments were conducted with a small sample
of surfactants from nonionic, anionic, and cationic catego-
ries with the Phase I procedures of filtration, aeration, and
C16 SPE (Ankley et al., 1990A). In these experiments,
filtration removed the toxicity of most of the surfactants
tested to some degree, and the degree of removal is most
probably dependent on sample matrix, especially solids
concentration. Aeration removed the toxicity of all the
surfactants tested to some degree while the C18 SPE
column removed the toxicity completely for all surfactants
regardless of class. Surfactants behave unpredictably
with regard to elution from C18 SPE columns. For ex-
ample, toxicity from surfactants of the nonionic and an-
ionic classes, eluted in all fractions 80% to 100% methanol/
water (Ankiey et al., 1990A). Elution in several fractions
rather than eluting in one or two fractions may be caused
by the polar/non-polar nature inherent in surfactants. The
toxicities from the cationic surfactants were either not
recovered in any of the fractions or were recovered to
only a small degree in the 100% methanol fraction.
Important indicators of surfactant toxicity are the
toxfcHy test results from aeration experiments. If volatility
can be eliminated and toxicity is reduced by aeration, this
is strong evidence that a surfactant might be contributing
to effluent toxicity (EPA, 1991 A). During aeration, surfac-
tants are most probably removed from solution by the
process of sublation. Sublation occurs because surfac-
tant molecules tend to congregate at the interface be-
tween the aqueous sample and the aerating nitrogen or
air bubbles and are brought to the surface of the liquid
sample by the bubbles. At the liquid surface the bubbles
break releasing the surfactant, which then adheres to the
aeration vessel walls. A compound that can be removed
by sublation is by definition a surfactant. It might be
possible to recover surfactants from glassware after the
sublation process. The glassware can be rinsed with a
solvent such as methanol, which can then be toxicity
tested and analyzed in the same manner as methanol
SPE fractions (Sections 2.2.8 and 2.3.8).
Overall, most surfactants exhibit some of the
behavior that is common to non-polar organic compounds
such as removal from the effluent by the C18 resin and
recovery in the methanol/water SPE fractions. While sur-
factants in general can be considered to be non-polar
organics, GC/MS analysis will probably not provide suc-
cessful surfactant identification. Most surfactants are not
readily chromatographed because of the polar segment
of the surfactant molecule. One exception is a class of
surfactants in common use that can be analyzed directly
by GC/MS, the alkylphenol ethoxylates. Gieger et al.
(1981), provides mass spectral data for the nonylphenol
mono-, di- and tri-ethoxylates, which can be used to help
identify these compounds. Techniques such as
derivatization can make some other specific surfactants
compatible with GC and GC/MS, but it is necessary to
know the specific identity of the surfactant.
It is difficult to positively identify an unknown
surfactant. Although there are many analytical methods
available for accurately quantifying specific surfactants,
these methods are useful only if the identity of the surfac-
tant is known, or at least suspected. It is not reasonable or
practical to analyze a sample using numerous intricate
methods, in the hope that one of these methods will
detect the surfactant in the sample. Unfortunately, there is
no analytical technique available that can readily provide
the identity and quantity of an unknown surfactant. Envi-
ronmental samples (such as municipal and industrial
effluents) contain numerous substances that can interfere
with available analytical methods. Also, pure surfactants
are actually mixtures of homologous and oligomers with
varying chain lengths and, in the case of many nonionic
surfactants, varying degrees of ethoxylation. The compo-
sition and therefore the toxicity of such a mixture might
vary. In the course of a TIE, it might become necessary
not only to identify the surfactant causing toxicity, but also
to learn which particular horriologue or oligomer is the
most toxic.
One approach to reducing the complexity of iden-
tifying an unknown surfactant is to determine whether the
unknown surfactant falls into the anionic or nonionic
class. APHA (1989) describes a method for determining
anionic surfactants as methylene blue active substances
(MBAS). MB AS method can successfully measure the
concentration of anionic surfactants of the sulfonate type,
the sulfate ester type, and sulfated nonionicstype. Unless
the identity of the anionic surfactant is known, the analyti-
cal measurement is calculated and expressed in terms of
the anionic surfactant linear alkylbenzene sulfonate (LAS).
APHA (1989) also describes a method for determining
nonionic surfactants as cobalt thiocyanate active sub-
stances (CTAS). This method is applicable to a wide
2-20
-------
range of polyether nonionic surfactants, which includes
the widely used alkyl and alkylphenol ethoxylated alcohols.
With these methods the relative amount of an-
ionic or nonionic surfactant can be estimated, but the
exact nature or molecular composition of the unknown
surfactant will not be determined. These analyses can be
conducted on the SPE fractions, HPLC fractions, fraction
concentrates, and the whole effluent. Determining the
class can be significant progress toward identifying the
unknown surfactant. With the class known, specific analy-
ses for the more common surfactants in that class can be
performed as a subsequent effort. Unless the identity of
the nonionic surfactant is known, the analytical measure-
ment is expressed in terms of an arbitrarily chosen refer-
ence nonionic surfactant.
The type of discharge being processed by the
wastewater treatment plant might provide information that
would enable one to target specific surfactants for analy-
sis. For example, industries feeding into the treatment
plant might be discharging certain surfactants or a par-
ticular kind of surfactant that is being used in the manu-
facturing or housekeeping processes. An analytical method
suitable for that particular surfactant could then be used
to determine whether toxic concentrations can be found in
the toxic effluent, fractions, or concentrates.
2.6 Alternate Fractionation Procedures
If toxicity is not recovered in the methanol proce-
dures described above (Sections 2.2 and 2.3), and toxic-
ity is not observed in the post-column effluent, alternative
elution procedures can be used. These procedures are
not as widely used as the methanol/water elutions dis-
cussed above but are effective for highly hydrophobic
compounds.
2.6.1 Modified Elution Method
The current Phase II method for fractionating
non-polar organic toxicants in aqueous samples does not
effectively fractionate compounds that are highly hydro-
phobic. Modifications made to the method have been
successful in overcoming this limitation (Schubauer-
Berigan and Ankley, 1991; Durhan et al., 1993). Hydro-
phobic compounds probably are more prevalent in
sediment pore waters than in treated effluents. Tracking
toxicity caused by these kinds of compounds will be more
difficult because of the potential for artifactual toxicity
from the solvents required to elute them. An elution
scheme incorporating water, methanol, and methylene
chloride has been designed that effectively fractionates
compounds over a log Kow range from 2.5 to 6.9. The
higher log Kow compounds, however, elute in the same set
of fractions. Further fractionation by HPLC might be nec-
essary to achieve better resolution of these kinds of
compounds. Substituting other sorbents for the currently
used C18 SPE resin have also produced encouraging
results. Both the C8 SPE and XAD-7 (Rhom and Haas,
Philadelphia, PA) sorbents might have utility with particu-
lar kinds of toxicants.
The modified elution scheme eliminates the 100%
methanol fraction used in the original method, and re-
places it with one 50% methylene chloride/methanol, and
three 100% methylene chloride fractions (v/v). The com-
position of the resulting eleven 3 ml (when using a
1,000 mg C18SPE column) fractions is shown in Table 2-
8. The methylene chloride containing fractions are com-
bined, then solvent exchange is conducted as described
below. The modified elution scheme would be used when
the original methanol/water elutions did not effectively
elute toxicity in the SPE fractions. In addition, if the
suspect toxicants were known to be highly hydrophobic,
as in sediment pore water, then the modified elution
scheme would be indicated. Blank toxicity should provide
insight concerning artifactual methylene chloride toxicity;
however, slight reductions in young production might
occur in both the blanks and sample fractions. Develop-
ment of this alternate procedure for chronic toxicity is
underway for the C. dubia and should be used with
caution at this time. If this procedure is used, it is impor-
tant to accompany the solvent exchanged methanol blank
with a methanol only blank.
When toxicity testing SPE fractions, it is always a
concern that the matrix of the effluent has been changed
and that chemicals might become bioavailable, whereas
they were not in the original sample. If this were to
happen, the fractions might be more toxic than expected
and chemicals might be added to the suspect toxicant list
erroneously. This kind of mistake should be caught by
obtaining a good toxicity value for the suspect toxicant in
an appropriate matrix. For instance, if the suspect toxi-
cant is highly insoluble in water, then when tested in an
effluent matrix it should have low toxicity because it is
unavailable to the organism. The alternate solvent elution
might enhance this problem because the solvent is more
likely to solubilize the more hydrophobic compounds than
Table 2-8. Composition of 11 Recommended Fractions in Modified
Elution Scheme
Composition of Eluting Solutions (% v/v)
Fraction
Water
Methanol
Methylene Chloride
1
2
3
4
5
6
7
8
9
10
11
75
50
25
20
15
10
5
0
0
0
0
25
50
75
80
85
90
95
50
0
0
0
0
0
0
0
0
0
0
50
100
100
100
2-21
-------
methanol. Therefore, additional confirmation steps might
be needed to eliminate the false suspects.
2.6.2 Solvent Exchange
Since methylene chloride is quite toxic to aquatic
organisms, even at very low concentrations (NOEC for C.
dubta Is 0.03%), it must be removed from SPE fractions
before the fractions can be tested for toxicity. The ex-
change of the methylene chloride fraction into methanol is
a relatively easy process because of methylene chloride's
volatility. The combined fractions to be exchanged (e.g.,
15 ml) are placed in a centrifuge tube with a teflon stir bar
and an additional 15 ml of methanol. The tube is placed in
a 30°C water bath and stirred while a gentle stream of
nitrogen is passed over the solution surface. When the
volume of the solution reaches 3 ml, the sides of the tube
are carefully rinsed with an additional 3 ml of methanol,
and the solution is reduced again to a final 3 ml volume.
Adjust the volume of methanol used in this procedure to
reflect the total volume of combined fractions. The final
volume of methanol may then be tested as suggested
previously In Sections 2.2.8. and 2.3.8. It is important to
obtain and toxicity test a methanol-only blank in addition
to the solvent exchanged methanol blank.
2.6.3 Alternative SPE Sorbents and
Techniques
In the SPE method described above, C18 bonded
silica is used as the solid phase for fractionating and
isolating non-polar organic toxicants. C18 bonded silica
was selected because, with proper conditioning, it does
not usually contribute artifactual toxicity to sample or
sample fractions, it often achieves the required degree of
separation and isolation of non-polar organic compounds,
and it Is commercially available in inexpensive, easy to
use, disposable columns. There is, however, no restric-
tion on the solid phase that is used in the TIE procedure,
as long as it results in the isolation and separation of non-
polar organic toxicants and at the same time does not
contribute artifactual toxicity. We have evaluated several
sorbents other than C bonded silica to use for this
purpose (Durhan et ai., 1993).
We evaluated two prepurified XAD sorbents, XAD-
4 and XAD-7 (Rohm and Haas, Philadelphia PA) and a C8
bonded silica sorbent. Of these sorbents, only XAD-4, a
non-polar styrene-divinyl benzene copolymer performed
as well as C18 bonded silica in the fractionation of non-
polar organic compounds. One disadvantage of using an
XAD sorbent such as XAD-4 is that it is not commercially
available in prepacked disposable columns. In addition, it
is important to obtain prepurified XAD-4 sorbent that is
free of toxic artifacts, otherwise extensive, time consum-
ing cleanup procedures are required before the sorbent
can be used in a toxicity based fractionation. We found
that on XAD-7, an acrylic ester copolymer, non-polar
organic compounds were inadequately fractionated be-
cause of resolution and co-elution problems. The C8
bonded silica yielded results that were similar but signifi-
cantly inferior to those obtained with C18 bonded silica.
Traditionally, SPE is carried out with the solid
phase particles packed in a cylindrical column or car-
tridge. An alternative form of SPE has been developed,
the Empore™ Extraction Disk, in which C18 bonded silica
particles are enmeshed in an inert PTFE matrix which is
then formed into a disk. The manufacturer (3M, St. Paul,
MM) claims good recovery of non-polar organics with flow
rates as high as 100 ml/min, which would make this an
attractive alternative form of SPE. We have evaluated this
technique to a limited degree with acutely toxic effluents
and sediment pore waters and feel it has great potential in
a toxicity based fractionation scheme. Especially attrac-
tive is the high flow rate which would allow for large
volumes of sample to be processed quickly. However, a
procedure for eluting non-polar organics from the disk into
several fractions has not yet been developed and could
prove to be a challenge.
2-22
-------
Section 3
Ammonia
3.1 General Overview
Unlike Phase II procedures for non-polar or-
ganic compounds or metals, the toxicant identification
methods described in this section are specific for ammo-
nia. The procedures used in this phase of the study
assume that Phase I tests and ammonia measurements
(see below) have implicated the pH sensitive toxicant,
ammonia as causing the acute or chronic toxicity (see
Phase I; EPA, 1991 A; EPA, 1992). Other compounds
with toxicities that increase directly with pH may lead to
confounding results or may give results similar to ammo-
nia. For instance, experiments at our laboratory have
shown that C. dubia are more acutely sensitive to cad-
mium, nickel, and zinc in acute tests at high pH levels
(Section 4). The testing in Phase II should help to
discern the toxicity caused by ammonia from that caused
by other compounds that might also become more toxic
as pH increases. The methods described below can be
used to identify ammonia as the toxicant and these data
could also be used in Phase 111 confirmation.
Ammonia is relatively unique in its behavior as
pH changes. When ammonia (NH3) dissolves in water,
some of the molecules react to form the ammonium ion
NH4+, and the equilibrium between these two species is
affected by both pH and temperature (EPA, 1985A). The
term "total ammonia" refers to the sum of the un-ionized
(NH3) and the ionized (NH4+) forms and is referred to as
N+. The toxicity of ammonia to some aquatic species
appears to be primarily caused by the un-ionized form.
The equilibrium shifts to increase the un-ionized ammo-
nia concentration with increasing pH and increasing
temperature. In a constant temperature situation, Table
3-1 shows that as pH increases by one unit, there is
nearly a 10-fold increase in the percent of un-ionized
ammonia NH3 present in aqueous solutions at pH 6.0-
8.5. The data in Table 3-1 are calculated using the
dissociation constants for ammonia (EPA, 1979). There
are two effects to consider for ammonia as the pH
increases; first, the concentration of NH3 increases (Table
3-1) and second, the toxicity of NH3 decreases (Tables
3-2 and 3-3). One possible explanation for the second
effect is that NH4+ is contributing to the toxicity (EPA,
1985A). Measuring and maintaining the pH of the test
solution and understanding the effect of pH on the
toxicity of ammonia are very important.
As discussed in EPA's ammonia water quality
criteria document (EPA, 1985A), the slope of the LC50-
pH curve for acute toxicity is similar for different aquatic
species (i.e., an average slope can be used for many
species). A model was developed to describe the pH
dependence of ammonia toxicity, primarily with data for
fishes and cladocerans (i.e., daphnids, fathead minnows,
rainbow trout, and coho salmon, see EPA, 1985A). This
model has been used with acute toxicity data generated
at a pH of 8 and a temperature of 25°C for both C. dubia
and fathead minnows, to predict the LC50 of NH3 at other
pH values (Tables 3-2 and 3-3). It is apparent that the
toxicity of NH3 is about seven times less at pH 7.0 than at
pH 6.0, but the amount of NH3 is ten times greater at pH
7.0 than at pH 6.0. Similarly, at pH 8.0, NH3 is three times
less toxic than at pH 7.0 but ten times more is available at
pH 8.0. Ammonia can be implicated as the cause of
toxicity if the effluent toxicity and the suspect toxicant
exhibit both of these pH effects. Acute toxicity test data
generated at ERL-D indicate that this model is not appro-
priate for all species. For example, the trend of pH-
dependence has not been observed in acute tests
conducted with the amphipod, Hyalella azteca, over a
range of pH values in reconstituted waters (EPA, 1991B)
until the hardness is greater than 160 mg/l. In hard or very
hard waters, H. azteca is more sensitive to NH3 at higher
pHs (P. Monson, personal communication, University of
Wisconsin, Superior, Wl). We recommend that the effect
of pH on the toxicity of ammonia be characterized for the
TIE organism, if it has not been done, so that accurate
predictions can be made for the organism.
