United States
                     Environmental Protection
                     Agency
Environmental Monitoring and
Support Laboratory
Cincinnati OH 45268
                     Research and Development
EPA-600/S4-84-018 Apr. 1984
&EPA          Project  Summary
                     EPA  Method  Study  19,  Method
                     609  (Nitroaromatics and
                     Isophorone)
                    Glenn Kinzer, Ralph Riggin, Thomas Bishop, Michelle A. Birts, Cory C.
                    Howard, and Robert Iden
                      An Intel-laboratory study in which 18
                     laboratories participated was conducted
                     to  provide precision  and accuracy
                     statements for  the proposed EPA
                     Method 609 for measuring concentra-
                     tions of the Category 4 chemicals nitro-
                     benzene, isophorone, 2,4-dinitrotoluene,
                     and 2,6-dinitrotoluene in municipal and
                     industrial aqueous discharges. Method
                     609 involves solvent extraction of the
                     pollutants with  methylene chloride.
                     followed by Florisil cleanup and subse-
                     quent gas chromatographic analysis of
                     the  four subject compounds, using
                     flame ionization  and electron capture
                     detection techniques.
                      The  study  design was  based  on
                     Youden's plan for collaborative tests of
                     analytical methods. Three Youden pair
                     samples of the test compounds were
                     spiked into six types of test waters and
                     then analyzed. The test waters were
                     distilled water, tap water,  a surface
                     water, and three different industrial
                     wastewater effluents. The resulting data
                     were statistically  analyzed  using the
                     computer program entitled "Interlabora-
                     tory Method Validation Study" (IMVS).
                     Mean recoveries  of the subject com-
                     pounds were  in  the range of 49-75
                     percent. Overall precision was  in the
                     range of 26-60  percent and single-
                     analyst precision  was in the range of
                     13-45 percent. In general, mean recov-
                     eries, overall standard deviations (S),
                     and the single-analyst standard devia-
                     tions (SR) were directly proportional to
                     the true concentration levels. There were
                     no discernible differences due to water
                     types among mean recoveries, overall
                     precisions, and sinale-analvst precisions.
  This Project Summary was developed
by EPA's Environmental Monitoring
and Support Laboratory. Cincinnati,
OH, to announce key findings of the
research project that is fully documented
in a separate report of the same title (see
Project Report ordering information at
back).


Introduction
  EPA  first promulgated  guidelines
establishing test procedures for the
analysis of pollutants in 1973, following
the passage of the Federal Water Pollution
Control Act in 1972 by Congress. Pur-
suant to the amendment and publication
of  these guidelines, EPA entered into a
Settlement Agreement—the so-called
Consent Decree-requiring it to study and,
if  necessary to regulate,  65 "priority"
pollutants and classes of pollutants of
known or suspected toxicity to the biota.
Subsequently,  Congress passed the
Clean Water Act of 1977, mandating the
control of toxic pollutants discharged into
ambient waters by industry.
  In order to facilitate the implementation
of the Clean Water Act, EPA selected for
initial study 129 specific toxic pollutants,
113 organic and 16 inorganic. The
organic pollutants were divided into 12
categories based on their chemical
structure.  Analytical methods were
developed for these 12 categories by EPA
through in-house and contracted research.
The use of these analytical methods may
eventually be required for the monitoring
of  the 113 toxic pollutants in industrial
wastewater effluents, as specified by the
Clean Water Act of 1977.

-------
  Method 609 was developed  in  the
Battelle-Columbus Laboratories under a
contract with the Physical and Chemical
Methods Branch, Environmental Moni-
toring and Support Laboratory of EPA.
The interim Method 609 is described in the
Federal  Register,  Vol. 44,  No. 233,
December 3,  1979. The method requires
extraction of the pollutants with methylene
chloride,  Kurderna Danish Concentration,
Florisil  cleanup,  and subsequent  gas
chromatographic analysis of  the four
subject compounds. Nitrobenzene  and
isophorone are measured  using  flame
ionization detection and the 2,4- and 2,6-
dinitrotoluenes  are  measured using
electron  capture detection.

Procedure
  The study design was patterned after
Youden's plan for collaborative evaluation
of precision and accuracy for analytical
methods in which samples are analyzed
in pairs,  each member of a pair having a
slightly  different concentration of the
constituent  of interest. The  analyst  is
directed  to do a single  analysis  and to
report one value for each sample, as for a
normal routine sample. Samples of three
Youden pairs used in this study contained
low, medium, and high concentrations of
the Category 4 compounds which were
spiked into  each of six different water
types and then analyzed.
  Prior to the start of the interlaboratory
method study, participants were familiar-
ized with both the study design and the
analytical procedure by analyzing one trial
Youden pair  sample followed by attend-
ance at a prestudy conference. After
resolving method interpretation  and
analytical problems there, participating
laboratories were supplied  with the test
materials required by the  study design
and instructed to begin the analyses.
  The test waters were:
  a. Distilled water
  b. A municipal drinking water
  c. A  surface  water,  for example, a
     river, vulnerable  to synthetic chem-
     ical contamination
  d. Three industrial  wastewaters from
     industries that were potential candi-
     dates for priority pollutant  control
     under  the  National Pollutant  Dis-
     charge Elimination System (NPDES)
     program.