It has not yet been determined whether the pH
dependence of ammonia toxicity described for acute tox-
icity is appropriate for chronic toxicity. The chronic toxicity
of ammonia to species typically used in effluent tests, at
temperatures similar to those used in TIEs and a variety
of pHs is presented in Table 3-4. If chronic ammonia
toxicity has not been characterized with respect to pH for
the TIE species, it is prudent for the investigator to
generate the ammonia toxicity data for at least three
distinct pH levels.
Generally, three procedures are used to implicate
ammonia in addition to measuring the ammonia in the
effluent. These are 1) the graduated pH test (in place of
3-1
-------
T«blo3
pH
6.0
6.1
6.2
6.3
6.4
6.5
6.6
6.7
6.8
6.9
7.0
7.1
7.2
7.3
7.4
7.5
7.6
7.7
7.8
7.9
8.0
8.1
8,2
8.3
8.4
8.5
8.6
8.7
8.8
8.9
9.0
•1. Percent Un
15
0.0274
0.0345
0.0434
0.0546
0.0687
0.0865
0.109
0.137
0.172
0.217
0.273
0.343
0.432
0.543
0.683
0.858
1.08
1.35
1.70
2.13
2.66
3.33
4.16
5.18
6.43
7.97
9.83
12.07
14.7
17.9
21.5
•Ionized Ammonia in Aqueous Solutions for Selected Temperatures and pH
Temperature (°C)
20
0.0397
0.0500
0.0629
0.0792
0.0865
0.125
0.158
0.199
0.250
0.314
0.396
0.497
0.625
0.786
0.988
1.24
1.56
1.95
2.44
3.06
3.82
4.76
5.92
6.43
9.07
11.16
13.6
16.6
20.0
24.0
28.4
21
0.0427
0.0537
0.0676
0.0851
0.107
0.135
0.170
0.214
0.269
0.338
0.425
0.535
0.672
0.845
1.06
1.33
1.67
2.10
2.62
3.28
4.10
5.10
6.34
7.85
9.69
11.90
14.5
17.6
21.2
25.3
29.9
22
0.0459
0.0578
0.0727
0.0915
0.115
0.145
0.182
0.230
0.289
0.363
0.457
0.575
0.733
0.908
1.14
1.43
1.80
2.25
2.82
3.52
4.39
5.46
6.78
8.39
10.3
12.7
15.5
18.7
22.5
26.7
31.5
23
0.0493
0.0621
0.0781
0.0983
0.124
0.156
0.196
0.247
0.310
0.390
0.491
0.617
0.776
0.975
1.22
1.54
1.93
2.41
3.02
3.77
4.70
5.85
7.25
8.96
11.0
13.5
16.4
19.8
23.7
28.2
33.0
24
0.0530
0.0667
0.0839
0.106
0.133
0.167
0.210
0.265
0.333
0.419
0.527
0.663
0.833
1.05
1.31
1.65
2.07
2.59
3.24
4.04
5.03
6.25
7.75
9.56
11.7
14.4
17.4
21.0
25.1
29.6
34.6
Values1
25
0.0568
0.0716
0.0901
0.113
0.143
0.180
0.226
0.284
0.358
0.450
0.566
0.711
0.893
1.12
1.41
1.77
2.21
2.77
3.46
4.32
5.38
6.68
8.27
10.2
12.5
15.2
18.5
22.2
26.4
31.1
36.3
26
0.0610
0.0768
0.0966
0.122
0.153
0.193
0.242
0.305
0.384
0.482
0.607
0.762
0.958
1.20
1.51
1.89
2.37
2.97
3.71
4.62
5.75
7.14
8.82
10.9
13.3
16.2
19.5
23.4
27.8
32.6
37.9
27
0.0654
0.0823
0.104
0.130
0.164
0.207
0.260
0.327
0.411
0.517
0.650
0.817
1.027
1.29
1.62
2.03
2.54
3.18
3.97
4.94
6.14
7.61
9.40
11.6
14.1
17.1
20.7
24.7
29.2
34.2
39.6
'Data from EPA, 1979.
the equftoxic solution test as described in the first Phase II
document; EPA, 1989A); 2) use of the zeolite resin to
remove the ammonia; and 3) air-stripping the ammonia
from the sample at a high pH (i.e., pH 11). For both the
zeolite resin method and the air-stripping method, subse-
quent toxicity tests and ammonia measurements are per-
formed on whole effluent and the post-treatment samples.
Depending on the presence of other toxicants in
the effluent, additional sample manipulations may be
needed before proceeding with the three basic tests. For
example, if toxic oxidants such as chlorine are also present
in the effluent, sodium thiosulfate must be added to the
sample before conducting the Phase II ammonia tests. To
date we have not seen ammonia and chlorine as co-
occurring toxicants in chronic tests, probably because
chlorine degrades rapidly in a test system at 25°C, while
ammonia does not. If the additional toxicant(s) can be
removed by the C18 SPE column, it may be possible to
conduct Phase II tests for ammonia on post-C18 SPE
column effluent sample. However, the problem of artifac-
tual toxicity associated with the post-C18 SPE column
effluent may prevent the use of the graduated pH test
(EPA, 1992) and/or the air-stripping test (see Section 8 of
EPA, 1991 A) on post-column samples.
The results of the graduated pH test, the post-
zeolite column test, and the air-stripping test, all will be
important in identifying ammonia as a toxicant in acutely
or chronically toxic samples. Use of pH changes where
graded responses are observed are particularly useful for
data evaluation in Phase III correlation steps. Some of the
Phase II tests for ammonia are the same steps that are
used for Phase III confirmation procedures; therefore,
tests such as spiking the effluent with ammonia and then
performing the graduated pH test or spiking the post-
zeolite effluent samples and then testing the samples
simultaneously with the Phase II tests will support the
confirmation steps in Phase III.
3.2 Toxicity Testing Concerns
A key issue in interpreting acute or chronic test
results for a pH dependent toxicant such as ammonia is
monitoring pH changes during the test period. Toxicity
differences in Phase I manipulations may be misinter-
preted simply because differences in NH3 toxicity can
occur with only a slight pH change. To illustrate, the
change in pH from 8.0 to 7.9 lowers the concentration of
NH3 20%, as does a change in pH from 6.1 to 6.0, but a
20% difference is much more important to the toxicity
3-2
-------
Table 3-2. Calculated Un-lonized Ammonia LCSOs (mg/l) Based on 24-h and 48-h Results of a Ceriodaphnia dubia Toxicity
Test Conducted at pH 8.0 and 25°C1
pH
6.0
6.1
6.2
6.3
6.4
6.5
6.6
6.7
6.8
6.9
7.0
7.1
7.2
7.3
7.4
7.5
7.6
7.7
7.8
7.9
8.02
8.1
8.2
8.3
8.4
8.5
8.6
8.7
8.8
8.9
9.0
Percent
Dissoc.
at 25°C
0.0568
0.0716
0.0901
0.1134
0.143
0.180
0.226
0.284
0.358
0.450
0.566
0.711
0.893
1.12
1.41
1.77
2.21
2.77
3.46
4.32
5.38
6.68
8.27
10.2
12.5
15.2
18.5
22.2
26.4
31.1
36.3
Un-ionized
Ammonia
Expected
24-h LC50
0.09
0.12
0.14
0.18
0.22
0.27
0.33
0.40
0.48
0.58
0.69
0.81
0.93
1.06
1.21
1.34
1.48
1.61
1.73
1.83
1.93
2.01
2.08
2.14
2.19
2.23
2.27
2.30
2.32
2.34
2.35
Total
Ammonia
24-h LC50
158
168
155
159
154
150
146
141
134
129
122
114
104
95
86
76
67
58
50
42
36
30
25
21
18
15
12
10
8.8
7.5
6.5
Un-ionized
Ammonia
Expected
48-h LC50
0.07
0.09
0.11
0.14
0.17
0.21
0.25
0.31
0.38
0.45
0.53
0.62
0.72
0.82
0.93
1.04
1.14
1.24
1.33
1.42
1.49
1.55
1.61
1.65
1.69
1.73
1.75
1.77
1.79
1.81
1.82
Total
Ammonia
48-h LC50
123
126
122
123
119
117
111
109
106
100
94
87
81
73
66
59
52
45
38
33
28
23
20
16
14
11
9.5
8.0
6.8
5.8
5.0
1 LCSOs for each pH interval were calculated using EPA's water quality criteria document formula (EPA, 1985A)
shown below.
(LC50[pH = 8.0])(1.25)
Formula LC50 = •
1 + 10
7.4-pH
2The 24 h and 48 h LC50s to C. dubia are 1.93 mg/l and 1.49 mg/l, respectively, at pH 8.0. The formula was used to
generated expected LC50s for pH values above 8, though the model is not recommended above pH 8, because
generally we have found C. dubia data to track with these predictions.
expressed by the ammonia at pH 8.0 than at pH 6.0. For
this reason frequent pH monitoring (at least daily) must be
performed on tests conducted to determine the trend of
ammonia toxicity. Ideally, continuous monitoring of pH is
desired. The pH should be measured on each test con-
centration and each replicate. Experience has shown that
the choice of pH meters and probes is critical to produce
reliable results. The pH meter used must read accurately
to two decimal places and should lock-on the stabilized
reading after the rate of change has diminished to a
specified rate. Routine cleaning of the probe and a stan-
dardized calibration procedure should be established.
The pH values can also be recorded after an elapsed time
of 60-90 sec. The pH readings should be made using a
constant and reproducible stirring rate. The stirring should
not result in excessive loss (or gain) of CO which will of
course change the pH. The choice of the pH electrode is
important. We have found that the glass-bodied combina-
tion electrodes provide the most consistent pH readings.
However, these should not be left in the test solutions for
longer than is needed to obtain constant readings of pH
because ions from the electrode reference solution can
leak into the test solution, potentially causing artifactual
toxicity.
For the Phase II ammonia toxicity tests more
replicates (at least double that used in Phase I) must be
used and tighter QA/QC procedures must be adhered to
than those described in the acute or chronic Phase I
manuals (see Section 1.2). For example, a control and at
least four effluent dilutions using concentrations that more
closely bracket the effect and no effect concentrations
(that were determined in Phase I) are used. While param-
eters such as time to mortality or onset of symptoms in
the acute and chronic tests are not an integral part of the
tests described below, these observations may be very
3-3
-------
Table 3-3, Calculated Un-lonized Ammonia LCSOs (mg/l) Based on 24-h, 48-h, 72-h, and 96-h Results of a Fathead Minnow (Pimephales
pnmelas) Toxicity Test Conducted at pH 8.0 and 25°C'
pH
6.0
6,1
6,2
6.3
6.4
6,5
6,6
6,7
6,8
6.9
7.0
7.1
7.2
7.3
7.4
7.5
7.6
7.7
7.8
7.9
8.0
Percent
Dissoc.
at253C
0,0568
0.0716
0.0901
0.1134
0.143
0.180
0.226
0.284
0.358
0.450
0.566
0.711
0.893
1.12
1.41
1.77
2.21
2.77
3.46
4.32
5.38
Un-ionized
Ammonia
Expected
24-h LC50
0.075
0.093
0.12
0.14
0.18
0.22
0.27
0.32
0.39
0.47
0.56
0.65
0.76
0.87
0.78
1.09
1.20
1.30
1.39
1.48
1.56
Total
Ammonia
24-h LC50
131
130
128
127
124
121
118
114
109
104
98
91
85
77
55
62
54
47
40
34
29
Un-ionized
Ammonia
Expected
48-h LC50
0.064
0.080
0.10
0.12
0.15
0.19
0.23
0.28
0.34
0.40
0.48
0.56
0.65
0.75
0.67
0.94
1.03
1.12
1.20
1.27
1.34
Total
Ammonia
48-h LC50
113
112
111
109
107
104
102
98
94
90
85
79
73
67
48
53
47
40
35
29
25
Un-ionized
Ammonia
Expected
72-h LC50
0.049
0.061
0.076
0.094
0.12
0.14
0.18
0.21
0.26
0.31
0.36
0.43
0.50
0.57
0.51
0.72
0.79
0.85
0.91
0.97
1.02
Total
Ammonia
72-h LC50
86
85
84
83
81
80
77
75
72
68 '
64
60
56
51
36
40
36
31
26
22
19
Un-ionized
Ammonia
Expected
96-h LC50
0.036
0.045
0.056
0.069
0.086
0.11
0.13
0.16
0.19
0.23
0.27
0.31
0.36
0.42
0.38
0.53
0.58
0.63
0.67
0.71
0.75
Total
Ammonia at
96-h LC50
63
63
62
61
60
58
57
55
53
50
47
44
41
37
27
30
26
23
19
17'
14
'LCSOs for each pH interval were calculated using EPA's water quality criteria document formula (EPA, 1985A) shown below. The 24-h, 48-h,
72-h, and 96-h LCSOs to fathead minnows are 1.56 mg/l, 1.34 mg/l, 1.02 mg/l, and 0.75 mg/I, respectively, at pH 8.0.
Formula LC5Q =
(Z.C50[pH = 8.0])(1.25)
1 + 107'4-""
Table 3-4. Un-Ionlzed Ammonia Toxicity Values for Species Frequently Used in Effluent Testing
Species Method1 pH Temp (°C)
LC502 (rng/1)
ChV3
C.dubisf
C. acanthla?-'
Simocephakis vstutus?-'
C.dubla'
C.dubia'
Daphnfamagnsf
P. pmmelas?
P. prome/as*
P. promalas?
C.dubtt?
C. <*/iwV
C. scanthia?-'
D. nwgna?
C.duW
P. pfomelasf
P. promefos*
P. pcomefas*
P. prome/os4
S, M
FT.M
FT.M
S.M
S,M
S,M
FT,M
FT, M
FT, M
4d-R, M
4d-R, M
7d-FT, M
NR
4d-R, M
FT, M
7d-R, M
7d-R, M
7d-R, M
Acute Data
6.2
7.1
7.1
7.2
8.2
8.2
7.8
8.0-8.3
8.1
Chronic Data
6.03
7.05
7.0-7.5
7.6
8.03
8.0
7.5-7.6
7.5-7.7
8.4
25
24
24
25
25
25
25.6
25.2
26.1
25
25
24-25
20.2
25
24.0
25.0
25.0
25
0.12
0.77
0.61
0.78
1.73
2.08
1.87
1.65
2.55
„
__
__
__
—
..
~"
—
_.
..
__
__
—
—
„
—
0.065
0.28
0.34
0.63
0.62
0.13
0.48
0.45
0.66
'FT - (tow-through; S « static; R m renewal of solutions at 24 or 48 h; M = measured concentration; NR = not reported.
*48-h LC50 for invertebrates and 96-h LC50 for fish.