  Analyses were  conducted on distilled
water to evaluate the analyst's proficiency.
Municipal drinking and surface  waters
were included as test waters since these
water types are subject to contamination.
Hence,  it was considered  important to
obtain information about the performance
of Method 609 in such matrices, as well
as those found in industrial wastewater
effluents.
  Statistical analyses of the data were
performed  using the  IMVS  computer
program. The IMVS  program which was
developed at Battelle's Columbus Labora-
tories is a  revised  version of the EPA
COLST program. The program is designed
to output the raw data in tabular form and
compile summary statistics including:
  • Number of data points
  • True value
  • Mean  recovery
  • Accuracy as percent relative error
  • Overall standard deviation
  • Overall percent relative standard
     deviation
  • Single-analyst standard deviation
  • Single-analyst percent relative
     standard deviation.
  The overall standard deviations indicate
the  dispersion expected among  values
generated from multiple laboratories. This
represents the broad error in any mass of
data collected in a  collaborative  study.
The  single-analyst  standard deviations
indicate the dispersion expected among
replicate determinations within a  single
laboratory.

Results and Discussion
  The  data collected during this  inter-
laboratory study were  statistically ana-
lyzed in order to establish the relationship
between precision and the true concen-
trations, and between accuracy and the
true concentration.  Those relationships
are  summarized by the linear regression
equations presented in Table 1.
  The results of the regression analyses
indicate apparent linear  relationships
between (1) overall standard  deviation
and  mean  recovery; (2) single-analyst
precision and mean recovery; and(3) true
concentration and mean recovery.
  The percent recoveries of  isophorone
and the nitroaromatic compounds were
in the range 49 - 75 percent. The overall
relative standard deviation ranged from
25  to  60  percent  and single-analyst
relative standard deviations varied from
13 to 45 percent.
  Evaluation of the  data in  Table 1
indicates that  if a  laboratory  performs
well with  the  method using distilled
water,  it should also  be able to obtain
comparable results with surface waters
and industrial wastewater, provided that
the  level  of interferences  does  not
overwhelm the  components  of interest.
However, it is important to recognize that
about 15 percent of the laboratories were
unable to achieve good results. A  major
contributing factor was the experience of
the laboratory in applying the method
(i.e., better  data will  be obtained as a
laboratory gains experience  with the
method). Also, certain experimental steps
in the  method may contribute errors in
the data. For example, improperly activated
Florisil results in selectively low recovery
for nitrobenzene,  while concentration
problems would  tend to selectively
decrease the recoveries for both nitro-
benzene and isophorone.  But, it is not
obvious from the data set which factors
are major contributors to  the analytical
errors.

  All  of  the laboratories  were  able to
achieve  satisfactory  chromatographic
performance. Several laboratories indi-
cated problems  with  "bumping" of the
extract in the Kuderna-Danish evaporator,
especially at low extract volume, which
can result in relatively uniform losses for
all four analytes. This phenomenon may
explain some  of the low recoveries
observed, since  recoveries were similar
for all  four analytes. A few laboratories
reported the formation of air bubbles in
the Florisil cleanup columns, but  this did
not appear to affect the results and was
not a widespread problem.
  One of the questions of interest in this
study was whether water types affected
the precision and accuracy of the method.
An analysis of variance procedure
(ANOVA) was used to test for the effect of
water  type  on precision and accuracy.
Based on the results of this analysisthere
was no indication that water type had a
significant  affect  on the precision or
accuracy of the method.