*ChV • chronic value which is the geometric mean of the no observed effect concentration (NOEC) and the lowest observed effect concentration
{LOEC)oranlC25.
••Data generated at ERL-D. >
*C. acanthia is equivalent to C. dubla.
«Data from EPA, 1985A.
'Data from 4-d C, dubfa tests conducted at ERL-D; the effect level is an IC25 (mg/l)
'Data from Betgger, 1990.
3-4
-------
useful in describing the identification steps used in con-
firming ammonia as the cause of toxicity.
3.3 Measuring Ammonia Concentration
We have found that the ammonia-selective elec-
trode method has been satisfactory for measuring the
ammonia concentrations in most samples, (EPA, 1983;
APHA, 1992). Other methods for measuring ammonia are
available (such as distillation, nesslerization, and titration)
and can be used successfully for determining ammonia
concentrations in effluents (EPA, 1983; APHA, 1992).
The level of detection for total ammonia generally need
not be below 0.5-1.0 mg/l, since concentrations of <1.0
mg/l of total ammonia have not been found to be toxic to
fathead minnows and C. dubia. If ammonia measure-
ments are below 1 mg/l and the sample is toxic, it is likely
that the toxicant is not ammonia and other identification
procedures should be pursued.
The most reliable ammonia measurements are
obtained on fresh samples. However, samples can be
preserved by adding concentrated sulfuric acid and stor-
ing the samples at 4°C. The pH of the preserved samples
should be in the range of 1.5 to 2.0 (EPA, 1983; APHA,
1992). In recent experiments, we have used samples that
were stored without acidification at 10°C or refrigerated at
4°C for short periods with good success.
During several effluent tests, the amount of am-
monia in the test solutions (see test details below) has
decreased over the duration of the test. When levels are
in the range of 0-30 mg/l, it is prudent to measure the
initial concentration of ammonia in the test solution and
again after animals were exposed.
3.4 Graduated pH Test
The purpose of the Phase II graduated pH test is
to provide more definitive toxicity test data to implicate
ammonia as the toxicant in Phase II. In turn, this data may
be used in Phase III to confirm the role of ammonia in the
toxicity of the effluent. More stringent pH control and pH
monitoring will be needed to interpret test results and
more precise toxicity estimates (i.e., more replicates,
more dilutions, larger number of organisms; see Section
1.2; EPA, 1991 A; EPA, 1992) are needed in Phase II than
in Phase I. When it is important to predict the impact of
the toxicant in the receiving water, the pH of the dilution
water should be maintained at receiving water pH. The
test procedures discussed below provide good pH control
for the graduated pH test. Greater detail is provided for
some of the procedures in Phase I (EPA, 1991 A; EPA,
1992).
The test chamber size, number of dilutions, species
to be tested, type of test (acute or chronic), and the
degree of toxicity of the effluent will dictate the volume of
effluent needed for the graduated pH test. As a general
guide for acute toxicity tests, 300 ml of effluent should
suffice for any of three pH adjustment tests described
below. The volume for chronic tests will vary based on the
type of chronic test performed, the species used, the
number of concentrations tested, and number of solution
renewals in addition to the items discussed above (Sec-
tion 1.2 and EPA, 1992).
The procedure for conducting the graduated pH
test is to evaluate and determine the toxicity of the
effluent at three different pHs (e.g., 6.0,7.0, and 8.0). The
pH should be measured in all of the chambers. If the pH
drifts 0.2 pH units or more, the results may not be usable
and better pH control must be achieved. However, if pH
fluctuates more than 0.2 pH units and toxicity is present
only, at one pH, the toxicity results may still be useful. The
pH levels selected must be within the physiological toler-
ance of the test species used (which generally is a pH
range of 6 to 9). We recommend use of two methods of
pH control and comparison of these results to determine
that the pH adjustment itself did not introduce an artifact
of toxicity. This type of testing may be critical to explaining
effects in Phase III (EPA, 1993A).
Regardless of the pH control method chosen, the
selection of the appropriate blank is difficult. The change
in pH of the dilution water or surface water is not compa-
rable to that of the effluent because the composition of the
solutions are different. For some effluents, the addition of
either acid or base can be used to adjust and hold the pH
within 0.1 pH unit. If this is possible, this technique can be
used to compare the results with either the CO2-pH
controlled test or the buffer-pH controlled test. Test re-
sults should be similar and these comparisons can be
used as a basis for identifying ammonia as a toxicant.
3.4.1 pH Control: Acid/Base Adjustments
The first method of pH adjustment is the acid/
base adjustment described in Phase I (EPA, 1991 A; EPA,
1992). For this manipulation, the adjustment of pH is
relatively easy and quick, and the loss of volatile com-
pounds is minimized. However, the drawbacks are that:
toxicity enhancement from the additives may occur (espe-
cially in a chronic TIE), the addition of strong acid or base
disrupts the carbonate system equilibrium, the effects of
the pH change in the blanks may not serve as a toxicity
control for the effluent, the pH stabilization time is lengthy,
and pH tends to drift in longer term tests. In the pH range
6 to 9, the amount of high quality acid or base added is
usually negligible, and the likelihood of toxicity caused by
increased salinity levels is low.
The pH of each concentration and replicate must
be frequently monitored because a constant pH during
the toxicity test must be maintained. Larger test volumes
may be useful to prevent rapid pH fluctuations. The
amount of acid and/or base added should be recorded for
each pH adjustment to track the additive amount in the
effluent samples and the blanks. If toxicity increases
dramatically, the concentrations of salts should be calcu-
lated to be sure the salinity has not increased above the
tolerance level for the TIE species.
3-5
-------
3.4.2 pH Control: CO, Adjustments
The second method uses CO2 to adjust and
control test solution pH. The pH is adjusted by varying
and controlling the CO2 concentration of the gas phase
over the water or effluent sample in closed headspace
test chambers. It is necessary to maintain a constant pH
throughout the test period. The pH of most natural waters
and some effluents is controlled by the bicarbonate buff-
ering system and surface waters normally contain <10
mg/I of COr Therefore, the amount of CO2 to add de-
pends on the desired pH and the chemistry of each test
solution. The CO2 adjustment has the advantages that the
pH is controlled without placing additives directly into the
effluent test solutions, the pH change is easy to make,
and the pH is generally stable for at least 24 h if the gas-
tight container is not opened. Frequent pH measurements
are still possible because the headspace can be ref lushed
with a predetermined concentration of CO2/air. The disad-
vantages are that toxicrty can occur from the CO2, the
concentration of CCyair varies for each dilution and efflu-
ent (which requires sample specific experimentation) and
the manipulations for chronic tests can be time-consum-
ing relative to the acid/base adjustment method. We have
not observed any increased toxicrty from the addition of
CO8 unless the concentration in the chamber is over 10%.
Adjustments of the pH to 6.0 or 7.0 can be made
by using CO2 with or without first adding HCI to the test
concentrations. The CO2 is purchased in pure form through
local commercial gas suppliers, and if particular concen-
trations of COj/air are frequently used, a cylinder of gas of
the desired concentration may prove to be resource-
efficient. The amount of CO2 needed to adjust the pH of
the solution Is dependent upon sample volume, the test
container volume, the desired pH, the temperature, and
the effluent constituents (e.g., dissolved solids). Some
preliminary work is needed to determine the concentra-
tion of CO2 to add to achieve the desired pH. When
dilutions of an effluent have the same hardness and initial
pH as the effluent, about the same amount of CO will
usually be needed for each dilution. Sometimes, higher
concentrations of CO2 are needed for the higher test
concentrations. Use of a dilution water of similar hardness
as the effluent may make the CO2 volume adjustments
easier. A different dilution water may only be used in
these tests if the toxicity has not been shown to be
dependent on water hardness at any pH.
In our laboratory, we have found that glass can-
ning jars with rubber seals and metal balers work well as
a gas-tight testing chamber. The testing chamber should
be large enough to hold the desired number of test cups,
with sufficient headspace to ensure proper DO levels. For
example, a 2-quart glass canning jar lying on its side will
easily hold 6-1 oz cups. We simultaneously test C. dubia
and fathead minnows in the same chamber using the test
solution volumes described in Section 1.2. Since many
plastics are permeable to CO2, glass containers are rec-
ommended. When CO,,/air is flushed into the headspace
of the test chamber, the pH of the test solutions will
usually reach equilibrium in about 1 h and a reliable pH
can be achieved. Generally, as the alkalinity increases,
the concentration of CO2 that is needed to maintain the
pH also increases. After 1 h, check the pH of the solutions
and flush the chambers again. Check the pH again after
2-3 h and from these data determine the concentration of
CO2 to add for initial pH adjustment for the actual toxicity
test and the amount needed for ref lushing after the cham-
ber is opened for feeding or pH measurements. In most
instances, the amount of CO2 produced by the test organ-
isms will not cause further pH shifts. When testing with
fish, which usually increase in size during the test, a pH
fluctuation may occur that would require flushing with
different (e.g., slightly lower) concentrations of CO2.
When the concentration of CO2 to inject for the
target pH values has been determined, prepare test solu-
tions, add test organisms (and food if necessary) and
inject the appropriate concentration of CO2 in air using a
1-liter gas tight syringe, and quickly close the test cham-
ber. The chambers should be flushed with the CO^air
mixture several times to ensure the displacement of air
currently in the chambers. Place the chamber out of direct
laboratory light, as temperatures tend to rise out of the
desirable test range in the closed chambers.
For effluents that have initial pH values from 7.8
to 8.5,0-10% CO2 concentration in the chamber has been
used to lower the pH to 6.0. Experiments in hard reconsti-
tuted water have shown that up to 8% CO2 can be
tolerated by C. dubia and fathead minnows in acute tests,
but 8% has been toxic to C. dubia and fathead minnows in
the 7-d tests. About 2-3.5% usually will lower the pH of
most effluents to 6.5-7.0. If more than 10% CO2 for acute
tests or 5% CO2 for chronic tests is needed to lower the
pH of the test solutions, before adding test animals adjust
the pH with high quality acid (EPA, 1991A; EPA, 1992)
and then flush the headspace with CO^air. The neces-
sary concentration of CO2 to use must first be determined
experimentally with effluent test solutions adjusted to the
appropriate pH with acid solutions. Sometimes >5% CO2
cannot be used for the dilution water pH adjustment test
without the CO2 causing toxicity.
The use of a single encloised test chamber for
controlling the pH at all test concentrations may allow the
transfer of volatile compounds among treatments. We
have experienced volatilization of ammonia in tests and
therefore, individual test chambers for each effluent con-
centration are preferable. Methods that use continuous
flow of a CO2/air mixture, such as tissue cell incubators,
may be preferable and give better pH control provided
that volatilization or cross contamination is not a problem.
At this time we have not attempted to use a continuous
flow of COj/air mixture and therefore cannot recommend
a system to use.
3-6
-------
Maintaining pH above the air equilibrium pH (gen-
erally above pH 8.3) is difficult without buffers (Section
3.4.3). The pH control in this high range is much more
difficult because the concentration of CO2 must be very
low and microbial respiration can increase the CO2 levels
in the test chamber. Use of CO2-free air in the headspace
may work or flushing a mixture of CO2-free air and normal
air through the headspace or test solution may be suc-
cessful. Because such small CO concentrations are
needed and because CO2 evolution by microorganisms or
test organisms can significantly alter the CO2 concentra-
tion, frequently flushing (two to four times a day) of the
headspace in, static tests will probably be required to
adequately control pH. For the chronic tests, we have not
attempted to use the CO2-free air bubbled through the
test solution, because more CO2 evolution tends to occur
during the chronic tests and the need for reflushing makes
the test labor intensive.
For the CO2-pH controlled tests, the pH should be
measured at least every 24 h for both acute and chronic
tests and ideally, continuously during pH controlled tests.
At each reading, flush the headspace with the CO,/air
mixture. A small amount of experimentation will confirm
whether the concentration of CO2 previously determined
is adequate, or whether the amount required for flushing
will be less than that used for the initial pH adjustment.
For chronic tests, daily renewal solutions should
be prepared, pH adjusted with HCI if necessary, dis-
pensed into test cups, and placed into a second glass jar
chamber and flushed with appropriate concentration of
CO2. These should be left to equilibrate at least 1-2 h.
Measure the pH quickly and transfer the animals to new
test cups and place them into the glass jar. Flush the
headspace again with the appropriate CCyair mixture.
Table 3-5. Percent Un-lonized Ammonia in Aqueous Solutions at
25°C and Various TDS Levels'
TSD
(mg/l)
0
250
500
750
1000
1500*
2000
3000
6.0
0.0568
0.0521
0.0505
0.0494
0.0485
0.0471
0.0460
0.0443
PH
7.0
0.566
0.519
0.503
0.492
0.483
0.469
0.458
0.441
8.0
5.38
4.96
4.81
4.71
4.63
4.50
4.40
4.24
9.0
36.2
34.3
33.6
33.1
32.7
32.0
31.5
30.7
'Data from Skarheim (1973).
For the 7-d tests with fathead minnows, the chambers
must be opened once more each day to accommodate
the feeding schedule. The experimenter can take advan-
tage of this by making a pH reading prior to placing food
into the test cups. CO/air must again be flushed into the
chamber. It is important to note that in the fathead min-
now test, the pH most likely will be lower after 24 h than in
the C. dubia test because of the food added and the
respiration of the fish which is considerably greater than
that of C. dubia.
Measurements of pH must be made rapidly to
minimize the CO2 exchange between the sample and the
atmosphere. Avoid vigorously stirring unsealed samples
because at lower pH values, the CO lost during the
measurement can cause a substantial pH rise. If possible,
measure the DO at the same time because ammonia may
have different toxicities when DO is decreased (EPA,
1985A). Keep in mind that if the test animals have been
dead for awhile, the pH and/or DO of the test water most
likely will have changed.
The controls in the CO2 chamber and the baseline
test act as checks on the general health of the test
organisms, the dilution water and most test conditions. If
the effluent pH in the baseline test is close to the pH of the
adjusted test solutions (at their respective LCSOs, IC25s
or ICSOs), the toxicity expressed in the two tests should
be similar. Significantly greater toxicity in the pH-adjusted
test may suggest interference from other factors such as
the ionic strength related toxicity if the pH was adjusted
with either HCI or NaOH, or possibly CO2 toxicity. Dilution
water blanks at the various pH levels may or may not be
appropriate since the effluent matrix may differ from that
of the dilution water. The dilution water blank will be
useful in checking the acids and bases that are added for
artifactual toxicity. Monitoring the acid and base additions
may be useful in determining if artifactual toxicity resulted
from the increase in salt content. Monitoring conductivity
of the effluent solutions after the addition of the acids and
bases may also be helpful in determining artifactual toxic-
ity. The ionic strength of hardwaters or saline waters
results in a decreased level of un-ionized ammonia (Table
3-5). For values of TDS from 0-500 mg/l, the dissociation
constants are expected to be more accurate than values
above 500 mg/l that were based on somewhat tenuous
assumptions (Skarheim, 1973; see Table 3-5).
3.4.3 pH Control: Buffer pH Adjustments
The third method of pH control uses the addition
of standard buffers to the effluent and dilution water to
adjust the pH. This method has the advantage in that pH
is stable with the buffer addition, the pH change during a
test is slow, frequent pH measurements are possible
because test vessels are not in air-tight chambers, and
the test method set-up is rapid. The disadvantages are
that toxicity enhancement or interference from buffers
may occur, not all buffers can be used without additional
3-7
-------
acfd/base adjustments, and the pH stabilization time may
be lengthy.
Hydrogen ion buffers are used to maintain the pH
level In the graduated pH test (EPA, 1991 A; EPA, 1992).
Three hydrogen ion buffers were used by Neilson et al.
(1990) to control pH in toxicity tests in concentrations
ranging from 2.5 to 4.0 mM. These three buffers were
chosen based on the work done by Ferguson et al.
(1980). These buffers are: 2-(N-morpholino) ethane-sul-
fonlc acid (Mes) (pK = 6.15), 3-(N-morpholinp) propane-
sulfonlc acid (Mops) (pK. = 7.15), and piperazine-N,N'-bis
(2-hydroxypropane) sulfonic acid (Popso) (pKa = 7.8). We
have also used two additional buffers: N-(2-hydroxyethyl)
pIperazine-N'-2-hydroxypropanesulfonic acid (Heppso)
(pKa » 7.8) and N-tris-(hydroxymethyl) methyl-3-
amlnopropanesulfonic acid (Taps) (pKa = 8.4). The Taps
buffer Is more frequently used than the Heppso buffer.
We have experienced problems of having to add an
excessive amount of base to obtain the desired pH with
the Popso buffer. The Taps buffer effectively maintains
the pH above 7.8. Keep in mind that pH is best main-
tained at or near the pKa of the buffer.
The acute toxicity of these buffers is low to both
C. dubfe and fathead minnows (EPA, 1991 A) and 4 mM
concentration or less of all five buffers has not caused
chronic toxicity to C. dubia or the fathead minnow. The
buffers are added at sublethal (e.g., NOEC) levels to
maintain the pH of test solutions. While these buffers
serve to prevent the pH from drifting during the test, pH
adjustment to the desired level is required in the prepara-
tion of the solution. A portion of the buffer compound is
weighed out and added to the aliquots of whole effluent
and dilution water, and both are then pH adjusted with
acid or base solutions to the desired pH values. Serial
dilutions are made, replicates prepared, and test organ-
isms are added. Care should be taken to ensure equilib-
rium of buffered solutions, which may take at least 1-2 h.
Dilutions should also be left to equilibrate and minor pH
adjustments should be made. In certain situations, it may
be desirable to prepare the solutions the day before tests
begin. At present, we have found we can use batch
solutions prepared ahead of time for solution renewals.
Our experience also indicates that the amount of any
buffer needed to hold any pH is effluent specific. Experi-
mentation with effluents will be required to determine the
lowest concentration of buffer needed to maintain the
desired pH. The test solutions need not be covered tightly
to maintain pH; however, pH should be measured at each
test reading at all dilutions.
Use of the buffers is still being developed and the
effects caused by interferences from the buffers them-
selves have not been fully studied. It is possible that the
buffers may reduce the toxicity of some toxicants, but this
has not generally been seen.
3.5 Zeolite Resin Method
Zeolite is composed of naturally occurring or
synthetically created crystalline, hydrated alkali-aluminum
silicates. The general formula is Mn+O*AI2O3-ySiO2*zH2O;
M = group IA or IIA element, n = +2. for group IA, +1 for
group IIA, y>2, and z = the number of water molecules
contained in the interconnected voids or channels within
the zeolite (Windholz, et al., 1983). When zeolite is placed
in aqueous solutions, the positively charged group IA or
IIA elements (Mn+) of the zeolite are mobile and can
undergo exchange with other cations in the water. As
such, zeolite has frequently been employed as ion ex-
change resins to remove the ammonium ion (NH4+) from
aqueous solutions in TIE work (Ankley et al., 1990S;
Burkhard and Jensen, 1993). Because of its ability to
exchange other cations such as heavy metals, and its use
as molecular sieves, filter adsorbents and catalysts, zeo-
lite was not suggested for use in Phase I, except as a
subsequent test (EPA, 1991 A). Zeolite can be effective in
Phase II, if Phase I results implicate ammonia as the
toxicant and establish that other types of toxicants (such
as non-polar organics and metals) play no role in the
effluent toxicity.
For the acute TIE procedure, zeolite particles
should be screened to be in the range of 32 to 95 mm, to
ensure efficient ion exchange while preventing channel-
ing or excessive resistance to flow. Extremely large or
small particles can be removed by screening the zeolite
with sieves or mesh screens. The zeolite column can be
prepared by taking 30 g of aquarium zeolite (Argent Chemi-
cal Laboratories, Redmond, WA) and adding it to 60 ml of
high-purity water. The zeolite slurry is poured into a
chromatography column (11 mm i.d. x 15 cm) and three
bed volumes of dilution water are passed through the
column. The last 10 ml of dilution water is collected for
use as a zeolite blank and should not be toxic. Next, 200
ml of 100% effluent is passed through the column at a
rate of 2 ml/min. The post-column effluent that is collected
will be toxicity tested and its ammonia concentration
measured. Temperature and pH should be recorded at
test initiation to provide the means to calculate both total
and un-ionized ammonia in the sample.
For chronic toxicity tests larger amounts of zeolite
should be used. This can be scaled up proportionally from
the amounts used in the acute zeolite work. The amounts
of solution needed for testing and ammonia measure-
ments will dictate the amount of sample to prepare.
Typically a slurry of 60 g of zeolite and 120 ml of high
purity water is sufficient for levels of ammonia in the range
of 5-50 mg/l and for processing 2,000 ml of effluent. The
post-zeolite effluent is collected in aliquots, then each is
toxicity tested. In this manner, break-through of ammonia
can be measured and toxicity of the various samples with
different ammonia levels can be estimated.
3-8
-------
Toxicity tests and ammonia measurements are
conducted on the effluent and post-zeolite column efflu-
ent. Removal of toxicity by the zeolite column and re-
moval of the ammonia concentration will add to the
evidence implicating ammonia as the toxicant. An aliquot
of the effluent sample (not having passed through the
zeolite column) is used for ammonia analysis and the
baseline toxicity test. These data will be compared with
the same data for the post-zeolite column effluent to
determine if the post-column reduction in effluent toxicity
is consistent with ammonia removal by the zeolite. The
control for test organism survival, dilution water quality
and other test conditions will be provided through toxicity
tests on dilution water. Dilution water (at the same hard-
ness as the effluent) should be passed through the zeolite
column, and will act as a blank for toxic artifacts leached
from the zeolite. Increased toxicity in the post-zeolite
effluent, relative to the whole effluent, indicates the pres-
ence of toxic artifacts. Since many cations will be ex-
changed, adding solids in the acute tests, such as the
YCT food (yeast-Cerophyl®-trout food) fed to C. dubia,
might improve control survival. Additional clean-up tech-
niques for the zeolite (such as Soxhlet extraction) or
alternate uncontaminated sources of zeolite might be
needed: Column packing, effluent pH, ammonia levels,
and flow rate through the column can all affect the effi-
ciency of the cation exchange process. Lowering effluent
pH prior to zeolite treatment and/or lowering flow rate
through the column might also result in greater removal of
ammonia. Occluded gas between zeolite particles might
also impair the column's capacity to remove ammonia. If
this appears to be a problem, the zeolite slurry should be
degassed by using a vacuum prior to pouring it into the
column.
Zeolite columns can be regenerated, but fresh
zeolite should be used to pack columns the first time. If
the graduated pH test and the zeolite test results are
consistent with ammonia toxicity, Phase III confirmation
procedures should be started.
Once ammonia is identified and confirmation is
initiated, the post-zeolite samples can be spiked with
ammonia at the same concentrations as are present in
the effluent. These tests are an integral part of the Phase
III confirmation process (EPA, 1993A).
3.6 Air-Stripping of Ammonia
This method of ammonia removal takes, advan-
tage of the fact that the relatively volatile un-ionized
ammonia (NH3) predominates in a solution with a pH
above 9.3. For this reason, one might expect that ammo-
nia would be removed during the Phase I pH 11 adjust-
ment/aeration test (acute testing) or the pH 10 adjustment
and aeration test (chronic testing). Based on our experi-
ence ammonia is not removed by this method, most likely
because the Phase I aeration manipulation is done in a
graduated cylinder, which has a low surface-to-volume
ratio. By stirring the sample for a longer period of time
(>1 h) at a high pH (pH 9.0 or higher) in a containerthat allows
a large surface area to volume ratio, most of the ammonia
can be removed from aqueous samples.
A measured amount of effluent for subsequent
analysis and testing is pH adjusted to 10 or 11 and placed
into a large shallow glass container (e.g., 1000 ml crystal-
lizing dish). The solutions are then agitated (stirred) con-
tinuously. The length of time the sample must be stirred is
dependent on the concentration of total ammonia in the
sample. We have found that for most samples of 10-100
mg/l of total ammonia, 1-6 h is adequate to remove most
of the ammonia. After air-stripping is completed, the
volume of effluent should be measured and any appre-
ciable loss replaced with high purity water or toxicity might
be caused by the concentration of other components in
the effluent. The ammonia concentration should be mea-
sured immediately after air-stripping and after volume
adjustment is complete to ensure ammonia levels are
reduced before toxicity tests are initiated. Toxicity tests on
the air-stripped solution can then be conducted for both
acute and chronic TIE work. Dilution water blanks at the
various pHs may or may not be appropriate since the
effluent matrix will probably differ from that of the dilution
water. Monitoring the acid and base additions may be
useful to determine if artifactual toxicity resulted from the
increase in salt content and subsequent evaporation that
occurred during the air-stripping process. Monitoring con-
ductivity of the effluent solutions after the addition of the
acids and bases may also be helpful in determining
artifactual toxicity. The dilution water blank should be
treated in the same manner as the effluent although it
may not serve as a true toxicity control for the effluent.
If the ammonia is decreased and the toxicity is
reduced or absent after air-stripping, ammonia is strongly
implicated as a contributing factor to the toxicity of the
effluent. The results of this test should be compared with
the aeration test results of Phase I, the baseline effluent
test and the other graduated pH tests. Other compounds
could precipitate as a result of the pH adjustment and
during the air-stripping procedure. Precipitates may form
during the air-stripping process and not dissolve after the
volume is readjusted, leaving these compounds unavail-
able.
3-9
-------
-------
Section 4
Metals
4.1 General Overview
This section contains procedures that can be
used to identify suspect metal toxicants. The initial evi-
dence used to implicate metallic toxicants is obtained
from the Phase I characterization tests, with the results
of the EDTA (ethylenediamine tetraacetic acid) addition
test providing the best indication of the presence of a
metal toxicant. When certain cationic metal toxicants are
present, a reduction in sample toxicity with the addition
of EDTA should be observed. Other Phase I manipula-
tions that remove or reduce sample toxicity and suggest
the presence of a cationic metal include the sodium
thiosulfate addition test, the use of a C18 SPE column,
and filtering the sample when combined with minor pH
adjustments. One additional indication of metal toxicity
may be when the organisms' response in the toxicity test
is atypical of the expected dose response relationship
(i.e., partial mortalities in several test concentrations
Schubauer-Berigan et al., 1993B).
Subsequent Phase I tests such as using ion
exchange resins might also lead one to the conclusion
that a metal is the toxicant. Toxicity removal or reduction
after a sample is treated with an anion exchange resin
might implicate toxicants that exist as anionic oxides in
water, such as arsenic, chromium, and/or selenium.
These anionic oxides will not be specifically removed or
rendered biologically unavailable by the routine Phase I
tests. Therefore, when the Phase I tests do not seem to
show any toxicity reduction, toxicants such as these
might be suspected and subsequent tests as discussed
above could be useful (see Phase I, EPA, 1991A; EPA,
1992). These situations should be approached on an
individual basis since other classes of toxicants might
demonstrate the same behavior in Phase I (e.q. total
dissolved solids (TDS)).
Further discussion and interpretation of the Phase
I results which would lead to the conclusion that a
cationic metal toxicant was present in a sample are
provided in the Phase I TIE documents (EPA, 1991 A-
EPA, 1992).
Other information, such as process details from
the discharger and information from past TREs and/or
TIEs, might also help to implicate cationic metals as the
toxicants/However, this type of information should be
interpreted and used with caution as it might bias the TIE
efforts.
If the EDTA addition test in Phase I showed that
toxicity was removed or reduced one should proceed to
the metal analysis section (Section 4.2). This section
provides guidance and recommendations for analyzing
samples for metals so that a list of suspect metal toxi-
cants can be obtained. This section also discusses clean
metal techniques, detection limits, a prioritization process
for analyzing for specific metals, dissolved vs biologically
available metals, and provides the rationale for assem-
bling the list of suspect metal toxicants. Prioritizing metals
to analyze from Phase I results is strongly recommended
in order to save money and time in the TIE process.
If other Phase I tests implicate a meta! but EDTA
does not, it may be helpful to acquire additional test
information through the use of EDTA addition tests so-
dium thiosulfate addition tests, graduated pH tests, and
ion-exchange resins. This additional toxicity testing (Sec-
tion 4.3) may be useful in certain situations before analyz-
ing for metals, even when EDTA additions reduced toxicity.
These situations include: when the addition tests of EDTA
and sodium thiosulfate in Phase I were performed using a
single sample concentration (i.e., no dilutions), when the
time it takes to obtain results of metal analyses is lengthy,
or when Phase I results indicate another type of toxicant
(non-metal) is present. The data obtained from the addi-
tional testing can then be included in the prioritization
process for metals analysis. Professional judgement is
required to decide when you have sufficient and appropri-
ate toxicity testing data to proceed to metals analysis.
After processing one sample, a list of suspects
may be generated. As future samples are evaluated, the
correlation between toxicity of a sample and the
concentration(s) of metal(s) over time may also be used
to narrow the list of suspect toxicant(s). In Phase III, the
suspect metal toxicant is implicated based upon the cor-
relation of effluent toxicity and metal concentrations ref-
erence suspect metal toxicity data, the use of additives
that chelate metal toxicants, and changes in toxicity ob-
served during manipulations of water quality characteris-
tics.
4-1
-------
The procedures in this chapter are generally ap-
plicable for both acute and chronic toxicity. The main
differences between the acute and chronic procedures
are the concentrations of additives used in the EDTA and
sodium thfosulfate addition tests, lower analytical detec-
tion limits, and generating non-toxic blanks for the ion
exchange resins for chronic toxicity testing. The use of
species other than C. dubia or fathead minnows will
require consideration of appropriate test volumes and
additive concentrations.
4.2 Analysis of Metals
4.2.1 Prioritizing Metals for Analysis
Many cationic metals can be analyzed in a spe-
cific sample, but to simplify the amount of analytical effort
needed for metals analysis, we suggest a prioritizing
process be performed before analyzing any samples. The
prioritization process is more valuable when the metal
analyses are performed by AA instrumentation since each
metal requires an individual analysis. Conversely, with
ICP (inductively coupled plasma) instrumentation, numer-
ous metals can be analyzed at once, and the prioritization
process is less valuable in this instance. With both ICP
and AA methods, a list of metals and required levels of
detection will be needed before the samples are ana-
lyzed.
This prioritization is based primarily upon acute
toxicity data with C. dubia. its applicability to chronic
toxicity and other species is expected to be similar but
has not yet been determined. The toxicity test results from
the EDTA additions, sodium thiosulfate additions, and
graduated pH tests performed in Phase I form the basis
for prioritization. When available, Phase II results from
using the procedures in Section 4.3, should be included in
this evaluation. Because we do not have a complete
understanding of the effects of these procedures for each
metal, the following should be taken as a starting point for
metals analysis.