Conclusions and
Recommendations
   Based on the results of the interlabora-
tory method study. Method 609 is a viable
analytical method for measuring concen-
trations of the Category 4 chemicals in
industrial wastewaters. Use of  Method
609 by  experienced analyst  should
enable  industries to meet the require-
ments of the NPDES  program  for dis-
charging the subject pollutants into the
environment.
   Certain laboratory operations in Method
609 have a primary impact on  method
performance. They are:
   a. Solvent extraction of the water
     sample.
   b. Activation  of the Florisil adsorbent
     and subsequent  preparation of the
     cleanup column.
   c. Concentration and exchange of the
     solvent extract and Florisil fractions

-------
Table 1.    Regression Equations for Accuracy and Precision of Method 609 by Compound and Water Type

Water Type                     Nitrobenzene       	  Isophorone	2.6-Dinitrotoluene
                                                                            2,4-dinitrotoluene
Distilled Water
Single-Analyst Precision
Overall Precision
Accuracy

Tap Water
Single Analyst Precision
Overall Precision
Accuracy

Surface Water
Single-Analyst Precision
Overall Precision
Accuracy

Wastewater (C-44)
Single-Analyst Precision
Overall Precision
Accuracy

Wastewater (C-45)
Single-Analyst Precision
Overall Precision
Accuracy

Wastewater (C-46)
Single-Analyst Precision
Overall Precision
Accuracy
(25-425 fjg/L]a

Sfl = 0.25X + 2.53
S =0.37X - 0.78
X = 0.06C + 2.00
SR = 0.30X + 1.63
S = O.45X 1.24
X = 0.71C - 0.34
SR = 0.37X - 2.20
S = 0.43X + 2.87
X = 0.69C - 0.99
SR = 0 18X + 2.10
S =0.26X + 0.54
X = 0.68C -0.38
SR = 0.20X + 3.10
S = 0.38X 0.58
X = 0.61C + 2.25
SR = 0.16X + 2.48
S = 0.26X + 1.88
X = 0.75C - 1.15
                                                        (25-
SR = 0.28X + 2.77
S = 0.46X + 0.31
X = 0.49C +2.93
SR =0.45X - 3.07
S =0.6OX - 3.27
X = 0.66C + 1.76
SR = 0.37X + 1.67
S = 0.46X -0.06
X = 0.59C -0.03
SR =0.27X + 2.06
S =0.33X +3.75
X=0.67C - 1 23
SR =0.22X + 7.15
S =0.52X - 1.16
X = 0.62C + 1.87
SR =0.26X + 6.86
S =0.54X + 1.64
X = 0.62C + 12.63
(1-60fjg/Lf

SR = 0.19X + 0.06
S = 0.36X - 0.00
X = 0.66C + 0.20
SR = 0.23X - 0.02
S = 0.37X - 0.06
X = 0.66C + 0.14
SR = 0.24X + 0.00
S = 0.34X + 0.03
X = 0.63C + 0.28
SR = 0.15X + 0.03
S = 0.25X + 0.01
X = 0.67C +0.18
SR = 0.13X + 0.11
S = 0.25X + 0.17
X = 0.68C + 0.09
SR = 0.20X - 0.01
S = 0.29X + 0.04
X = 0.67C + 0.12
f1-55ug/Lf

SR = 0.20X + O.OS
S = 0.37X - 0.07
X = 0.65C + 0.22
SR = 0.25X + 0.03
S = 0.35X - O.06
X = 0.65C + 0.17
SR = 0.27X + 0.08
S = 0.34X + 0.21
X = 0.59C + 0.25
SR = 0.18X + 0.09
S = 0.28X + 0.04
X = 0.64C + 0.21
SR = 0.16X + 0.09
S = 0.32X + 0.09
X = 0.60C +0.07
SR = 0.22X - 0.05
S = 0.29X + 0.03
X = 0.64C + 0.16
(at Concentration range of compound for which regression equations are generally applicable.
X = Mean Recovery
C = True Value for the Concentration
  d. Operation of the  gas chromato-
     graphic flame  ionization  detection
     and  gas chromatographic electron
     capture detection systems.
The user of the method must exercise
care in conducting  these operations in
order to obtain accurate and reproducible
data.
              Glenn Kinzer. Ralph Riggin.  Thomas Bishop, Michelle A. Bins, Cory C. Howard,
                and Robert Iden  are with Battelle-Columbus Laboratories,  Columbus  OH
                43201.
              Edward L. Berg and Robert L. Graves are the EPA Project Officers (see below).
              The complete report, entitled "EPA Method Study 19, Method 609 (Nitroaromatics
                and Isophorone)." (Order No. PB84-176 908; Cost: $11.50, subject to change)
                will be available only from:
                       National Technical Information Service
                       5285 Port Royal Road
                       Spring field, VA 22161
                       Telephone:  703-487-4650
              The EPA Project Officers can be contacted at:
                       Environmental Monitoring and Support Laboratory
                       U.S. Environmental  Protection Agency
                       Cincinnati, OH 45268
                                                                               U.S GOVERNMENT PRINTING OFFICE; 1984 — 759-015/7693

-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
     BULK RATE
POSTAGE & FEES PAID
        EPA
   PERMIT No. G-35
Official Business
Penalty for Private Use $300

-------