Information regarding historical discharge moni-
toring data, past or current TRE and/or TIE information, or
process information may be useful in prioritizing metals
for analysis. For example, if a discharger uses zinc in their
manufacturing process and EDTA removed the toxicity, it
would be logical to analyze for zinc first. If zinc was
present at nontoxic concentrations or at concentrations
too low to cause the observed toxicity, analysis for addi-
tional metals would be performed. If zinc was present at
concentrations high enough to cause the observed toxic-
ity, Phase III procedures (EPA, 1993A) should then be
started to confirm zinc as the identified suspect toxicant.
When EDTA additions reduce or remove the
toxicily of the sample, initially copper, lead, cadmium,
nickel, and zinc should be measured. When sodium thio-
sulfate additions reduce or remove the toxicity of the
sample, copper, cadmium, and silver should be mea-
sured.
Phase I results would not normally lead to the
conclusion that an anionic toxicant was present (i.e.,
cationic metals that exist in aqueous samples as anionic
oxides). If additional Phase I tests had been performed
which characterized anionic toxicants or other specific
discharger information was available, measurements of
arsenic, chromium, and selenium should be made.
As stated above, these metals should be a start-
ing point for metals analysis. Further interpretation of the
Phase I results could be done by including the results of
the graduated pH test and by jointly examining the results
of the EDTA addition, thiosulfate addition, and graduated
pH tests.
When multiple toxicants of different classes are
present, Phase I data are often difficult to interpret. One
should try to identify and confirm as soon as possible the
role of one toxicant when multiple toxicants are present.
By defining the role of one toxicant, efforts can be better
focused on the remaining unidentified toxicants.
4.2.2 Metal Analysis Methods
There are three types of chemical instrumenta-
tion available for the analysis of cationic elements; these
are AA, inductively-coupled plasma-atomic emission spec-
troscopy (ICP-AES), and inductively-coupled plasma-mass
spectrometry (ICP-MS).
EPA methods using ICP-AES, ICP-MS, and AA
(EPA, 1983; EPA, 1991D) are available for quantifying
cationic metals in aqueous samples. Tables 4-1 and 4*2
summarize method detection limits for the analysis of
cationic metals in aqueous samples using AA with direct
aspiration, AA with the furnace procedure, ICP-AES, and
ICP-MS.
The detection limits required in Phase II for the
identification of suspect cationic metal toxicants will be
determined by the toxicity of metals for the TIE species. In
some cases, especially for chronic toxicity, the effect level
might be lower than the detection limits listed in Tables 4-1,,
and 4-2. Detection limits should be improved to obtain
optimal levels of detection (i.e., at least two times lower
than the effect level).
Toxicity data for some species and test types for
many metals have not been determined, especially for 7-d
chronic toxicity tests. Therefore, to determine the needed
levels of detection, effect levels for specific metals may
have to be determined.
The required level of detection will often dictate
the method heeded for performing the metal measure-
ment. It will be beneficial for laboratories to compile a
database containing method detection limits and toxic
effect levels for cationic metals using data,from their
organisms, analytical methods, and toxicity testing condi-
tions. These data are not necessary in advance but fthis
4-2
-------
Table 4-1. Atomic Absorption Detection Limits and Concentration Ranges1
Direct Aspiration
Furnace Method2
Metal
Aluminum
Antimony
Arsenic3
Beryllium
Cadimum
Calcium
Chromium
Cobalt
Copper
Lead
Magnesium
Manganese
Mercury4
Molybdenum (p)
Nickel(p)
Potassium
Selenium2
Silver
Sodium
Tin
Vanadium (p)
Zinc
Detection
Limit
(mg/l)
0.1
0.2
0.002
0.005
0.005
0.01
0.05
0.05
0.02
0.1
0.001
0.01
0.0002
0.1
0.04
0.01
0.002
0.01
0.002
0.8
0.2
0.005
Optimum
Concentration
Range( mg/l)
5-50
1-40
0.002 - 0.02
0.05 -2
0.05 -2
0.2-7
0.5-10
0.5 -5
0.2 -5
1-20
0.002 - 0.5
0.1 -3
0.0002 - 0.01
1 -40
0.3 - 5
0.1-2
0.002 - 0.02
0.1-4
0.03 - 1
10-300
2-100
0.05 - 1
Detection
Limit
(W/l)
3
3
1
0.2
0.1
1
1
1
1
0.2
1
•|
2
0.2
5
4
0.05
Optimum
Concentration
Range (ng/l)
20 - 200
20 - 300
5-100
1 -30
0.5 - 10
5-100
5 -inn
1 VU
5-100
5-100
. 1-30
3-60
5Cf\
- OU
5-100
1 - 95
1 Ł.\J
20 -300
10-200
0.2-4
2-rir °oul"°"=" «=K»-UUII iimns ana concentration ranges were taken from EPA 1983
The listed furnace values are those expected when using a 20 ul injection and normal gas flow except in
gX«h~^^ intermpt is used The
3Gaseous hydride method.
"Cold vapor technique.
Table 4-2.
Element •
Estimated Instrumental Detection Limits for ICP-MS and
ICP-AES
Estimated Detection
Limit, ICP-MS' ftig/l)
Estimated Detection
limit, ICP-AES2 (pg/l)
Aluminum
Arsenic
Antimony
Beryllium
Cadmium
Calcium
Chromium
Cobalt
Copper
Lead
Magnesium
Manganese
Molybdenum
Nickel
Potassium
Selenium
Silver
Sodium
Vanadium
Zinc
0.05
0.9
0.08
0.1
0.1
_
0.07
0.03
0.03
0.08
0.1
0.1
0.2
_
5
0.05
_
0.02
6.2
45
53
32
0.3
4
10
7
7
6
42
30
2
8
15
.3
75
7
29
8
2
The estimated instrumental detection limits are taken from EPA,
1991D. They are given as a guide for instrumental limits, the actual
detection limits are sample dependent and may vary as the sample
matrix varies.
2The estimated instrumental detection limits as shown are taken from
EPA, 1983. They are given as a guide for an instrumental limit. The
actual method detection limits are sample dependent and may vary
as the sample matrix varies.
3Highly dependent on operating conditions and plasma.
type of information will be very useful for future TIE
efforts.
When toxicity effect levels appear to be below the
detection limits of current analytical methods, the use of
"clean" analytical techniques may be required through all
steps in the analysis of the sample because background
contamination is the major cause of inadequate levels of
detection. Some general principles of clean metal tech-
niques include the use of contamination free reagents,
acid cleaned plastic labware, acid cleaned membrane
filters (not glass fiber), class 100 benches for sample
preparation, proper sample collection, preservation, and
storage procedures; and proper QA/QC procedures using
blanks, spiked matrixes, and replicate analyses. A sum-
mary of clean metal techniques and procedures for lower-
ing the levels of detection can be found in Nriagu et al
1993; Patterson and Settle, 1976; and Zief and Mitchell'
1976.
Some cationic metals, such as arsenic, selenium,
and chromium, have different stable oxidation states in
aqueous samples and more importantly the different oxi-
dation states may have different toxicities. In Section
4.2.3, procedures to determine the concentration of the
different oxidation states are provided.
In some TIEs, a measurement of the metals
associated with the suspended solids may be needed
4-3
-------
(Section 4.2.4). Procedures for preparing suspended sol-
ids removed by filtration for metals analysis are available;
see EPA method 200 (EPA, 1983) and EPA method
200.2 (EPA, 1991D).
4.2.3 Metal Speclatlon
The procedures suggested above (Section 4.2.2)
are used to determine the total concentration of a metal in
an effluent. Many metals exist in water in different forms
due to the various stable oxidation states of the metal.
Arsenic (As3*, As5*), chromium (Cr3*, Cr6*), and selenium
(Se4*, Se**) are important metals that exist in different
forms in water. Determining the speciation for these met-
als may be important in the TIE since the toxicities are
different for the various forms of each metal. For example,
Cr5* is the form that is of toxicological concern while Cr3*
is generally not toxic (EPA, 1985D).
For chromium, methods for measuring the
hexavalentform (Ct6*) such as method 218.5 (EPA, 1983)
are available. The amount of the trivalent form of chro-
mium (less toxic form) is determined by taking the differ-
ence between the concentrations for total and hexavalent
chromium.
For arsenic, the method of Ficklin (1983) is sug-
gested for speciation measurements. This method uses
an anion-exchange resin to separate the arsenite (As3*)
and arsenate (As5*, more toxic form) species. Graphite
furnace atomic-absorption spectroscopy is then used to
measure the concentrations of each form.
For selenium, the method of Oyamada and Ishizaki
(1986) is recommended for speciation. This method (like
that for arsenic) uses column chromatography with an
anion-exchange resin to separate the selenite (Se4*) and
selenate Se6* (more toxic form) and graphite furnace
atomic-absorption spectroscopy to measure each form.
Ion chromatography can also be used to deter-
mine the different forms of the above metals (EPA, 1991 D),
but we have not used this technique to date.
4.2.4 Identification of Suspect Metal Toxicants
Initial implication of suspect metals based on a
comparison of total metal analyses data and effluent
toxiclty test results should be made. Then analysis for
dissolved and suspended metals can be made if neces-
sary. These metal values should be compared to avail-
able toxicity values, but tests on reference metals might
have to be conducted with matching effluent conditions,
such as pH and hardness to obtain comparable toxicity
values. Side-by-side tests with individual reference metal
standards and effluent samples might prevent being mis-
lead by different test designs and are worth the effort.
Literature summaries of metal toxicity data are also avail-
able (EPA, 1980; EPA, 1985B; EPA, 1985D; EPA, 1985E;
EPA, 1985F; EPA, EPA, 1986; EPA, 1987; EPA, 1988B;
and AQUIRE, 1992). In addition to matching the hardness
and pH of the dilution water to the effluent sample by the
addition of the appropriate ratios of magnesium carbon-
ate and calcium carbonate, it might be possible in some
cases to mimic the wastewater total suspended solids
(SS) and total organic carbon (TOG) in the water used to
test the metal. For example, TOC and SS from the
addition of the YCT food can be at levels such that the
total SS level in the dilution water might be similar to that
found in the effluent. TOC may also be modified by the
addition of humic acid. If the dilution water does not
closely match the effluent, nonstandard dose-response
relationships are observed in the toxicity test, i.e., several
test concentrations exhibit partial mortality. In addition, a
trend is noticed that as metal concentrations decrease at
the effluent LCSOs or IC25s, toxicity of the effluent in-
creases.
If a sample is to be filtered, a membrane filter(s),
such as a 0.45 u.m polycarbonate filter should be pre-
pared by rinsing with high purity water, followed by an
appropriate volume of dilution water needed for blank
toxicity tests and analysis. The toxicity test guidance is
described in Section 1.2 and in the Phase I documents
(EPA, 1991 A; EPA, 1992). An appropriate quantity (<50
ml) of the last portion of the high purity water passing
through the filter should be collected as an analytical
blank to check for metals contamination from the filter and
the filtration apparatus. An aliquot of the effluent is then
filtered through the 0.45 urn membrane filter(s). If more
than one filter is required for the effluent, a portion of each
can be combined for testing.
The filtered and unfiltered effluent samples and
the filtration blank should be tested for toxicity to measure
the effect of filtration on sample toxicity. The toxicity test
techniques are described in the non-polar organic section
(Sections 2.2.3 and 2.3.3) unless data are needed for
Phase III confirmation and then, greater replication and
randomization will be needed (see Section 1.2). The
toxicity tests should be performed for the test species
using a dilution water (e.g., reconstituted water) of a
similar hardness and pH to that of the effluent. If toxicity is
reduced or removed upon filtration (and effluent toxicity
has not previously been affected by C18 SPE or filtration
through a glass fiber filter), it is possible that metals were
retained by the 0.45 urn filter. Analysis for metals retained
by the filter may help in interpreting sample data.
Metals analyses should be performed on the
analytical blank collected from the filter and on the filtered
and unfiltered effluent samples. The choice of metals to
measure will be determined by the prioritization process
described above. As stated previously, the level of detec-
tion for the metal of interest should be lower than the
effect concentration for the metal.
Biologically Available Metals: Traditionally, dis-
solved metals for aqueous samples have been defined as
those that pass through a 0.45 urn membrane filter, i.e.,
4-4
-------
polycarbonate filter. The dissolved metals are in no way
synonymous with the biologically available metals. Other
than the use of an aquatic organism there is no technique
to determine the biologically available fraction of the total
metal. Furthermore, only rudimentary techniques are avail-
able to specifically identify the individual species of a
metal (e.g., free charged metal ions [Mn+], inorganic ion
pairs or complexes such as aquoions, [M(HaO)n+m],
hydroxoions [M(OH) "-P+], oxoions [MO n-2<<+], organic com-
plexes and chelates [M x EDTA], metal species bound to
high molecular weight organic material [M x lipid] or metal
species in the form of highly dispersed colloids or sorbed
on colloids [M x clay]). Stumm and Morgan (1981) have
listed some general methods for assisting in identification
of individual species. In some cases, binding of metals to
inorganic and organic ligands in effluents will reduce the
bioavailability of the metals and cause the metal concen-
tration at the effluent LC50, IC50, or IC25 to be larger
than the metal concentration determined in the metal
dilution water toxicity test. For a set of effluent samples
with a wide range of toxicities, better agreement should
occur between the effect concentration of the metal in a
dilution water toxicity test and the more toxic effluent
samples (where the toxicity testing matrix of the effluent
more closely matches that of the dilution water). Methods
for determining the bioavailable fraction of the total metal
are limited.
Some indication of the binding of metals to organ-
ics in the effluent may be arrived at through hexane
extraction of an aliquot of the sample (Stary, 1964).
Theoretically, metals bound to organic materials that are
soluble in hexane should be extracted from the effluent.
The hexane can then be evaporated and the residue
reconstituted and analyzed for metals. Additionally, the
loss of metals can be estimated by repeating the metal
analysis on the extracted effluent and comparing this
result to the hexane extract results. The toxicity attributed
to metals associated with organics might be estimated by
performing a toxicity test on the solvent extracted effluent.
Traces of hexane must be removed from the extracted
eff|uent by aeration prior to toxicity testing. The effects of
aeration on sample toxicity must also be considered in
this analysis. In any case, metals strongly suspected of
causing or contributing to sample toxicity should be tested
in dilution water as described above with the TIE test
species.
The effects of variable water quality characteris-
tics on metal toxicity must be evaluated over the effluent
sampling period. One way to assess this is to collect
several samples over a short time span. As an example,
for an acutely toxic effluent, collect six grab samples in 24 h, and
calculate the correlation coefficient for sample metal concen-
tration (or summed toxic units of metals) versus sample
toxicity for each sample. The set of correlation coeffi-
cients for multiple sampling events might give results less
affected by hardness, SS, and TOC, assuming that water
quality characteristics affecting metal toxicity will vary less
during short time periods. For chronic toxicity, it might be
useful to measure concentrations of metals in several
daily samples and conduct separate chronic tests on
each sample. Obviously, metal concentrations must vary
enough to provide a sufficient range for correlation. When
one reaches this stage, Phase III work should start using
Phase III methods. Symptoms, species sensitivity, spik-
ing, water quality adjustments and correlation are all
applicable Phase III approaches to confirm the cause of
toxicity.
4.3 Additional Toxicity Testing Methods
Guidance on EDTA addition tests, sodium thio-
sulfate addition tests, graduated pH tests, and the use of
ion-exchange resins for use in Phase II are presented in
this section. These procedures might be used before
performing analyses for cationic metals, but most often
they will be used to refine a list of suspect metal toxicants
and to provide data to support the identified suspect in
Phase 111.
In the acute Phase I, EDTA and sodium thiosul-
fate addition tests can be conducted by adding incremen-
tal amounts of EDTA or sodium thiosulfate to a single
effluent concentration. To provide further evidence in
Phase II, these two tests should be conducted with efflu-
ent dilutions to assess the toxicity reduction (see EPA,
1992). The data generated from these procedures pro-
vide a powerful tool for identifying the cause of toxicity in
samples containing mixtures of cationic metals. For ex-
ample, toxicity caused by either copper or zinc could be
determined by using the following test information: toxicity
of both metals would be removed by EDTA addition
(Section 4.3.1), sodium thiosulfate can remove toxicity
caused by copper but not zinc (Section 4.3.2), and copper
is more toxic at higher pH levels while zinc is not (Section
4.3.3). Depending on how the toxicity of the sample
changes with these tests, one could eliminate one of
these metals from the list of suspect metal toxicants.
Results of this type of testing will be used to
develop evidence implicating the identified suspect metal.
These tests would be performed on a number of samples
over time to demonstrate the consistency of the cause of
toxicity. In addition, when a mixture of toxicants is present,
additions of EDTA or thiosulfate could be used to remove
the cationic metal toxicity after performing other Phase II
manipulations, e.g., C18 SPE.
4.3.1 EDTA Addition Test
Any reduction in effluent toxicity effected by the
addition of EDTA suggests that certain cationic metals
might be present in the effluent at toxic levels. Back-
ground information and discussion of the behavior of
EDTA and cationic metals can be,found in Phase I (EPA,
1991 A; EPA, 1992).
ideally, the amount of E/DTA added would be just
enough to chelate the toxicant(s) without causing EDTA
4-5
-------
toxteHy or substantially changing the water quality. For
either C. dubia or fathead minnows, we have found it
useful to add two different EDTA concentrations to two
separate effluent tests (with dilutions). Controls without
EDTA must be included. The EDTA stock solution is
added after the effluent dilutions are prepared so that the
EDTA concentration is the same at each effluent dilution
(see Phase I, EPA, 1991A; EPA, 1992).
In Phase II, conducting simultaneous EDTA addi-
tion tests on effluent and the suspect metal in matching
test water can provide evidence supporting the suspect
metal as the toxicant if the results of these two tests are
similar. If the metal is chelated by EDTA in the dilution
water test but not in the effluent test then either there is a
strong matrix effect from the effluent or it is the incorrect
suspect metal. It is important to use the same pH in both
tests in case there is any pH effect on the metal's toxicity.
In addition to removing toxicity caused by metals,
EDTA reduces the acute toxicity of some cationic surfac-
tants. This reduction of toxicity might also occur in chroni-
cally toxic effluents, and the toxicity reduced by EDTA
should not be assumed to be due only to cationic metals.
4.3.2 Sodium Thlosulfate Addition Test
The acute Phase I oxidant reduction test (EPA,
1991A) or the chronic sodium thiosulfate addition test
(EPA, 1992) is used to determine to what extent constitu-
ents reduced by the addition of sodium thiosulfate
(NajS.O,) are responsible for the effluent toxicity. Al-
though the use of the sodium thiosulfate test was de-
signed to determine if oxidative compounds (such as
chlorine) were responsible for effluent toxicity, experience
has also shown that thiosulfate can also form a stable
non-toxic complex with some metals. Since the complexing
ability of thiosulfate is more metal specific than EDTA, this
reagent can be used to determine if a specific metal is
responsible for the effluent toxicity. Recent work by Mount
Tablo 4-3. Metal LCSOs with Respect to Test pH1
LC50 (ug/1)
Metal
Species
pH6.2
pH7.2
pH8.2
Zn
Ni
Pb
Cu
Cd
O. dubia
P. promelas
C. dub/a
P. promelas
C. dubia
P. promelas
C. dubia
P. promelas
C. dubia
P. promelas
>530
830
>200
>4000
280
810
10
15
563
54
360
333
137
3360
>2700
>5400
28
44
350
74
95
502
13
3080
>2700
>540
201
>200
121
<5
'LC50 values were determined at 48-h for C. dubia and 96-h for
P. promelas. Data taken from Schubauer-Berigan etal., 1993A.
(1991) has shown that in acute toxicity tests with C. dubia
in moderately hard water that Cu2*, Cd2+, Hg2+, Ag+, and
Se6* can be complexed using sodium thiosulfate (see ,
EPA, 1991A for more details). This complexing ability
might not be applicable to chronic toxicity. For example, in
a C. dubia 7-d test with copper, the toxicity was not
reduced with sodium thiosulfate addition but was reduced
with EDTA addition.
If the addition of sodium thiosulfate does not
reduce the effluent toxicity thought to be related to metals,
the use of SO (EPA, 1991A) additions followed by the
addition of sodium thiosulfate is recommended. In some
situations, the thiosulfate concentration may be reduced
by non-toxic oxidants and thus, not be available for
complexing the toxic metal. The addition of SO should
preferentially reduce these non-toxic oxidants which will
allow the now available thiosulfate to complex the toxic
metals. Depending upon the complexation ability of so-
dium thiosulfate for a specific metal, it might or might not
complex the toxic metal. If the suspected metal toxicant
can be complexed (e.g., cadmium, copper, selenium (as
selenate), mercury; see EPA, 1991 A; EPA, 1992), then a
reduction in sample toxicity should occur with the addition
of sodium thiosulfate. If the suspected metal cannot be
complexed (e.g., zinc, lead, manganese, and nickel), then
no reduction in sample toxicity should occur with the .
addition of the sodium thiosulfate.
As with the EDTA addition test, sodium thiosul-
fate additions should be conducted concurrently on the
effluent and on dilution water spiked with the suspect
metal toxicant. Care must be used to conduct these tests
at similar pH levels. When toxicity test results are consis-
tent with the expected behavior, strong evidence relating
the suspected metal toxicant to the cause of the effluent
toxicity has been obtained. These results in conjunction
with the ion exchange test, analytical measurements for
toxic metals, and the EDTA addition test provide evidence
sufficient for one to proceed to toxicant confirmation,
Phase III, of the TIE.
Both sodium thiosulfate and EDTA can reduce
the toxicity of some metals and this information can be
helpful in identifying the toxicant. However to date, this
effect of thiosulfate/metal complexation has not been
demonstrated for chronic toxicity. Knowing which metals
are bound by both sodium thiosulfate and EDTA and
which metals are complexed with only one or the, other
additive can be very helpful in narrowing down the pos-
sible toxicant.
4.3.3 Metal Toxicity Changes with pH
In Phase I, the graduated pH test is performed to
evaluate the presence of compounds whose toxicity var-
ies with pH. For ammonia, toxicity is greatest at pH 8.5
and least at pH 6.5 for some species. Therefore, as
suggested in the first Phase II document, that for samples
in which toxicity is enhanced at elevated pH, the identifi-
cation effort should focus on ammonia. However, some
4-6
-------
effluent and sediment pore water TIEs have indicated that
some toxicity caused by metals can be affected by pH
within the range of pH 6 to 9 (Schubauer-Berigan et al.,
1993A). Some metals, notably zinc, nickel and cadmium,
exhibit greater toxicity at elevated pH, which could con-
fuse their characterization with that of ammonia (Table 4-
3), while copper and lead show elevated toxicity at pH
6.2. These pH-dependent toxicities can be used as a tool
for the identification (and confirmation) of toxicity caused
by these metals. For example, toxicity to C. dubia in a
sediment pore water sample was completely removed by
additions of EDTA. The sample also exhibited greater
toxicity at pH 6.5 than at 8.5, and metal concentrations
indicated that only copper was present at toxicologically
significant concentrations. The pH dependent toxicity of
the sample along with the EDTA addition results and
metal analysis supported the identification of copper as
the toxicant.
4.3.4 Ion-Exchange Test
Ion-exchange resins have been used in TIEs for
generating supporting evidence for identifying the cause
of toxicity in effluents (Doi and Grothe, 1989; Phase II
zeolite test). For cation exchange resins, removal of toxic
cations such as NH4+, Cd2+, and Pb2+ from the effluent
occurs with the corresponding release of cations (i.e.,
counter ions) such as H+ and Na+ into solution. Similarly,
for anion exchange resins, removal of toxic anions such
as Cr2O/- and AsO/- from the effluent occurs with the
corresponding release of anions such as OH- and CI" into
solution. For both cation and anion exchange resins,
charge neutrality exists between the resin and aqueous
phase and therefore, if the resins remove 5 u.moles of
Cd2+ from solution, 10 u,moles of H+ would be released
into solution. The exchange process is concentration
dependent and is reversible. Cations removed from the
solution may then be recovered from the exchange resin
by passing an acidic solution over the resins (e.g., 1 N
HCI for analysis of metals).
We have had limited experience with ion ex-
change resins but the following general guidance can be
provided. First, ion exchange resins are not chemical
specific but rather remove a wide range of cations or
anions, metallic and non-metallic. We have observed that
anion exchange resins can remove cations (e.g., Zn2+)
from solution quite efficiently. The reasoning that only
cationic materials are removed by cation ion exchange
resins is not always reliable. Experimental verification of
which materials were removed by the resin will be neces-
sary on a case-by-case basis. Second, wide changes in
the pH of the post-column effluent can occur depending
upon the type of cation or anion released by the resin.
These changes in pH will cause problems in interpreting
toxicity tests if the pH is not adjusted prior to the toxicity
test. Third, many of the ion exchange resins are based
upon a styrene or acrylic divinylbenzene backbone and
this material can remove other types of toxicants such as
non-polar organics. Consequently, because of its non-
specificity the removal of toxicity by an ion exchange
column should not be used as the only piece of evidence
to implicate a metal as the toxicant.
Resins under evaluation and/or those which have
been used include IRA-35, IRA-68, IRA-94, IRA-900,
IRC-718, and GT-73 (Rohm and Haas, Philadelphia PA)
and aquarium zeolite (Argent Chemical Laboratories,
Redmond WA). The key to obtaining useable data from
an ion exchange test is to obtain non-toxic blanks. Since
numerous ion exchange resins exist, guidance for prepar-
ing all resins for TIE work cannot be provided. A variety of
procedures have been used in our laboratory to condition
the columns and to obtain non-toxic blanks.
Effluent volumes ranging from 1,000 to 10,000 ml
have been used, and the volume is dependent on the
hardness of the dilution water, bed volume of the column,
strength and type of the ion change resin, which ions
were being exchanged, the toxicity of the effluent, and the
species being tested. For example, for acutely toxic efflu-
ents, glass chromatography columns (11 mm i.d.) are
packed with about 10 cm of resin and the solutions are
pumped up through the column at a flow rate of 4 to 5 ml/
min. First, a small volume of high purity water (e.g., 200
ml) is passed through the column, and discarded. Next,
the dilution water (volumes are variable, i.e., 1,000-
5,000 ml) is passed through the column until the pH of the
post-column dilution water is above 7.0.
Following this procedure, the necessary volume
of dilution water to use for toxicity testing is passed
through the column and collected. The type of dilution
water to use is effluent specific and in general, should be
the same as the dilution water used in the toxicity test for
the effluent. The pH of the post-column dilution water
should be monitored and the pH adjusted to return the
water solution to its original pH. Toxicity tests are then
performed on the post-column dilution water sample (col-
umn blank). After obtaining non-toxic blanks for a particu-
lar batch of resin, the conditioning process can be used
on other aliquots of the resin with a similar procedure;
however, column toxicity blanks must always be tested.
To identify acute toxicity, we generally begin by
using 200 ml of effluent (filtered or unfiltered) and collect
the post-column effluent. The pH of the post-column
effluent is checked and if necessary the pH is adjusted to
that of the baseline test, and tested for toxicity. For
chronic toxicity, the volume of effluent needed for the
toxieity test will dictate the amount of resin and the size of
the column. When evaluating a new resin, use propor-
tions of water, effluent, and resin, similar to those de-
scribed above for acutely toxic effluents. New aliquots of
resins should be prepared and used for each ion ex-
change test. By doing so, artifactual toxicity problems
4-7
-------
from other effluents and sample manipulations can be
avoided.
We have had limited success in the elution of the
Ion exchange resin to recover the exchanged toxicant(s);
therefore, we cannot provide specific guidance. In theory,
catfons and anions can be eluted from ion exchange
resins using a strong acidic (HCI) or basic (NaOH) solu-
tion. Performing successful toxicity testing on these solu-
tions is extremely difficult because of artifactual toxicity
problems.
When toxicity is removed by the ion exchange
test, useful information about the toxicant(s) may be
obtained. However, as discussed above, the removal of a
toxicant by the column may not be as straightforward as
first perceived. The use of other manipulations and ana-
lytical measurements on the pre- and post-column efflu-
ents will be required to establish the significance of the
results of the manipulation.
When toxicity is not removed by the ion exchange
test and non-toxic blanks are obtained, the conclusion
that the toxicant is not a cation or anion can be made.
However, the slight possibility exists that the resin may
not be able to exchange the toxicant because of steric
and size considerations.
4-8
-------
Section 5
Chlorine
5.1 General Overview
One of the first analytical measurements recom-
mended in the Phase I documents (EPA, 1991A; EPA,
1992) upon arrival at the laboratory is for total residual
chlorine (TRC) in each effluent sample. Chlorine is a
commonly used biocide and oxidant and is frequently
found at acutely toxic concentrations in municipal efflu-
ents (EPA, 1985C). Sublethal chronic toxicity from chlo-
rine in effluent samples is not as likely to occur due to the
degradation of chlorine (see below) with holding of the
sample. Chlorine is unstable in aqueous solutions and
decomposition is more rapid in solutions when chlorine is
present at low concentrations. From the TRC measure-
ment and the Phase I tests (sodium thiosulfate addition
and aeration tests), further steps to identify the effects
that might be due to chlorine can be taken. Oxidants
other than chlorine occur in effluents and the removal of
toxicity by the addition of sodium thiosulfate does not
prove that chlorine was the cause of effluent sample
toxicity.
Molecular chlorine or hypochlorite dissociates
into free aqueous chlorine, hypochlorous acid, and hypo-
chlorite ion when added to effluents. Chlorine can also
combine with ammonia to form chloramines, i.e., mono-,
di-, and tri-chloramines and with organic compounds,
especially organic nitrogen (APHA, 1992). The measured
total residual chlorine (TRC) of an effluent is the concen-
tration of free and combined forms (mentioned above)
added together. The portion of the TRC associated with
an individual form is matrix dependent. Chlorinated in-
dustrial and wastewater effluents normally contain only
the combined form of chlorine (APHA, 1992).
These various forms of combined chlorine may
have different effect concentrations for toxicity, and the
toxicities of these individual forms are not all known for
acute or chronic toxicity to C. dubia or fathead minnows.
However, while the TRC level in the effluent samples
may be the same, the concentration of the various forms
may be different because of the matrix inherent to the
effluent. This matrix of TRC may also be variable from
sample to sample for the same discharger.
Another complication is that Current analytical
methods for measuring TRC are not chlorine specific:
Other oxidizing compounds, e.g., bromine, iodine, hydro-
gen peroxide, ozone, and manganese, will be quantified
as chlorine by the analytical methods for measuring TRC
and may provide the analyst with a false positive for
chlorine.
5.2 Tracking Toxicity and TRC Levels
Several methods are available for measuring to-
tal TRC (EPA methods 330.1, 330.2, 330.3, 330.4, and
330.5 (EPA, 1983)). Measurements of TRC in the effluent
upon arrival of the sample at the laboratory should always
be made. If TRC is not detected, chlorine should not be
considered a suspect toxicant since the analytical meth-
ods do not yield false negatives.
For acutely toxic effluents, grab samples both
before and after the chlorination process should be col-
lected simultaneously (i.e., within minutes of each other).
Upon arrival of these samples at the laboratory, a baseline
toxicity test should be initiated and at pre-determined
intervals after day 1 (e.g., day 2, day 3, day 5, day 8) to
evaluate whether the toxicity is degrading. TRC determi-
nations should be performed in conjunction with each
toxicity test.
Generally the TRC in most effluent samples stored
at 4°C degrades in 2 to 5 d after collection. Therefore, if
residual chlorine is a toxicant the toxicity of the post-
chlorination sample should decrease as TRC levels de-
crease, and pre- and post-chlorination samples should
have the same toxicity after the decay of TRC.
The toxicity of chlorine in an effluent sample will
be dependent on the matrix of the effluent and the spe-
cies tested. If chlorine toxicity data does not exist for the
species being used, it will be necessary to measure the
LC50 or IC25 of chlorine using the TIE organisms and
dilution water. Using those LC50 and/or IC25 values, the
comparison of TUs of the effluent to the TUs of residual
chlorine is useful to evaluate the effects of the TRC.
When the TU comparison data and pre- and post-chlori-
nation toxicity data indicate TRC as a suspect toxicant
Phase III procedures should be initiated.
With the measurable levels of TRC at sample
collection, the loss of toxicity with the corresponding
decreasing levels of TRC, and the pre- and post-chlorina-
5-1
-------
tion samples exhibiting similar toxicity with the decrease
!n TRC, Phase 111 confirmation should begin (EPA, 1989B;
EPA, 1993A). However, these steps will not insure that
the toxicant is chlorine since other oxidants will be de-
tected by the TRC measurement techniques.
5-2
-------
Section 6
Identifying Toxicants Removed by Filtration
6.1 General Overview
If the results of Phase I.tests indicate that the
filtration manipulation removed or reduced toxicity, the
investigator should carefully compare these results to
those of the other manipulations before trying to identify
the toxicants that might be on the filter. We have ob-
served that metals, non-polar organic compounds and
volatile compounds can all be removed under certain
filtering conditions, but these observations have been
dependent on the individual effluent or the sediment pore
water samples. Other Phase I manipulations (e.g., EDTA,
C18 SPE extraction) can lead to subsequent Phase II
identification steps. However, for toxicity reductions ef-
fected by filtration, more intermediate steps of Phase I
type manipulations must be done before analytical proce-
dures are used to identify the toxicant(s). In addition,
some other manipulations may. provide specific informa-
tion regarding the identity of toxicants that may have been
removed by filtration; these include additions of PBO
(Section 2.5.1), the graduated pH test (Phase I tests for
determining toxicity caused by ammonia, metals and
ionizable organic compounds), and the sodium thiosulfate
test (Phase I test for detecting toxicity caused by volatile
oxidants such as chlorine or metals). If one or more of
these manipulations removes toxicity, then identification
work should proceed as described in the previous sec-
tions to identify the cause of toxicity.
It is important to consider that all toxicity removed
by filtration may not be actually removed by the process
of filtering. For example, when the pH of the sample is
altered, the mechanism(s) for removal by filtration can
change. While ammonia is predominantly ionized at a
sample pH of 8.3, the ammonia would not tend to be
removed through volatilization if a vacuum was applied
for filtration purposes. Yet by adjusting the sample pH to
11, the ionized ammonia concentration decreases to 1.7%
at 25°C. When a sample is adjusted to a pH of 11,
volatilization of a toxicologically significant amount of the
un-ionized ammonia could occur and the toxicity results
would indicate that filtration removed toxicity. Also, changes
in speciation at elevated pH render many metals in-
soluble, which could result in their removal by the filter at
pH 11 (Schubauer-Berigan etal., 1993B).
If the toxicant is thought to be a non-polar organic
toxicant, and filtration partially removes toxicity, it may be
useful and save toxicity testing time to eliminate the
filtration step altogether before applying the sample to the
C SPE column (discussion in Section 2.2.2 and Section
2.3.2).
6.2 Filter Extraction
When filtration has been the only manipulation to
affect the toxicity, then extraction of the filters and track-
ing the toxicity of the extracts should be attempted. In
addition, the use of other types of filters should be evalu-
ated (i.e., nylon, teflon, and polycarbonate) to see if
toxicity removal is a function of the filter type. In using the
extraction procedures, the idea is to separate the toxic
compounds associated with the filter by extracting them
into a solvent. Next, efforts are made to concentrate the
toxic compounds in the filter extract and test them at a
concentration that can be related to the original sample
and evaluate the efficiency of the extraction. Identifying
the filter-removable contaminants without additional infor-
mation can be difficult because of the lack of specificity of
the filtration process. But once a suspect candidate has
been discovered, then measurements can be made to
determine whether a lexicologically significant concentra-
tion of the suspect toxicant(s) had indeed been removed
by filtration. If this is the case, then it may not be neces-
sary to consider further extractions of the filters. If, how-
ever, the concentrations of the suspect toxicant(s) are not
decreased after filtration then it may be useful to attempt
additional identifications by solvent extraction of the filters
as described below.
One technique we have used with filterable toxic-
ity is to extract the filters with either polar or non-polar
solvents. To remove toxicants from the filter we have
used either organic solvents (methanol, methylene chlo-
ride) or pH 3 high purity water as the extraction solvent.
The solvent is then toxicity tested to track toxicity (methyl-
ene chloride must first be exchanged into methanol),
additional Phase I tests are performed to characterize the
filter extract, and then chemically analyzed using Phase II
procedures. It is important to remove all of the methylene
chloride before toxicity testing a filter extract and these
procedures are described in detail in Section 2.6.2. To
date, methanol has been used to extract toxicity from
filters used with effluent samples and methylene chloride/
methanol solutions have been used to extract filters from
6-1
-------
sediment pore water. The experiences described below
are based on acute toxicity experiments, and efforts to
recover filterable toxicity for chronically toxic effluent have
not yet been needed.
To Isolate a toxicant removed through filtration,
several filters can be combined and extracted simulta-
neously if necessary. The volume of sample passed
through the filters is important for calculating concentra-
tion factors, and should be recorded. The filtrate should
also be reserved for toxicity comparisons and analytical
testing. Sufficient sample should Be passed through the
filter to allow for both toxicity testing and chemical analy-
sis on both the filtered sample and the filter-extract solu-
tion (generally >200 ml). Carefully move the filters to a
glass (acid leached) or plastic beaker, then soak the
filters (1-5) in 20 ml of solvent for 1 h. Cool water sonica-
tfon is optional to attempt to recover particle-associated
compounds. Carefully remove the filters and save (store
at 4°C) in case additional extractions are necessary. If pH
3 high purity water is used as the extraction solvent, the
extract should be readjusted to the initial pH of the
sample, then toxicity tested. If methanol is used, it is
evaporated to ~2ml under a stream of nitrogen. Be
careful to rinse the sides of the containers with methanol
to ensure complete solubilization of organic compounds.
This methanol solution can then be toxicity tested using
SPE fraction testing procedures (Section 2.2.8). Alterna-
tively, if a methylene chloride/methanol solvent is used,
the solvent should first be exchanged into pure methanol
(Section 2.6.2), then treated as the methanol extract
described above. The concentration of the solvent extract
will depend on the volume originally passed through the
filters, which depends on the desired high test concentra-
tion, and the volume of extract and filtered sample re-
quired for analytical purposes. Blank filters (through which
has been passed a volume of dilution water) should be
extracted and tested identically to the sample filters to
ensure that the solvents do not introduce artifactual toxic-
ity.
For any of the extraction techniques, the solu-
tions should be tested at the same time as the baseline
test (unfiltered) and filtered sample test to compare the
toxicity recovered by the filter extraction with that re-
moved from the sample.
Another option for toxicants removed by filtration
is to try other techniques to remove the toxicants which
avoid filtration. For example, sediment pore water samples
have been centrifuged at relatively high speeds (10,000-
20,000 g) for 30 min prior to passing the sample over the
C18SPE column and filtration could thus be eliminated.
Filter extractions (EPA, 1991 A) have been used
in several sediment TIE studies, with procedures sug-
gested for both non-polar organics and metals (Schubauer-
Berigan et al., 1990; Schubauer-Berigan and Ankley,
1991). In some effluent and pore water samples, toxicity
thought to be caused by non-polar organic compounds
(e.g., PAHs and polymers) has also been removed by
filtration. These compounds may be associated with par-
ticulate material, and be physically filtered from the sample,
or removed by association with oil and grease that sorbs
to the filter.
In some cases, binding of metals or organic
compounds to inorganic and organic ligands in effluents
or sediment pore waters will reduce their bioavailability
and when toxicity testing filter extracts, it is always a
concern that the matrix of the sample has been removed
and that chemicals might become available when they
were not in the original sample. If this were to happen, the
extracts might be more toxic than expected and chemi-
cals might be added to the suspect toxicant list errone-
ously. This kind of mistake should be caught by obtaining
a good toxicity value for the suspect in an appropriate
matrix (more detailed discussion in Section 2.6.1), There-
fore, additional confirmation steps might be needed to
eliminate the false suspects.
6-2
-------
Section 7
References
Amato, J.R., D.i. Mount, E.J. Durban, M.T. Lukasewycz,
G.T. Ankley, and E.D. Robert. 1992. An Example of
the Identification of Diazinon as a Primary Toxicant
in an Effluent. Environ. Toxicol. Chem. 11:209-216
Ankley, G.T., M.T. Lukasewycz, G.S. Peterson and
D.A. Jenson. 1990A. Behavior of Surfactants In
Toxicity Identification Evaluations. Chemosphere.
21:3-12.
Ankley, G.T., A. Katko, and J.W. Arthur. 1990B. Identi-
fication of Ammonia as an Important Sediment-
Associated Toxicant in the Lower Fox River and
Green Bay, Wisconsin. Environ. Toxicol. Chem.
9:313-322.
Ankley, G.T., J.R. Dierkes, D.A. Jensen and G.S.
Peterson. 1991. Piperonyl Butoxide as a Tool in
Aquatic Toxicological Research with Orgahophos-
phate Insecticides. Ecotoxicol. Environ. Safety.
21:266-274.
Ankley, G.T. and L.B. Burkhard. 1992. Identification of
Surfactants as Toxicants in a Primary Effluent.
Environ. Toxicol. Chem. 11:1235-1248.
APHA, 1989. Standard Methods for the Examination of
Water and Wastewater, 17th Edition. American Pub-
lic Health Association, Washington, D.C.
APHA, 1992. Standard Methods for the Examination of
Water and Wastewater, 18th Edition. American Pub-
lic Health Association, Washington, D.C.
AQUIRE, 1992. Aquatic Toxicity Information Retrieval
Database, Technical Support Document. Prepared
for Environmental Research Laboratory, Duluth,
MN.
Beigger, T. 1990. Interactions of Ammonia and Chlorine
on Fathead Minnows and Japanese Medaka. The-
sis for University of South-Western Louisiana,
Lafayette, LA.
Burkhard, L.P., E.J. Durhan, and M.T. Lukasewycz. 1991.
Identification of Non-Polar Toxicants in Effluents us-
ing Toxicity Based Fractionation with Gas Chroma-
tography/Mass Spectrometry. Anal. Chem.
63:277-283.
Burkhard, L.P., J.J. Jenson. 1993. Identification of Ammo-
nia, Chlorine, and Diazinon as Toxicants in a Munici-
pal Effluent. Arch. Environ. Contam. Toxicol., In Press.
DeGraeve, G.M., J.D. Cooney, T.L. Pollock, N.G.
Reichenbach, J.H. Dean, M.D. Marcus, and D.O.
Mclntyre. 1988. Fathead Minnow 7-day test: Round
Robin Study. Intra- and Interlaboratory Study to De-
termine the Reproducibility of the Seven-day Fathead
Minnow Larval Survival and Growth Test. Battelle,
Columbus, OH. Available from: American Petroleum
Institute, 1220 L Street, NW, Washington, DC 20005,
Report Number 4486.
Doi, J. and D.R. Grothe. 1989. Use of Fractionation/
Chemical Analysis Schemes for Plant Effluent Toxic-
ity Evaluation. In: G.W. Suter II (Editor), Aquatic
Toxicology and Environmental Fate. 11:123 ASTM
STP 1007, American Society for Testing and Materi-
als, Philadelphia, PA.
Durhan, E.J., M.T. Lukasewycz, and J.R. Amato. 1990.
Extraction and Concentration of Nonpolar Organic
Toxicants From Effluents Using Solid Phase Extrac-
tion. Environ. Toxicol. Chem. 9:463-466.
Durhan, E.J., M.T. Lukasewycz, and S. Baker. 1993.
Alternatives to Methanol-Water Elution of Solid-Phase
Extraction Columns for the Fractionation of High Log
Kow Organic Compounds in Aqueous Environmental
Samples. J. Chromatogr. 629:67-74.
EPA. 1979. Aqueous Ammonia Equilibrium - Tabulation
of Percent Un-ionized Ammonia. EPA/600/3-79/091.
Environmental Research Laboratory, Duluth, MN.
7-1
-------
EPA. 1980. Ambient Water Quality Criteria for Silver.
EPA-440/5-80-071. Environmental Research Labo-
ratory, Duluth, MN.
EPA. 1982. Methods for Organic Chemical Analysis of
Municipal and Industrial Wastewater. EPA-600/4-82/
057. Environmental Monitoring and Support Labora-
tory, Cincinnati, OH.
EPA. 1983. Methods for Chemical Analysis of Water and
Wastes. EPA-600/4-79-020. Environmental Monitor-
ing and Support Laboratory, Cincinnati, OH.
EPA. 1985A. Ambient Water Quality Criteria for Ammo-
nia. EPA-440/5-85-001. Environmental Research
Laboratory, Duluth, MN.
EPA. 1985B. Ambient Water Quality Criteria for Cad-
mium. EPA-440/5-84-032. Environmental Research
Laboratory, Duluth, MN.
EPA. 1985C. Ambient Water Quality Criteria for Chlorine.
EPA-440/5-84-030. Environmental Research Labora-
tory, Duluth, MN.
EPA. 1985D. Ambient Water Quality Criteria for Chro-
mium. EPA-440/5-84-029. Environmental Research
Laboratory, Duluth, MN.
EPA. 1985E. Ambient Water Quality Criteria for Copper.
EPA-440/5-84-031. Environmental Research Labora-
tory, Duluth, MN.
EPA. 1985F. Ambient Water Quality Criteria for Lead.
EPA-440/5-84-027. Environmental Research Labora-
tory, Duluth, MN.
EPA. 1986. Ambient Water Quality Criteria for Nickel.
EPA-440/5-86-004. Environmental Research Labora-
tory, Duluth, MN.
EPA. 1987. Ambient Water Quality Criteria for Zinc. EPA-
440/5-87-003. Environmental Research Laboratory,
Duluth, MN.
EPA. 1988A. Methods for Aquatic Toxicity Identification
Evaluations: Phase I Toxicity Characterization Proce-
dures. EPA/600/3-88/034. Environmental Research
Laboratory, Duluth, MN.
EPA. 1988B. Ambient Water Quality Criteria for Alumi-
num. EPA-440/5-86-008. Environmental Research
Laboratory, Duluth, MN.
EPA. 1989A. Methods for Aquatic Toxicity Identification
Evaluations: Phase II Toxicity Identification Proce-
dures. EPA/600/3-88/035. Environmental Research
Laboratory, Duluth, MN.
EPA. 1989B. Methods for Aquatic Toxicity Identification
Evaluations: Phase III Toxicity Confirmation Proce-
dures. EPA-600/3-88/036. Environmental Research
Laboratory, Duluth, MN.
EPA. 1989C. Short-Term Methods for Estimating the
Chronic Toxicity of Effluents and Receiving Waters to
Freshwater Organisms. Second Edition. EPA/600/4-
89/001 and Supplement EPA/600/4-89/001 A. Envi-
ronmental Monitoring and Support Laboratory,
Cincinnati, OH.
EPA. 1991 A. Methods for Aquatic Toxicity Identification
Evaluations: Phase I Toxicity Characterization Proce-
dures. Second Edition. EPA/600/6-91/003. Environ-
mental Research Laboratory, Duluth, MN.
EPA. 1991B. Sediment Toxicity Identification Evaluation:
Phase I (Characterization), Phase II (Identification),
Phase III (Confirmation) Modifications of Effluent Pro-
cedures. EPA-600/6-91/007. Environmental Research
Laboratory, Duluth, MN.
EPA. 1991C. Methods for Measuring the Acute Toxicity of
Effluents to Freshwater and Marine Organisms. Fourth
Edition. EPA-600/4-90/027. Environmental Monitor-
ing and Support Laboratory, Cincinnati, OH.
EPA. 1991D. Methods for the Determination of Metals in
Environmental Samples. EPA-600/4-91/010. Environ-
mental Monitoring and Support Laboratory, Cincin-
nati, OH.
EPA. 1992. Toxicity Identification Evaluations: Character-
ization of Chronically Toxic Effluents, Phase I. EPA-
600/6-91/005F. Environmental Research Laboratory,
Duluth, MN.
EPA. 1993A. Methods for Aquatic Toxicity Identification
Evaluations: Phase III Toxicity Confirmation Proce-
dures for Samples Exhibiting Acute and Chronic Tox-
icity. EPA-600/R-92/081. Environmental Research
Laboratory, Duluth, MN.
EPA. 1993B. Short-term Methods for Estimating the
Chronic Toxicity of Effluents and Receiving Waters to
Freshwater Organisms. Third Edition. EPA-600/4-91/
002. Environmental Monitoring and Support Labora-
tory, Cincinnati, OH.
EPA. 1993C. Short-term Methods for Estimating the
Chronic Toxicity of Effluents and Receiving Waters to
Marine and Estuarine Organisms. Second Edition,
EPA-600/4-91/021. Environmental Monitoring and
Support Laboratory, Cincinnati, OH.
Fickiin, W.H.. 1983. Separation of Arsenic (III) and Ar-
senic (V) in Ground Waters by Ion-exchange. Talanta.
30:371-373.
7-2
-------
Ferguson, W.J., K.I. Braunschweiger, W.R.
, B,raunschweiger, J.R. Smith, J.J. McCormick, C.C.
,>.'.-. Wasmann, N.P. Jarvis, D.H. Bell, and N.E. Good.
1980. Hydrogen Ion Buffers for Biological Research.
Anal. Biochem. 104:300-310.
Flaschka, H.A. and A.J. Barnard, Jr. (Eds.). 1967. Che-
....... lates in Analytical Chemistry. Marcel Dekker, Inc.,
.•', New York, NY. 418 p:
Giger, W., C. Stephanou, and C. Schaffner. 1981. Persis-
tent Organic Chemicals in Sewage Effluents: I. Identi-
, ficationsofNonylphenols and Nonlyphenol Ethoxylates
, by Glass Capillary Gas Chromatography/Mass Spec-
". trorrietry. Chemosphere. 10:1253-1263.
Giles, M.A. and R. Danell. 1983. Water Dechlorination by
t. Activated Carbon, Ultraviolet Radiation and Sodium
'Sulphite. Water Res. 17:667-676.
Jpp, K,M,, T.Z. Kendall, A.M. Askew, and R.B. Foster.
1991. Use of Fractionation Procedures and Extensive
Chemical Analysis for Toxicity Identification of a
Chemical Plant Effluent. Environ. Toxicol. Chem.
.-;, .: iO'981-990. .','..
Lukasewycz, M.T. and EJ. Durhan. 1992. Strategies for
the Identification of Non-Polar Toxicants in Aqueous
...: Environmental. Samples using Toxicity-Based Frac-
tionation and Gas Chromatography/Mass Spectrom-
etry. J._Chromatogr. 580:215-228.
Mount, D.R. 1991. A Toxicity-Based Approach to Pollut-
•.,-, ant Identification. In:, Proceedings of the Thirtieth
Annual EPA Conference on Analysis of Pollutants in
v the:Environment, May 9 and 10, 1990. 21W-7005.
Office of Water, Washington, D.C.
Neilson, A.J., A.S; •Allard, S. Fischer, M. Malmberg, and
, T. Viktor. 1990, Incorporation of a Subacute Test with
...'.- Zebra Fish into a Hierarchical System for Evaluating
the Effect of Toxicants in the Aquatic Environment.
Ecotox. and Environ. Safety. 20:82-97.
..Norberg-King, T.J., E.J. Durhan, G.T. Ankley, and E.
Robert. 1991. Application of Toxicity Identification
i Evaluation Procedures to the Ambient Waters of the
, , Colusa Basin Drain. Environ. Toxicol. Chem. 10:891-
901.
- Nprberg-King, T.J. 1993. A Linear Interpolation Method to
Estimate Sublethal Toxicityr The Inhibition Concen-
tration (ICp) Approach. Release 2.0. National Effluent
Toxicity Assessment Center Technical Report 03-93.
Environmental Research Laboratory, Duluth, MN.
Nriagu, J.O., G. Lawson, H.K.T. Wong, and J.M. Azcue.
' v 1993. A Protocol for Minimizing Contamination in the
Analysis of Trace Metals in Great Lakes Waters. J.
Great Lakes Res. 19:175-182.
Oyamada, N. and M. Ishizaki. 1986. Fractional Determi-
nation of Dissolved Selenium Compounds of
. Trimethylselenonium Ion, Selenium (IV) and Selium
(VI) in Environmental Water Samples. Analyt. Sci.
2:365-369.
Patterson, C.C. and D.M. Settle. 1976. The Reduction of
Orders of Magnitude Errors in Lead Analysis of Bio-
logical Materials and Natural Waters by Evaluating
and Controlling the Extent and Sources of Industrial
Lead Contamination Introduced During Sample Col-
lection, Handling, and Analysis. In: Accuracy in Trace
Analysis: Sampling, Sample Handling, Analysis. P.O.
LaFleur (Ed.). National Bureau of Standards Spec.
Publ. 422. UvS. Government Printing Office, Wash-
ington D.C.
Skarheim, H.P. 1973. Tables of the Fraction of Ammonia
in the Undissociated Form for pH 6 to 9, Temperature
0 to 30'C, TDS 0-3000 mg/l, and Salinity 5-35 g/kg.
Sanitary Engineering Research Laboratory (SERL)
Report No. 73-5. University of California-Berkeley,
Berkeley, CA.
Schubauer-Berigan, M.K., G.T. Ankley, and J.R. Dierkes.
1990. Toxicity Identification Evaluations of Contami-
nated Sediments in the Buffalo River, NY and the
Saginaw River, Ml. Final Report to the Great Lakes
National Program Office ARCS Program. NETAC
Technical Report 20-90.
Schubauer-Berigan, M.K. and G.T. Ankley. 1991. The
Contribution of Ammonia, Metals, and Nonpolar Or-
ganic Compounds to the Toxicity of Sediment Intersti-
tial Water from an Illinois River Tributary. Environ.
Toxicol. Chem. 10:925-940.
Schubauer-Berigan, M.K., J.R. Dierkes, P.O. Monson,
and G.T. Ankley. 1993A. The pH-dependent toxicity
of Cd, Cu, Ni, Pb, and Zn to Ceriodaphnia dubia,
Pimephales promelas, Hyalella azteca, and
Lumbriculus variegatus. Environ. Toxicol. Chem.,
12:1261-1266.
Schubauer-Berigan, M.K., J.R. Amato, G.T. Ankley, S.E.
Baker, L.P. Burkhard, J.R. Dierkes, J.J. Jenson, M.T.
Lukasewycz, and T.J. Nprberg-King. 1993B. The Be-
havior and Identification of Toxic Metals in Complex
Mixtures: Examples from Effluent and Sediment Pore
Water Toxicity Identification Evaluations. Arch.
Environ. Contam. Toxicol. 24:298-306.
Smith, R.M. and A.E Martelh 1981. Critical Stability Con-
stants. Volume 4: Inorganic Complexes. Plenum Press,
NY. 87 p.
Stary, J. 1964. The Solvent Extraction of Metal Chelates.
Pergamon Press Ltd., Oxford, England. 240 p.
7-3
-------
Stumm, W. and JJ. Morgan. 1981. Aquatic Chemistry -
An Introduction Emphasizing Chemical Equilibria in
Natural Waters. John Wiley & Sons, Inc. New York,
NY. 583 p.
Walsh, G.E., and R.L Gamas. 1983. Determination of
Bfoactrvfty of Chemical Fractions of Liquid Wastes
using Freshwater and Saltwater Algae and Crusta-
ceans. Environ. Sci. Technol. 17:180-182.
Windholz, M., S. Budavari, R.F. Blumetti and E.S. Otterbein
(Eds.). 1983. The Merck Index: An Encyclopedia of
Chemicals, Drugs, and Biologicals. Tenth Edition.
Merck & Co., Inc., Rahway, NJ.
Zief, M., and J.W. Mitchell. 1976. Contamination Control
in Trace Element Analysis. Chemical Analysis Series,
Vol. 47. John Wiley & Sons, Inc., New York, NY.
7-4
-------
Appendix A
Effluent Volume Calculation Worksheets
A-1
-------
Table A-1. Effluent Volume Calculation Worksheets
SPE Fractlonatlon of the Effluent
1) Volume of effluent: ve ml
See Table 2-6 for initial suggestions
for the volume of effluent.
2) SPEfractionation:
Eluate volume from the SPE column: a _ ml
See Table 2-1 for approximate eluate volume
or measure volume.
Concentration factor for eluate: bt = ve •*- a b _ x
3) Testing organism and conditions:
Toxicity test volume/replicate: . c _ ml
C.dubia 10-15 ml/replicate (acute/chronic)
D.magna 10-1 5 ml/replicate (acute)
D.pulex 10-1 5 ml/replicate (acute)
P. promelas 1 0-200 ml/replicate (acute)
P. promelas 50-250 ml/replicate (chronic)
Number of replicates: d _
Initial sample + number of renewals e _
Highest test concentration: f x
Is the methanol concentration okay?1
ml eluate = f x c +• bn . ,
%methanol » ml eluate x 100 *c
4) Volume of eluate needed for toxicity testing:
If no dilutions: g^cxdxexf-i-b, g __ _ ml
If using 0.5 dilution factor:
h = 2xcxdxexf^-b1 h ml
If using dilutions by spiking each concentration2 directly:
J-ixcxdxe + b, j _ ml
Total volume of eluate used: k = g, h, or j k _ ml
5) Volume of eluate remaining after toxicity testing:
n>! = a - k m _ ml
'Acceptable levels of methanol for C. dubia and fathead minnows are <0.6% and 1%, respectively.
2An example of using four test concentrations, the number of dilutions may vary.
A-2
-------
Table A-1. Continued
Toxlclty Testing and GC/MS Analysis of the SPE Concentrate
1) SPE Concentration of the SPE fraction:
Eluate volume from the SPE column: a — ml
See Table 2-3 for approximate eluate volume
or measure volume.
Concentration factor for the eluate: b2 = fy x m1 + a b2 x
2) Testing organism and conditions: :
Toxicity test volume/replicate: c ml
C.dubia 10-15 ml/replicate (acute/chronic)
D.magna 10-25 ml/replicate (acute)
D.pulex 10-25 ml/replicate (acute)
P.promelas 10-200 ml/replicate (acute)
P. promelas 50-250 ml/replicate (chronic)
Number of replicates: d
Initial sample + number of renewals: e
Highest test concentration: * x
3) Volume of eluate needed for toxicity testing:
If no dilutions: g = cxdxexf-s-b2 9 ml
If using 0.5 dilution factor:
h = 2xcxdxexf*b2 n m
If using dilutions by spiking each concentration directly:
j = ixcxdxe*b2 I ml
Total volume of eluate used: k = g,h, or j k ml
4) Amount of eluate used for GC/MS analysis: I ml
5) Volume of eluate remaining after toxicity testing:
mz = a - k -1 m2 ml
A-3
-------
Table A-1. Continued
HPLC Fractlonatlon of the SPE Concentrate
1) HPLC Fractfonation of the SPE concentrate:
HPLC Fraction volume: a ml
See Sections 2.2.10 and 2.3.10 —
Concentration factor for the eluate: b, = b,xm,-s- a b Y
32 2 **g __ A
2) Testing organism and conditions:
Toxicity test volume/replicate: c m(
C.dubia 10-15 ml/replicate (acute/chronic) ~~
D.magna 10-25 ml/replicate (acute)
D.pulex 10-25 ml/replicate (acute)
P.promelas 10-200 ml/replicate (acute)
P.promelas 50-250 ml/replicate (chronic)
Number of replicates: $
Initial sample + number of renewals: e
Highest test concentration: f x
3) Volume of eluate needed for toxicity testing:
If no dilutions: g = cxdxexf-3-b3 g m(
If using 0.5 dilution factor:
h-2xcxdxex
If using dilutions by spiking each concentration directly:
h-2xcxdxexf-s-b3 h ml
j _ m,
Total volume of eluate used: k = g, h, or j k _ ml
5) Volume of eluate remaining after toxicity testing:
m3 _ ml
A-4
-------
Table A-1. Continued
Concentration of the HPLC Fraction for Toxicity Testing and GC/MS Analysis
1) SPE Concentration of the SPE fraction: .
Eluate volume from the SPE column: a
See Table 2-3 for approximate eluate volume -
or measure volume.
Concentration factor for the eluate: b4 = b3xm3 - a D4 *
2) Testing organism and conditions: ,
Toxicity test volume/replicate: c
C.dubia 10-15 ml/replicate (acute/chronic)
D. magna 10-25 ml/replicate (acute)
D.pulex 10-25 ml/replicate (acute)
P.promelas 10-200 ml/replicate (acute)
P. promelas 50-250 ml/replicate (chronic)
Number of replicates: o
Initial sample + number of renewals: ®
Highest test concentration: T
3) Volume of eluate needed for toxicity testing:
If no dilutions: g = cxdxexf *b4 9
If using 0.5 dilution factor: * .
h = 2xcxdxexf*b4 n ml
If using dilutions by spiking each concentration directly:
• 1 i xr/xriVe^b' J m'
Total volume of eluate used: k = g,h, or j k ml
4) Amount of eluate used for GC/MS analysis: ' ml
5) Volume of eluate remaining after toxicity testing:
m4 = a-k-l m4 ml
A-5
-------
------- |